This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
, <w>. In Esperanto fehlen die Umlaute sowie, <w>, und. Es benutzt einige diakritische Zeichen: [o], [D], [x], <Óˆ> [z], <sˆ> [s] und [w]. In Esperanto und Volapük ist die Orthographie phonema-
tisch, in Interlingua dagegen historisch orientiert. Die Morpheme der drei Plansprachen sind den europäischen Hauptsprachen entnommen und in unterschiedlichem Maße verändert worden. Die Morpheme des Volapük stammen vor allem aus dem Englischen, Deutschen und Französischen sowie aus dem Latein. Sie wurden meist bis zur Unkenntlichkeit verändert: (1) engl. knowledge J nol ‘Wissen’, speak J pük ‘Sprache’; dt. Mund J mud ‘Mund’, Dank J dan ‘Dank’; lat. animal J nim ‘Tier’, angulus J gul ‘Ecke’; frz. universite´ J niver ‘Universität’, compliment J plim ‘Kompliment’. Die Morpheme des Esperanto der ersten zwei bis drei Jahrzehnte seiner Existenz entstammen zu ca. 75 Prozent latiniden Sprachen, zu 20 Prozent germanischen und zu 5 Prozent anderen (vgl. Janton 1993: 51). Sie wurden nach graphematischen, phonematischen und semantischen Gesichtspunkten ausgewählt und z. T. verändert (vgl. Szerdahelyi 1983: 285⫺290). Dabei geht es um Morpheme als Teile der synchronen Wortbildung: (2) lat. augurare J au˘gur-i ‘wahrsag-inf’, certus J cert-a ‘sicher-adjr’; frz. accoucher J akusˆ-i ‘entbind-inf’, crayon J krajon-o ‘Bleistift-nr’; engl. in spite of J spit-e ‘trotz-advr’, team J team-o ‘Mannschaft-nr’; ital. ancora J ankorau˘ ‘noch’, atunno J au˘tun-o ‘Herbst-nr’; dt. Bild J bild-o ‘Bild-nr’; russ. kolbasa J kolbas-o ‘Wurst-nr’, krome J krom ‘außer’. Die Morpheme in Interlingua sind Durchschnittsformen des Italienischen, Spanischen, Portugiesischen und Französischen. Fehlt Übereinstimmung in mehr als einer dieser Sprachen, die als Kontrollsprachen gelten, so werden die Morpheme aus dem Lateinischen entnommen (vgl. Gode 21971: XXVI). Die Äquivalente des Beispiels (1) lauten in Interlingua: (3) cognoscentia (saper), lingua, bucca (ore), regratiamento, animal, angulo, universitate, complimento. Die Interlingua-Äquivalente für (2) lauten: (4) auguar, certe, assister in le parto, stilo de graphite, malgrado, equipa, anque (etiam), autumno, pictura, salsicia, foris de. Wortstammvarianten fehlen in Volapük und Esperanto, sind aber in Interlingua häufig, da
144. Plansprachen
diese Plansprache das Wortbildungssystem ihrer Quellsprachen übernommen hat (z. B. zu face-r ‘mach1-inf (machen)’, fact-o ‘mach2-nr (Tatsache)’, effect-o ‘wirk1-nr (Wirkung)’, effic-ace, ‘wirk2-adjr1 (wirksam)’, effect-ive, ‘wirk1-adjr2 (wirklich)’). Nur in diachroner Sicht auf die Quellsprachen gibt es auch im Esperanto solche Varianten (far-i ‘mach1-inf (machen)’, fakt-o ‘mach2-nr (Tatsache)’, efekt-o ‘wirk1-nr (Wirkung)’, efik-a ‘wirk2-adjr (wirksam)’, efektiv-e ‘wirk3-advr (wirklich)’). Sie sind synchron jedoch selbständige Basen für weitere Wortbildungen.
4.
Wortbildung
4.1. Kompatibilität der Morpheme Die Morpheme in Volapük und Esperanto zeichnet eine hohe Kompatibilität aus. Die syntagma-internen Beziehungen der unmittelbaren Konstituenten einer Konstruktion werden durch die Faktoren der phonologischmorphologischen und der semantischen inneren Valenz gesteuert (vgl. Stepanova 1971). Das Wirken der semantischen Faktoren ist bei Volapük und Esperanto stärker möglich als bei Interlingua, deren Morphemkonstruktionen ihren Prototypsprachen folgen müssen und durch zahlreiche Ausnahmen gekennzeichnet sind. Volapük und insbesondere Esperanto zeichnen sich somit durch eine leistungsfähige autonome Wortbildung aus, die natürlich auch durch ethnosprachige Einflüsse, Analogiebildungen, die Morphemstruktur, Wirkungen der Sprachtradition und andere Faktoren beeinflußt wird. Die wichtigsten morphologischen Eigenschaften des Esperanto sind durch Unveränderbarkeit, hohe Kompatibilität und eine relativ einfache Struktur der vorwiegend romanischen Morpheme umrissen. Das unterstreicht das folgende Zitat des Begründers des Esperanto: “Ich habe die bestehenden Begriffe vollständig zergliedert, so dass die ganze Sprache nur aus unveränderlichen Wörtern besteht, anstatt aus Wörten mit einer unendlichen Anzahl von grammatikalischen Formen ... Verschiedene grammatikalische Formen aber, so wie die gegenseitigen Beziehungen zwischen den Wörtern u. dergl., werden durch Anreihung unveränderlicher Wörter ausgedrückt ... er [der Lernende-DB] ahnt sogar nicht, dass das, was er Endung, Präfix oder Suffix nennt, ein völlig selbständiges Wort ist, das immer dieselbe Bedeutung behält, sei es am Anfang oder am Ende eines anderen Wortes, oder auch als selbständiges Wort; dass jedes Wort mit gleichem Rechte als Stammwort
1565 oder als grammatikalisches Anhängsel gebraucht werden kann.” (Dr. Esperanto 1887: 13)
Zamenhofs Auffassung von der Zergliederung der Ideen in selbständige Elemente impliziert das Konzept vom lexikalisch-kategorialen Charakter der autosemantischen Grundmorpheme (vgl. Kalocsay & Waringhien 51985: 375⫺390), wonach diese substantivische, adjektivische oder verbale Grundbedeutung haben und erst durch Anfügen der wortkategorialen Suffixe auf die Wortebene gehoben werden. So hat das substantivische wortkategoriale Morphem {-o}, angefügt an das substantivische Morphem {film} (film-o, ‘Film-nr’), in lexikalischer Hinsicht pleonastische Bedeutung, trägt aber grammatische Information (Singular). An das verbale Grundmorphem {skrib} (skrib-i ‘schreib-inf’) angefügt, substantiviert es dieses (skrib-o, ‘Schrift’ oder ‘Schreiben’, in Abhängigkeit vom Kontext). Das bedeutet, daß in Esperanto Komposition als Wortbildungsverfahren vorherrscht. Diese Theorie wurde durch den Bruder von Ferdinand de Saussure begründet (vgl. Rene´ de Saussure 1910; Couturat 1907). Sie hat sich u. a. für kontrastivlinguistische Untersuchungen (vgl. Blanke 1981) und für Zwecke der automatischen Übersetzung (vgl. Schubert 1997: 129 f.) bewährt. 4.2. Einige Wortbildungstypen und -prozesse Volapük, Esperanto und Interlingua unterscheiden sämtliche traditionellen Wortarten. In Esperanto können substantivische, adjektivische und verbale Grundmorpheme (Träger des Wortbedeutungskerns und Ausgangspunkt der Wortbildung) auf der Wortebene nur als gebundene Morpheme auftreten. Die Autosemantika sind morphologisch gekennzeichnet. Das betrifft in Volapük nur Adjektiv, Verb und abgeleitetes Adverb und in Interlingua ebenfalls Verb und abgeleitetes Adverb. Die folgende Tabelle macht das für drei Wortarten deutlich: Abgeleitete Adverbien werden in Esperanto durch {e}, in Volapük durch {o} und in Interlingua durch {(a)mente} gekennzeichnet: Esperanto rapid-e, Interlingua rapidamente, Volapük vif-o ‘Schnelligkeit-advr (schnell)’. Da Interlingua den Wortbildungsmustern der romanischen Sprachen folgt, macht es dem romanisch Gebildeten ⫺ gemessen an seinen Sprachgewohnheiten ⫺ einen natürli-
1566
XVI. Systeme morphologischer Struktur: Sprachskizzen
Substan- Adjektive Verben tive (Inf.) Volapük
böd plan nam
gret-ik ful-ik gol-ik
vob-ön vok-ön pön-ön
Interlingua
ave planta mano
grande plen prematur
labora-r puni-r voce-r
Esperanto
bird-o plant-o man-o
grand-a plen-a fru-a
labor-i vok-i pun-i
(Deutsch
Vogel Pflanze Hand
groß voll früh
arbeiten rufen strafen )
Tab. 144.1: Substantive, Adjektive und Verben in Volapük, Interlingua und Esperanto
cheren Eindruck als Esperanto und insbesondere als Volapük. Diese “Natürlichkeit” erhöht die passive Verständlichkeit der Texte, geht aber zu Lasten der Regelmäßigkeit in der Wortbildung. Diese wiederum erleichtert die aktive Beherrschung und von Vorbildsprachen unabhängige Abbildung von Begriffen, wie die folgende Tabelle für die von Substantiven abgeleiteten Adjektive zeigt: In Esperanto ist das Hauptverfahren die Komposition. Denominale Komposita sind häufig: (5) dom-o ⫹ pord-o J dom-pord-o ‘Haus-Tür-nr (Haustür)’ Weniger häufig sind Deadjektiva: (6) rapid-a ⫹ trajn-o J rapid-trajn-o ‘schnell-Zug-nr (Eilzug)’ und deverbale Komposita:
Nation lun-a mar system-a fratr-e
(8) skrib-i ⫹ il-o ‘schreib-inf ⫹ Instrumentnr (Schreibstift)’; mis-skrib-i ‘negschreib-inf (verschreiben)’ sind syntaktisch auflösbar, wie die auf (5)⫺(8) bezogenen Beispiele zeigen: (9) pord-o de dom-o ‘Tür-nr von Haus-nr’, rapid-a trajn-o ‘schnell-adjr Zug-nr’, plor-(-ant-)-e lament-i ‘wein-(-part.akt. pr‰s-)-advr klag-inf’, il-o por skrib-i ‘Instrument-nr für schreib-inf’, mis-e skrib-i ‘neg-advr schreib-inf’. Esperanto verfügt traditionell über 10 Präfixe und 31 Suffixe, von denen der größte Teil als Affixoide angesehen werden kann, da sie auch als selbständige Basen für Wortbildungsprozesse funktionieren. Das ist in Volapük selten der Fall und fehlt in Interlingua. Konstruktionen nur aus Affixoiden sind häufig: (10) estr-ar-an-o ‘Leiter-Menge-Mitglied-nr (Vorstandsmitglied)’; re-dis-igˆ-i ‘wiederauseinander-inch-inf (sich wieder trennen)’. Aber auch Adjektive, Präpositionen und andere Wortarten können den Status eines Affixoids erhalten, wenn sie reihenbildend sind: (11) plen-a ‘voll-adjr’, dolor-o ‘schmerz-nr’ J dolor-plen-a ‘schmerz-voll-adjr (schmerzvoll)’, labor-plen-a ‘arbeit-volladjr (arbeitsvoll, arbeitsreich)’; en ‘in’, ir-i ‘geh-inf (gehen)’ J en-ir-i ‘in-gehinf (hineingehen)’, ˆÓet-i ‘werf-inf (werfen)’, en-Óˆet-i ‘in-werf-inf (hineinwerfen)’. Explizite Derivativa entstehen durch Kombinationen von Grundmorphemen (labor-o ⫹ temp-o J labor-temp-o ‘arbeit-Zeit-nr (Ar-
(7) plor-i ⫹ lament-i J plor-lament-i ‘wein-klag-inf (wehklagen)’
Interlingua
Diese und auch mit Affixoiden gebildete Komposita:
Esperanto nation-a lun-ar mar-in system-atic frat-erne
naci-o lun-o mar-o sistem-o frat-o
Volapük naci-a lun-a mar-a sistem-a frat-a
net mun mel sit blod
net-ik mun-ik mel-ik sit-ik blod-ik
Tab. 144.2: Adjektivableitungen von Substantiven in Interlingua, Esperanto und Volapük
144. Plansprachen
beitszeit)’) oder durch Konstruktionen mit wortkat. Suffixen (blu-a ⫹ okul-o ⫹ -a J bluokul-a ‘blau-Auge-adjr (blauäugig)’). Die hohe Kompatibilität der Morpheme des Esperanto erlaubt oft eine weitestgehende Auffüllung von Modellen der Wortbildung. Es sei z. B. das Modell Präposition ⫹ Substantiv ⫹ faktitives Suffix -ig- angeführt: (12) sur-tabl-ig-i ‘auf-Tisch-fakt-inf (auf den Tisch bringen, auftischen)’, subtabl-ig-i ‘unter-Tisch-fakt-inf (unter den Tisch bringen)’, sur-ter-ig-i ‘aufErde-fakt-inf (auf die Erde bringen, landen)’, en-ter-ig-i ‘in-Erde-fakt-inf (in die Erde bringen, begraben)’, el-terig-i ‘aus-Erde-fakt-inf (aus der Erde bringen, ausgraben)’. Diese Konstruktionen sind dann wieder Basen für weitere Wortbildungen, z. B. für Adjektive und Substantive: (13) Adjektive: sur-tabl-ig-a ‘auf-Tisch-faktadjr’, sub-tabl-ig-a ‘unter-Tisch-faktadjr’, sur-ter-ig-a ‘auf-Erde-fakt-adjr’, en-ter-ig-a ‘in-Erde-fakt-adjr’, el-terig-a ‘aus-Erde-fakt-adjr’; Substantive: sur-tabl-ig-o ‘auf-Tischfakt-nr’, sub-tabl-ig-o ‘unter-Tischfakt-nr’, sur-ter-ig-o ‘auf-Erde-faktnr’, en-ter-ig-o ‘in-Erde-fakt-nr’, elter-ig-o ‘aus-Erde-fakt-nr’ usw. Die Faktoren der Sprachökonomie bei Wortbildungsprozessen in Esperanto wirken durch die Prinzipien der Sprachbequemlichkeit und der Redundanz (s. Art. 31). Sie regeln, daß in ein Wortsyntagma alle die Morpheme eingefügt werden, die erforderlich sind, aber nicht mehr als ausreichen, um möglichst klar und genügend redundant die Bedeutung zu motivieren, die durch das Wort bezeichnet werden soll (vgl. Rene´ de Saussure 1910: 74; Kalocsay & Waringhien 51985: 385⫺388). So kann z. B. in Abhängigkeit vom Kontext ‘Essen’ (als Nahrungsmittel oder Handlung) durch mangˆ-o ‘ess-nr’ bezeichnet werden. Fehlen jedoch Kontextinformationen, so ist für ‘konkrete Sache’ das Suffix -aÓˆ- (mangˆ-aÓˆ-o ‘esskonkrete.Sache-nr’) oder für ‘Dauer der Tätigkeit’ das Suffix -ad- (mangˆ-ad-o ‘ess-andauernde.Tätigkeit-nr’) anzufügen. In Esperanto entstehen zahlreiche Synonyme oder Teilsynonyme durch das parallele Wirken von zwei Möglichkeiten der Wortschatzerweiterung. Zum einen werden komplette (meistens romanische, aber zunehmend auch
1567 englische) Konstruktionen übernommen und in Esperanto als Grundmorpheme eingepaßt: redaktor-o ‘Redakteur-nr (Redakteur)’. Zum anderen werden vorhandene Wortbildungsmorpheme genutzt: redakt-ist-o ‘redigier-beruflich. Tätiger-nr (Redakteur)’. Volapük kennt Komposition und Derivation. Derivation überwiegt. Es gibt 56 Präfixe und 85 Suffixe, Wortbildungsaffixe und grammatische Affixe inklusive. Darunter sind 20 nur für die Fachsprache der Medizin und der Chemie vorgesehen (vgl. Schmidt 1933: 17). Bei denominalen Komposita wird in die Kompositionsfuge das Genitivmorphem {a} eingefügt: vin-a-flad ‘Wein-gen-Flasche (Weinflasche)’, vol-a-pük ‘Welt-gen-Sprache (nom.sg) (Weltsprache)’. Bei denominalen Komposita bezeichnet Volapük die syntagma-internen Beziehungen der unmittelbaren Konstituenten manchmal genauer als Esperanto: (14) Esperanto: hom-am-o ‘Mensch-lieb-nr (Menschenliebe, Liebe des [zum] Menschen)’. Volapük: men-a-löf ‘Menschgen-Liebe (des Menschen Liebe)’, mel-ilöf ‘Mensch-akk-Liebe (Liebe zum Menschen)’. In Interlingua ist das Hauptverfahren der Wortbildung die Derivation, die jedoch mehr eine Lehre der Etymologie der romanischen Quellsprachen ist als ein produktives Wortbildungsverfahren. Es werden 124 aktive Wortbildungsaffixe aufgeführt, 44 Präfixe und 80 Suffixe (vgl. Gode & Blair 21971: XLVI⫺XLVII). Statt nominaler Komposition wie in Esperanto oder Derivation wie in Volapük, bevorzugt Interlingua syntaktische Ausdrücke: (15) Esperanto: bus-staci-o ‘Bus-Station-nr (Busstation)’; sun-lum-o ‘Sonne-Lichtnr (Sonnenlicht)’; akv-o-fort-o ‘Wasser-nr-Kraft-nr (Wasserkraft)’; akv-ovapor-o ‘Wasser-nr-Dampf-nr (Wasserdampf)’; sen-akv-a ‘ohne-Wasser-adjr (wasserlos)’ Interlingua: station de omnibus; lumine solar; fortia hydraulic; vapor de aqua; sine aqua. Syntaktische Ausdrücke können bei Bedarf auch im Esperanto gebildet werden, und zwar sowohl attributive als auch präpositionale. Die Äquivalente zu (15) sind: (16) bus-a staci-o ‘Bus-adjr Station-nr’, staci-o de bus-o ‘Station-nr von Bus-
1568
XVI. Systeme morphologischer Struktur: Sprachskizzen
nr’; sun-a lum-o ‘Sonne-adjr Licht-nr’, lum-o de sun-o ‘Licht-nr von Sonne-nr’; akv-a fort-o ‘Wasser-adjr Kraft-nr’, fort-o de akv-o ‘Kraft-nr von Wasser-nr’; akv-a vapor-o ‘Wasser-adjr Dampf-nr’, vapor-o de akv-o ‘Dampfnr von Wasser-nr’; sen akv-o ‘ohne Wasser-nr’. Mit Wortbildungsmitteln können in Volapük und Esperanto (jedoch nicht in Interlingua) besonders regelmäßig und systemdurchgängig einige Aktionsarten bezeichnet werden. Dazu folgende Beispiele: (17) Intensiva: Volapük (le-): le-dun-ön ‘ints-Tat-inf (eifrig tun)’; le-fäl-ön ‘ints-Fall-inf (stürzen)’. Esperanto (-eg-): rid-eg-i ‘lach-ints-inf (schallend lachen)’; krieg-i ‘schrei-ints-inf (brüllen)’. (18) Faktitiva: Volapük (-ükön): red-ükön ‘Röte-fakt. inf (rot machen, röten)’; fin-ükön ‘Ende-fakt.inf (beenden)’. Esperanto (-ig-): rugˆ-ig-i ‘rot-fakt-inf (röten)’, pligrand-ig-i ‘kmpr-groß-fakt-inf (vergrößern)’. In Interlingua selten: z. B. purifica-r ‘sauber-fakt-inf (säubern)’, neutral-isa-r ‘neutral-fakt-inf (neutralisieren)’. (19) Kausativa: Volapük (aJi; aJä): seat-ön ‘lieg-inf’ J seit-ön ‘lieg:kaus-inf (legen)’; lag-ön ‘häng(intr.)-inf’ J läg-ön ‘häng(intr.): kaus-inf (hängen, tr.)’. Esperanto (ig-): kusˆ-ig-i ‘lieg-kaus-inf (etw./jmd. hin legen)’; pend-ig-i ‘häng(intr.)-kaus-inf ((auf)hängen, trans.)’. (20) Inchoativa: Volapük (-ikön): red-ikön ‘rot-inch.inf (rot werden)’; fin-ikön ‘Ende-inch.inf (zu Ende gehen)’. Esperanto(-igˆ-): rugˆigˆ-i ‘rot-inch-inf (rot werden, erröten)’; pal-igˆ-i ‘bleich-inch-inf (erbleichen)’. Volapük und Esperanto können die Aspektkategorie Perfektiv lexikalisch bezeichnen: (21) Volapük (fi-): Infinitiv dun-ön ‘Tat-inf (tun)’, Perfektiv fi-dun-ön ‘pfv-Tat-inf (zu Ende tun, erledigen)’; pen-ön ‘Schreibstift-inf (schreiben)’, fi-pen-ön ‘pfv-Schreibstift-inf (fertig schreiben)’. (22) Esperanto (fin- von fin-i ‘Ende-inf (beenden)’): fin-skrib-i ‘Ende-schreib-inf (zu Ende schreiben)’, fin-far-i ‘Endetun-inf (zu Ende tun, erledigen)‘.
5.
Grammatische Kategorien und Paradigmen
Die drei Plansprachen kennen kein Genus. Beim Numerus wird der Singular durch Nullmorphem und der Plural für alle Kasus durch ein Formativ (Volapük: {-s}, Esperanto: {-j}, Interlingua: {-s} bzw. {-es}) morphologisch bezeichnet: (23) Volapük: ob-ik gud-ik buk ‘ich-adjr gutadjr Buch (mein gutes Buch)’, ob-ik gud-ik buk-s ‘ich-adjr gut-adjr Buch-pl (meine guten Bücher)’. Esperanto: mi-a bon-a libr-o ‘ich-adjr gut-adjr Buchnr’, mi-a-j bon-a-j libr-o-j ‘ich-adjr-pl gut-adjr-pl Buch-nr-pl’. Interlingua: mi bon libro ‘mein gut Buch’, mi bon libro-s ‘mein gut Buch-pl’. In Volapük werden die Kasus Genitiv, Dativ und Akkusativ morphologisch bezeichnet. Es kennt außerdem den Vokativ. In Esperanto wird nur der Akkusativ durch das monofunktionale {-n} bezeichnet, das ggf. hinter dem Plural-Marker {-j} folgt: mi vid-as mi-aj-n bon-a-j-n amik-o-j-n ‘ich seh-akt.pr‰s ichadjr-pl-akk gut-adjr-pl-akk Freund-nr-plakk (ich sehe meine guten Freunde)’. Für den Nominativ kann ein Nullmorphem (la amiko-ø ‘art Freund-nr-nom’) angenommen werden. Interlingua kennt keine morphologisch bezeichneten Kasus, hat aber Varianten zur Bezeichnung der Objektformen für einige Personalpronomen (io ‘ich’ ⫺ me ‘mich’, ille ‘er’ ⫺ le ‘ihn’). Sämtliche in anderen Sprachen markierten Kasus werden in Esperanto und Interlingua durch Präpositionalphrasen wiedergegeben (Genitiv: Esperanto: de la teknik-o ‘von art Technik-nr’, Interlingua: del technica ‘von:def Technik’; Dativ: Esperanto: al la teknik-o ‘zu art Technik-nr’, Interlingua: al technica ‘zu Technik’). Das Deklinationsparadigma für Substantive in Volapük ist regelmäßig (fat ‘Vater’):
Nominativ Genitiv Dativ Akkusativ (Vokativ
Singular
Plural
fat fat-a fat-e fat-i o fat
fat-s fat-a-s fat-e-s fat-i-s o fat-s)
Tab. 144.3: Deklination von Substantiven in Volapük
1569
144. Plansprachen
Aktiv
Passiv
Volapük
Präs. Prät. Fut.
vok-öl ä-vok-öl o-vok-öl
pa-vok-öl pä-vok-öl po-vok-öl
Esperanto
Präs. Prät. Fut.
vok-ant-a vok-int-a vok-ont-a
vok-at-a vok-it-a vok-ot-a
Präs. Prät.
voca-nte
Futurpartizipien Aktiv (-ont-a) und Passiv (-ot-a) gebildet werden: (24) mi est-as vok-ont-a ‘ich sein-pr‰s rufpart.akt.fut-adjr (ich bin im Begriff zu rufen)’, analog dazu: Präteritum estis vok-ont-a, Futur est-os vok-ont-a und Konditional est-us vok-ont-a. Dazu gehören die Passivformen:
Tab. 144.4: Partizipien in Volapük, Esperanto und Interlingua
(25) mi est-as vok-ot-a ‘ich sein-pr‰s rufpart.pass.fut-adjr (man ist im Begriff, mich zu rufen)’, analog dazu: Präteritum est-is vok-ot-a, Futur est-os vok-ot-a und Konditional est-us vok-ot-a.
Nach dem gleichen Paradigma werden in Volapük auch Adjektive (wenn sie nicht unmittelbar hinter dem Substantiv stehen), der Artikel, einige Pronomen und Numeralien dekliniert. Die Konjugation ist in allen drei Plansprachen regelmäßig. Für die Bildung der komplexen Zeiten werden Partizipien verwendet. {-öl} dient in Volapük zur allgemeinen Bezeichnung von Partizipien (Infinitive: Volapük vok-ön, Esperanto vok-i, Interlingua voca-r ‘ruf-inf’): Volapük kann auch Partizipien vom Perfekt, Plusquamperfekt und Futur II bilden, die hier nicht aufgeführt sind. Interlingua kennt nur die Partizipien Präsens Aktiv und Präteritum Passiv. Esperanto hat in Übereinstimmung mit seinen Tempi ein symmetrisches System. Interlingua übernimmt eine Reihe von Unregelmäßigkeiten aus den romanischen Sprachen, so drei Verbalstämme: auf -a- (voca-r ‘ruf-inf’), auf -e- (batte-r ‘schlag-inf’) und auf -i- (puni-r ‘straf-inf’). Bei e-Stämmen wird im Passivpartizip -e- zu -i- (batti-te ‘schlag-part.pass.pr‰t’). Sowohl Volapük als auch Esperanto verfügen über die sprachlichen Mittel, Tempi in verschiedenen Modi sehr differenziert darzustellen. Dazu dienen im Esperanto die Partizipien und das Modalverb est-i (‘sein-inf’). Neben dem Indikativ kennen Esperanto und Interlingua als Modi nur noch Imperativ und Konditional. Volapük unterscheidet zusätzlich Konjunktiv und Volitiv. Volapük geht synthetisch, Esperanto und Interlingua sowohl synthetisch als auch analytisch vor. Die wichtigsten Zeitformen des Verbs im Indikativ zeigt Tab. 144.5. In Esperanto können in den verschiedenen Zeitstufen auch komplexe Ausdrücke mit den
Auch die Bildung synthetischer Formen durch die Anfügung des finiten Hilfsverbs est-as ‘sein-pr‰s’, est-is ‘sein-pr‰t’, est-os ‘sein-fut’ an das jeweilige Partizip (-ant-a, -int-a, -ont-a bzw. -at-a, -it-a, -ot-a) ist möglich, kommt aber selten vor (mi labor-as ‘ich arbeit-akt.pr‰s’ J mi est-as labor-ant-a J mi labor-ant-as ‘ich arbeit-part.akt.pr‰s-akt. pr‰s (ich arbeite)’ als Verlaufsform). Analog dazu gibt es Formen im Passiv: la leter-o skrib-at-is ‘art Brief-nr schreib-part.pass. pr‰s-akt.pr‰t (der Brief wurde geschrieben)’ usw. Nur in Volapük werden die Stufen der Komparation morphologisch bezeichnet: gletik (‘groß-adjr’), glet-ik-um ‘groß-adjr-kmpr (größer)’, glet-ik-ün ‘groß-adjr-sup (am größten)’ (Esperanto: grand-a ‘groß-adjr’, pli grand-a ‘mehr groß-adjr’, plej grand-a ‘am meisten groß-adjr’; Interlingua: grande, plus grande, le plus grande). Unter Esperantologen umstritten ist die Frage, ob die zusammengesetzten Zeiten im Passiv mit estis -at-a/estis it-a Aspektualität ausdrücken oder nur temporale Bedeutung vermitteln (vgl. La Zamenhofa 1961; Lötzsch 1991). Die in Esperanto erfolgte regelmäßige morphologische Markierung der Verben, Substantive, Adjektive und der abgeleiteten Adverbien, der grammatischen Kategorien des Verbs und des Akkusativs ermöglichen eine klare Markierung der Satzglieder und damit der Satzstrukturen. Die Satzstruktur la ~o ~e ~as ~an ~on en la ~o ist durch semantisch kompatible Wörter leicht auffüllbar: (26) la student-o rapid-e art Student-nr schnell-advr leg-as interes-a-n les-akt.pr‰s interessant-adjr-akk libr-o-n en la bibliotek-o Buch-nr-akk in art Bibliothek-nr ‘der Student liest schnell ein interessantes Buch in der Bibliothek’
Interlingua
voca-te
1570
XVI. Systeme morphologischer Struktur: Sprachskizzen
Aktiv
Passiv
Präsens: Volapük Esperanto Interlingua
vok-ob mi vok-as (mi est-as vok-ant-a) io voca
pa-vok-ob mi est-as vok-at-a io es (⫽ esse) voca-te
Präteritum: Volapük Esperanto Interlingua
ä-vok-ob mi vok-is (mi est-is vok-ant-a) io voca-va
pä-vok-ob mi est-is vok-ata io esse-va voca-te
Perfekt: Volapük Esperanto Interlingua
e-vok-ob mi est-as vok-int-a io ha (⫽ habe) voca-te
pe-vok-ob mi est-as vok-it-a io ha essi-te voca-te
Plusquamperf.: Volapük Esperanto Interlingua
i-vok-ob mi est-is vok-int-a io habe-va voca-te
pi-vok-ob mi est-is vok-it-a io habe-va essi-te voca-te
Futur I: Volapük Esperanto Interlingua
o-vok-ob mi vok-os (mi es-tos vok-ant-a) io voca-ra´
po-vok-ob mi estos vokata io esse-ra´ voca-te
Futur II: Volapük Esperanto Interlingua
u-vok-ob mi est-os vok-int-a io habe-ra´ voca-te
pu-vok-ob mi est-os vok-it-a io habe-ra´ essi-te voca-te
Tab. 144.5: Konjugationsbeispiele (Indikativ) für Volapük, Esperanto und Interlingua
Seine klare morphologische Struktur macht Esperanto geeignet für computerlinguistische Zwecke (vgl. Schubert 1997). Vergleichbare Aussagen lassen sich zu Volapük und Interlingua wegen fehlender Untersuchungen nicht machen. Typologisch ist Volapük eine lexikalisch apriorisch-aposteriorische Mischsprache (vgl. 1) des agglutinierenden Typs. Interlingua ist lexikalisch rein romanisch und hat wie seine Quellsprachen flektierende Züge. Esperanto ist ebenfalls hochgradig agglutinierend, hat aber auch isolierende Züge. Einige Besonderheiten seiner Struktur ergeben sich aus der Spannung zwischen europäischer Herkunft des Morphemmaterials und seiner Verarbeitung, die asiatische Züge aufweist (vgl. Piron 1980: 33 f.; Nuessel 2000: 55⫺57).
6.
Linguistische Bedeutung
Die wissenschaftliche Auseinandersetzung mit Plansprachen kann nicht nur helfen, neue Wege zur Lösung praktischer internationaler
Kommunikationsprobleme zu finden, sondern hat auch heuristische Bedeutung für linguistische Erkenntnisse. Das betrifft sowohl die Reflexion linguistischer Einsichten und Fakten sowie sprachphilosophischer Probleme der jeweiligen Epoche, in der die einzelnen Projekte seit dem Mittelalter entstanden sind, als auch die Möglichkeit, an (wenn auch unterschiedlich lange und intensiv funktionierenden) Plansprachen Theorien zu überprüfen und weiterzuentwickeln. Erwähnt sei z. B. die Frage der Planbarkeit sprachlicher Strukturen und Prozesse und die Rolle des Bewußten in der Entwicklung von Sprachen, die Bewertung sprachlicher Eigenschaften (was ist “leicht” oder “schwer” an einer Sprache) u. a. Quasi unter experimentellen Bedingungen lassen sich Fragen des Sprachwandels an in ihrer Entstehung und Entwicklung genau datierbaren Sprachen verfolgen. Bei vielen Autoren von Plansprachen waren Aspekte der komparativen und konfrontativen Linguistik intuitiv vorhanden, oder es wurden entsprechende Untersu-
1571
144. Plansprachen
chungen durchgeführt. Das schließt auch häufig Reflexionen über die allen Sprachen gemeinsamen Züge (Universalien) ein. Für die Entstehung und Entwicklung der Terminologiewissenschaft z. B. waren plansprachliche Impulse von großer Bedeutung (vgl. Blanke 1998 b: 136). Erfahrungen mit Plansprachen haben sich auf die Entwicklung der Theorien mancher Linguisten ausgewirkt (vgl. Martinet 1991). Eine qualitativ und quantitativ ständig wachsende interlinguistische und esperantologische Fachliteratur erleichtert die Auswertung der mit Plansprachen gewonnenen theoretischen und praktischen Erkenntnisse und Erfahrungen (vgl. Blanke 2003: 155⫺192).
7.
Illustrative Texte
Von Volapük über Esperanto zu Interlingua zeigen die Texte für den mit europäischen Sprachen Vertrauten eine zunehmende Verständlichkeit. Der romanische Charakter der Morpheme nimmt deutlich zu. Der synthetische Charakter von Volapük kontrastiert mit dem analytischen von Interlingua. Esperanto weist sowohl synthetische als auch analytische Züge auf. 7.1. Volapük (27) Vol-a-pük bin-on pük Welt-gen-Sprache sein-3.sg.n Sprache mek-av-ik Kunst-Wissenschaft-adjr sek-o pük Resultat-advr Sprache le-din-o aug-Sache-advr ned-öf-ik. Bedürfnis-Eigenschaft-adjr ‘Die Weltsprache ist (eine) künstliche Sprache (die) folglich (eine) Sprache (ist, die) vor allem benötigt (wird).’ (Sirkülapenäd lü Volapükaflens Valöpo. Scunthorpe, 2001, 1, 1) 7.2. Esperanto (28) La kanton-o-j est-as art Kanton-nr-pl sein-pr‰s suveren-a-j kaj rajt-as souverän-adjr-pl und recht-akt.pr‰s re-form-i si-a-n wieder-form-inf refl-adjr-akk lern-ej-sistem-o-n konform-e lern-Ort-System-nr-akk konform-advr
al lok-a-j bezon-o-j. zu Ort-adjr-pl bedürfen-nr-pl ‘Die Kantone sind souverän und haben das Recht, ihr Schulsystem zu reformieren, entsprechend (zu) den örtlichen Bedürfnissen.’ (Monato. Antwerpen, 2001, 22.2, 15) 7.3. Interlingua (29) Post le trovata de un tumba pro nach def Fund von indef Grab für un infante neanderthal, le indef Kind Neanderthal def experto-s ha deb-ite Expert-pl hab müss-part.pass.pr‰t revide-r lor theoria-s. revidier-inf def.pl Theorie-pl ‘Nach dem Fund eines Grabes für ein Neanderthaler-Kind, die Experten haben gemußt, ihre Theorien zu revidieren’ (Panorama: In Interlingua. Beekbergen, 2000, 13.6,17)
8.
Zitierte Literatur
Back, Otto (1996), “108. Plansprachen”. In: Goebl, Hans et al. (Hrsg.), Kontaktlinguistik. Contact Linguistics. Linguistique de contact. Berlin, New York: de Gruyter (Handbücher zur Sprach- und Kommunikationswissenschaft 12.1), 881⫺888 Becker, Ulrich (2001), “Interlinguistik im Internet”. In: Fiedler, Sabine & Liu, Haitao (Hrsg.), Interlingvistikaj studoj. Interlinguistische Studien. Festlibro por/Festschrift für Detlev Blanke. Dobrˇichovice: KAVA-PECH, 244⫺253 Bishop, Brian R. (1983), “Sˆangˆoj en Volapuko”. Planlingvistiko 2.5, 2⫺8 Blanke, Detlev (1981), Plansprache und Nationalsprache: Einige Probleme der Wortbildung des Esperanto und des Deutschen in konfrontativer Darstellung. Berlin: Akad. d. Wiss. (Linguistische Studien 85) Blanke, Detlev (1985), Internationale Plansprachen. Berlin: Akademie-Verlag (Sammlung Sprache 34) Blanke, Detlev (1997), “The Term ‘Planned Language’”. In: Tonkin (Hrsg.), 1⫺20 Blanke, Detlev (1998 a), “Interlinguistik und Plansprachen”. Sitzungsberichte der Leibniz-Sozietät 21.2, 45⫺76 Blanke, Detlev (1998 b), “Terminology Science and Planned Languages”. In: Oeser, Erhard & Galinski, Christian (Hrsg.), Eugen Wüster (1898⫺1977).
1572
XVI. Systeme morphologischer Struktur: Sprachskizzen
Leben und Werk: Ein österreichischer Pionier der Informationsgesellschaft. Vienna: TermNet, 133⫺ 168
Lapenna, Ivo & Lins, Ulrich & Carlevaro, Tazio (1974), Esperanto en perspektivo. Rotterdam: Universala Esperanto Asocio
Blanke, Detlev (2000), “Vom Entwurf zur Sprache”. Interface ⫺ Journal of Applied Linguistics ⫺ Tijdschrift voor toegepaste linguistı¨ek 15.1, 37⫺89
Large, Andrew (1985), The Artificial Language Movement. Oxford, New York: Blackwell, London: Deutsch
Blanke, Detlev (2003), “Interlinguistics and Esperanto studies. Paths to the Scholarly Literature”. Language Problems & Language Planning 27.2, 155⫺192
Lötzsch, Ronald (1991), “Tempus, Passiv und Aspektualität im Esperanto”. Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung 44, 402⫺414
Cherpillod, Andre´ (1995), Konciza Gramatiko de Volapuko: Gramat brefik Volapüka. Courgenat: Autor Couturat, Louis (1907), E´tude sur la de´rivation en Esperanto. Coulommiers: Brodard
Martinet, Andre´ (1991), “Sur quelques questions d’interlinguistique: Une interview de Franc¸ois Lo Jacomo et Detlev Blanke”. Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung 44, 675⫺687
Couturat, Louis & Leau, Le´opold (32001), Histoire de la langue Universelle ⫹ Les nouvelles langues internationales. Hildesheim, New York: Olms [Histoire: 11903, Les nouvelles: 11907]
Nuessel, Frank (2000), The Esperanto Language. New York, Ottawa, Toronto: LEGAS
Dr. Esperanto 1887 (⫽ Ludwig L. Zamenhof), Internationale Sprache: Vollständiges Lehrbuch und Vorrede. Warschau: Gebethner et Wolff Dulicˇenko, Aleksandr D. (1990), Mezˇdunarodnye vspomogatel’nye jazyki. Tallinn: Valgus Gledhill, Christopher (22000), The Grammar of Esperanto: A Corpus-Based Description. München: Lincom Europa [11998] Gode, Alexander (1953), “The Case for Interlingua”. The Scientific Monthly 77.8, 80⫺90 Gode, Alexander (21971), Interlingua-English: A dictionary of the International Language. New York: Ungar [11951] Gode, Alexander & Blair, Hugh E. (21971), Interlingua, a Grammar of the International Language. New York: Ungar [11951] Gopsill, Peter (1990), International Languages: A Matter for Interlingua. Sheffield: British Interlingua Society Haupenthal, Reinhard (1982), “Volapük-Bibliographie”. In: Schleyer, Johann Martin (21982, Hrsg.), Volapük. Die Weltsprache. Hildesheim: Olms [11880] Janton, Pierre (1993), Esperanto: Language, Literature, and Community (ed. by Humphrey Tonkin). New York: State University of New York Press Jong, Arie de (1931), Wörterbuch der Weltsprache: Vödabuk Volapüka. Leiden: Brill Kalocsay, Ka´lma´n & Waringhien, Gaston (51985), Plena analiza gramatiko de Esperanto. Rotterdam: Universala Esperanto Asocio [11935] La Zamenhofa (1961) ⫽ La Zamenhofa Esperanto: Simpozio pri ata/ita. La Laguna: Stafeto
Philippe, Benoıˆt (1991), Sprachwandel bei einer Plansprache am Beispiel des Esperanto. Konstanz: Hartung-Gorre Piron, Claude (1980), Esperanto: eine linguistische Standortbestimmung. A. d. Esperanto übers. v. Günther Becker. Saarbrücken: Saarl. EsperantoBund [Original in Esperanto: 1977] Sakaguchi, Alicja (1998), Interlinguistik. Gegenstand, Ziele, Aufgaben, Methoden. Frankfurt/M. usw.: Peter Lang (Duisburger Arbeiten zur Sprachund Kulturwissenschaft 36) Saussure, Rene´ de (1910), La construction logique des mots en Espe´ranto. Gene´ve: Kündig Schmidt, Johann (1933), Lehrbuch der Weltsprache Volapük für Deutschland und die deutschsprachigen Länder. Frankfurt/M.: Brönner Schubert, Klaus (1989), “Interlinguistics ⫺ its Aims, its Achievements, and its Place in Language Science”. In: Schubert (Hrsg.), Interlinguistics: Aspects of the Science of Planned Languages. Berlin, New York: Mouton de Gruyter (Trends in Linguistics. Studies and Monographs 42), 7⫺44 Schubert, Klaus (1997), “Ausdruckskraft und Regelmäßigkeit: Was Esperanto für automatische Übersetzung geeignet macht”. In: Tonkin (Hrsg.), 116⫺139 Stepanova, M[arija]. D. (1971), “Die ‘innere Valenz’ des Wortes und das Problem der linguistischen Wahrscheinlichkeit”. In: Helbig, Gerd (Hrsg.), Beiträge zur Valenztheorie. Halle: Niemeyer, 133⫺142 Szerdahelyi, Istva´n (1983), “Entwicklung des Zeichensystems einer internationalen Sprache: Esperanto”. In: Fodor, Istva´n & Hage´ge, Claude (Hrsg.), Language Reform: History and Future, Bd. III. Hamburg: Buske, 277⫺308
144. Plansprachen Tonkin, Humphrey (1997, Hrsg.), Esperanto, Interlinguistics, and Planned Language. Lanham usw.: Univ. Press of America (Papers on the Center for Research and Documentation on World Language Problems 5)
1573 Tonkin, Humphrey & Fettes, Mark (1996), Esperanto Studies: An Overview. Rotterdam: Universala Esperanto-Asocio (Esperanto Document 43A)
Detlev Blanke, Berlin (Deutschland)
XVII. Morphologischer Wandel I: Theoretische Probleme Morphological change I: Fundamental issues 145. Fundamental concepts 1. 2.
6. 7. 8. 9.
Introduction Processes of language acquisition and language use Morphological change proper Morphologization Morphological change due to extragrammatical forces Morphology and contact-induced change Conclusions Uncommon abbreviations References
1.
Introduction
3. 4. 5.
For the purpose of this article morphological change is defined as any change that results in a different morphological system of a language. There are a number of traditional concepts regarding morphological change, such as leveling, rule generalization, and, probably the most wideknown among these, analogy. These concepts are not very satisfying. On the one hand, they show a great deal of overlap; on the other hand, these terms do not shed light on the general forces underlying the different types of morphological change since they refer to the effects rather than to the processes of change as such. In the following discussion of the fundamental concepts regarding morphological change, the existing terminology will therefore not constitute the starting-point. Following Koefoed (1978), the view is taken that a language is never an active force in itself. The causes of change must be sought in processes of language acquisition and language use. Any change is the result of such a process acting or reacting upon a given ‘state of affairs’ in the language in question. By a language we do not only mean the grammar (or the linguistic knowledge of a native speaker), but also the output of that grammar. In many cases a language may show characteristics, either in its rule system or in the
system’s output, that may be said to ‘trigger’ a morphological change. Examples of such triggers are for instance a rule the conditions of which are little transparent, or a set of output strings that may easily be reanalyzed. As was pointed out above, such characteristics of a language are not considered the cause of a change. Rather, we consider them only a factor leading to change. An insightful discussion of morphological change, then, must show in which way the different processes of language acquisition and language use are triggered by specific linguistic situations, i. e. in which way the characteristics of a language and the processes of language acquisition and language use interact. The relevant processes of language acquisition and language use are discussed in section 2. As to the configurations that may give rise to morphological change, we will distinguish between (i) morphological change proper, i. e. changes originating in the morphological system and/or its output; (ii) morphologization, i. e. the incorporation of (mor)phonological and syntactic phenomena into morphology; and (iii) morphological change due to extragrammatical forces. These will be discussed in sections 3, 4 and 5, respectively. In section 6 the effects of language contact on morphological systems will be briefly dealt with, while 7 contains the conclusions.
2.
Processes of language acquisition and language use
With regard to morphological change, especially the types subsumed under the label of analogy, it has been a long tradition to assign a prominent role to the language learning child. Leveling, overgeneralization, reinterpretation are often exclusively ascribed to the child’s ‘language acquisition device’ or to
145. Fundamental concepts
some specific heuristic principles that form part of it. The child, having no direct access to the grammar of his parents’ language, looks for patterned relations within the set of utterances he hears and may arrive at a different grammar than that of his parents’ generation. In this way, language acquisition, and particularly the process of rule discovery, may lead to rule change. This picture, with its stereotypical distinction between the language learning child and the language using adult speaker, is too simplistic. No doubt, processes of rule discovery play a role in many types of morphological change (cf. 2.1). However, it is clearly wrong to associate all processes of rule discovery or rule creation exclusively with language acquisition by young children. Since these activities may also be rooted in language use (cf. 2.2), rule discovery and/or rule creation may also play a role in the speech of adult speakers. 2.1. Language acquisition, particularly rule discovery The first step in rule discovery is to assign a morphological structure to a given set of words. That is, to observe a formal-semantic relation between these words and another set of words, so that the first can be seen as in some way formed on the basis of the latter. The relation can then be interpreted as the product of a rule of word formation. The second step is to establish the precise nature of that rule and the conditions which determine its domain of application. In his search for morphological relations and word formation rules the learner/user is guided by what has been called the Humboldtian principle, that is a preference for one-to-one relations between forms and meanings. Both alternation in form and variation in meaning diminish the transparency of a rule and may trigger a change. As to the domain of application, the language learner has a preference for general, clear-cut conditions. Therefore, the transparency or ‘discoverability’ of a rule depends on three factors: the analyzability of the derived words, the regularity of its effects and the ‘straightforwardness’ of its conditions. If a rule is less than optimal in one or more of these respects, language acquisition can lead to a change. The rule in question may not only be ‘missed’ by the language learner, but the complex forms that derive from it may be reinterpreted, the rule may be altered in such a way that its effects become more regular,
1575 and the conditions delimiting the rule’s domain may be adjusted. In other words, rule discovery may lead to regularization. In 3.1 a number of such changes will be discussed. It is important to see, however, that not every formal-semantic relation between words implies the existence of a morphological ‘rule’. That is, not all relationships between words that language users may be aware of, are necessarily translated into rules with which new words can regularly be formed. Some of these relationships, for instance, bear upon categories that are unproductive. Besides, in order to learn a language it is not necessary to discover all the morphological relations between the existing words within the lexicon of that language. One can learn complex words as unstructured entities. As a matter of fact, some complex words must be learned in this way before the learner can start to discover morphological structure. This implies that not all morphological rules have the same function and status with regard to the linguistic knowledge of the speakers. Morphological rules may vary in necessity (from the viewpoint of language acquisition) and saliency (from the viewpoint of language use). Rules that apply very often, among them those that define the obligatory categories of a language, cannot be ‘missed’, of course. That is, they must be learned by all language learners in order to achieve an acceptable level of competence. Evidently, this applies to most inflectional rules. In addition, it may also apply to derivational morphology in that at least a passive knowledge is called for of those derivational rules that are frequently used to derive new words. However, at the other end of this scale one finds rules which only pertain to a handful, not too frequently used words, the occurrence of which is by and large restricted to the formal registers. As to the level of competence that a learner must acquire, these rules represent a kind of luxury; knowledge of these rules is not obligatory to obtain ‘full’ competence of the language involved. 2.2. Language use Within the rather heterogeneous domain of ‘language use’, three domains of processes or activities should be distinguished: (a) the psycho-physiological mechanisms of speech production and perception; (b) the expression of thoughts and intentions by individual speakers by means of the
1576
XVII. Morphologischer Wandel I: Theoretische Probleme
conventional language system that they have acquired; (c) the establishing of linguistic conventions, essentially a process of negotiation about social contracts. 2.2.1. Speech production and Speech perception In general, the mechanisms of speech production and perception (domain a) do not directly affect the morphological system of a language. Indirectly, however, they often have repercussions on the morphological system of a language, since these mechanisms may affect the phonological shape of words. This will be briefly discussed in 5. 2.2.2. ‘Looking for words’ Words, their mutual relations and their internal structure, are directly involved when people try to express their thoughts in the language they know (domain b). In expressing his thoughts, in ‘looking for words’ for what he wants to say, a speaker has essentially two sources at his disposal. First, the set of conventional names for concepts and situations (what Saussure 1916 called the tre´sor of the language community) and, second, the regular morphological system of his language, which by means of productive rules provides for new, non-conventional expressions. When a speaker, in expressing a concept, applies a productive rule of his language, this can be considered a manifestation of ‘rule-governed creativity’ (Chomsky 1964). The speaker does not exceed the existing possibilities of the conventional language system; his (knowledge of the) language provides for the expressions he needs, the language remains intact. However, as will be pointed out below, in looking for words a speaker may also exceed the limits of the regular system. That is, he may also turn to ‘rule-changing creativity’ (Chomsky 1964). With regard to the existing complex words, it is important to see that within the activity to express one’s thoughts two processes are at work. The first process leads to the loss of word structure due to conventionalization and lexicalization. This will be discussed in 2.2.3. The second process leads to the systematization of the lexical stock (i. e. the set of existing complex words). In formulating their thoughts speakers not only use their language, but constantly ‘reexamine’ it. As was already clearly observed by Von Humboldt, the lexical stock of lan-
guages should not be equated with “eine fertig daliegende Masse”, since it is “ein fortgehendes Erzeugniss und Wiedererzeugniss des wortbildenden Vermögens” (Von Humboldt 1836: 109⫺110). It is this activity of reflection on the set of existing expressions (the language as the speaker’s repertoire of linguistic expressions) that lies at the root of different types of reinterpretation (among which back formation, restructuring, folk-etymology, etc.) and adaptation. What these processes have in common is that they involve the systematization of the lexical stock. The inspection of the lexical stock of a language by its speakers does not only lead to systematization of the existing complex words, but this activity may also underlie ‘creative coining’. Words formed in this way exceed the limits of the regular system, meaning that they represent instances of rulechanging creativity. That is, the activity of expressing one’s thoughts and intentions does not only have repercussions on the existing words, but it may also lead to the formation of new words. Precisely because of the fact that the lexical stock of a language is not a given morphologically structured whole, the inspection of the lexical stock may result in all sorts of discoveries, which in their turn may underlie all kinds of ‘creative’ coinings. Put differently, the lexicon of a language also contains the germ of all kinds of creative coining. Thanks to this inspection of the lexical stock, unproductive relations may be revitalized, new relations between words may be discovered, a relation between not more than two existing words may be extrapolated to other words, a rule may be ‘inverted’ so that back formations are produced, and so on. One type of situation in which this activity is called for is the one in which the speaker’s knowledge of his language does not provide a conventional expression for the concept he wants to refer to. This situation is typical of children and that is why manifestations of creative coining are more frequent in the speech of children than in that of adults. However, it is certainly not so that children have a monopoly on rule-changing creativity. It may also be the case that there is a referential gap in relation to the concept in question, i. e. the language does not provide conventional means to express that concept. In that case, adult speakers may also take recourse to creative coining. Creative coining, finally, can also be invoked because of socio-stylistic
1577
145. Fundamental concepts
reasons. Evidently, the desire to be original can also be the motive to coin new expressions instead of using ready-made conventional names. Like in the preceding case, creative coining because of socio-stylistic reasons presupposes relatively high level of competence, meaning that it is particularly common in the language of adults. 2.2.3. The establishing of linguistic conventions As was pointed out in 2.2.2, looking for and using usual expressions (expressions that one has heard before and has remembered) may lead to lexicalization, through an intermediate step which we call ‘conventionalization’ (domain c). Conventionalization implies that a linguistic expression, although regularly formed by means of productive rules, acquires a kind of ‘currency’ within the usage of the community. More and more speakers use the expression as a ready-made form, instead of coining it (or another new word) to express the concept. The essential moment in this process of conventionalization is that, in a growing number of speech situations, the expression functions as a whole. That is, its internal structure, although for most speakers still recognizable, is no longer essential for its referential function. In this respect, the regular complex expression already functions in the same way as an unanalyzable simplex word. Once conventionalized, the expression becomes susceptible to processes affecting its systematic properties. Sound changes may obscure its internal structure and the relation with the original base, while shifts in meaning may obscure its semantic transparency (of these, the sound changes originate in domain a, the shifts in meaning in extralinguistic social reality). The result will be a lexicalized word with unpredictable phonological and/or semantic properties, which has to be learned separately. Eventually, lexicalized words may lose their internal structure completely. Crucial to this development is that it is the process of conventionalization that makes it possible that processes of speech production and perception do their ‘destructive’ work with regard to morphological structure. The establishing of linguistic conventions, however, is not only relevant to the conventionalization of individual complex words, but it is also relevant to the morphological system. The fact is, that morphological rules may have a specific social status as well. It is not uncommon, for instance, that a deriva-
tional rule is only used in a specific register of the language (cf. Art. 33). Besides, whether or not a morphological rule is considered productive is not a question of its systematic properties alone, but also of its conventional status. Changes in productivity or in the stylistic properties of morphological categories or rules must therefore also be considered morphological changes. These changes lead to different morphological system as well. 2.2.4. Summary In the preceding, two opposite processes were distinguished within the central activity to express one’s thoughts and intentions. First, the inspection of the lexical stock may result in the systematization of existing words. In addition, the same activity may underlie creative coining, too. Second, the use of complex words may also result in conventionalization. This process entails that the internal structure of the words in question is no longer essential for their referential functioning. Subsequently, this may lead to lexicalization: the conventionalized words develop unpredictable semantic and formal properties.
3.
Morphological change proper
Under this heading those types of morphological change are discussed which originate in a language’s morphology and/or the words it has produced. That is, in all instances of morphological change that will be treated in this section the starting point, i. e. the configuration that triggers the change, is to be situated either in the morphological rule system or in its output, i. e. the set of existing (‘conventional’) complex words. 3.1. Changes triggered by the rule system: regularization This class of morphological changes comprises those developments which somehow or other involve the ‘streamlining’ of the morphological rule system. As a general term for these changes we propose regularization. In many cases, regularization of a rule implies generalization, i. e. the extension of its domain of application. This, however, is not always the case since regularization may be more dramatic than the mere streamlining of the morphological rule system (cf. below). The most obvious cases of regularization are those in which a given rule becomes less restricted (more general) or more transpar-
1578
XVII. Morphologischer Wandel I: Theoretische Probleme
ent. Note that rule generalization may both relate to the elimination of accidental exceptions and to the elimination of systematic ‘gaps’ in the system. Generalization, finally, may also come into effect in the delimitation of rival categories. An example of the elimination of an accidental exception is the coming into existence of regular preterites such as scheerde next to irregular (‘strong’ schoor in Dutch (cf. scheren ‘to shave’). Irregular forms of strong verbs, so to say, run the risk of being ousted by their regular (‘weak’) counterparts. In many cases, these regular forms are not in common use, of course, but they may be formed any moment by the regular application of the rule of preterite formation. The elimination of a systematic gap can be illustrated by means of the female personal names in -e in Dutch. When these forms became a more or less regular pattern in the present-day language, they refrained from pluralization. In the course of time, however, they gradually came under the reach of the plural system, resulting in plurals such as studente-s ‘female students’. (Note, however, that this process has not yet been completed, meaning that there are still quite some female personal names in -e which do not readily underlie a plural form.) In the case of rival categories ⫺ particularly when more than one is productive ⫺ there can be question of gradual changes in the division of labour between the categories involved, leading to a greater transparency of the system. An example of such a shift is the pluralization of simplex words ending in -ing in modern Dutch, examples being haring ‘herring’, ketting ‘chain’, and koning ‘king’. In standard Dutch these words pluralize in -en, irrespective of the fact that the syllable in -ing does never bear main stress. In all kinds of non-prestigious varieties of Dutch (dialects, child language), and also in Afrikaans, however, these words have a plural in -s, which is completely in line with the general pattern of Dutch (and Afrikaans) plural formation, according to which words with non-final main stress generally have -s plurals. Another type of regularization involves the streamlining of the semantic properties of morphological rules. An interesting example of the latter development is presented by Dutch deverbal adjectives in -baar ‘-able’, which represent a relatively recent pattern in the language. During the period of their in-
troduction, adjectives in -baar were little transparent as far as their meaning was concerned. Specifically, they could have both active and passive readings. In the nineteenth century, however, the passive meaning became more and more dominant. In contemporary Dutch the productive rule invariably produces words with a passive meaning (cf. e. g. converteer-baar ‘able to be converted’). It is evident that the process of ‘semantic leveling’ that the category in -baar has undergone had consequences for the way in which coining in -baar is conditioned in the present-day language. In modern Dutch, adjectives in -baar can only be regularly derived from transitive verbs (or, more precisely, from verbs with a transitive reading), meaning that the rule’s domain has become more restricted. The above examples, particularly the latter one, already made clear that regularization is a phenomenon which is more comprehensive than the mere elimination of exceptions to a given rule. That is, regularization is more comprehensive than rule generalization. In the case of the semantic streamlining of the adjectives in -baar, there was no question of the elimination of gaps (neither accidental nor systematic) in the rule’s morphological domain. The following example, although of a completely different nature, illustrates the same fact: regularization is a much wider phenomenon than the mere elimination of gaps in the input of a rule. In Van Marle (1978) attention was drawn to the regularization of formations expressing ‘to be able/to be unable’ in the eastern dialects of Turkish and in closely related Azeri (or Azerbaijani). In modern Turkish the notion ‘to be able’ is expressed by means of compounding: the verb bil ‘to know’ is added to the required verb stem which as such is extended by the element -e/-a, e. g. gel-ir-im ‘come-pres-1.sg (I come)’ versus gel-e-bil-ir-im ‘come-øknow-pres-1.sg (I can come)’. A remarkable feature of modern Turkish is, that the negative counterparts of the above verb forms ⫺ i. e. the forms expressing ‘to be unable’ ⫺ are not formed in a parallel way, as is the case in other negative verbs. As can be inferred from the following example, the so-called impotential verb does not involve compounding at all but suffixation instead. The impotential verb is formed by means of adding the suffix -eme/-ama (to the verb stem: e. g. gel-emedi-m ‘come-impot-past-1.sg (I could not come)’ versus gel-di-m ‘come-past-1.sg (I
145. Fundamental concepts
have come)’. As a result, in the impotential verb the element bil- ‘to know’ is lacking. In closely related Azeri/Azerbaijani and in the adjacent Turkish dialects, this anomaly in the Turkish verb has been regularized (Caferogˇlu & Doerfer 1959: 304). In these varieties the impotential verb is no longer characterized by a unique pattern ⫺ i. e. a pattern completely independent of the potential verb ⫺ but it is formed in a way that is completely parallel to the potential verb, e. g. impotential gäl-ä-bil-mä-di-m ‘come-ø-can-neg1.sg (I have not been able to come)’ which, except for the negative suffix -mä, is similar to potential gäl-ä-bil-di-m ‘come-ø-can-past1.sg (I have been able to come)’ Crucial to this regularized pattern is, that the impotential verb, like the potential verb, involves the compounding of bil- ‘to know’ to the verb stem. In conclusion, regularization involves the replacement of an idiosyncrasy in the system by a form or pattern which directly conforms to the system. This process may vary from the straightforward elimination of accidental exceptions to much more dramatic developments leading to fundamental changes in the system (see also Koch 1995 for other types of regularization with far-reaching consequences). 3.2. Changes triggered by characteristics of existing words: systematization Many morphological changes have their origin in the constant, never ending inspection of lexical stock from the part of speakers of the language (cf. 2). This activity of the speakers may manifest itself in essentially two ways: reinterpretation and adaptation. The common characteristic of both processes is that they both involve the systematization of the lexical stock (however limited this may be). Eventually, reinterpretation may give rise to rule creation. 3.2.1. Reinterpretation Reinterpretation involves the assigment of a structure to existing words which involves a change with respect to the original situation. There are many different types of reinterpretation. First, we have to do with at least three possibilities: (i) a simplex word is assigned a morphological structure, (ii) a complex word is considered simplex, and (iii) a complex word is assigned a structure that differs from its original structure. Second, the process of reinterpretation may be governed
1579 by all kinds of additioinal factors which may bear upon phonology, semantics and ‘use’. Consequently, the notion of reinterpretation bears upon a rather heterogeneous set of phenomena, ranging from folk etymology to affix telescoping. A well-known instantiation of the first type of reinterpretation mentioned above, i. e. a simplex word is considered complex, is the phenomenon of back formation (Rückbildung). In Bloomfield (1933: 412) this phenomenon is illustrated with the example of the reinterpretation of Middle English cheris ‘cherry’ (from Old French cherise). The final s of cheris was associated with the formally identical plural ending, which resulted in the coming into existence of a new singular form cherry. In this new system, the form cherries ⫺ which was the original singular ⫺ functions as plural form. Similar examples are not hard to come by. In Dutch, singular forms ending in -en, so to say, run the risk of being reinterpreted as plural forms. The fact is, that the most frequently used plural ending in Dutch reads -en. In this way, cyclamen ‘id.’ became cyclaam, varken ‘pig’ became vark (in child language, also in Afrikaans), and molen ‘mill’ became mool (in American Dutch). The second type of reinterpretation distinguished above, involves the loss of internal structure of a complex word. Most instances of complex words losing their internal structure involve conventionalization/lexicalization rather than reinterpretation (see 2 and 5). However, reinterpretation, too, may result in loss of structure. An example is formed by words such as biels ‘(railroad) sleeper/tie’ or rails ‘rail’ in modern Dutch. For many speakers of the present-day language, these words are simplex. Diachronically, however, these words are plurals, be it plurals which did not join in with the general characteristics of Dutch plural formation. According to this system, monosyllabic words that are consonant-final take the plural ending -en, which means that regular plural forms are never monosyllabic. That is, in terms of the living structure of Dutch, the plurals biels and rails looked like singulars, and that is how they were reinterpreted. The third type of reinterpretation, i. e. the type which leads to a change in the original structure of complex words, can best be illustrated with the help of the phenomenon that is known as affix telescoping. In Paul (51920: 245) it was already pointed out that
1580
XVII. Morphologischer Wandel I: Theoretische Probleme
derivatives of the second degree may develop into derivatives of the first degree, in consequence of the fact that the original derivative of the second degree is directly related to the simplex base. This process results into the amalgamation of the two suffixes that are involved (Van Marle 1984; 1990 b; Haspelmath 1995). Evidently, the reinterpreted words may develop into a new derivational pattern, on the basis of which new words may be directly formed by means of the amalgamated suffix. This may be illustrated by means of Dutch female personal names in -er-es, e. g. dans-er-es ‘female dancer’. Words of this type are generally looked upon as derivatives in -es which have neutral personal names in -er (dans-er ‘dancer’) as their starting point. However, the possibility should not be excluded that female personal names of the dans-er-es type are considered to be directly derived from the verb dans-en ‘to dance’, the result of which is the amalgamation of -er and -es into -eres. In this way, the formations in -eres become identical to their productive rivals in -ster, by means of which deverbal female personal names are coined. Cf. roei-ster ‘female rower’ which has the verb roei-en ‘to row’ as its base. That the formations in -er-es are really reinterpreted as formations in -eres, is proven by the fact that we find words in -eres for which there is no ready base in -er at hand. Examples are posteres/postbod-eres ‘female postman’ and stripteas-eres ‘female stripper’ which can best be looked upon as derivatives in -eres which have post/postbode ‘postman’ and striptease ‘id.’ as their starting point. In the example just given, the reinterpretation of a derivative of the second degree as a derivative of the first degree was, to a certain extent, determined by the morphological system. In Van Marle & Koefoed (1980) instances of this type of reinterpretation were characterized as ‘short cuts’ in the morphological system. A derivative of the second degree is directly related to the simplex base, which results in the coming into existence of a derivative of the first degree. Complex words may also fall victim to processes of reinterpretation of a completely different kind. They may also be reinterpreted on the basis of their ‘phonological’ properties, i. e. on the basis of the phoneme sequences of which they consist. This type of reinterpretation can best be characterized as a process of resegmentation. The prime force behind this process does not involve the
form-meaning characteristics of the words in question, but the classification (and concomitant segmentation) of these words on the basis of their shared phonological properties. Consider the following personal names in Dutch: leugen-aar ‘liar’, molen-aar ‘miller’ and Nijmegen-aar ‘inhabitant of Nijmegen’. In Dutch neutral personal names take -aar when the base ends (among other things) in en. This pattern has given rise to a new suffix -enaar, which we meet in e. g. bult-enaar ‘hunch back’ and Delft-enaar ‘inhabitant of Delft’, which are based on bult ‘hunch’ and Delft respectively. The coming into existence of the new suffix -enaar is interesting in that it makes clear that complex words may be classified and segmented on the basis of their phonological properties while their morphological properties are glossed over. A salient characteristic of Dutch personal names, apparently, is that they often share the final sequence enaar. Originally, this ‘surface characteristic’ of the words in question had no morphological status, -aar being the suffix and en only being the final part of the simplex stem (without any meaning whatsoever). However, this phonological property of the personal names was assigned morphological importance, meaning that enaar became a suffix. This is evidenced by the fact that new words were coined by means of ‘resegmented’ -enaar. The rise of -enaar in Dutch is also interesting in that it makes clear that this phenomenon of resegmentation need not be governed by syllable boundaries, as is sometimes suggested in the literature. The syllable boundary, to be sure, never lies before the schwa of -enaar, the syllable boundary always lies before the n, i. e. ‘within’ the suffix. That is, as far as prosody is concerned, -enaar is no unit. This is not to say that syllable structure is never relevant. A classic example of resegmentation which is generally considered to be governed by syllable structure, is the coming into existence of the suffix -ling, which used to be -ing. In formations such as Karol-ing the syllable boundary lies before or ‘within’ the l preceding -ing. It has often been suggested that this fact may have contributed to the coming into existence of the suffix -ling. Evidently, in the case of -ling the position of the syllable boundary may have played a part. However, Dutch -enaar makes clear that syllable structure is no necessary factor in the phonologically governed resegmentation of complex words, in that the initial
145. Fundamental concepts
schwa of -enaar exceptionlessly forms part of another syllable than remaining naar. Still, the phonologically determined resegmentation of words such as molenaar ‘miller’ has resulted in a new suffix -enaar. To conclude, Dutch -enaar illustrates that the phonological classification and segmentation of complex words may be governed by phonological similarity alone. As was pointed out above, it may well be that there are instances of phonologically determined resegmentation in which syllable structure plays a part, but this is certainly no general condition for this process to take place. 3.2.2. Adaptation Adaptation is the effect of paradigmatic pressure. Individual words may undergo a change which makes them more ‘fit’ into the morphological patterns of the language as a whole. Like reinterpretation, morphological adaptation may adopt quite different shapes. The common trait of adaptive processes is that they are primarily triggered by the surface forms and that they cannot be linked to the underlying rule system in a direct way. (See 3.2.1 where the relevance of ‘surface characteristics’ came up for discussion as well.) A well-known type of adaptation is hypercharacterization (Van Marle 1978). This kind of morphological change represents the formal adaptation of a given word to the regular patterns that it is semantically related to, but not formally. Consider e. g. an irregular (‘strong’) preterite such as liep (cf. lopen ‘to walk’). This irregular preterite lacks the formal characteristics of regular preterites which take -te or -de (depending on the voice quality of the stem-final segments). The adaptive change that such forms may undergo, involves the ‘superfluous’ addition of -de/-te to the irregular preterites, resulting in ‘hypercharacterized’ forms such as liep-te. The latter forms are sometimes referred to as ‘double preterites’, but this characterization is somewhat misleading in that these adapted preterites are normal preterites as far as their meaning is concerned. Hypercharacterization is not only prominent in inflection ⫺ cf. also ‘double plurals’ such as Dutch musea’s ‘museums’ besides ‘correct’ musea ⫺ but also in derivation. Throughout the history of Dutch, personal names which are not formed by means of -er (-der), have been assigned this suffix superfluously. Cf. e. g. dragon-der ‘dragoon’ and medicijn-er ‘doctor’ (from French
1581 dragon and me´decin respectively) which represent older examples, while aio-er, uhd-er and postdoc-er for aio (an acronym based on assistent in opleiding ‘research assistant’), uhd (an acronym based on universitair hoofddocent ‘university lecturer’) and post-doc ‘idem’ represent recent adaptations. Crucial to morphological adaptation is that it cannot be explained in terms of the rule system proper. Morphological adaptation can only be understood in terms of the ‘output’ of this system. That is, regular preterites are formed by means of -de/-te. Preterites which are not formed in this way, are assigned the formal characteristics of their regular counterparts all the same. Morphological adaptation, then, is a process that is triggered by the canonical forms that the rule system produces, but not by the rule system as such. 3.2.3. Summary In general, systematization (i. e. reinterpretation and adaption) contribute to the coherence of the lexical stock of a language. In addition, reinterpretation may result in new affixes which may become the formal exponent of new productive patterns (Dutch -enaar). In that way reinterpretation may ⫺ indirectly ⫺ contribute to the ‘lexical enrichment’ of a language. 3.3. Changes triggered by characteristics of inflectional paradigms As we have seen in 3.2, morphological change need by no means be triggered by the rule system, since in many cases morphological change is rooted in properties of the output. Particularly morphological adaptation makes this clear (see 3.2.2). Another phenomenon which illustrates this point, is paradigm leveling (which, in contrast to morphological adaptation, is well-known and frequently discussed). The forms constituting an inflectional paradigm either tend to give up formal differences that do not correspond with differences in meaning (leveling), or tend to generalize them throughout the paradigm (extension). Well-known, for instance, is the elimination of the vowel alternation between singular and plural in the preterite of Dutch ‘strong’ verbs such as binden ‘to bind’. Originally, the singular had a as stem vowel (bant) vis-a-vis the o in the plural (bonden). This difference between singular and plural has been eliminated (leveled), resulting in modern Dutch bont (spelled bond) -bonden.
1582
XVII. Morphologischer Wandel I: Theoretische Probleme
Crucial to changes of this kind is that they, like the instances of adaptation (cf. 3.2.2), bear upon the surface characteristics of the forms in question. However systematic the above discussed vowel alternation may have been, the difference at the surface between the singular and the plural constituted the trigger for the change in question. This change involves the elimination of the meaningless difference in form between the singular and the plural. Consequently, paradigm leveling represents an adaptive process: the singular forms (with a) have adapted their form to the plural forms (with o), a development which results in new singulars (with o).
4.
Morphologization
Generally speaking, morphologization involves the incorporation of originally (mor)phonological or syntactic phenomena into morphology. On the whole, the incorporation of extra-morphological phenomena into the morphological system involves complex changes in which all kinds of factors interact. Morphologization, finally, forms part of the more comprehensive phenomenon of grammaticalization. 4.1. The incorporation of (mor)phonological phenomena The classic example of the morphologization of a non-morphological phenomenon is the occurrence of umlaut in cases such as Bruder ‘brother’ versus Brüder ‘brothers’ or Vater ‘father’ versus Väter ‘fathers’. As is wellknown, the history of umlaut represents a development from a phonetic to a phonological, and from a phonological to a morphonological process. In present-day German, umlaut is predominantly morphonological in nature, since its occurrence is determined by morphology. In Baum ‘tree’ ⫺ Bäum-e ‘trees’ or Mann ‘man’ ⫺ Männ-er ‘men’, for instance, umlaut is concomitant to pluralization by means of -e and -er, respectively. However, in the case of Bruder ‘brother’ ⫺ Brüder ‘brothers’ or Vater ‘father’ ⫺ Väter ‘fathers’, things are different, in that in these cases umlaut is the only bearer of the notion ‘plural’. That is, in these cases umlaut is no longer concomitant to pluralization. In the words in question umlaut has become plural marker itself, meaning that it has developed into a purely morphological phenomenon. The morphologization of originally morphonological phenomenona is not particu-
larly rare. Morphonological phenomena are concomitant, i. e. they are associated with a particular morphological process. If the overt marker of that process ⫺ i. e. the suffix ⫺ disappears, the concomitant phenomenon may take over the function of overt morphological marker. This also happened to plural formation in certain varieties of Spanish, particularly in the dialect of Granada. In certain nouns the notion of plurality is exclusively expressed by vowel laxing. Cf. e. g. [kabe´ha] ‘head’ versus [ka¸bee¸´ ha¸] ‘heads’ or [see¸´ lva] ‘forest’ versus [see¸´ lva¸] ‘forests’ (all vowels that are marked are lax). In other nouns the notion of plural is expressed by a combination or vowel laxing and aspiration. Cf. e. g. [lo´bo] ‘wolf’ versus [lo o¸´ bo¸h] ‘wolves’ or [pı´so] ‘floor’ h ´ versus [pı¸ıso¸ ] ‘floors’ (all the above data stem from Alonso et al. (1950) and are cited in Hooper (1976: 36)). Like in German, in this variety of Spanish the plural ending (-s) has disappeared. In Spanish this happened via a process of aspiration, which in its turn functioned as the phonetic trigger of vowel laxing. Exactly like in German, the phonological phenomena which were originally concomitant to pluralization have developed into the markers of plurality when the suffix disappeared. In the case of Spanish this implies that in the words in question vowel laxing and aspiration have lost their concomitant morphonological character and have developed into true morphological phenomena. Morphologization of originally non-morphological phenomena may come into existence in another way as well. Consider the morphologization of tone in the dialects spoken in the Dutch and Belgian Provinces of Limburg. As is pointed out in Goossens (1964), apocope of final schwa in bisyllabic words may lead to a tone shift in the first syllable. Specifically, the original dragging tones is replaced by a falling tone. In the dialects in question, this phenomenon cannot only be observed in certain plural forms, but also in adjectival declension and in a few gerunds. In some nouns a difference in tone is the only marker of plurality, e. g. ste: n * ‘stone’ versus ste:*n ‘stones’ where the difference in tone is the sole marker of plurality (*-subscript indicates a dragging tone, whereas *-superscript indicates a falling tone). The same difference can be found in the verb system. Verbs like du: n ‘to do’ and * zi: n ‘to see’ have a dragging tone in the in* finitive, whereas they have a falling tone in the gerund (the form used after the preposi-
145. Fundamental concepts
tion te), cf. te du*n and te zi:*n (in all cases Goossens’ notation has been used). Originally, these gerunds ended in -e. When this ending got lost, there was a tone shift, meaning that the new monosyllabic gerunds got a falling tone. Goossens suggests, then, that in these examples it is not the case that an originally concomitant phenomenon has taken over the primary function, but that an originally non-morphological phenomenon is invoked to signal a morphological distinction, thus becoming a morphological phenomenon itself. 4.2. From word to affix It is not only (mor)phonological phenomena which can be incorporated into morphology. Syntactic phenomena may be incorporated into morphology as well. This latter development typically results in the rise of new affixes (as was pointed out in 3.2.1, reinterpretation may also result in new affixes). The prototypical form that this development adopts is that words lose their individual status and turn into ‘affix-like’ elements and eventually become true affixes. However, this is not the only way in which morphologization may give rise to new affixes. Another way in which new suffixes may arise is the gradual transition of the second constituent of compounds into suffixes. This development, too, may be considered the result of morphologization. An interesting example of this gradual shift of a word into an affix is provided by the Turkish postposition ile ‘with’ which has developed into a suffix (but which still functions as a postposition as well). This change in status is reflected by the fact that the new suffix -le (after consonants)/-yle (after vowels) is, like all regular suffixes, subject to vowel harmony (but see below), whereas postpositions are not. So, in modern Turkish we have both the original construction vapur ile ‘boat with1 (by boat)’ with bisyllabic ile and the new suffixed variant vapur-la ‘boatwith2 (by boat)’ (with reduced, i. e. monosyllabic -la). The fact that developments like these are gradual by definition is evidenced by the fact that there are cases where the original postposition appears already as an affix, but where it is ‘not yet’ subject to vowel harmony. (The affix-status of -yle/-le is clear from its condensed form: it has lost its initial vowel, i. e. i, meaning that it has lost its bisyllabic character.) This is the case when it is used after a third person possessive suffix,
1583 e. g. karI-sI-yle ‘wife-poss:3.sg-with1 (with his wife)’ or omuz-u-yle ‘shoulder-poss:3.sg-with1 (with his shoulder)’. Note, however, that the latter phenomenon ⫺ after a third person possessive suffix, -yle is invariable ⫺ is a characteristic of, as Lewis (1967: 87) calls it, ‘educated pronunciation’ which is neglected by many speakers of modern Turkish. In the language of these speakers, yle is no longer invariable after the third person possessive suffix as is evidenced by e. g. karI-sI-yla and omuz-u-yla where it has become subject to vowel harmony. Evidently, the language of the latter speakers reflects a stage that can best be considered one step further into the direction of ‘true suffix’ that the postposition ile has taken. The development of ile not only illustrates the formal change from word to affix, but also the incorporation of syntactic phenomena into the morphological system: a function word gradually develops into a case suffix (5.1). The change from word to affix is also prominent in a completely different context, viz. in compounds. The second constituent of compounds may develop into a suffix-like element fairly easily. This phenomenon underlies the fact that in many languages suffixes can be found that have word-like properties. In Dutch there is a whole series of words that end in -boer. Cf. e. g. groente-boer ‘greengrocer’, melk-boer ‘milkman’, patat-boer ‘seller of French fries’, vis-boer ‘fishmonger’, vodden-boer ‘old-clothes-man’, etc. This element is formally identical to the word boer ‘farmer’. However, in many of the compounds with boer as a second constituent, this element cannot simply be equated with the independently occurring and formally identical word. That is, in these cases there is no direct semantic link between the second constituent -boer in compounds and the word boer ‘farmer’. What has happened, is that the word boer in compounds such as groenteboer and melkboer has developed into a suffix-like element, meaning ‘seller of X’. This ‘seller of’ interpretation has gained a certain popularity as far as the coining of new words is concerned. In these newly coined words (e. g. patat-boer ‘seller of French fries’, sigarenboer ‘seller of cigars’, etc.) there need not be question of goods produced by farmers any longer, which means that the link between the suffix-like element -boer in the compounds in question and the word boer ‘farmer’ has become weaker and weaker.
1584
XVII. Morphologischer Wandel I: Theoretische Probleme
Both in the case of Turkish ile vis-a-vis -yle/-le and in the case of Dutch boer vis-avis -boer, the original word and the newly arisen affix still form part of the language. This need not be the case, however. Not infrequently, the original word has disappeared from the language. The Dutch suffix -loos, for example, has no doubt developed from the adjective loos ‘deprived of’. However, this word no longer forms part of modern Dutch, its sole remnant (apart from some frozen expressions) is the word-like suffix -loos. Complex adjectives in -loos still display a compound-like character. This is, among other things, evidenced by its phonological characteristics: -loos is still a phonological word. As far as grammar is concerned, however, -loos is a suffix, since it does not occur independently. In other cases, the affix has undergone such drastic changes that the link with the word it has developed from, is no longer clear. In Dutch, many prefixes have also originated from independent words. This is for instance the case with the prefixes beand ver-, both with schwa. However, for the speakers of present-day Dutch, these prefixes are no longer related to the words they have developed from, viz. bi and for respectively (in the modern language bij ‘with’ and voor ‘for’). Morphologization, finally, is a very important process ⫺ particularly when it is followed by (or interacts with) regularization ⫺ because of its potential drastic character. Morphologization is the force underlying dramatic developments such as the development of postpositions into case endings (see above) or the development of auxiliaries into tense markers. Particularly in the case that such changes lead to large-scale regularization, they may result in morphological systems that are fundamentally different from those that preceded them. It has been suggested that all affixes stem from words. The correctness of this farreaching hypothesis concerning the origin of affixes is doubtful, however (see also 3.2.1). What has been overlooked, among other things, is that phonological phenomena may be assigned a morphological interpretation as well (also 4.1). One way in which this incorporation of purely phonological (i. e. meaningless) elements into morphology may come about, is the development of ‘phonological chunks’ into suffixes. This phenomenon is known as secretion (Jespersen 1922: 384⫺ 386). A particularly interesting example of
this development is presented in Jespersen (1939). In that study it is pointed out that the variation in English between forms with and without final n sometimes resulted in the addition of n/en to words, without changing their meaning. In this way, Jespersen points out, forms such as hidden (next to hid) and olden (next to old) came into being. Verbs, too, were candidates for extension by means of en. This extension is particularly interesting in the case of verbs derived from adjectives. Originally, these verbs were formally identical to their adjectival base, i. e. they were formed by means of conversion. In this stage, there was no formal difference between the adjective hard and the verb hard which is formed on the basis of the adjective. These deadjectival verbs could undergo en extension as well and in this way verbs such as darken, deepen, harden, ripen and sharpen came into being. Originally, the element en in these latter verbs was not linked to any grammatical and/or semantic value, meaning that it is completely comparable to en in e. g. often (from older oft) (Jespersen 1939). However this may be, the extended deadjectival verbs became subject to reanalysis. These verbs were no longer regarded as extended verbs, but they were considered to be directly derived from the adjectival base by means of the suffix (!) -en. This new status of the verbs in -en is evidenced by the fact that the newly established derivational pattern led to new coinings. An example is the verb biggen. There has never been a verb big, which means that the verb biggen cannot be conceived of as an extended version of it. Put differently, the verb biggen can only be considered to be directly derived from the adjective big, meaning that the element en has developed into a derivational suffix -en. The same holds for verbs such as cheepen, coarsen, laten, louden, and ruden. These verbs make clear that it is not necessarily the case that all ‘bound forms’ (i. e. affixes) stem from originally free forms. As said, according to Jespersen (1939), the suffix -en in the de-adjectival verbs in question does not stem from a free form: en has never been a word. In older English en was only the phonological extension of the original verb. This makes clear that purely phonological entities may also be incorporated into morphology, in this case in consequence of the reinterpretation of the verbs they are part of. The net result of this reinterpretation is that a purely phonological sequence develops into a suffix.
145. Fundamental concepts
5.
Morphological changes due to extragrammatical forces
In this section two distinct sets of phenomena will be dealt with. In both cases the phenomena in question relate to different aspects of language use (see 2.2). First of all, changes will be discussed that relate to the indirect repercussions of speech production and speech perception on the morphological system. Subsequently, it is the effects of conventionalization that will be discussed. Clearly, these are the effects of, respectively, the first and the third domain of language use that were discerned in 2.2. (The relevance of the second domain of language use (‘looking for words’) for morphological change was discussed in section 2.2.2, where it was associated with the never-ending inspection of the lexical stock by the speakers of a language (also 3.2)). 5.1. The effects of speech production and speech perception As was pointed out in section 2.2, the processes of speech production and speech perception particularly affect the phonological form of words. This means that they affect morphology only indirectly. In general, it is particularly the final parts of words that fall victim to these processes, the net result of which is the weakening and eventual loss of word-final elements. Since these latter elements are often (part of) suffixes, the net result is that these phonological forces are particularly devastating with respect to suffixes. In the case of complex words it is not only the word form that may be affected by these forces, but the grammatical information associated with the affected suffixes may get lost as well. Particularly in the case that suffixes do not bear stress, they easily fall victim to phonological weakening and loss. It has been suggested that the phonological factors in question may become operative more easily when the semantic content of the endings is opaque. Case systems, for instance, are considered to resist phonological weakening better when each case has a transparent semantic content and, consequently, has a neatly defined position within the case system as a whole. However this may be, even cases whose semantic characteristics are clear, may fall victim to phonological erosion when they do not bear stress. One of the reasons why the case system of a language like Turkish is so strong, is that the case endings in Turkish
1585 are stressed. As a matter of fact, the Turkish case system is even expanding: the above discussed suffix -yle/-le (which stems from the postposition ile, see 4.2) clearly develops into the direction of a new case suffix (viz. that of instrumental/comitative). 5.2. The effects of conventionalization As will be clear by now, the processes of conventionalization and lexicalization first of all affect individual words. Once formed, complex words may become part and parcel of the conventional lexical stock, which means that their internal morphological properties are no longer relevant to their functioning. As was pointed out in section 2, as to their functioning conventionalized complex words are similar to simplex words. In consequence of the ‘irrelevance’ of their structural properties, conventionalized complex words are an easy target for phonological and semantic forces affecting and obscuring their internal structure. However, the conventionalization of complex words may in the long run also affect the morphological system. Morphological processes may lose their productivity when their semantic properties get obscured. This may be the case when too many of the words that they have produced have become semantically opaque. This happened for instance with the Dutch deverbal adjectives in -lijk (e. g. wenselijk ‘desirable’, cf. wens-en ‘to desire’). In Van Marle (1988) it is suggested that this once very productive pattern has lost its productivity in consequence of the semantic diversity that the words in -lijk exhibited. Another type of change which should be discussed in this section on conventionalization, are changes in the register-dependent character of morphological processes. As was pointed out in Art. 33, many morphological processes are conventionally linked to a specific (set of) register(s). The category of deverbal adjectives in -baar ‘-able’, for instance, is a typical representative of a process with a clear High-character (Van Marle 1990 a), and so is the category of deverbal abstract nouns in -ing (Van Haeringen 1971). In Art. 33 it was pointed out, too, that categories with a High-character tend to have morphological domains which are clearer defined (and, not infrequently, more restricted) than categories with a Low-character. This implies that changes in the register-dependent character of morphological processes may have consequences for the range of words which may
1586
XVII. Morphologischer Wandel I: Theoretische Probleme
potentially serve as input for the process in question. In Van Marle (1990 a) it is pointed out that this is indeed the case. The category in -baar, for instance, which is the successor of the category in -lijk, has a much more restricted domain than its predecessor. In contrast to the category in -lijk, the category in -baar is a typical exponent of the High-registers. In Afrikaans, the daughter language of Dutch which ‘copied’ Dutch derivational morphology during its standardization, the category in -baar lost its High-character, the consequence of which is that in Afrikaans deverbal adjectives in -baar can be formed much more freely than in Dutch (Van Marle 1996). Similarly, there are dialects in Dutch where coining in -ing is remarkably popular. In these dialects, and different from the standard language, the category in -ing has no clear High-character, a fact which directly fits in with the observation that in general there is no strong trend in dialects to distinguish between distinct registers. Not surprisingly, in these dialects coining in -ing runs much more smoothly than in the standard language. A quite uncommon type of morphological change which also belongs to this category, finally, is the development of a morphological process into the direction of a register marker. This is what happened to Afrikaans diminutive formation. Apart from normal diminutives, Afrikaans has a whole series of atypical diminutives, the occurrence of which is restricted to ‘affective usage’ (see Art. 33). Interesting about these a-typical diminutives is, that they may be based on pronouns, prepositions and verbs and that the attachment of the diminutive suffix is not associated with a change in word class. Specifically, if an a-typical diminutive is formed on the basis of a pronoun or verb, the resulting word is still a pronoun or verb. Cf. hom ‘him’ versus hompie, where the latter diminutivized form is still a pronoun. Typical of these atypical diminutives in Afrikaans is, that they have lost their referential function, in consequence of which the development underlying their coming into existence can best be considered an instance of ‘demorphologization’.
6.
Morphology and contact-induced change
In our above discussion of morphological change we have concentrated on monolingual situations. All forces discussed operate
language-internally, i. e. in a situation in which a given language is transferred from one generation to another without there being other languages spoken in the same community at the same time. That is, contact-induced change was left out of consideration completely. Evidently, it is beyond the scope of this article to discuss the effect of language contact on morphological systems in detail. Language contact may adopt various shapes, and the effects on the morphology of a language may vary accordingly. In this connection we confine ourselves to some brief remarks. 6.1. Borrowing In 4 we discussed the fact that morphological systems may expand due to the incorporation of phonological and syntactic phenomena. Another well-known type of morphological expansion involves language contact, i. e. the incorporation of loan affixes (cf. also Art. 152). Note, however, that as a rule this type of expansion is ⫺ as far as morphology is concerned ⫺ only an indirect effect of the bilingual situation in question. More often than not, these so-called loan affixes have not been directly borrowed at all. In most cases it is not the affixes that are borrowed but the complex words that these affixes are part of. Put differently, in most cases the borrowing of words is a predominantly (or even exclusively) lexical process which results in an expansion of the lexicon in the first place. However, since the lexical stock ⫺ native or nonnative ⫺ is subjected to a constant process of inspection from the part of the speakers of the language (see 2), the affixes figuring in complex words may be recognized and subsequently ‘isolated’ from the borrowed words in question. The implication of this is that in many cases the occurrence of loan affixes is no directly contact-induced phenomenon and neither is their ‘recognition’. What is contactinduced, is the lexical process of borrowing (lexical expansion). In cases of long-term cultural contact, borrowing may result in a lexical stock and a morphological system in which two strata can be distinguished, a native and a nonnative one. This is the synchronic reflex of the contact-induced process of borrowing. 6.2. Morphological erosion In 5.1 it was discussed that morphological systems may experience erosion, since the processes of speech production and speech
145. Fundamental concepts
perception often affect the phonological form of words. As a rule, the erosion of morphological systems due to performance factors represents a process with a gradual character. However, morphological erosion may also have a contact-induced character, since certain types of language contact may lead to morphological erosion as well. This loss of morphological distinctions may affect both the (original) first language of the speakers due to language shift and language loss, and the (original) second language due to imperfect learning. Note that the decay of a morphological system due to language contact is much more abrupt than the gradual erosion of a morphological system due to the effects of speech production and speech perception. In American Dutch, for instance, both the plural suffixes of nouns and the verb endings are spontaneously dropped, simply because of the fact that the present-day speakers of American Dutch do no longer have full command of their language (Smits 1996). Evidently, the effects of language contact on morphological systems may be quite dramatic, depending among other things on the intensity of the contact situation. Morphology, particularly inflection, appears to be highly vulnerable to language contact, since one of the most radical contact-induced developments typically affects this part of a language. In cases of intense contact, (the greater part of) morphological systems may fall victim to large-scale erosion, meaning that large parts of the original system are lost more or less completely.
7.
Conclusions
In the preceding sections an attempt has been made to arrive at a classification of morphological changes based on the idea than any morphological change is the result of a certain type of process acting upon a specific domain or component of a language. By way of conclusion the various types of morphological change discussed so far will be briefly characterized in terms of the language domains in question and the processesd of language acquisition and use that are involved. The first two types of morphological change that were discussed are regularization and systematization. These two types of change originate in the morphological system itself, i. e. in the set of existing words, in the
1587 relations between them, or in the rules by means of which they are formed. Regularization bears upon those changes that are triggered by idiosyncracies in the rule system. As its name indicates, this process is directed towards more regular and transparent morphological rules. This may result in the well-known phenomenon of rule generalization, i. e. the increase of a rule’s domain by the elimination of its exceptions. On the other hand, the effect may be the restriction of a rule’s domain as well, as a consequence of the fact that its conditions may also become more strictly delimited. Systematization comes in two types: reinterpretation and adaptation. Reinterpretation comprises a number of changes in the morphological analysis of existing words. The factor at work is the speaker’s tendency to look for transparent and clearly recognizable structure and segmentation of the words of his or her language. Well-known subtypes of reinterpretation are back formation, restructuring, folk etymology, and the isolation of affixes out of complex loan words. (In the latter case this may lead to the coming into existence of a new productive morphological rule.) Adaptation refers to the morphological adjustments of existing words, due to the paradigmatic pressure of semantically similar but formally different words. The best known subtypes are leveling and hypercharacterization. In the first case, a word (or class of words) loses its (their) characteristics which makes it (them) deviant from semantically similar words, in the second case a word (or class of words) acquires the formal characteristics of the category whose meaning it shares. As said, both regularization and systematization have their roots in the morphological system. In terms of the processes involved, they are different, of course. Regularization is primarily associated with language acquisition (particularly with rule discovery), whereas systematization is primarily considered to be the outcome of language use (viz. the activity of ‘looking for words’). The view that systematization is a manifestation of language use is further corroborated by the fact that the processes of reinterpretation and adaptation presuppose that the relevant systematic properties of the language have already been acquired. On a deeper level, however, regularization and systematization are
1588
XVII. Morphologischer Wandel I: Theoretische Probleme
related processes. Both processes are the result of essentially the same activity: the ongoing search by language learners/language speakers for structure, patterns and regularity within the lexical stock and its underlying system. In addition, both processes tend to have a ‘streamlining effect’ on the domain they bear upon: regularization involves the streamlining of the rule system, whereas systematization involves the streamlining of the lexical stock. The second group of changes that were discussed can be subsumed under the heading of morphologization. These changes do not originate in the morphological system of a language itself; rather they arise as reactions upon originally (mor)phonological or syntactic configurations in the output of a language (which, in their turn, may have come into existence due to the efficacy of performance factors). The essential process is the same as in the first group of morphological changes: the search for transparent morphological patterns. A common type of morphologization is the incorporation of (mor)phonological distinctions into morphology, which involves, among other things, the reinterpretation of a meaningless formal characteristic of a given set of words as meaningful. A second type of morphologization is the transition of words to affixes. Particularly well-known is the development of function words into affixes. Another change that may be mentioned here is the development of ‘part of a compound’ into suffix. Evidently, due to the latter developments new affixes come into existence. Note, however, that this does not necessarily imply that all affixes derive from independent words. Phonological chunks in the output may be morphologized as well, which means that new affixes may also have a phonological origin. In general, morphologization results in the growth of morphological systems. As was pointed out above, however, this does not imply that the forces underlying morphologization are fundamentally different from those in which morphological change proper is rooted. On the contrary, both morphological change proper and morphologization are largely rooted in the same source: the speaker’s disposition to look for meaningful and transparent word structure. As a matter of fact, this seems to be the most fundamental force in morphology, in that it creates, maintains, extends and recreates morphological systems.
The remaining types of morphological change are due to performance factors. In general, these forces lead to the loss of morphological structure. First, the mechanisms of speech production and speech perception often bring about the loss of morphologically relevant distinctions and finally to deflection. Second, spontaneous use of language in concrete situations easily leads to the conventionalization of complex words, a phenomenon which opens the way to lexicalization. In its turn, lexicalization may result in the loss of productivity of morphological rules, due to the fact that a rule may lose its transparency when too many of the words it has produced have become lexicalized and have become opaque, either formally or semantically. Finally, a type of change which is also due to performance factors but which does not necessarily lead to the loss of morphological structure, are changes in the conventional status of morphological rules. Since many morphological categories, particularly categories belonging to derivational morphology, are to a certain extent sensitive to stylistic conditions, various social and cultural processes may lead to a change in their conventional status. Such changes may lead both to the increase and the decrease of the morphological potential of these rules, depending on the direction that the change in conventional status has taken. By way of conclusion, a brief remark on the overall nature of morphological change. More often than not, morphological change displays a little mechanistic nature. This is very clear in the case of systematization, but even in the case of regularization ⫺ i. e. in the case of morphological change proper! ⫺ morphological change may be remarkably idiosyncratic. This non-mechanistic and often unpredictable character of morphological change is due to two factors. First, many types of morphological change do not react upon characteristics of the morphological system as such, but upon properties of individual words, specific groups of words, or output strings. Second, the main force in morphological change itself is not a mechanistic but rather a creative process, stemming from the speaker’s aptness to assign ⫺ i. e. to recognize or to create ⫺ internal structure to words.
8.
Uncommon abbreviations
impot impotential
145. Fundamental concepts
9.
References
Alonso, Da´maso & Vicente, Alonso Zamora & Canellada, Marı´a Josefa (1950), “Vocales Andaluzas”. Nueva Revista de Filologı´a Hispa´nica 4, 209⫺230 Bloomfield, Leonard (1933), Language. London: Allen & Unwin Caferogˇlu, Ahmet & Doerfer, Gerhard (1959), “Das Aserbeidschanische”. In: Deny, Jean & Gronbech, Karel & Scheel, Helmuth & Tongan, Zeki Velidi (eds.), Philologiae Turcicae Fundamenta. Wiesbaden: Steiner, 280⫺307 Chomsky, Noam (1964), Current Issues in Linguistic Theory. The Hague etc.: Mouton Goossens, Jan (1964), “Een tweede Limburgs gerundium”. Taal en Tongval 16, 103⫺106 Haeringen, Coenraad Bernardus van (1971), “Het achtervoegsel -ing: mogelijkheden en beperkingen”. De Nieuwe Taalgids 64, 449⫺468 Haspelmath, Martin (1995), “The Growth of Affixes in Morphological Reanalysis”. In: Booij, Geert & Marle, Jaap van (eds.), Yearbook of Morphology 1994. Dordrecht: Kluwer, 1⫺29 Hooper, Joan B. (1976), An Introduction to Natural Generative Phonology. New York etc.: Academic Press Humboldt, Wilhelm von (1836), Über die Verschiedenheit des menschlichen Sprachbaues und ihren Einfluss auf die geistige Entwicklung des Menschengeschlechts, ed. by E. Wasmuth. Berlin: Schneider Jespersen, Otto (1922), Language, its Nature, Development & Origin. London: Allen & Unwin Jespersen, Otto (1939), “The History of a Suffix”. Acta Linguistica 1, 48⫺56 Koch, Harold (1995), “The Creation of Morphological Zeroes”. In: Booij, Geert & Marle, Jaap van (eds.), Yearbook of Morphology 1994. Dordrecht: Kluwer, 31⫺71 Koefoed, Geert (1978), “Taalverandering in het licht van taalverwerving en taalgebruik”. In: Koefoed, Geert & Marle, Jaap van (eds.), Aspecten van
1589 taalverandering. Groningen: Wolters-Noordhoff, 11⫺70 Lewis, Geoffrey L. (1967), Turkish Grammar. Oxford: Oxford University Press Marle, Jaap van (1978), “Veranderingen in woordstructuur”. In: Koefoed, Geert & Marle, Jaap van (eds.), Aspecten van taalverandering. Groningen: Wolters-Noordhoff, 127⫺176 Marle, Jaap van (1984), “Morfologische veranderingen in breder perspectief”. Tijdschrift voor Nederlandse Taal- en Letterkunde 100, 131⫺153 Marle, Jaap van (1988), “On the Role of Semantics in Productivity Change”. In: Booij, Geert & Marle, Jaap van (eds), Yearbook of Morphology 1988. Dordrecht: Foris, 139⫺154 Marle, Jaap van (1990 a), “A Case of Morphological Elaboration: the History of Dutch -baar”. Folia Linguistica Historica 9, 213⫺234 Marle, Jaap van (1990 b), “Rule-creating Creativity: Analogy as a Synchronic Morphological Process”. In: Dressler, Wolfgang U. & Luschützky, Hans C. & Pfeiffer, Oskar E. & Rennison, John R. (eds.), Contemporary Morphology. Berlin, New York: Mouton de Gruyter, 267⫺273 Marle, Jaap van (1996), “On the Interplay of Inherited and Non-inherited Features in Afrikaans Derivational Morphology”. In: Nielsen, Hans F. & Schøsler, Lene (eds.), The Origins and Development of Emigrant Languages. Odense: Odense University Press, 103⫺115 (RASK Supplement Vol. 6, NOWELE Supplement Vol. 17) Marle, Jaap van & Koefoed, Geert (1980), “Humboldtiaanse taalveranderingen, morfologie en de creativiteit van taal”. Spektator 10, 111⫺147 Paul, Hermann (51920), Prinzipien der Sprachgeschichte. Tübingen: Niemeyer [81970] Saussure, Ferdinand de (1916), Cours de linguistique ge´ne´rale. Paris: Payot Smits, Caroline (1996), Disintegration of Inflection: The Case of Iowa Dutch. The Hague: HAG (HIL Dissertations 22)
Geert Koefoed, Utrecht (The Netherlands) Jaap van Marle, Heerlen (The Netherlands)
1590
XVII. Morphologischer Wandel I: Theoretische Probleme
146. Grammaticalization: from syntax to morphology 1. 2. 3. 4. 5.
Introduction An overview of grammaticalization The process of grammaticalization Uncommon Abbreviations References
1.
Introduction
1.1. History of grammaticalization studies That bound morphological formatives often have their ultimate origin in independent lexical items has been a commonplace observation since the early 19th century (cf. Bopp 1816; Humboldt 1825). The phenomenon was not of primary interest to historical linguists more concerned with morphological constructions which could be reconstructed for a proto-language than with secondary developments, but its importance is explicitly recognized in discussions of general principles of diachronic linguistics such as Whitney (1875), Paul (51920 [11880]), and von der Gabelentz (1891). Meillet (1912) first applied the term grammaticalization to the process by which lexical items enter into the grammatical system (“le passage d’un mot autonome au roˆle d’e´le´ment grammatical”), a process whose endpoint is the development of new morphological constructions. Already in this discussion Meillet anticipates one of the most significant facts about the process, viz. that it is a gradual process rather than a sudden categorial shift. He distinguishes four degrees of grammaticalization of the French copula, from its lexical use in equational sentences (je suis celui qui suis) to its use as a tense auxiliary (je suis parti), and points out the ambiguous category of French faire, which is both a lexical verb ‘do, make’ and a causative: “laissez peut eˆtre un mot principal, dans laissez cela par example; mais ici [in laissez venir a` moi les petits enfants] laissez venir forme un ensemble, ou` laissez est, en quelque mesure, un auxiliaire.” (Meillet 1912 [1926: 134])
Since Meillet’s recognition and naming of grammaticalization as a distinct phenomenon worthy of study, the topic has attracted the attention of a few scholars, notably Kuryłowicz, who defined it in similar terms: “Grammaticalization consists in the increase of the range of a morpheme advancing from a lexical to a grammatical or from a less grammatical to a more
grammatical status, e.g. from a derivative formant to an inflectional one.” (Kuryłowicz 1965: 69)
However, the phenomena of grammaticalization have not been of central interest to most scholars of historical or synchronic linguistics, and it is only since about 1970 that it has begun to be systematically studied; it became the subject of sustained cross-linguistic investigation by the scholarly community only in the 1980’s. Important contemporary works and collections include Givo´n (1979), C. Lehmann (1986; 1995), Heine & Reh (1984), Heine et al. (1991), Heine (1993; 1997), Hopper & Traugott (1993), Traugott & Heine (1991, eds.), Rissanen et al. (1997, eds.), Ramat & Hopper (1998, eds.), and Fischer et al. (2000, eds.). The word grammaticalization (grammaticization and grammatization are also used in the same sense) implies a process of becoming “grammatical”. The reference can be taken as being to lexical morphemes becoming grammatical ones, or, more broadly, to any linguistic construct (a morpheme, syntactic construction, or discourse pattern) becoming part of the grammatical system of a language. Recently scholars have begun to use the term to refer to shifts from more pragmatic to more grammatical function of syntactic constructions, e.g. the development of subject from topic constructions (see e.g. Givo´n 1979; 1989). Grammaticalization involves changes in each of the three basic areas of linguistic structure: semantics/pragmatics, morphology/syntax, and phonology. The shift of a lexical form to a grammatical function involves, first, some shift in its semantic and/ or pragmatic function. This is the necessary precondition for a shift in syntactic category, a reanalysis of the syntactic construction. These two shifts in turn result in destressing of the grammaticalized morpheme, resulting in phonological reduction and cliticization, which in turn can lead to morphologization. The process is essentially unidirectional (Haspelmath 1999); the development of lexical from grammatical forms, while attested, is rare. Two fundamental questions outline the topic of grammaticalization: when, how, and why does a lexical form grammaticalize, and what specific types of grammatical formative develop from what specific types of lexical item?
146. Grammaticalization: from syntax to morphology
1.2. Grammatical and lexical meaning The phenomenon of grammaticalization has important implications for the traditional notion of “lexical” vs. “grammatical” morphemes. On the one hand, the traditional conception of grammaticalization as the passage of a form from the first to the second of these categories relies crucially on this distinction. On the other, the actual phenomena which we discover in studying cases of grammaticalization show that, like many other dichotomous categorizations in linguistics, this distinction is in fact a gradient rather than a clear split, and that forms must be thought of as more or less grammatical rather than simply as grammatical or lexical. While it is easy to find unambiguously lexical and unambiguously grammatical morphemes, it is notoriously difficult to draw a clear dividing line. Thus prepositions and subordinating conjunctions are more lexical than case inflections, but less lexical, and more likely to have syntactic as opposed to lexical value, than ordinary nouns and verbs. Grammaticalization research shows that historical change is almost always from more lexical to more grammatical status. All such change is grammaticalization, the shift of a relational noun or serialized verb into an adposition just as much as the shift of a adposition to a case desinence. The gradual nature of grammaticalization also calls into question the common conception of purely grammatical meaning. There is a strong correlation between the semantics of lexical items and their potential for grammaticalization ⫺ for example, future tense constructions develop from verbs originally meaning ‘want’ or ‘go’. If the meanings of grammatical forms derive by regular processes from lexical meaning, then grammaticalization provides a key to our understanding of the notion of grammatical meaning (Bybee 1988; Sweetser 1988; Heine et al. 1991; Heine 1993). The fact that not only do grammatical morphemes develop from lexical morphemes, but that specific types of grammatical morpheme regularly develop from lexical forms with particular meanings, suggests that grammatical meaning must to some degree have the same sort of semantic content as lexical meaning, rather than being purely structural. 1.3. Theoretical significance of grammaticalization studies Already in 1912 Meillet points out that grammaticalization, though equally as important as analogy in the development of new gram-
1591 matical forms, had received much less attention in historical linguistics. From a modern perspective, informed by knowledge of a wider range of languages, we can assert that grammaticalization is in fact much more important than analogy; nevertheless until recently it has continued to receive less attention than it merits in historical and general linguistics. One reason for this neglect is that the facts of grammaticalization pose a serious challenge to a fundamental aspect of structuralist synchronic analysis, an aspect which remains fundamental to omit much contemporary work in the generative paradigm. A structuralist or generative model consists of an inventory of syntactic categories and a set of rules for combining them into larger structures. The initial problem posed by examples such as Meillet’s is that in such data we seem to find one and the same morpheme as a member of two or more categories (e.g. in Meillet’s example as both copula and tense auxiliary). On further examination of such data the problem worsens, as it often is difficult or impossible, at least on any non-ad hoc basis, to assign some uses of an etymon to one or another category. Grammaticalization sensu strictu involves the shift of morphemes from one form class to another, and often involves the innovation of a new form class. Thus, in grammaticalization we find evidence bearing on the nature of morphosyntactic categories and their place in the organization of grammar. We regularly find cases in which grammaticalizing forms occupy intermediate categorial status. For example, in modern English we find a range of more and less verb-like characteristics among grammaticalizing forms such as used to, want to, ought to, etc. (Bolinger 1980). Such data call into serious question the adequacy of any model of linguistic structure which takes the notion of discrete, clearly defined morphosyntactic categories as a theoretical given. Another traditional distinction which requires reevaluation in the light of modern studies of grammaticalization is the opposition of synchronic and diachronic analysis. Some scholars (especially Hopper 1987; 1991; Givo´n 1989), are now suggesting that our traditional notion of a static synchronic “state” of a language in which every morpheme and construction can be unambiguously assigned a categorial place in the grammar is not only an idealization, but an unre-
1592
XVII. Morphologischer Wandel I: Theoretische Probleme
alistic one, and that viewing grammar in terms of the fluid categories and indeterminately syntacticized constructions which are the stuff of the diachronic study of grammaticalization provides a more adequate basis for understanding the “synchronic” structure as well as the diachronic changes in language.
2.
An overview of grammaticalization
2.1. Grammaticalization exemplified The study of grammaticalization reveals recurrent patterns for the origin of particular grammatical structures. For example, causative morphemes regularly develop from serialized or complement-taking verbs with meanings like ‘make’, ‘give’, or ‘send (on an errand)’. Dative case markers originate in ‘give’ verbs or in locative/allative markers, and locative and allative markers derive diachronically from relational nouns or from verbs of location, position, and motion. One important aspect of ongoing research on grammaticalization is the cataloguing of typical sources for various grammatical constructions, which show common patterns throughout the world. As a typical example, we may consider the origin of grammatical tense categories. While Bopp is still criticized for his promiscuous identification of Sanskrit conjugational endings as grammaticalized copulas, the mechanism which he suggested as the origin of verbal desinences in Sanskrit is one which is widely attested in languages of the world (e.g. Givo´n 1971; Haas 1977; Heine & Reh 1984). A well-known example is the development of the modern French, Spanish, and Italian synthetic future conjugations from fusion of auxiliary habere ‘have’ with the infinitive, so that e.g. Fr. chanterai ‘I will sing’ reflects a Vulgar Latin morphologization of an earlier infinitive ⫹ auxiliary construction cantare habeo (cf. Benveniste 1968; Hopper & Traugott 1993). While in the world’s languages we can find many tense affixes for which no etymology is recoverable from available data, and there are occasional examples of tense morphology with other origins (e.g. in morphologized adverbs; cf. Art. 110), in the vast majority of cases, if the origin of morphological tense markers can be traced, they will be found to originate in auxiliary verb constructions of some sort. Auxiliary verbs, in turn, represent grammaticalizations of originally biclausal syntac-
tic constructions, in which the potential auxiliary is a finite verb with a complement clause, or the highest verb in a serialization construction (cf. Art. 78). We can illustrate the sequence with a grammaticalization series from modern Central Tibetan (DeLancey 1991; 1997). Tibetan, like many verb-final languages, makes abundant use of a clausechaining structure in which only the last of a sequence of clauses has tense/aspect/evidentiality marking. Preceding clauses are marked with a subordinator, here glossed ‘nf’ for “non-final”, which functions only to mark them as non-final clauses in a chain. It is common for two chained clauses to share all arguments, with the result that their verbs occur in sequence, separated only by the ‘nf’ morpheme: (1) khos kha⫽lags zas-byas phyin-song he:erg meal ate-nf went-pf ‘He ate and left.’ There is a small set of verbs which can occur as the second member of such a sequence without intervening ‘nf’ marking. These thus constitute a distinct syntactic subcategory, which we can categorize as auxiliaries. The difference between the two constructions can be illustrated using the same etymon, bzhag, which as a lexical verb means ‘put’, and as a grammaticalized auxiliary forms a perfect construction: (2) kho phyin-byas bzhag-pa_red he went-nf put-pf ‘He went and put it there.’ (3) kho phyin bzhag-pa_red he went put-pf ‘He has gone.’ In the auxiliary construction (3), bzhag is still clearly the main verb; it carries full lexical tone, and takes tense/aspect marking. The same etymon occurs also in an even more grammaticalized construction, as a suffix marking inferential perfect: (4) kho phyin-zhag he went-infr.pf ‘He has left (I infer).’ Here the lexical verb phyin is the main verb. Zhag is unstressed and phonologically reduced, and represents still another morphosyntactic category, that of tense/aspect/evidentiality suffix.
146. Grammaticalization: from syntax to morphology
2.2. Stages of grammaticalization Discussion of processes of historical change traditionally present a series of “stages” leading from an initial to a final state, with the implication that these stages are themselves distinct states of linguistic structure. This can be a useful idealization, as long as it is recognized as an idealization of what is in fact a continuous phenomenon (cf. C. Lehmann 1985, which develops a scale for assessing the degree of grammaticalization of a morpheme). But the overall process of grammaticalization is perhaps better conceptualized in terms of a number of processes, some of which facilitate, precondition, or promote others, but which do not necessarily proceed strictly serially. The starting point of the process is a productive syntactic construction: noun phrase with genitive dependent, matrix with complement clause, conjoined or chained clauses, etc. The precondition for grammaticalization is that there be some lexeme or lexemes which occur frequently in this construction for some semantic/pragmatic reason: potential relational nouns like ‘top’, ‘face’, ‘back’, regularly used to provide further locational specification in noun phrases used as locatives, phasal or other complement-taking verbs like ‘finish’ or ‘want’, semantically nonspecific transitive verbs like ‘use’ or ‘hold’ conjoined or serialized with more specific verbs. This usually involves a lexeme with a very general meaning, which can therefore be used in a wide range of contexts. Thus a verb meaning ‘finish’ is much more likely to metamorphose into an aspectual marker, and thus potentially to enter the morphological system, than one meaning ‘buy’ or ‘repair’, which are initially usable in a much smaller range of semantic/pragmatic contexts. This situation, in which a particular construction ⫺ a productive syntactic structure with a specific lexeme in a specific slot ⫺ is a useful and regularly-used locution in the language, i.e. where speakers regularly refer to ‘the face of [noun phrase]’, ‘finish [verb phrase]’, etc., is the initial point of grammaticalization. We can refer to this situation as “functional specialization” of the construction. The next step is for the lexeme to undergo a certain amount of semantic bleaching, to use Givo´n’s term, or, put another way, for the locution to be used in an extended set of contexts, including some in which the literal meaning of the lexeme is not applicable:
1593 “... without an external change of its exponent a category may undergo important internal (functional) changes due simply to an extension of a limitation of its range. The logical principle of the mutual relation of range and content has to be applied in such a case: the increase of the range of a given category entails the impoverishment of its content, and vice versa.” (Kuryłowicz 1965: 57 f.)
This can often be observed even in forms which have not undergone any formal grammaticalization. For example, finish in English is often used to mean simply ‘stop’, so that one can say I’ve finished writing for today, even when the project on which the speaker is working is far from completed. A more grammaticalized example is the Tibetan verb sdad ‘sit’, which is currently developing into a progressive auxiliary. It can now occur in sentences such as (5): (5) kho rgyugs⫽shar_slod(-ni) sdad-zhag he run(-nf) sit-pf ‘He was running/kept running/was always running.’ The semantic incompatibility of ‘run’ and ‘sit’ is such that sdad cannot possibly be interpreted with its original lexical sense here. The structure thus cannot be a simple case of verb serialization, but represents an early stage of grammaticalization. The next stage of grammaticalization occurs when this lexeme begins to “decategorialize” (Heine et al. 1991; for further examples see Matisoff 1991), i.e. to lose the morphosyntactic characteristics of its original category. (This is a broad notion which includes more specific concepts such as C. Lehmann’s (1985: 307) “morphological degeneration ... the loss of ability to inflect”). For example, in the English construction on top of [noun phrase], top, while clearly a noun in origin, is un-nounlike in several respects. It lacks an article, and it cannot pluralize: we can say on top of all the houses, with top as a relational noun, or on the tops of all the houses, with top as an ordinary noun, but we cannot pluralize the relational noun: *on tops of all the houses. Note that top in its relational noun use is already semantically bleached, in that the top of [noun phrase] necessarily refers to a specific part of the object, while on top of [noun phrase] simply refers to whatever side of it is uppermost at the moment. (E. g. if a refrigerator is lying on its back, something resting on the door, which is the uppermost surface, is on top of the refrigerator, but is not on the
1594
XVII. Morphologischer Wandel I: Theoretische Probleme
top of the refrigerator). This is typical, and apparently universal; it is easy to find lexemes, like English finish, which have undergone some semantic bleaching without any morphosyntactic decategorialization, but decategorialization does not occur without some prior functional shift, toward either grammaticalization or lexicalization. From this point we have a continuous process of further decategorialization and phonological reduction. One possible outcome is recategorialization (cp. C. Lehmann’s 1985 “paradigmaticization”). There may already exist a morphosyntactic category into which the grammaticalizing form will fit and which it can enter. For example, the English preposition atop represents one endpoint of the grammaticalization of a noun ‘top’. Or, if similar changes are occurring involving several functionally related morphemes, they may become a new paradigmatic category ⫺ a famous example is the English modals. The final stages of grammaticalization are cliticization and morphologization, in which a grammaticalized form becomes increasingly bound phonologically and syntactically to a lexical head. While outlines of grammaticalization sometimes present this as a necessary final step, it is not; grammaticalized forms may remain as auxiliaries or “particles”, never becoming bound. The exact conditions which determine whether and to what extent morphologization will occur are as yet not well-understood, though they appear to have to do with stress patterns and with word-order typology ⫺ morphologization appears to be more common in verb-final than in verbmedial languages, for example. (For some discussion see W. Lehmann 1973; Donegan & Stampe 1983.) 2.3. The cycle The best-known version of the idea of discrete ordered stages in historical change is the hypothesis of the “morphology-syntax cycle”. The traditional version describes an essentially mechanical sequence. Morphology is by its nature subject to phonological erosion, which over time reduces the distinctiveness of affixes to the point where essential distinctions are lost. In the face of this languages resort to grammaticalization as a therapeutic measure, creating new periphrastic grammatical constructions to replace lost morphological forms. Since grammatical for-
matives, whether free or bound, do not normally carry accent, newly-grammaticalized forms are now subject to phonological reduction, and over time cliticize, and eventually develop into new morphological constructions. These now are by nature subject to phonological erosion, and over time the new formation loses distinctiveness, and a new cycle takes place. It has sometimes been suggested that this is a typological cycle, i.e. that every language passes through successive analytic and synthetic stages (cf. Hodge 1970). This is certainly open to doubt, and in any case would be impossible to document for most languages and families. A more plausible claim is that the cycle operates at the level not of whole-language typology, but of the instantiation of particular functions, so that, for example, tense marking, or even more specifically the expression of a particular tense category, in a language will alternate between periphrastic and morphological encoding. Examples at this level are easier to find; consider for example the contemporary competition of the French morphological future ⫺ itself the end-point of grammaticalization of an earlier periphrastic construction with ‘have’ ⫺ with a new periphrastic future in aller ‘go’: je chanterai vs. je vais chanter ‘I will sing’. While there is no doubt that the cycle is a descriptively useful schema for interpreting historical change, the traditional explanatory account which accompanies it is empirically inadequate. In the traditional description, the rise of new periphrastic constructions is motivated by, and thus follows, the loss of older morphological ones: “Das, was man Aufbau nennt, kommt ja, wie wir gesehen haben, nur durch einen Verfall zu Stande, und das, was man Verfall nennt, ist nür die weitere Fortsetzung dieses Prozesses. Aufgebaut wird nur mit Hilfe der Syntax. Ein solcher Aufbau kann in jeder Periode stattfinden, und Neuaufgebautes tritt immer als Ersatz ein da, wo der Verfall ein gewisses Mass überschritten hat.” (Paul 51920: 351)
Such interpretations of diachronic development effectively assume an informationtheoretic functional model of language in which there is a fixed set of functions which linguistic structure must be able to handle, and some inherent pressure to avoid redundancy. The simple picture of the cycle is one
1595
146. Grammaticalization: from syntax to morphology
in which a language at any given synchronic stage has one mechanism to carry out each essential function, and as this mechanism wears out it must be replaced by a new one so that the function will not be lost. Thus, for example, Benveniste (1968), in discussing the innovation of periphrastic grammatical constructions which constitutes the first step of what we call grammaticalization, labels it conservative mutation, explicitly assuming that periphrastic constructions develop to replace earlier morphological constructions in the same function. However, there is abundant evidence that this model is too simple. It is not the case that a language will have only one functioning means of expressing a particular meaning at a given stage, or that serious phonological erosion of one construction is a necessary condition for the rise of another (DeLancey 1985; Hopper 1991; Hopper & Traugott 1993). Consider, as a simple example, the gradual replacement of the French inflected past and future verb forms by periphrastic constructions with avoir and aller. There is clearly a stage in this development (i.e. Modern French) where the inflectional and periphrastic constructions co-exist, where the new construction is already grammaticalized while the older one remains functional. Arguably we could claim that the traditional account has the direction of causation backwards, that it is in fact the development of the new construction which leads to the loss of the old, rather than the decay of the older one leading to the development of a new replacement (Bybee 1985; DeLancey 1985). 2.4. Sources and pathways Grammaticalization processes are not random; particular types of grammatical formative tend to develop from specific lexical sources. Recent studies have made considerable progress in developing a catalogue of such pathways of development (Traugott 1978; 1988; Ultan 1978; Givo´n 1979; Heine & Reh 1984; Bybee 1988; Bybee et al. 1994; Bybee & Dahl 1989; Heine et al. 1991; various papers in Traugott & Heine 1991, eds.). The best studied category of nominal morphology is case inflection. Since the functions of case inflection and adpositional marking show considerable overlap ⫺ indeed in many languages case marking is accomplished by adpositions rather than inflections ⫺ we should expect to find a diachronic connection between the two. Indeed, case inflection does
ordinarily develop from cliticization and subsequent morphologization of postpositions (Kahr 1975); examples of morphologization of prepositions appear to be rare. Postpositions and prepositions both develop either from serial verbs or from relational noun constructions. Other categories of nominal inflection have not been as well-studied, but see for example Greenberg (1978) on the origin of gender marking. Verbs, of course, show a much wider range of common inflectional categories than nouns (cf. Art. 72), but most of these come from the same proximate source, auxiliary verbs, which are the typical source of tense/ aspect and modality marking, deictic specification, and causative, benefactive, and other “applicative” constructions. The other common morphological category in verbs, person and number agreement, arises most commonly from the morphologization of unstressed, non-contrastive pronouns. Auxiliary verbs develop from lexical verbs through two types of construction: either complementation or clause-chaining. Since European languages are not prone to clausechaining, many of the best-known examples of the development of auxiliaries and hence verbal inflection arose from complement constructions; for example, the French synthetic future represents reanalysis and subsequent morphologization of the infinitive form as a complement of the verb habere ‘have’. The path from clause-chaining through verb serialization to auxiliarization is also widely attested. For example, the Lhasa Tibetan perfect/perfective paradigm includes, among other forms, an inferential perfect -zhag /-sˇa/ and a volitional perfective -pa yin /-payı˜˜ı/, reduced in running speech to /-pı˜˜ı/. The first of these developed from a serial verb construction with the verb bzhag ‘put’, originally a grammaticalization of clause-chaining constructions along the lines of (6): (6) sha btsabs-nas bzhag meat chop-nf put ‘chop the meat and put it (aside)’ The second developed from a nominalized clause (-pa is a nominalizer) as the complement to a higher copula (yin), so that the modern perfective construction V-pı˜˜ı reflects an older structure *[S-pa] yin. Special note should be made of this second pattern, which exemplifies a cross-linguistically very com-
1596
XVII. Morphologischer Wandel I: Theoretische Probleme
mon phenomenon of tense/aspect and/or evidentiality marking through the grammaticalization of copulas.
3.
The process of grammaticalization
3.1. Functional aspects of grammaticalization We have referred to the notion of “functional specialization” as prerequisite to grammaticalization. The essential precondition for grammaticalization is that a lexical form have some special functional status which distinguishes it from other members of its syntactic category. We can describe three paths to grammaticalization: semantic specialization through metaphor, reanalysis through pragmatic inference, and what we may call referent conflation. It must be emphasized that these are not presented here as competing explanations; all are clearly attested as occurring. Moreover, they are not mutually exclusive, and particular cases of grammaticalization may be best explained in terms of some combination of these processes. Many examples of shift from a lexical to a more grammatical meaning for a morpheme involve metaphorical extension. One example that has been well-explored is the “body-part model” for the development of spatial expressions. In a wide range of languages some or most of the relational nouns which express spatial relations such as ‘front’ and ‘back’, and eventually adpositions derived from these relational nouns, are drawn from bodypart vocabulary (see e.g. Friedrich 1969; Brugman 1983; Heine et al. 1991). Among the commonest such semantic transfers are expressions for ‘on’ derived from nouns meaning ‘head’, for ‘behind’ from ‘back’, for ‘in front of’ from ‘face’, and for ‘inside’ from ‘belly’. Another well-explored metaphorical field is the wide range of “localist” phenomena in which spatial expressions are used with temporal, logical or other meaning (Anderson 1973; Traugott 1978). The most prevalent example of this is the use of forms meaning ‘at’, ‘from’, and ‘to’ to express temporal as well as spatial relations, which is so universal as to leave open to question in what sense it is appropriately considered to be metaphorical. Another is the common development of perfect or past tense constructions from con-
structions with ‘come’, and of futures from ‘go’ (Traugott 1978). A well-explored type of abstraction from one cognitive domain to another is the development of grammatical from spatial (or “local”) case forms (Anderson 1971; DeLancey 1981); the most widely attested development is the origin of dative case markers from locative or allative adpositions (e.g. English to, French a`, and their cognates in other Germanic and Romance languages). Another line of explanation for semantic shift, sometimes referred to as metonymic (Heine et al. 1991), finds the initial shift in the transfer of the primary meaning of the construction from one to another aspect of the situation to which it applies. For example, we often find perfective constructions developing from verbs meaning ‘go’, ultimately deriving from serial verb constructions meaning ‘did X and then went’; the obvious inference that the doing was completed before the doer left ceases to be merely an inference and becomes the primary semantic content of the construction. As we will note below, grammaticalization typically involves the flattening of a syntactic structure ⫺ a biclausal construction is reanalyzed as a single clause, or a noun phrase with another noun phrase embedded within it is reinterpreted as a single adpositional phrase. This syntactic reanalysis is driven by a semantic reinterpretation in which two conceptually distinct referents are reinterpreted as one. For example, in an ordinary genitive construction like the leg of the table, each lexical noun refers separately; although the leg is part of the table, the noun phrase makes distinct reference to the table and its leg. An adpositional construction such as on the table, in contrast, refers only once. Thus in the development of an adposition from a relational noun, e.g. atop from *on the top of, we see the conflation of two referents into one. 3.2. Syntactic aspects of grammaticalization Functional specialization is an ubiquitous phenomenon; lexemes with certain semantic characteristics tend to be almost automatically specialized in this sense to some degree. While these functional factors are the engine which drives the process, most linguists do not recognize grammaticalization per se until actual changes of grammatical structure have occurred (but cf. Hopper 1987; Givo´n 1989
1597
146. Grammaticalization: from syntax to morphology
for suggestions that this distinction is to some extent artificial). The critical patterns of structural change relevant to the development of morphology are concerned with changes in constituent structure, particularly head-dependent relations, and changes in the status of the morphosyntactic boundaries setting off the grammaticalizing morpheme. Consider again the three stages of grammaticalization illustrated by the Lhasa Tibetan examples (2)⫺(4) (see 2.1). In (2), we can identify bzhag as both the syntactic head and the semantic/pragmatic center of the construction. Syntactically, it is the finite verb, and in the rightmost position which in Tibetan is characteristic of heads. Semantically, the ‘putting’ is the primary foregrounded information; the informational contribution of ‘went’ is essentially adverbial. In (3) the syntactic and functional analyses are no longer congruent: while bzhag remains the syntactic head, as noted previously, phyin is the lexical verb and the information focus. By the stage illustrated in (4), the shift of informational focus is complete, and by many analyses the head-dependent relations have also shifted. Some modern interpretations of the head-dependent relation would identify zhag as still the head of its construction; in any framework with this feature the description of the changes involved in grammaticalization is simpler, since it does not have to involve reanalysis of the head-dependent relations in the construction. Accompanying each of these shifts is a downgrading of the boundary between the two verbs. In (2), the two verbs are in two distinct clauses, and a clausal boundary separates them. In (3) there is only one clause, but the lexical verb phyin and the auxiliary bzhag remain separate words. In (4) -zhag is a suffix, with a morpheme boundary corresponding to the word boundary in (3) and the clause boundary in (2). The mechanics of this sort of boundary downgrading and erasure are discussed at length by Langacker (1977). 3.3. Grammaticalization and lexicalization A lexical morpheme which loses its autonomy may follow various career paths, which we can broadly categorize as grammaticalization and lexicalization. (I use “lexicalization” here to refer to the process by which originally independent lexemes become parts of
new lexical items; the term has several other uses, including one, essentially the converse of grammaticalization, in which it refers to an inflected form leaving its paradigm and becoming a distinct lexeme; the best-known category of example is the development of adverbials from oblique case-forms of nouns; cf. Art. 150). The difference is in whether the output of the process is a new lexeme or a new construction. For example, consider the Early New English and now dialectal progressive prefix a- in English, as in: (7) I’m a-wanting for to go. This represents grammaticalization of an earlier preposition. Exactly the same process of phonological reduction and cliticization of a preposition is the source of the a- in atop. In this case, the end product of the process is, not a new construction, but a single lexical item, and thus an example of lexicalization. (As discussed above, the shift of top out of the open class of nouns, and the creation of a new member of the closed class of prepositions, are more profitably discussed as grammaticalization.) Obviously this distinction depends on the distinction between lexical and grammatical morphemes, and thus like that distinction must be viewed as a continuum rather than a dichotomy. For example, the same reduction of a preposition discussed in the preceding paragraph also produced the odd family of a- adjectives (awake, asleep, alone, etc.). In terms of productivity this development lies somewhere between that of atop and the entirely productive progressive prefix. Here the process created a new subcategory; the resulting forms differ from most adjectives in that they cannot occur in prenominal position (*an alone man). But as a new category this has no other repercussions in the grammar, since there is no new grammatical construction associated with it, and it is not particularly useful to consider this as an example of grammaticalization. Occupying the middle ground between lexicalization and pure grammaticalization is the development of derivational morphology, which as Paul points out is not systematically distinguishable from the origin of inflection: “Auf die gleiche Weise wie die Ableitungssuffixe entstehen Flexionssuffixe. Zwischen beiden gibt es ja überhaupt keine scharfe Grenze.” (Paul 51920: 349)
1598
XVII. Morphologischer Wandel I: Theoretische Probleme
For the purposes of grammaticalization theory the most useful criterion for identifying grammaticalization is the degree to which the output of the process is a new productive construction, i.e. a new element of grammatical structure, as opposed to simply a new set of (one or more) lexical forms.
4.
Uncommon abbreviations
nf non-final clause marker infr.pf inferential perfect
5.
References
Anderson, John (1971), The Grammar of Case: Towards a Localistic Theory. Cambridge: Cambridge University Press Anderson, John (1973), An Essay Concerning Aspect. The Hague: Mouton Benveniste, Emile (1968), “Mutations of Linguistic Categories”. In: Lehmann, Winfred, & Malkiel, Yakov (eds.), Directions for Historical Linguistics: a Symposium. Austin: University of Texas Press, 83⫺94 Bolinger, Dwight (1980), “Wanna and the Gradience of Auxiliaries”. In: Brettschneider, Gunter & Lehmann, Christian (eds.), Wege zur Universalienforschung: Sprachwissenschaftliche Beiträge zum 60. Geburtstag von Hansjakob Seiler. Tübingen: Gunter Narr, 292⫺299 Bopp, Franz (1816), Über das Conjugationssystem der Sanskritsprache in Vergleichung mit jenem der griechischen, lateinischen, persischen und germanischen Sprachen. Frankfurt: Andreäische Brugman, Claudia (1983), “The Use of Body-Part Terms as Locatives in Chalcatongo Mixtec”. In: Report #4 of the Survey of Californian and Other Indian Languages. Berkeley: University of California, 235⫺290 Bybee, Joan (1985), Morphology: A Study of the Relation Between Meaning and Form. Amsterdam, Philadelphia: Benjamins Bybee, Joan (1988), “Semantic Substance vs. Contrast in the Development of Grammatical Meaning”. Berkeley Linguistics Society: Proceedings of the Annual Meeting 14, 247⫺264 Bybee, Joan & Dahl, Östen (1989), “The Creation of Tense and Aspect Systems in the Languages of the World”. Studies in Language 13, 51⫺103 Bybee, Joan & Perkins, Revere & Pagliuca, William (1994), The Evolution of Grammar: Tense, Aspect, and Modality in the Languages of the World. Chicago: University of Chicago Press
DeLancey, Scott (1981), “An Interpretation of Split Ergativity and Related Patterns”. Language 57, 626⫺57 DeLancey, Scott (1985), “The Analysis-synthesislexis Cycle in Tibeto-Burman: A Case Study in Motivated Change”. In: Haiman, John (ed.), Iconicity in Syntax. Amsterdam: Benjamins, 367⫺89 DeLancey, Scott (1991), “The Origins of Verb Serialization in Modern Tibetan”. Studies in Language 15, 1⫺23 DeLancey, Scott (1997), “Grammaticalization and the Gradience of Categories: Relator Nouns and Postpositions in Tibetan and Burmese”. In: Bybee, Joan L. & Haiman, John & Thompson, Sandra A. (eds.), Essays on Language Function and Language Type. Amsterdam: Benjamins, 51⫺69 Donegan, Patricia, & Stampe, David (1983), “Rhythm and the Holistic Organization of Language Structure”. Papers from the Parasession on the Interplay of Phonology, Morphology, and Syntax. Chicago: Chicago Linguistic Society, 337⫺353 Fischer, Olga & Rosenbach, Anette & Stein, Dieter (2000, eds.), Pathways of Change: Grammaticalization in English. Amsterdam: Benjamins Friedrich, Paul (1969), On the Meaning of the Tarascan Suffixes of Space. (Memoir 23 of the International Journal of Linguistics; supplement to the IJAL 35.4) Gabelentz, Georg von der (1891), Die Sprachwissenschaft: Ihre Aufgaben, Methoden, und bisherigen Ergebnisse. Leipzig: Weigel Givo´n, Talmy (1971), “On the Verbal Origin of the Bantu Verb Suffixes”. Studies in African Linguistics 2.2, 145⫺163 Givo´n, Talmy (1979), On Understanding Grammar. New York, London: Academic Press Givo´n, Talmy (1989), Mind, Code, and Context: Essays in Pragmatics. Hillsdale/NJ: Erlbaum Greenberg, Joseph (1978), “How Does a Language Acquire Gender Markers?”. In: Greenberg et al. (eds.), Vol. III, 47⫺82 Greenberg, Joseph & Ferguson, Charles & Moravcsik, Edith (1978, eds.), Universals of Human Language, Vol. 1⫺4. Stanford: Stanford University Press Haas, Mary R. (1977), “From Auxiliary Verb Phrase to Inflectional Suffix”. In: Li (ed.), 525⫺ 537 Haspelmath, Martin (1999), “Why is Grammaticalization Irreversible?”. Linguistics 37.6, 1043⫺ 1068
146. Grammaticalization: from syntax to morphology
1599
Heine, Bernd (1993), Auxiliaries: Cognitive Forces and Grammaticalization. New York: Oxford University Press
Lehmann, Christian (1995), Thoughts on Grammaticalization. Munich: Lincom Europa (LINCOM Studies in Theoretical Linguistics 1)
Heine, Bernd (1997), Possession: Cognitive Sources, Forces, and Grammaticalization. Cambridge: Cambridge University Press
Lehmann, Winfred P. (1973), “A Structural Principle of Language and its Implications”. Language 49, 47⫺66
Heine, Bernd & Claudi, Ulrike & Hünnemeyer, Friederike (1991), Grammaticalization: A Conceptual Framework. Chicago: University of Chicago Press
Li, Charles (1977, ed.), Mechanisms of Syntactic Change. Austin, London: University of Texas Press
Heine, Bernd & Reh, Mechthild (1984), Grammaticalization and Reanalysis in African Languages. Hamburg: Helmut Buske Hopper, Paul (1987), “Emergent Grammar”. Berkeley Linguistics Society: Proceedings of the Annual Meeting 13, 139⫺157 Hopper, Paul (1991), “On some Principles of Grammaticization”. In: Traugott & Heine (eds.), Vol. I, 17⫺35 Hopper, Paul, & Traugott, Elizabeth (1993), Grammaticalization. Cambridge: Cambridge University Press Humboldt, Wilhelm von (1825), “Über das Entstehen der grammatikalischen Formen und ihren Einfluß auf die Ideenentwicklung”. Abhandlungen der Königlichen Akademie der Wissenschaft zu Berlin, 401⫺430 Kahr, Joan (1975), “The Renewal of Case Morphology: Sources and Constraints”. Stanford Working Papers in Linguistic Universals 20, 107⫺ 151 Kuryłowicz, Jerzy (1965), “The Evolution of Grammatical Categories”. Diogenes 51, 55⫺71 [reprinted in: Kuryłowicz, Jerzy (1975), Esquisses linguistiques II. München: Fink] Langacker, Ronald (1977), “Syntactic Reanalysis”. In: Li (ed.), 57⫺139 Lehmann, Christian (1985), “Grammaticalization: Synchronic Variation and Diachronic Change”. Lingua e Stile 20.3, 303⫺318 Lehmann, Christian (1986), “Grammaticalization and Linguistic Typology”. General Linguistics 26.1, 3⫺23
Matisoff, James A. (1991), “Areal and Universal Dimensions of Grammaticalization in Lahu”. In: Traugott & Heine (eds.), Vol. II, 383⫺453 Meillet, Antoine (1912), “L’e´volution des formes grammaticales”. Scientia (Rivista de scienza) 12.26 [reprinted in: Meillet, Antoine (1926): Linguistique historique et linguistique ge´ne´rale, 130⫺148] Paul, Hermann (51920), Prinzipien der Sprachgeschichte. Halle: Niemeyer [11880] Ramat, Anna, & Paul Hopper (1998, eds.), The Limits of Grammaticalization. Amsterdam: Benjamins (Typological Studies in Language 37) Rissanen, Matti & Kytö, Merja & Kirsi Heikkonen (1997, eds.), Grammaticalization at Work. Berlin: Mouton de Gruyter Sweetser, Eve (1988), “Grammaticalization and Semantic Bleaching”. Berkeley Linguistics Society: Proceedings of the Annual Meeting 14, 389⫺405 Traugott, Elizabeth (1978), “On the Expression of Spatio-Temporal Relations in Language”. In: Greenberg et al. (eds.), Vol. III, 369⫺400 Traugott, Elizabeth (1988), “Pragmatic Strengthening and Grammaticalization”. Berkeley Linguistics Society: Proceedings of the Annual Meeting 14, 406⫺416 Traugott, Elizabeth, & Heine, Bernd (1991, eds.), Approaches to Grammaticalization, Vol. 1⫺2. Amsterdam: Benjamins (Typological Studies in Language 19) Ultan, Russell (1978), “The Nature of Future Tenses”. In: Greenberg et al. (eds.), Vol. III, 83⫺ 123
Scott DeLancey, Eugene (U.S.A.)
1600
XVII. Morphologischer Wandel I: Theoretische Probleme
147. Morphologisierung: von der Phonologie zur Morphologie 1. 2. 3. 4.
Was ist Morphologisierung? Zur Geschichte des Konzepts Parameter der Morphologisierung Zitierte Literatur
1.
Was ist Morphologisierung?
Von Morphologisierung spricht man im allgemeinen, wenn im Laufe der Sprachgeschichte nichtmorphologische grammatische Erscheinungen zu morphologischen Erscheinungen werden. Will man den Begriff der Morphologisierung etwas genauer umschreiben, so geschieht das am zweckmäßigsten, indem man von der Bestimmung der Morphologie, genauer gesagt der Flexionsmorphologie, ausgeht. Die Flexionsmorphologie ist im Rahmen des Sprachsystems dadurch gekennzeichnet, daß sie erstens grammatische Kategorien formal durch Marker symbolisiert, und daß sie zweitens innerhalb der Grenzen des Wortes operiert. Eine Morphologisierung findet immer dann statt, wenn eine grammatische Erscheinung, der bisher nur eine dieser beiden für die Flexionsmorphologie konstitutiven Eigenschaften zukam, infolge von Sprachwandel auch die jeweils zweite dieser Eigenschaften herausbildet. Damit ergeben sich grundsätzlich zwei unterschiedliche Möglichkeiten der Morphologisierung. Zum einen können syntaktisch-analytische Marker (Marker mit Wortcharakter), die ja bereits grammatische Kategorien symbolisieren, durch Tilgung von Wortgrenzen zu morphologischen Markern werden. Zum anderen können phonologische Alternationen, die innerhalb des Wortes operieren und die Formen von Flexionsparadigmen betreffen, durch Ersetzung ihrer phonologischen Bedingungen durch morphologische zu morphologischen Markern werden. Allen Morphologisierungen ist gemeinsam, daß sich im Wort, der Domäne der Morphologie, neue Zeichenverhältnisse herausbilden, neue Relationen zwischen Zeicheninhalten, also grammatischen Kategorien, und Zeichenformen, also grammatischen Markern. Dennoch unterscheiden sich die beiden Typen von Morphologisierung hinsichtlich ihres Status und ihrer Ergebnisse beträchtlich. Im ersten Fall ist die Morphologisierung nur eine Etappe eines umfassenden Grammatikalisierungsprozesses; sie führt zur Herausbildung von additiven Markern (Markern mit Morphemcharakter)
und ist entsprechend in den Zusammenhang der Grammatikalisierung einzuordnen (vgl. Art. 145). Im zweiten Fall hingegen stellt die Morphologisierung selbst einen eigenständigen, meist relativ komplexen Wandelprozeß dar; sie führt typischerweise (wenn auch nicht ausschließlich; vgl. 3.3) zur Herausbildung von modifikatorischen Markern (Markern mit Alternationscharakter). Die Morphologisierung phonologischer Alternationen (im weiteren kurz Morphologisierung) konstituiert einen spezifischen Bereich der historischen Morphologie. Da phonologischen und morphologischen Alternationen jeweils entsprechende Regeln zugrundeliegen, läßt sich auch die Morphologisierung am angemessensten bezogen auf den Regelbegriff definieren: Morphologisierung liegt genau dann vor, wenn an die Stelle einer Regel/mehrerer Regeln, die in einem phonologischen Kontext eine formale Operation vornimmt/vornehmen, eine Regel tritt/mehrere Regeln treten, die die entsprechende Operation oder deren Inversion in einem morphologischen Kontext vornimmt/vornehmen. Es findet also, kurz gesagt, ein Übergang von phonologischen Regeln zu morphologischen Regeln statt, wobei dieser Übergang kein unmittelbarer sein muß, sondern auch vermittelt über morphonologische Regeln erfolgen kann (vgl. 3.2). Eine notwendige Voraussetzung für die Morphologisierung einer phonologischen Regel ist, daß diese formale Alternationen innerhalb von Flexionsparadigmen bewirkt. Der resultierende morphologische Kontext besteht aus grammatischen Kategorienmerkmalen (wie Numerus-, Kasus-, Tempusmerkmalen); in Sprachen mit Flexionsklassen treten auch entsprechende Klassenmerkmale auf. Bei Morphologisierungen gibt es nicht notwendigerweise ein Eins-zu-eins-Verhältnis zwischen einer phonologischen und einer morphologischen Regel. Das zeigen z. B. die verschiedenen Morphologisierungen des Umlauts im Deutschen; im Laufe der Sprachgeschichte wird eine einheitliche phonologische Umlautregel durch eine ganze Anzahl von morphologischen Umlautregeln ersetzt. Ein Beispiel dafür, daß eine resultierende morphologische Regel die inverse Operation der ursprünglichen phonologischen Regel vornimmt, ist der morphologische ‘Rückumlaut’ der deutschen Verben
147. Morphologisierung: von der Phonologie zur Morphologie
des Typs brennen ⫺ Präteritum brannte mit Umkehrung der ursprüglichen Ableitungsrichtung von ‘a J e’ zu ‘e J a’ (vgl. 3.4). Mit der Morphologisierung von phonologischen Regeln ändert sich deren Status grundsätzlich. Eine Regel, deren (vollständig oder doch partiell erhaltene) Funktion darin bestand, eine Klasse von Lautfolgen an den phonologischen Kontext anzupassen, erhält die neue Funktion der formalen Symbolisierung morphologischer Kategorien, wie sie allen morphologischen Regeln zukommt. Damit wird die phonetische Motivierung der Regel durch eine semiotische Motivierung ersetzt; aus lautlichen Alternationen werden sprachliche Zeichen; vgl. dazu z. B. urnordisch landa ‘Land’ ⫺ Nom./Akk. Pl. lo˛ndu mit Assimilation [a] > [c] an das [u] der Folgesilbe und neuisländisch land ⫺ Nom./Akk. Pl. lönd mit morphologischer a/ö-Alternation als alleinigem Numerus-Kasus-Marker.
2.
Zur Geschichte des Konzepts
Bereits die traditionelle Sprachwissenschaft der zweiten Hälfte des 19. Jahrhunderts stößt bei ihren sprachhistorischen Untersuchungen auf die Problematik der Morphologisierung, ohne daß freilich der Terminus auftritt. So stellt Paul (1880) fest, daß es normalerweise kein “lautgesetz” gibt, “das nicht, sobald es einmal in einer anzahl von fällen das etymologisch eng zusammenhängende lautlich differenziert hat, auch eine reaction gegen diese differenzierung hervorriefe” (Paul 1880: 104). Paul konstatiert, daß eine solche Entwicklung jedoch nicht eintritt, wenn die formale Differenzierung mit einem Funktionsunterschied, d. h. einem morphologischen Unterschied, zusammenfällt. Dieses Zusammenfallen kann dann nämlich “die ursache zu dauernder bewahrung eines lautlichen unterschiedes sein, und dies vor allem, wenn er zugleich [...] durch die formale analogie widerstandsfähig gemacht wird” (Paul 1880: 114), d. h. wenn die Alternation morphologisch systematisiert wird: “Bei dem zusammentreffen dieser beider umstände kann sich die vorstellung von dem lautlichen unterschiede so fest mit der von dem functionsunterschiede verbinden, dass dem sprachgefühl beides unzertrennbar erscheint. Auf diese weise wird allmählig der zufällig entstandene bedeutungslose unterschied zu einem bedeutungsvollen. Er wird es um so mehr, je weniger die bedeutungsverschiedenheit durch sonstige unterschiede in der lautgestaltung deutlich gekennzeichnet ist. So vermag sich die sprache einen ersatz zu schaffen für den in folge
1601
des lautlichen verfalls eintretenden verlust der charakteristischen merkmale des functionsunterschiedes.” (Paul 1880: 114; Hervorhebung im Original)
Um die theoretische Einordnung der Morphologisierung haben sich dann bereits zur Zeit der Junggrammatiker vor allem namhafte Vertreter des frühen russischen Strukturalismus verdient gemacht. Baudouin de Courtenay, der offenbar auch den Terminus “Morphologisierung” prägt, unterscheidet in seiner “Theorie phonetischer Alternationen” (Baudouin de Courtenay 1895) u. a. drei für den Morphologisierungszusammenhang einschlägige Typen von Alternationen, nämlich (i) “neophonetische”, d. h. phonologisch reguläre Alternationen, (ii) “traditionelle”, d. h. phonologisch und morphologisch irreguläre Alternationen und (iii) “psychophonetische”, d. h. phonologisch irreguläre, aber morphologisch reguläre Alternationen (Baudouin de Courtenay 1895: 110). Die Morphologisierung wird bei Baudouin in umfassendere grammatische Zusammenhänge gestellt. Für ihn hat die Sprache eine phonetische, eine semasiologische (semantische) und eine morphologische Seite, wobei sich letztere auf die grammatische Struktur der Sprache bezieht. Dabei verbinden sich phonetische Vorstellungen mit semasiologischen oder morphologischen; Lautsegmente oder auch phonetische Merkmale werden semasiologisiert oder morphologisiert (Baudouin de Courtenay 1908: 10 f.). Semasiologisierung und Morphologisierung haben jedoch einen recht unterschiedlichen Status. Alle (distinktiven) Unterschiede werden generell im Rahmen von Morphemen semasiologisiert, d. h. zur Unterscheidung der Morpheme als Bedeutungsträger genutzt. “Demgegenüber ist die Morphologisierung bestimmter Unterschiede zwischen artikulatorischauditiven Elementen in der Geschichte der Sprachen eine Übergangserscheinung. Sie wird gewöhnlich durch bestimmte historisch-phonetische Prozesse hervorgerufen, die zur Auflösung eines Phonems in zwei oder mehr führen; und diese neu entstandenen Abwandlungen eines ehemals einzigen Phonems [...] können psychophonetische Alternationen werden, die mit einem festen Unterschied von Formen verknüpft sind.” (Baudouin de Courtenay 1922: 67; deutsche Übersetzung nach Mugdan 1984: 75)
Die Morphologisierung erfolgt also in zwei Schritten: “Neophonetische” Alternationen verlieren zunächst durch Lautwandel ihre phonologische Motivierung und werden zu “traditionellen” Alternationen (i > ii). Wenn
1602
XVII. Morphologischer Wandel I: Theoretische Probleme
diese dann morphologisch systematisiert werden, bilden sich “psychophonetische” Alternationen heraus. Es tritt Morphologisierung ein (ii > iii). Wenn das nicht geschieht, bleiben sie als rein traditionelle, irreguläre Alternationen erhalten (für Details vgl. Mugdan 1984: 73 ff.). Auf Baudouin aufbauend legt dann sein Schüler Kruszewski (der auch Baudouins erst später publizierte Auffassungen zur Morphologisierung kennt) ein geradezu erstaunlich modernes Konzept der unterschiedlichen Typen von “Lautabwechslungen”, d. h. von Alternationen, und der sprachhistorischen Übergänge zwischen ihnen vor (Kruszewski 1881), das in seinen Grundzügen noch heute gilt. Er unterscheidet als erster zwischen (i) phonetisch bedingten Alternationen, (ii) morphonologisch bedingten Alternationen und (iii) morphologisch bedingten Alternationen. Es gibt einen sprachhistorischen Entwicklungsprozeß von (i) zu (iii), wobei der Wandel von (i) zu (ii) “den Anfang der Verbindung von phonetischen Erscheinungen mit morphologischen Kategorien” darstellt (Kruszewski 1881: 23). Da die morphonologisch bedingten Alternationen eine Art von Umbauperiode zwischen (i) und (iii) markieren, stellen sie sich zunächst als scheinbar unregelmäßig dar; sie tendieren aber quasi gesetzmäßig zu einer neuen Regularität. Für die weitere Entwicklung gibt es zwei Möglichkeiten. Entweder erfolgt ein Wandel von (ii) zu (iii) in Form einer “Differenzierung”, d. h. einer Funktionalisierung der Alternation, wie beim deutschen Umlaut als Pluralmarker in Fällen wie Loch ⫺ Löcher; die Alternation ist dann morphologisiert. Oder es findet eine morphologische “Assimilation”, ein Ausgleich innerhalb des Paradigmas, statt, wie im Konsonantismus bestimmter starker Verben des Deutschen, vgl. ahd. lesan ⫺ las ⫺ la¯rum > nhd. lesen ⫺ las ⫺ lasen. In diesem Fall verschwindet die Alternation aus der Grammatik (Kruszewski 1881: 24 ff.; vgl. dazu Klausenburger 1994: 2562). Nach einer langen Zeit, in der die Morphologisierung in der theoretischen Linguistik faktisch keine Rolle spielte, wird die Problematik in den siebziger Jahren wieder aufgegriffen. Ausgangspunkt dafür ist die zusammenfassende Darlegung der generativen phonologischen Standardtheorie durch Chomsky & Halle (1968). Diese Theorie betrachtet sämtliche Alternationen von Lauteinheiten ungeachtet ihrer jeweiligen Bedingtheit und Funktion einheitlich als phonologi-
sche Regeln. In den dadurch hervorgerufenen Diskussionen wird auch die Frage nach den zu unterscheidenden Typen von Alternationsregeln und den Kriterien ihrer Abgrenzung voneinander thematisiert. Damit wird dann auch wieder die Problematik der Genese und Entwicklung von Alternationsregeln theoretisch relevant. So entsteht eine ganze Anzahl von Arbeiten, die sich unter unterschiedlichen Gesichtspunkten mit der Morphologisierung auseinandersetzen. Dabei steht meist die Frage nach den Triebkräften von Morphologisierungsprozessen im Mittelpunkt. Unmittelbaren Bezug auf die Schwachstellen der generativen Phonologie nimmt Skousen (1975). Bei der Behandlung der finnischen Konsonantengradation kommt er zu dem Ergebnis, daß diese durch (vermeintliche) phonologische Regeln und die entsprechenden abstrakten Repräsentationen nicht in auch nur einigermaßen angemessener Weise erfaßt werden kann, denn ihre Domäne ist im modernen Finnischen durch nichtphonologische, nämlich morphologische Bedingungen eingeschränkt bzw. ausgeweitet. Die ursprüngliche phonologische Regel ist im Laufe der Sprachgeschichte zu einer morphonologischen Alternation geworden. “The motivating force” für die Morphologisierung von Regeln ist “a desire for surface regularity”, das Bestreben nach morphologischer Regularität (Skousen 1975: 125). Hooper (1976) hebt die semiotische Seite von Morphologisierungsprozessen hervor: “After a phonetic alternation appears in the language, the tendency is for this alternation to walk its way up towards the meaning end of the grammar, moving from a purely phonetic function to a semantic function” (Hooper 1976: 86). Hier setzt dann auch Dressler (1977) an, der Morphologisierungen generell damit erklärt, daß “morphologische Funktionen semiotisch wichtiger sind als phonologische” und demgemäß in entsprechenden Fällen “die morphologische Funktion [...] immer wichtiger, die phonologische Funktion hingegen immer unwichtiger wird” (Dressler 1977: 24). Klausenburger (1979) macht auf das Faktum aufmerksam, daß phonologische Regeln nur dann morphologisiert werden können, wenn sie über ein “preexisting morphological conditioning” verfügen (Klausenburger 1979: 32), d. h. wenn die Alternationen von Anfang an (zufälligerweise) an morphologische Kategorien gebunden sind. Alle diese Arbeiten haben gemeinsam, daß Morphonologisierungen und Morphologisierun-
147. Morphologisierung: von der Phonologie zur Morphologie
gen ausschließlich als durch morphologische Triebkräfte bedingt angesehen werden. Einen gegensätzlichen Standpunkt vertritt Robinson (1975) bei einer historischen Analyse des deutschen Umlauts. Er setzt sich (ähnlich wie Skousen 1975 hinsichtlich der finnischen Gradation) mit der abstrakt-generativen Behandlung des neuhochdeutschen Umlauts als phonologischer Regel auseinander und wertet diesen (Wurzel 1970 folgend) als morphologisch bedingt. Die Morphologisierung der phonologischen Umlautregel findet seiner Meinung nach einheitlich zu dem Zeitpunkt statt, zu dem die phonologischen Umlautfaktoren durch phonologische Tilgungen und Neutralisierungen abgebaut worden sind (Robinson 1975: 4 f.). Die Morphologisierung des deutschen Umlauts ist damit ausschließlich phonologisch bedingt, was im übrigen der gängigen germanistischen Lehrmeinung entspricht (vgl. jedoch dazu 3.1). Des weiteren wird angenommen, daß im Fall des deutschen Umlauts die Morphologisierung ohne morphonologische Zwischenstufe erfolgt. In Wurzel (1980) wird dann anhand von einschlägigen Beispielfällen gezeigt, daß sowohl morphologisch als auch phonologisch ausgelöste Morphologisierungen anzunehmen sind. Dabei wird auch die oft sehr große innere Komplexität von scheinbar einfachen Morphologisierungsprozessen im Spannungsfeld von Phonologie und Morphologie herausgearbeitet. Schließlich wird eine Klassifizierung von Morphologisierungen nach den Parametern der Bedingtheit, des Typs der Ausgangsregel, der Anzahl der Schritte im Gesamtprozeß, der Anzahl der Ausgangsregeln und der der resultierenden Regeln sowie des Auftretens bzw. Nichtauftretens von Regelinversion vorgeschlagen (Wurzel 1980: 457 ff.). Die Beiträge zur Morphologisierung aus den achtziger und neunziger Jahren thematisieren speziell die Unidirektionalität von Morphologisierungsprozessen und deren Erklärbarkeit. Dressler (1985) ordnet die Morphologisierung in einen semiotischen Theorierahmen ein: Wörter sind primäre sprachliche Zeichen; Morpheme und morphologische Regeln sind Zeichen von Zeichen (den Wörtern), d. h. damit sekundäre Zeichen; und Phoneme sind Zeichen von Zeichen (den Morphemen) von Zeichen (den Wörtern), d. h. tertiäre Zeichen. Damit hat das Lexikon semiotische und epistemologische Priorität gegenüber der Morphologie und die Morphologie gegenüber der Phonologie (Dressler
1603
1985: 283). Das bedeutet, daß auch morphologische Indexikalität höher zu bewerten ist als phonologische Indexikalität. Wenn man weiter berücksichtigt, daß (grammatisch bedingter) Sprachwandel zu effizienteren grammatischen Strukturen von Subsystemen führt, so ergibt sich daraus die Tendenz des Übergangs von phonologischen Regeln zu morphonologischen Regeln und weiter zu morphologischen Regeln. Die umgekehrte Entwicklungsrichtung ist dagegen nicht möglich (Dressler 1985: 311). Diese auf Kruszewskis Konzept zurückgehende Gerichtetheit der Übergänge zwischen den drei Regeltypen kann heute weitgehend als allgemein akzeptiert gelten. Anders argumentieren jedoch Morin et al. (1990), daß im Fall der Herausbildung des französischen ‘oTensing’ im Auslaut die eingetretene Sprachveränderung überhaupt nur dann erklärbar ist, wenn man annimmt, daß morphonologische Regeln ‘rephonologisiert’ werden können, was bedeuten würde, daß die Unidirektionalität unhaltbar wäre. Klausenburger (1994) arbeitet jedoch heraus, daß in diesem Fall nicht eine (ohnehin fragwürdige) morphonologische Regel, sondern ein phonetischer bzw. phonotaktischer Sachverhalt der französischen Aussprache phonologisch generalisiert wird (Klausenburger 1994: 2563 f.). Um zu ermitteln, ob ein ‘Weg zurück’ unter entsprechenden Bedingungen möglich ist, überprüft er Fälle aus der romanischen Sprachgeschichte, in denen die Sprecher in ihrer Sprache gegebene Alternationen prinzipiell entweder phonologisch oder morphologisch werten könnten. Der weitere Verlauf der Entwicklung erweist in allen Fällen, daß jeweils die morphologischen Bedingungen für die Alternation gewählt wurden. Das gleiche Ergebnis zeigen auch Experimente, in denen sowohl phonologisch als auch morphologisch interpretierbare Vokalalternationen in spanischen Nonsens-Verben von nativen Sprechern eindeutig an morphologischen Kategorien festgemacht wurden (Bybee 1985: 66 ff.). Der Weg von einer morphonologischen Regel zu einer phonologischen Regel, der der semiotischen Priorität der Morphologie gegenüber der Phonologie widerspricht, ist aller vorliegenden Evidenz nach ausgeschlossen.
3.
Parameter der Morphologisierung
Alle Morphologisierungen sind (per definitionem) durch den Übergang von phonologischen zu morphologischen Regeln gekenn-
1604
XVII. Morphologischer Wandel I: Theoretische Probleme
zeichnet, doch darüber hinaus haben sie kaum durchgängige Gemeinsamkeiten. Die vorkommenden Morphologisierungen unterscheiden sich in einer ganzen Reihe von relevanten Eigenschaften, die zudem auch noch jeweils weitgehend unabhängig voneinander sind. Eine wirklich erschöpfende Typologisierung von Morphologisierungen stellt sich unter diesen Bedingungen als sehr kompliziert dar und wäre nur im Rahmen einer umfassenden Darstellung anhand einer großen Anzahl von Beispielen möglich. Hier können demgegenüber nur einige wenige einschlägige Beispielfälle vorgestellt und aufgrund der wichtigsten Parameter der Morphologisierung zueinander in Bezug gesetzt werden. Diese Parameter sollen sein: die Initiierung der Morphologisierung, der Weg von der phonologischen zur morphologischen Regel, der Typ der resultierenden morphologischen Regel, die Ableitungsrichtung der morphologischen Regel und der Status des neuen Kategorienmarkers. 3.1. Die Initiierung von Morphologisierungen: phonologisch versus morphologisch Jede Morphologisierung wird durch einen Sprachwandel ausgelöst. Das kann entweder ein phonologischer oder ein morphologischer Wandel sein, womit sich ein erster Klassifizierungsparameter für Morphologisierungen ergibt. Vgl. dazu zunächst die folgenden Beispielfälle. Beispiel 1: Im Althochdeutschen zeigen starke Verben mit einem hinteren Stammvokal im Infinitiv wie graban ‘graben’ bedingt durch die phonologische Umlautregel im Präsens Indikativ entsprechende Alternationen: 1. Sg. grabu, 2. Sg. grebis, 3. Sg. grebit; 1. Pl. grabe¯n, 2. Pl. grabet, 3. Pl. grabent. Vor einem [i] der Folgesilbe erscheint der umgelautete Vokal, sonst der unumgelautete. Beim Übergang zum Mittelhochdeutschen tritt ein phonologischer Wandel ein, durch den die unbetonten Endsilbenvokale zu [e] (phonologisch /e/) neutralisiert werden; vgl. Infinitiv graben; 1. Sg. grabe, 2. Sg. grebest, 3. Sg. grebet; 1. Pl. graben, 2. Pl. grabet, 3. Pl. grabent. Der Umlaut hat seine phonologische Bedingtheit verloren (vgl. 3. Sg. grebet vs. 2. Pl. grabet). Er tritt in den Präsensparadigmen der starken Verben jetzt im morphologischen Kontext ‘2./3. Sg. Präs. Ind.’ auf und ist damit zum Kategorienmarker geworden. An die Stelle der phonologischen Umlautregel ist bei
diesen Verben eine morphologische Umlautregel getreten, die noch heute gilt; vgl. ich grabe, du gräbst usw. Beispiel 2: Einen parallelen Ausgangspunkt (wenn auch einen anderen Verlauf) hat eine Morphologisierung im Nordischen. Hier alternieren im späteren Urnordischen entsprechend einer phonologischen i-Umlautregel nichtumgelautete und umgelautete Flexionsformen im Präsens Indikativ der langsilbigen starken Verben auf gleiche Weise wie im Althochdeutschen; vgl. Infinitiv blo¯ta(n) ‘opfern’; 1. Sg. blo¯tu, 2. Sg. blø¯tiR, 3. Sg. blø¯tiR; 1. Pl. blo¯tum usw. Die Morphologisierung beginnt in diesem Fall mit einem morphologischen Wandel, der Systematisierung des Verhältnisses zwischen Funktion und Form im Paradigma. Die 1. Person Singular wird im Vokalismus an die beiden anderen Singularformen angeglichen; es ergibt sich das Teilparadigma: Inf. blo¯ta(n); 1. Sg. blø¯tu, 2. Sg. blø¯tiR, 3. Sg. blø¯tiR; 1. Pl. blo¯tum, 2. Pl. blo¯teÎ, 3. Pl. blo¯ta(n). Die Umlautregel gilt weiterhin für die Vokale vor [i], aber zusätzlich dazu morphologisch bedingt für den Vokal der 1. Person Singular. Damit erfaßt die Alternation durch Ausdehnung ihrer Domäne außer den i-Instanzen anderer Paradigmen jetzt auch den gesamten Singular Präsens Indikativ der Verben des Typs blo¯ta(n). Aufgrund von phonologischen Tilgungen verschwindet dann später das (ursprünglich) umlautbewirkende [i]; vgl. das resultierende altisländische Paradigma: Inf. blo´ta; 1. Sg. blø´t, 2. Sg. blø´tr, 3. Sg. blø´tr; 1. Pl. blo´tum usw. Die Alternation ist für diese Verben zu einer morphologischen Regel geworden. Dieser schließen sich dann auch die kurzsilbigen starken Verben an, vgl. taka ‘nehmen’, tek, tekr usw., wodurch die eine morphologische Regel belastende phonologische Einschränkung der Langsilbigkeit verschwindet. Beispiel 3: Im frühen Althochdeutschen zeigen die Maskulina der i-Deklination wie gast ‘Gast’ die folgenden Flexionsformen: Sg. Nom./Akk. gast, Gen. gastes, Dat. gaste, Instr. gestiu; Pl. Nom./Akk. gesti, Gen. gestio, Dat. gestim. Die Alternation ist bedingt durch die erwähnte phonologische Umlautregel. Im späteren Althochdeutschen tritt dann in den entsprechenden Paradigmen ein ‘analogischer’ Ausgleich ein, durch den bei weiterhin vorhandenem [i] die Instrumental Singular Form gestiu durch gastiu ersetzt wird. Die Domäne der phonologischen Umlautregel wird morphologisch eingeschränkt; der
147. Morphologisierung: von der Phonologie zur Morphologie
Umlaut tritt nur noch, und zwar durchgängig, im Plural auf. Beim Übergang zum Mittelhochdeutschen werden auch hier die Vokale der Flexionsendungen zu [e] reduziert. Des weiteren verschwindet die Kategorie des Instrumentals aus dem Flexionssystem. Das Substantiv gast flektiert dann in folgender Weise: Sg. Nom./Akk. gast, Gen. gastes, Dat. gaste; Pl. Nom./Akk./Gen. geste, Dat. gesten. Die phonologische Umlautregel ist durch eine morphologische Regel ersetzt worden, die im Kontext der grammatischen Kategorie Plural den Vokal verändert, wenn das Wort zur Umlautklasse gehört. Diese Klasse ist durch ein Flexionsmerkmal spezifiziert, das diese von der Klasse der sonst gleich flektierenden nichtumlautenden starken Maskulina des Typs tac ‘Tag’ unterscheidet; vgl. gast ⫺ Nom. Pl. geste vs. tac ⫺ Nom. Pl. tage. Im ersten Beispiel wird die Morphologisierung also durch einen phonologischen Wandel ausgelöst, in den beiden anderen Beispielen durch einen morphologischen Wandel. Das bedeutet jedoch nicht, daß diese Morphologisierungen jeweils ausschließlich phonologisch bzw. ausschließlich morphologisch bedingt sind. In allen drei betrachteten Fällen wirken phonologische und morphologische Bedingungen zusammen. Im Beispiel 1 folgt dem phonologischen Abbau des umlautbewirkenden [i] eine morphologische Reanalyse, indem das Auftreten des Umlautvokals mit den entsprechenden grammatischen Kategorien verbunden wird. Und in den Beispielen 2 und 3 werden nach dem jeweils auslösenden morphologischen Wandel die umlautbewirkenden Vokale durch phonologische Reduzierungen bzw. Tilgungen abgebaut. Da weder eine phonologisch ausgelöste Morphologisierung ohne morphologische Reanalyse, noch eine morphologisch ausgelöste Morphologisierung ohne phonologische Tilgung des ursprünglichen phonologischen Kontextes vorstellbar ist, ergibt sich die Schlußfolgerung, daß alle Morphologisierungen zugleich phonologisch und morphologisch bedingt sind. Deshalb soll hier nicht (wie häufig) zwischen phonologisch und morphologisch bedingten, sondern genauer zwischen phonologisch und morphologisch initiierten Morphologisierungen unterschieden werden. Phonologisch initiierte Morphologisierungen kommen morphologisch gesehen zufällig zustande; die phonologischen Bedingungen der Alternation verschwinden und werden notgedrungen durch morphologische Bedin-
1605
gungen, d. h. grammatische Kategorien, ersetzt. Demgegenüber werden morphologisch initiierte Morphologisierungen von den Sprechern ‘gezielt’ in die Wege geleitet. Eine Alternation wird trotz ihrer fortbestehenden phonologischen Bedingtheit von den Sprechern als zusätzlicher Marker derjenigen Kategorie oder Kategorien reanalysiert, in der oder denen sie typischerweise auftritt. Das morphologisch unsystematische Nichtauftreten der Alternation wie im Fall des urnordischen 1. Sg. Präs. Ind. blo¯tu bzw. ihr unsystematisches Auftreten wie im Fall des althochdeutschen Instr. Sg. gestiu wird durch Ausdehnung bzw. Einschränkung der Domäne der Regel beseitigt. Bei den morphologisch initiierten Morphologisierungen zeigt sich damit deutlich die (von verschiedenen Linguisten konstatierte) semiotische Priorität der Morphologie gegenüber der Phonologie. Zu ergänzen ist, daß beide Typen der Morphologisierung von phonologischen Regeln die im System vorhandenen Flexionskategorien voraussetzen. Durch sie entstehen neue Marker für bereits gegebene Kategorien, aber (anders als bei der Morphologisierung syntaktischer Marker) niemals neue Flexionskategorien. 3.2. Der Weg von der phonologischen zur morphologischen Regel: direkt vs. indirekt Im Rahmen von Morphologisierungsprozessen sind (wie gesagt) drei Klassen von Regeln zu unterscheiden. Phonologische Regeln nehmen formale Operationen im Kontext phonologischer Einheiten wie Merkmale, Segmente und Silben vor. Sie enthalten grundsätzlich keine grammatischen Kategorienmerkmale wie ‘Plural’, ‘Genitiv’, ‘Präsens’ usw. und keine Flexionsklassenmerkmale. Morphologische Regeln vollziehen demgegenüber solche Operationen im Kontext von grammatischen Kategorien, in Sprachen mit Flexionsklassen bezogen auf diese. Morphonologische Regeln nehmen formale Operationen in einem Kontext vor, der sowohl phonologische Einheiten als auch grammatische Kategorien enthält. Ein Blick auf die Beispiele zeigt, daß im Prozeß der Morphologisierung der Übergang von phonologischen zu morphologischen Regeln entweder direkt oder indirekt, d. h. über eine morphonologische Regel, erfolgen kann. Im Beispiel 1 vollzieht sich die Morphologisierung direkt. Aufgrund des phonologischen Abbaus sämtlicher umlautbewirkender [i]-Instanzen im Paradigma werden diese voll-
1606
XVII. Morphologischer Wandel I: Theoretische Probleme
ständig durch grammatische Merkmale ersetzt. Anders im Beispiel 3. Hier wird der im Prinzip weiter funktionierende phonologische Umlaut für die Instrumental Singular Formen von Substantiven des Typs gast blokkiert. Der althochdeutsche Umlaut erfährt darüber hinaus noch eine morphologisch bedingte Blockierung, d. h. eine weitere Morphonologisierung. Beispiel 4: Im älteren Althochdeutschen zeigen die schwachen Maskulina wie z. B. hano ‘Hahn’ die folgenden Flexionsformen: Sg. Nom. hano, Gen./Dat. henin, Akk. hanun; Pl. Nom./Akk. hanun, Gen. hano¯no, Dat. hano¯m. Vor folgendem [i] tritt also ‘lautgesetzlich’ Umlaut auf. Dieser Umlaut im Genitiv und Dativ Singular wird schon sehr früh durch einen morphologischen Wandel beseitigt; henin wird durch hanin ersetzt, wodurch im Paradigma jetzt durchgängig der unumgelautete Vokal erscheint. Beim Übergang zum Mittelhochdeutschen wird auch das [i] des Markers -in phonologisch zu [e] reduziert. Die Flexionsformen des Paradigmas lauten jetzt Sg. Nom. hane, Gen./Dat./Akk. hanen; Pl. hanen. Die gemeinsame Zwischenstufe dieser Entwicklungen, die spätalthochdeutsche Umlautregel, läßt sich damit wie folgt skizzieren:
Ï Mask ¸ (a) Ô ⬃ i-Dekl Ô Ô -Pl Ô V J [-hinten] / Ì ˝ Ô ⬃ [ n-Dekl ] Ô (b) Ô Ô Ó ___ K1 i ˛ (c) Abb. 147.1: Spätalthochdeutsche Umlautregel
Diese Regel ist eine morphonologische Regel; ihre Anwendung ist partiell (noch) phonologisch und partiell (schon) morphologisch bedingt. Die morphologischen und phonologischen Bedingungen der Alternation lassen sich hier nicht unabhängig voneinander formulieren. Die Teilregeln sind extrinsisch geordnet, denn die blockierenden morphologischen Teilregeln (a) und (b) müssen der phonologischen Teilregel (c) vorausgehen. Im Mittelhochdeutschen ist dann der phonologische Kontext der Regel abgebaut. Die Teilregel (a), die den Vokalwechsel im Paradigma gast systematisiert, findet ihre Entsprechung in einer morphologischen Regel, die dann den Vokal im inversen (positiven) Kategorienkontext ‘i-Dekl ⫹ Pl’ umlautet. Die ursprüngliche phonologische Regel ist auf indi-
rektem Wege morphologisiert worden. Die Teilregel (b), die den Wechsel im gesamten Paradigma hano blockiert, verschwindet aus der Grammatik, da sie keine Kategorienmerkmale enthält. In solchen Fällen tritt ein morphologisch bedingter Ausgleich im Paradigma (eine ‘Assimilation’ in Kruszewskis Sinne), aber keine Morphologisierung einer Alternation ein. Nicht jede Morphonologisierung führt also zur Morphologisierung. Auch im Beispiel 2 erfolgt die Morphologisierung über eine morphonologische Zwischenstufe. Die zunächst nur vor einem [i] der Folgesilbe eintretende Alternation wird auf die Umgebung ‘Stark, ⫺ Prät, ⫺ Konj, ⫺ Pl’ ausgedehnt; sie folgt wiederum teils (schon) morphologischen, teils (noch) phonologischen Bedingungen. Die morphologische und die phonologische Teilregel sind zwar nicht untrennbar miteinander verzahnt wie im vorangehenden Beispiel (da die morphologische Teilregel einen positiven Kontext hat, ließen sich die beiden Teilregeln mit etwas größerem Aufwand auch als selbständige Regeln formulieren und zeigen keine extrinsische Ordnung), doch sie nehmen die gleiche formale Operation vor und haben sich überschneidende Anwendungsbereiche. Mit der Tilgung der [i]-Instanzen im Paradigma verschwindet die phonologische Teilregel, übrig bleibt eine morphologische Regel. Die diskutierten (wie auch viele weitere) Beispiele scheinen dafür zu sprechen, daß phonologisch initiierte Morphologisierungen direkt zu morphologischen Regeln führen (Beispiel 1), während morphologisch initiierte (Beispiele 2 und 3) den Weg über morphonologische Regeln nehmen. Doch vgl. dazu den folgenden Fall: Beispiel 5: Bedingt durch die phonologische u-Umlautregel tritt im späteren Urnordischen u. a. auch in den Paradigmen der maskulinen und neutralen a-Substantive eine Alternation zwischen [a] und [c] (o˜ ) auf. Vor einem [u] im Flexiv erscheint im Stamm [c], sonst [a]. Vgl. z. B. Maskulinum armaR ‘Arm’; Sg. Nom. armaR, Gen. armas, Dat. arme¯, Akk. arma; Pl. Nom. armo¯R, Gen. armo¯, Dat. o˛rmumR, Akk. arma(n) und Neutr. landa ‘Land’; Sg. Nom./Akk. landa, Gen. landas, Dat. lande¯; Pl. Nom./Akk. lo˛ndu, Gen. lando¯, Dat. lo˛ndumR. In der Entwicklung zum Altisländischen werden die Flexionsendungen aufgrund von eindeutig phonologischen Veränderungen getilgt bzw. reduziert, so daß sich die Formen Sg. Nom.
147. Morphologisierung: von der Phonologie zur Morphologie
1607
armr, Gen. arms, Dat. armi, Akk. arm; Pl. Nom. armar, Gen. arma, Dat. o˛rmum, Akk. arma und Sg. Nom./Akk. land, Gen. landas, Dat. landi; Pl. Nom./Akk. lo˛nd, Gen. lando, Dat. lo˛ndum ergeben. Während der Umlaut im Dativ Plural bei erhaltenem [u] weiterhin phonologisch bedingt ist, wird er (notgedrungen) im Nominativ und Akkusativ Plural der Neutra mit dem entsprechenden morphologischen Kategorienkontext assoziiert (ähnlich übrigens auch in anderen Paradigmen). Der u-Umlaut ist im Altisländischen damit zu einer morphonologischen Regel geworden. Die weitere Entwicklung ist dadurch gekennzeichnet, daß zum Neuisländischen hin durch phonologische Epenthese neue [u]-Instanzen in den Flexiven entstehen, vgl. altisländisch armr > neuisländisch armur, wobei (erwartungsgemäß) kein Umlaut auftritt. Es erfolgt weiterhin der Übergang von [u] zu [y] (orthographisch weiter u) und von [c] (o˛) zu [ø] (ö). Umlaut erscheint also in den Paradigmen nicht mehr im phonologischen Kontext ‘vor [y]’, sondern im morphologischen Kontext ‘vor dem Dat. Pl. Flexiv -um [ym]’; vgl. Dativ Plural örmum und löndum, aber Nominativ Singular armur. Daß hier tatsächlich der Kategorienkontext entscheidend ist, zeigen auch Neubildungen wie 4 arum ‘darüber’ (aus 4 ar ‘dort’ und um ‘um, über’), die ebenfalls umlautlos bleiben. Damit ist auch die phonologische Teilregel der bisherigen morphonologischen Regel morphologisiert worden. Im Laufe des Morphologisierungsprozesses sind an die Stelle der ursprünglichen phonologischen u-Umlautregel u. a. (die Regel unterliegt auch weiteren Morphologisierungen) zwei für die betrachteten Paradigmen geltende morphologische Regeln getreten, eine für den Nominativ/Akkusativ Singular Neutrum, vgl. lönd, und eine für den Dativ Plural beider Genera, vgl. örmum, löndum (für De´ rnason 1985). tails vgl. A
Beispiel 6: Diese bedingt in den Paradigmen der ir-Stämme des Typs hrind ‘Rind’ Alternationen zwischen [ir] und Ø der folgenden Art: Sg. Nom./Akk. hrind, Gen. hrind-ir-as, Dat. hrind-ir-a, Instr. hrind-ir-u; Pl. Nom./Akk. hrind-ir-u, Gen. hrind-ir-o, Dat. hrind-ir-um. Später tritt ein phonologischer Wandel ein, durch den ein auslautendes [u] unter bestimmten Bedingungen beseitigt wird. Die frühalthochdeutsche Entsprechung des Paradigmas sieht dann so aus: Sg. Nom./Akk. hrind, Gen. hrind-ir-es, Dat. hrind-ir-e, Instr. hrind-ir-u (hier ist die Erhaltung des [u] durch Ausgleich bedingt); Pl. Nom./Akk. hrind-ir, Gen. hrind-ir-o, Dat. hrind-ir-um. Es ergeben sich also Instanzen von [ir] im Auslaut; das Nichtvorkommen bzw. Vorkommen von [ir] im Paradigma ist nicht mehr phonologisch faßbar und wird statt dessen an die grammatischen Kategorien gebunden. Da das Element [ir] nicht in der Nominativ Singular Form des Paradigmas, wohl aber in den meisten abgeleiteten Formen auftritt, wird es als ein Flexiv interpretiert, das in den Kategorien des Genitivs, Dativs und Instrumentals Singular und des Plurals durch eine morphologische Regel eingeführt wird. Die phonologische Regel ist durch eine morphologische Regel ersetzt worden. In einem weiteren Schritt wird [ir] im Verlauf des Althochdeutschen dann aufgrund seiner Verteilung im Paradigma (sowie weiterer Faktoren; vgl. Wurzel 1980: 445 ff.; 1992: 282 ff., 291 f.) als Pluralmarker reanalysiert und entsprechend in seinen Singularvorkommen beseitigt, vgl. ‘normalahd.’ Sg. Nom./Akk. hrind, Gen. hrind-es, Dat. hrind-e, Instr. hrind-u; Pl. Nom./Akk. hrind-ir, Gen. hrind-ir-o, Dat. hrind-ir-um. Die unsystematische morphologische Regel wird so in eine Pluralregel umgewandelt, die als er-Pluralregel bekanntlich noch heute existiert; vgl. Rind ⫺ Rinder.
Hier führt eine phonologisch initiierte Morphologisierung anders als im Beispiel 1 zunächst zu einer morphonologischen Regel. Dieser Unterschied ergibt sich daraus, daß durch den auslösenden phonologischen Wandel die umlautbewirkenden [u]-Instanzen nicht vollständig, sondern nur partiell getilgt werden. Es läßt sich somit generalisieren: Phonologisch initiierte Morphologisierungen, bei denen im ersten Schritt sämtliche bedingenden phonologischen Kontexte abgebaut werden, führen direkt zu morphologischen Regeln, alle anderen Morphologisierungen dagegen über morphonologische Regeln.
Hier ist das Ergebnis der Morphologisierung einer phonologischen Regel keine modifikatorische, sondern eine additive morphologische Regel. Für diese Entwicklung gibt es zwei Gründe: Erstens ist die Ausgangsregel eine phonologische Tilgungsregel und keine Alternationsregel, und zweitens operiert sie am Wortende. Daraus läßt sich verallgemeinernd der Schluß ziehen, daß die Morphologisierung von phonologischen Regeln, die am Wortende phonologische Substanz tilgen oder einführen, zu additiven morphologischen Regeln führt. Das gilt in dieser Form natürlich nur für Suffixsprachen; es ist jedoch
1608
XVII. Morphologischer Wandel I: Theoretische Probleme
zu erwarten, daß das gleiche auf am Wortanfang operierende Tilgungs- bzw. Epentheseregeln in Präfixsprachen zutrifft. 3.3. Die Ableitungsrichtung der resultierenden morphologischen Regel: Beibehaltung vs. Inversion Bei Morphologisierungen von phonologischen Regeln kann die Ableitungsrichtung der Regeln beibehalten oder umgekehrt werden. In den Beispielfällen 1, 2, 3 und 5 bleibt die Ableitungsrichtung auch nach der Morphologisierung die gleiche. Unumgelautete Vokale werden in bestimmten Kontexten in umgelautete Vokale überführt. Anders im Beispiel 6. Hier wird bei der Morphologisierung eine phonologische Regel mit der Operation ‘ir J Ø’ durch eine morphologische Regel mit der inversen Operation ‘Ø J ir’ ersetzt. Auch über die Beispiele hinaus ist die Beibehaltung der Ableitungsrichtung weitaus häufiger als ihre Inversion, was darauf hindeutet, daß die Inversion spezifische Bedingungen voraussetzt, was durch einen weiteren Fall von Regelinversion, der Morphologisierung des sogenannten Rückumlauts im Deutschen, bestätigt wird: Beispiel 7: In einer Teilklasse der schwachen Verben stehen sich im frühen Althochdeutschen bedingt durch die phonologische Umlautregel umgelautete Infinitiv- und Präsensformen einerseits und unumgelautete Präteritalformen andrerseits gegenüber. Vgl. z. B. Infinitiv brennjan ‘brennen’, 1. Sg. Präs. Ind. brennju, 2. Sg. Präs. Ind. brenis usw. ⫺ 1. Sg. Prät. Ind. branta, 2. Sg. Prät. Ind. branto¯s usw. und Infinitiv stelljan ‘stellen’, 1. Sg. Präs. Ind. stellju, 2. Sg. Präs. Ind. stelis usw. ⫺ 1. Sg. Prät. Ind. stalta, 2. Sg. Präs. Ind. stalto¯s usw. (Die Geminaten sind dabei durch das folgende unsilbische [i4] hervorgerufen.) In der weiteren Entwicklung wird durch phonologischen Wandel das [i4] getilgt, wobei die Infinitivendung in [en] übergeht. Es ergeben sich die Formen brennen, brennu, brenis ⫺ branta, branto¯s bzw. stellen, stellu, stelis ⫺ stalta, stalto¯s. Das Auftreten des Umlautvokals [e] im Infinitiv und in den meisten Präsensformen (allen außer der 2./3. Person Singular) ist nicht mehr phonologisch faßbar. Da bei diesen Verben im Infinitiv und im Präsens durchgängig umgelautete Vokale und im Präteritum durchgängig unumgelautete Vokale erscheinen, wird der Vokalwechsel (als eine Art ‘neuer Ablaut’) mit diesen Kategorien assoziiert und damit morphologisiert.
Dabei wird als Ausgangsvokal nicht das unumgelautete [a] des Präteritums, sondern das umgelautete [e] des Infinitivs und des Präsens gewählt; die Ableitungsrichtung der Regel wird also von ‘a J e’ zu ‘e J a’ umgekehrt. Daß die Annahme der Inversion korrekt ist, zeigt der spätere weitgehende Abbau der Vokalalternation, der stets zugunsten des Präsensvokals erfolgt; vgl. stellen ⫺ stellte. Weshalb tritt nun in den beiden letzten Beispielen eine Inversion der Ableitungsrichtung ein? Die anderen Morphologisierungen betrafen jeweils phonologische Regeln, die innerhalb des Paradigmas bestimmte abgeleitete Flexionsformen, aber nicht die Grundform erfaßten; vgl. z. B. ahd. graban ⫺ 2. Sg. Präs. Ind. grebis und urnord. landa ⫺ Nom. Pl. lo˛ndu. Die formale Ableitungsrichtung der Regeln entspricht der grammatischen Ableitung der Kategorien in den Paradigmen. In solchen Fällen gibt es keinen Grund zur Inversion; die Ableitungsrichtung der Regel bleibt auch nach der Morphologisierung erhalten. Dagegen sind die beiden hier relevanten Beispiele dadurch charakterisiert, daß jeweils eine phonologische Regel morphologisiert wird, die gerade die Grundform des Paradigmas, aber nicht bestimmte abgeleitete Flexionsformen erfaßt; vgl. hrind mit ir-Tilgung, aber Nom. Pl. hrindiru ohne ir-Tilgung sowie Infinitiv brennjan mit Umlaut, aber Präteritum branta ohne Umlaut. Die formale Ableitungsrichtung der Regel ist der grammatischen Ableitung der Kategorien in den Paradigmen gegenläufig. In solchen Fällen wird bei Verlust der phonologischen Bedingungen (zumindest normalerweise) die Ableitungsrichtung der Regel umgekehrt, so daß dann die in der Grundform auftretende Alternante die Eingabe der neuen morphologischen Regel darstellt; vgl. Nom. Sg. hrind ohne ir-Einführung, aber Nom. Pl. hrindir mit ir-Einführung sowie Infinitiv brennen ohne Vokalwechsel, aber Präteritum branta mit Vokalwechsel ‘e J a’. Damit spezifiziert die morphologisierte Regel dann ⫺ wie in den Fällen ohne Inversion ⫺ einen Marker für eine abgeleitete Kategorie, in den Beispielen den Pluralmarker [ir] bzw. den Präteritalmarker ‘Vokalwechsel’. Daß Inversionen bei Morphologisierungen relativ seltener vorkommen als die Beibehaltung der Ableitungsrichtung resultiert einfach daraus, daß phonologische Alternationen typischerweise durch Segmente in Flexiven hervorgerufen werden, und daß in abgeleiteten (markierten)
147. Morphologisierung: von der Phonologie zur Morphologie
Kategorien eher Flexive auftreten als in (unmarkierten) Basiskategorien (Greenberg 1963: 74 f.). Man vgl. dazu noch einmal die [i]- und [u]-haltigen Flexive des Germanischen. 3.4. Der Status des neuen Kategorienmarkers: Hauptmarker vs. Nebenmarker Die Kategorienmarker der Flexionsformen haben nicht alle den gleichen Status. Ein Marker kann die einzige Symbolisierung (Hund ⫺ Hund-e, Vater ⫺ Väter), die primäre Symbolisierung (Wolf ⫺ Wölf-e) oder eine sekundäre Symbolisierung (Wolf ⫺ Wölf-e) einer Kategorie am Wort sein. In den ersten beiden Fällen liegt ein Hauptmarker, im letzten Fall ein Nebenmarker der entsprechenden Kategorie vor. Welcher Marker bei Mehrfachsymbolisierung den Status des Hauptmarkers hat, ergibt sich aus der Systematik des Sprachsystems. Wenn, was für Sprachen wie das Deutsche typisch ist, eine Kategorie durch einen additiven und einen modifikatorischen Marker symbolisiert wird, so ist im allgemeinen der additive Marker der Hauptmarker, weil dieser in allen Wörtern der betreffenden Flexionsklasse vorhanden ist, während der modifikatorische Marker nur eingeschränkt auftritt. Man denke etwa an die Substantive der deutschen er-Pluralklasse wie Kalb ⫺ Kälber und Rind ⫺ Rinder, wo das Suffix bei allen Wörtern, der Umlaut (bedingt durch seine phonologische Genese) aber nur bei Wörtern mit hinterem Stammvokal erscheint. Doch durch Morphologisierungen von phonologischen Regeln können sowohl Hauptmarker als auch Nebenmarker entstehen, wie die Beispielfälle im einzelnen erweisen. Morphologisch initiierte Morphologisierungen wie die Fälle 2 und 3 sind immer dem Bestreben der Sprecher geschuldet, formale Alternationen an die grammatischen Kategorien anzupassen. Im Beispiel 2 wird die 1. Person Singular an die anderen Singularformen angeglichen, vgl. 1. Sg. blo¯tu > blø¯tu wie 2./3. Sg. blø¯tiR; im Beispiel 3 der Instrumental Singular an die übrigen Singularformen, vgl. Instr. Sg. gestiu > gastiu wie Nom./Akk. Sg. gast usw. Im ersten Fall resultiert daraus ein Präsens Singularmarker, im zweiten ein Pluralmarker. Da in beiden Fällen vorher keine einheitlichen Marker für die betreffenden Kategorien existierten, erhalten die neuen Marker jeweils den Status von Hauptmarkern. Hingegen ist die Herausbildung
1609
von Nebenmarkern durch morphologisch initiierte Morphologisierungen nur schwer vorstellbar. Dazu müßte bereits ein einheitlicher Marker für die betreffende Kategorie vorhanden sein. Ein solcher einheitlicher Marker würde sich (zumindest in Fällen, die den hier diskutierten Beispielen vergleichbar sind) aber dann auch in einheitlicher Weise auf die Form des Stammes auswirken. Alternationen zwischen den verschiedenen Formen der Kategorie (etwa den einzelnen Kasusformen des Plurals) und damit morphologisch initiierte Morphologisierungen (im Sinne von Ausgleichen) sind unter solchen Bedingungen überhaupt nicht möglich. Unter den phonologisch initiierten Morphologisierungen sind zwei Gruppen von Fällen zu unterscheiden. Bei der Morphologisierung der Alternationen ahd. 1. Sg. grabu ⫺ 2. Sg. grebis > mhd. 1. Sg. grabe ⫺ 2. Sg. grebest (s. Beispiel 3) und frühahd. Inf. brennjan ⫺ Prät. Ind. 1. Sg. branta > ahd. Inf. brennen ⫺ Prät. Ind. 1. Sg. branta (Beispiel 7) bleiben trotz der phonologischen Reduktionen die alten additiven Marker als Hauptmarker erhalten. Die neuen modifikatorischen Marker erhalten folglich den Status von Nebenmarkern. Davon abweichend verschwinden bei der Entwicklung von vorahd. Nom./Akk. Sg. hrinda ⫺ Nom./Akk. Pl. hrindiru > ahd. Nom./Akk. Sg. hrind ⫺ Nom./ Akk. Pl. hrindir die ursprünglichen Marker, so daß das Element [ir] zum alleinigen Pluralmarker und damit zum Hauptmarker wird. Interessant ist in diesem Zusammenhang das Beispiel 5, die Morphologisierung des u-Umlauts in Paradigmen des Typs urnord. Nom./ Akk. Sg. landa ⫺ Nom./Akk. Pl. lo˛ndu, Dat. Pl. lo˛ndum > nisld. Nom./Akk. Sg. land ⫺ Nom./Akk. Pl. lönd, Dat. Pl. löndum. Hier wird aufgrund der phonologischen Reduzierung der alte Marker des Nominativs und Akkusativs Plural beseitigt; der u-Umlaut wird zum alleinigen Marker, d. h. zum Hauptmarker. Dagegen bleibt im Dativ Plural der alte Marker erhalten, der neue Marker ‘u-Umlaut’ ist entsprechend der Nebenmarker. Es sei noch darauf verwiesen, daß durch Morphologisierung entstandene Marker im Laufe der weiteren Entwicklung unter entsprechenden Bedingungen durchaus Veränderungen in ihrem Status erfahren können, wie das Beispiel 3 anschaulich zeigt: Mit dem Übergang von Wörtern des Typs gast zur mittelhochdeutschen Flexion Sg. Nom./Akk. gast, Gen. gastes, Dat. gaste; Pl. Nom./Akk./
1610
XVII. Morphologischer Wandel I: Theoretische Probleme Initiierung
Weg resultierender phon. Regel > Regeltyp morph. Regel
Ableitungsrichtung
B. 1: mhd. grebest
phonologisch
direkt
modifikatorisch Beibehaltung Nebenmarker
B. 2: urnord. blø¯tu
morphologisch
indirekt
modifikatorisch Beibehaltung Hauptmarker
B. 3: ahd. gastiu
morphologisch
indirekt
modifikatorisch Beibehaltung Hauptmarker ⫺
Markerstatus
B. 4: ahd. hanin
(morphologisch) (indirekt)
⫺
B. 5: isld. lönd; löndum
phonologisch
indirekt
modifikatorisch Beibehaltung Hauptmarker; Nebenmarker
⫺
B. 6: ahd. hrind; hrindir
phonologisch
direkt
additiv
Inversion
Hauptmarker
B. 7: ahd. brennen; branta phonologisch
direkt
modifikatorisch Inversion
Nebenmarker
Tab. 147.1: Übersicht
Gen. geste, Dat. gesten entsteht auf phonologischem Weg ein neuer, einheitlicher additiver Pluralmarker -e [e] (vgl. Dat. Pl. gest-e-n) mit dem Status des Hauptmarkers; der Umlaut wird so zum Nebenmarker. Doch noch während des Mittelhochdeutschen wird der Umlaut u. a. auf Maskulina übertragen, die über keinen Pluralmarker am Wort (mehr) verfügen, vgl. vater ⫺ veter, nagel ⫺ negel. In diesen Wörtern ist der Umlaut damit wieder alleiniger Marker, also Hauptmarker, für den Plural; bei den übrigen Substantiven bleibt er Nebenmarker. 3.5. Zusammenfassung Abschließend soll in Tab. 147.1 noch kurz zusammenfassend dargestellt werden, wie sich die diskutierten Fälle von Morphologisierungen hinsichtlich der einzelnen Parameter verhalten. Auf diese Weise ergibt sich eine Kreuzklassifikation von Morphologisierungsprozessen, die zwar bei weitem kein vollständiges, aber doch wohl ein recht instruktives Bild des Gesamtphänomens der Morphologisierung von phonologischen Regeln und damit des Weges von der Phonologie zur Morphologie bietet.
4.
Zitierte Literatur
´ rnason, Kristja´n (1985), “Morphology, PhonolA ogy and u-umlaut in Modern Icelandic”. In: Gußmann, Edmund (Hrsg.), Phono-Morphology: Studies in the Interaction of Phonology and Morphology. Lublin: Catholic University, 9⫺22 Baudouin de Courtenay, Jan (1895), Versuch einer Theorie phonetischer Alternationen: Ein Capitel aus der Psychophonetik. Straßburg: Trübner [Im Faksi-
mile abgedruckt in: Jan Baudouin de Courtenay (1984), Ausgewählte Werke in deutscher Sprache (hrsg. von J. Mugdan), München: Fink] Baudouin de Courtenay, Jan (1908), “O zwia˛zku wyobrazen´ fonetycznych z wyobrazenami morfologicznemi, syntaktycznemi i semazjologicznemi”. In: Sprawozdania z podziedzen´ Towarzyswta Naukowego Warszawskiego, r. 1, Wydział je˛zykoznawstwa i literatury (z. 5/5), 9⫺28 Baudouin de Courtenay, Jan (1922), Zarys historji je˛siku polskiego. Warszawa: Polska Składnica Pomocy Szkolnych Bybee, Joan (1985), Morphology: A Study of the Relation Between Meaning and Form. Amsterdam: Benjamins Chomsky, Noam & Halle, Morris (1968), The Sound Pattern of English. New York, Evanstone: Harper and Row Dressler, Wolfgang Ulrich (1977), Grundfragen der Morphonologie. Wien: Verlag der Österreichischen Akademie der Wissenschaften Dressler, Wolfgang Ulrich (1985), Morphonology: The Dynamics of Derivation. Ann Arbor: Karoma Greenberg, Joseph H. (1963), “Some Universals of Grammar with Particular Reference to the Order of Meaningful Elements”. In: Greenberg, Joseph H. (Hrsg.), Universals of Language. Cambridge/ MA: MIT Press, 59⫺90 Hooper, Joan B. (1976), An Introduction to Natural Generative Phonology. New York: Academic Press Klausenburger, Jürgen (1979), Morphologization: Studies in Latin and Romance Morphophonology. Tübingen: Niemeyer Klausenburger, Jürgen (1994), “Morphologization”. In: Asher, R. E. & Simpson, J. M. Y. (Hrsg.),
1611
148. Analogical change The Encyclopedia of Language and Linguistics, Bd. V. Oxford, usw.: Pergamon Press, 2562⫺2567
Skousen, Royal (1975), Substantive Evidence in Phonology. The Hague: Mouton
Kruszewski, N[ikolaj] (1881), Ueber die Lautabwechslung. Kasan: Universitätsbuchdruckerei
Wurzel, Wolfgang Ullrich (1970), Studien zur deutschen Lautstruktur. Berlin: Akademie-Verlag (Studia grammatica 8)
Morin, Yves-Charles & Langlois, Marie-Claude & Varin, Marie-Eve (1990), “Tensing of Word-final [c] to [o] in French: The Phonologization of a Morphophonological Rule”. Romance Philology 43, 507⫺528 Mugdan, Joachim (1984), Jan Baudouin de Courtenay (1845⫺1929): Leben und Werk. München: Fink Paul, Hermann (1880), Principien der Sprachgeschichte. Halle: Niemeyer
Wurzel, Wolfgang Ullrich (1980), “Ways of Morphologizing Phonological Rules”. In: Fisiak, Jacek (Hrsg.), Historical Morphology. The Hague: Mouton, 443⫺462 Wurzel, Wolfgang Ullrich (1992), “Morphologische Reanalysen in der Geschichte der deutschen Substantivflexion”. In: Folia Linguistica Historica 8.1⫺ 2, 279⫺307
Wolfgang Ullrich Wurzel †, Berlin (Deutschland)
Robinson, Orrin W. (1975), “Abstract Phonology and the History of Umlaut”. Lingua 37, 1⫺29
148. Analogical change 1. 2. 3. 4. 5. 6.
Introduction Definition and exemplification Previous analyses Optimality Theory Conclusion References
1.
Introduction
Analogical change is a topic which has received relatively little attention in recent handbooks on phonology, morphology, or historical linguistics. The reason for this may well be that it touches on so many different aspects of our linguistic capacity and on different components of linguistic theory that it is not clear where it is most properly treated. This also makes it one of the most interesting topics of investigation, and testing grounds for particular linguistic theories, especially those which seek to unite such different aspects in a unified model. Analogical change is a specific type of sound change, simply because the sounds of a word change in the process. However, it has been assumed for centuries that the motivation for analogical changes is to be found not so much in the phonetic/phonological environment of the sound(s) concerned, but rather that other words in the language to which the changing word is morphologically related, play a role. This is where morphol-
ogy and the organisation of the lexicon come in. Finally, all processes of change must be examined in the light of theories of acquisition and variation, which have now developed to such an extent that we may test them against new sets of data. The sometimes erratic process of analogical change is then test material just like more regular sound changes.
2.
Definition and exemplification
The concept of analogy has been used in various branches of science for ages. In linguistics, the Alexandrian grammarians used it to establish grammatical correctness in Homer. More recently, analogy has been used as a shorthand term for the historical phenomenon of analogical change, and, loosely, as a way of indicating the explanation for such historical language changes (cf. Lehmann 1995). Let us first give a simple example (cited in Joseph 1998; cf. also Meillet & Vendrye`s 1924: 77; Schindler 1974: 3, 6). In Early Latin, the word for ‘small’ was parwos ‘small. nom.sg’, with genitive parwı¯ (see 1a). A process then became active in the language which deleted the rounded semivowel w before round vowels (1b).
1612
XVII. Morphologischer Wandel I: Theoretische Probleme
(1) (a) Early Latin parwos ‘small.nom.sg’ parwı¯ ‘small.gen.sg’ (b) w J Ø /___ [V, ⫹round] (c) Intermediate: paru:s, parvi (d) Classical Latin: parvus, parvı¯ As a result, an alternation between w and Ø was introduced into the paradigm of the word for ‘small’ (1c). By the time of Classical Latin, this alternation was destroyed again, or levelled out, by restoring the original consonant sound in the nominative, resulting in parvus, parvı¯ (1d). The latter change is usually referred to as paradigmatic levelling, as its motivation is generally felt to be a restoration of uniformity in the paradigm of this word.
3.
Previous analyses
The assumption that the cause for analogical levelling is connected to regularity in morphological paradigms dates back at least to the Neogrammarians (cf. Art. 10). Hermann Paul (51920: 205), for instance, concluded that paradigmatic pressure (“Festigkeit des Zusammenhangs der etymologischen Gruppen”), i.e. the pressure to keep a paradigm as uniform as possible, played an important role in this type of change. Note that paradigmatic pressure usually operates in retrospect: it restores regularity after regular phonetic/ phonological changes have created alternations in a paradigm: it does not prevent such irregularities from occurring (cf. Art. 32). For the Neogrammarians, as historical linguists, the linguistic structure of words had very much still a historical outlook: for them a word was defined as its history, and not as its role and function in the synchronic system. This is the reason why the Neogrammarians could not yet gain a full insight into its mode of occurrence. To attain this, comprehension of the Saussurean concept of language as a synchronic system was an indispensable precondition (see Jankowsky 1972: 137). Saussure would also develop the so-called proportional solution to analogical changes, already mentioned by Paul (e.g. Paul 51920: 117), for instance to shed light on another case of analogical levelling, the well known (partial) levelling of the [s] ~ [r] alternation caused by rhotacism in Latin (see Kiparsky 1982 for further discussion, among many other sources):
(2) o¯ra¯to¯rem : o¯ra¯tor ⫽ hono¯rem : x x ⫽ honor Saussure also notes the capricious character of analogy. In German, for instance, the plurals in (3a) are regular i.e. etymological Umlaut formations. The plurals in (3b) have become umlauted analogically. However, the plurals in (3c) have resisted analogy for some reason or another. It is therefore impossible to say in advance how far the imitation of a particular model will extend, or which patterns are destined to provoke it (Saussure 1983: 222). (3) (a) Gast/Gäste ‘guest.sg/pl’ Balg/Bälge ‘skin.sg/pl’ (b) Kranz/Kränze ‘wreath.sg/pl’ (from earlier plural Kranza) Hals/Hälse ‘neck.sg/pl’ (from earlier plural Halsa) (c) Tag/Tage ‘day.sg/pl’ (not *Täge) Salz/Salze ‘salt.sg/pl’ (not *Sälze) Proportional solutions as explanations of analogical changes are insightfully intuitive and easy to work with. Still, this line of thinking about analogical change came under attack in the 1960s, under influence of the generative paradigm (Chomsky & Halle 1968), which of course had to redraw its own position with regard to language change in general and analogy in particular. One of the critics of the proportional model as an explanatory device was King (1969). One question which must be asked with respect to proportional analogies like that in (2) is just how similar the inputs have to be to give rise to a possible analogical proportion. In this case, it makes intuitive sense to “compare” o¯ra¯to¯rem and hono¯rem, but the question remains if satisfactory conditions on possible inputs can be formulated. Hence, such proportions are useless as an explanatory device, i.e. a device which will predict what analogical changes are possible and what are impossible changes. A second piece of criticism was that in some cases different proportions must be drawn up for different words, which seem to be undergoing the same levelling process, as in the case of the spread of the plural s in English: a different proportion must be drawn up for words ending in a vowel, an obstruent or any other consonant. Much effort has therefore gone into introducing more
1613
148. Analogical change
structure into analogical proportions, so that its effects are better understood, primarily already by Kuryłowicz (1949) and Man´czak (1958). These efforts show that, without morphological analysis, analogical proportions are useless as predictive devices. In Chomsky & Halle (1968), grammars were conceived of as lexicons complemented by lists of rules. Language change could therefore be conceived of as changes in lexical entries or in changes of the rules of the language. One important idea of historical generative linguistics was that languages become simpler formally. One way of simplifying grammars is to simplify the rules themselves, e.g. in terms of the number of distinctive features affected by the rule. Another way was to lose rules completely from the grammar. Finally, there was much discussion on the question whether some orderings of rules were simpler than others. Kiparsky (1972) argued that grammars evolved into simplicity and that as a result analogical effects were observed. For instance, the levelling which took place after the rule in (1b) had created an alternation in the paradigm, can be simply formalised as the loss at a later stage of that same rule. Similar effects can be obtained by rule reorderings of different types (see Kiparsky 1972). In more complicated cases of analogical levelling, however, grammar simplification could not be achieved. Take the Latin rhotacism case briefly mentioned above. Because the change from the earlier underlying form /hono:s/ to the later /honor/ was only partial, it could not be expressed as being motivated by the grammar at all, as Kiparsky (1982: 100) notes, although intuitively the change to a uniform paradigm represents a simplification. If the grammar must be complicated to describe processes of analogical levelling, then we must surely be on the wrong track. In response to this, Kiparsky (1971) proposed that functional constraints should also help to establish whether a grammar was simpler than an older grammar. To this effect he proposed the constraint given in (4): (4) Paradigmatic Uniformity (Kiparsky 1971: 598) Allomorphy tends to be minimized in a paradigm. This constraint evaluates the output of a grammar as a whole, and is in effect a restatement of Hermann Paul’s “paradigmatic pressure” mentioned above.
4.
Optimality Theory
There are a number of reasons to believe why the recently introduced framework of Optimality Theory (OT) (Prince & Smolensky 1993, and much subsequent work) may shed further light on analogical change. First, the traditional assumption is that analogical change is caused by relations between different output forms. Optimality Theory is a theory of grammar in which outputs are selected on the basis of how well they observe general constraints, which are formalised over the outputs. No single output form will observe all constraints, but constraint interaction will ensure that an output form is mapped onto an input. Given this concern for the output, Optimality Theory may be a good candidate for dealing with an evidently output-related phenomenon such as analogical change. Second, analogical change represents a case of conflicting interests: phonological rules will bring about an alternation, and morphological wellformedness conditions (such as (4)) will try to level these out. In principle, such a state of affairs should be readily amenable to an Optimality-theoretic treatment, especially since in recent years proposals have been made to extend this framework with modules describing acquisition and variation (see e.g. Hayes 2000; Boersma 2000; Tesar 2000). In the case of analogical levelling, exemplified in (1), a number of stages must be distinguished. The first stage is the introduction of an alternation as a result of the loss of the rounded semivowel before round vowels. This can be regarded as the effect of a constraint against the adjacency of two elements which are too much alike, the Obligatory Contour Principle (OCP, see e.g. McCarthy 1986). This constraint conflicts with a Faithfulness constraint which demands that underlying forms are changed as little as possible. Possibly after a stage of variation, in which the two constraints are somehow “balanced”, the Obligatory Contour Principle constraint takes precedence and the semivowel will not be pronounced. Later, a labial consonant is restored. This could be formalised as a reversal of the ranking of the two constraints, on the assumption that underlying forms have remained the same, quite parallel to rule reordering in generative grammar. This begs the question of why the two constraints should be reordered, however. Recent analyses (e.g. Kenstowicz 1996) have
1614
XVII. Morphologischer Wandel I: Theoretische Probleme
sought the explanation in the interplay of the two constraints mentioned above with a third one, a correlate of Paradigmatic Uniformity (4). The operation of such a constraint would favour outputs which have a uniform base form. An issue here is which direction of levelling is taken: there are always multiple ways in which a paradigm could be regularised, and the question is if we can predict which way will be taken in individual cases.
5.
Conclusion
The term analogy is used for a wide variety of purposes. Here we focused on one of the most common types of analogical change, viz. analogical levelling. Since the times of the Neogrammarians, it has been suggested that pressure to keep paradigms as uniform as possible plays a role in eradicating phonologically-induced alternations. It turned out to be difficult to formalise this idea in traditional generative grammar. The framework of Optimality Theory may be better suited for the purpose in the light of its insistence on surface constraints and its explicit way of dealing with conflicting tendencies.
6.
References
Boersma, Paul (2000), “Learning a Grammar in Functional Phonology”. In: Dekkers et al. (eds.), 465⫺523 Chomsky, Noam & Halle, Morris (1968), The Sound Pattern of English. New York: Harper and Row Dekkers, Joost & van der Leeuw, Frank & van de Weijer, Jeroen (2000, eds.), Optimality Theory: Phonology, Syntax, and Acquisition. Oxford: Clarendon Press Hayes, Bruce P. (2000), “Gradient WellFormedness in Optimality Theory”. In: Dekkers et al. (eds.), 88⫺120 Jankowsky, Kurt R. (1972), The Neogrammarians: A Re-Evaluation of their Place in the Development of Linguistic Science. The Hague: Mouton Joseph, Brian D. (1998), “Diachronic Morphology”. In: Spencer, Andrew & Zwicky, Arnold (eds.), The Handbook of Morphology. Oxford: Blackwell, 351⫺373 Kenstowicz, Michael (1996), “Base-Identity and Uniform Exponence: Alternatives to Cyclicity”. In: Durand, Jacques & Laks, Bernard (eds.), Current
Trends in Phonology: Models and Methods, Vol. II. Manchester: University of Salford Press, 365⫺395 King, Robert D. (1969), Historical Linguistics and Generative Grammar. Englewood Cliffs/NJ: Prentice Hall Kiparsky, Paul (1971), “Historical Linguistics”. In: Dingwall, William Orr (ed.), A Survey of Linguistic Science. College Park: University of Maryland Linguistics Program, 576⫺649 [reprinted in: Kiparsky (1982)] Kiparsky, Paul (1972), “Explanation in Phonology”. In: Peters, Stanley (ed.), Goals of Linguistic Theory. Englewood Cliffs/NJ: Prentice Hall, 189⫺ 227 [reprinted in: Kiparsky (1982)] Kiparsky, Paul (1982), Explanation in Phonology. Dordrecht: Foris Publications Kuryłowicz, Jerzy (1949), “La nature des proce`s dits ‘analogiques’”. Acta Linguistica 5, 15⫺37 Lehmann, Christian (1995), “Synsemantika”. In: Jacobs, Joachim et al. (eds.), Syntax: Ein internationales Handbuch zeitgenössischer Forschung, Vol. II. Berlin: de Gruyter (Handbücher der Sprachund Kommunikationswissenschaft, 9.2), 1251⫺ 1266 Manc´zak, Witold (1958), “Tendences ge´ne´rales des changements analogiques”. Lingua 7, 298⫺325, 387⫺420 McCarthy, John J. (1986), “OCP Effects: Gemination and Antigemination”. Linguistic Inquiry 17, 207⫺263 Meillet, Antoine & Vendrye`s, Joseph (1924), Traite´ de grammaire compare´e des langues classiques. Paris: Champion Paul, Hermann (51920), Prinzipien der Sprachgeschichte. Halle: Niemeyer [11880; 21886] Prince, Alan S. & Smolensky, Paul (1993), Optimality Theory: Constraint Interaction in Generative Grammar. Piscataway/NJ: Rutgers University (Technical Report of the Rutgers Center for Cognitive Science 2) Saussure, Ferdinand de (1983), Course in General Linguistics, translated and annotated by Roy Harris. London: Duckworth [11916] Schindler, Jochem (1974), “Fragen zum paradigmatischen Ausgleich”. Die Sprache 20, 1⫺9 Tesar, Bruce (2000), “Optimality and Strict Domination in Language Learning”. In: Dekkers et al. (eds.), 592⫺620
Jeroen van de Weijer, Leiden (The Netherlands)
1615
149. Remotivation and reinterpretation
149. Remotivation and reinterpretation 1. 2. 3. 4. 5. 6.
1.
Characterization of the phenomena and definition of terms Subclassifications and typical examples Causes and hindering factors of secondary motivation Effects of secondary motivation Different approaches to secondary motivation References
Characterization of the phenomena and definition of terms
Words which for some reason (non-native origin, nonce-morpheme, phonological structure) cause problems for memory processing and storage sometimes acquire secondary morphological and/or semantic structure; they thus change from opaqueness to transparency. This stands in contrast to the reverse process of demotivation and obscuration, by which originally motivated words lose their transparency (cf. Art. 150; see 3.1 for the reasons for the existence of these opposite processes). The process of secondary motivation or remotivation may take a number of different forms, for which different terms are used (there is, however, no universal agreement on the use of terminology). Reinterpretation is a type of remotivation where no formal change has occurred; reinterpreted words show secondary semantic association with etymologically unconnected elements, which may effect semantic change (Fr. jour ouvrable, from Lat. operari ‘[to] work’, now reinterpreted as connected with ouvrir ‘[to] open’, 2.2.3). The term is also used for certain processes involving the analysis of inflexional affixes as word formation affixes (or vice versa), and for the reanalysis of syntagmatic structures (2.2.6, 2.2.7). Metanalysis is another process which does not involve formal change. Here, the internal morpheme boundary of a word is shifted, so that semantic change as well as secondary analogies may result (standard example: Eng. hamburg-er J ham-burger, with resulting suffix {-burger}, which becomes productive in analogous formations like cheese-burger, onion-burger, etc.). Re-motivation (as a specific process of remotivation) occurs when the opaqueness of the word, too, is secondary, i.e. when a word loses its original transparency and after a
period of opaqueness re-acquires it (e.g. through spelling pronunciation, as Eng. forehead > /fArid/) > /fc:hed/) or (rarely) when a word acquires a new transparency different from the secondary one (⫽ tertiary motivation, e.g. Germ. Maul-wurf, see 2.2.2). Folk etymology or popular etymology are terms frequently used for secondary motivation, particularly of non-native words (e.g. Latinisms) or obsolete words or words with one obsolete element. The terms are, however, also used synonymously with secondary motivation. The term folk-etymology was coined by Ernst Förstemann in his article “Über deutsche Volksetymologie” (Förstemann 1852), but has often been criticized as inappropriate (Ullmann 21966: 34 f.; 1972: 101; Mayer 1962: 5⫺20; Miettinen 1965: 30; Koziol 21974: 118) because it disregards “erroneous” interpretation or intentional misetymologizing of words by medieval scribes, Renaissance humanists, etc. Such suggestions as a posteriori motivation (Ullmann 21959: 91), associative etymology (Ullmann 21966: 35), or synchrone Etymologie (Bergenholtz 1975) have however not been successful (cf. also Mayer 1962: 11, Fn. 2). Phenomena related to secondary motivation are malapropism (the confusion of Latinisms), agglutination/deglutination (see 2.2.7), back-formation and analogization (see 2.2.8; cf. also Mayer 1962: 25⫺27), and certain types of initialization (Ungerer 1991; see also 2.2.8). All processes of secondary motivation were for a long time seen as erroneous etymologies or (later) as rarely occurring reactions against the arbitrariness of the linguistic sign and therefore as exceptional if not “pathological” (this word is used in Saussure 1967: 210, orig. 1916: 241). Following Gillie´ron (1922), they were later and are now usually viewed as normal and frequent processes which have their cause in the storage problems of human memory (Plank 1981: 196 f.). These processes are now also seen in connection with certain stages of child (first) language acquisition, at which children tend to “wrongly” analyze or overinterpret newly acquired words (Leopold 1949: 144 f. with additional literature; Berko 1958: 168⫺ 170; Karpf 1990: 129 f.; Art. 165). Secondary motivation seems to be a fairly universal phenomenon in language change
1616
XVII. Morphologischer Wandel I: Theoretische Probleme
and language contact. Examples quoted in the literature range from Ancient Greek, Latin, and Old English to most modern European and some African languages. The literature listed in a comprehensive bibliography (Schreiner 1987) concerns 48 different languages (see also the commented bibliography in Olschansky 1996).
2.
Subclassifications and typical examples
2.1. Criteria of subclassification A more systematic subclassification than the one implied in the different terms defined in 1 may use the following criteria (see also Olschansky 1996: 178⫺220 and Blank 1997: 304 f.): (a) etymological origin of the remotivated word. This word can either be of foreign origin (2.2.1) or a native word of which (usually) one element has become obsolete and is replaced by a still existing morpheme (2.2.2); (b) total or partial remotivation. If the remotivated word was originally opaque, total remotivation as compound or derivative is the normal case. This is the usual process with words of foreign origin, which however sometimes remain partially opaque (2.2.1). With words of native origin, which have usually remained semitransparent, partial remotivation occurs, since only the obsolete or obscured element has to be remotivated (2.2.2); (c) change of phonetic and/or graphemic form, which may (2.2.4) or may not occur (2.2.3; cf. 1, “reinterpretation”; see also Mugdan 1984: 113 on Baudouin de Courtenay’s distinction between “inner” and “outer” folk etymology); (d) change of morpheme-structure, which may or may not occur. The remotivation of foreign words usually involves the acquisition of a new morphological structure (usually as compound, 2.2.1), while partial remotivation of indigenous words normally retains the old morphemestructure (2.2.2). However, the following changes may occur: shift of morphemeboundary (metanalysis), reanalysis of affix as root (2.2.4), reanalysis of root syllable or word formation affix as inflexional affix or vice versa (2.2.6);
(e) re-motivation as simplex or as complex word (the different possibilities are listed in Koziol 21972: 317⫺320); (f) other criteria (see the examples in 2.2.7). 2.2. Typical examples In the following lists of examples only the information necessary for understanding the process of remotivation is given. Naturally, each word underwent a number of stages in pronunciation, spelling, and semantic interpretation, and some words show different stages of remotivation in the different varieties (e.g. dialects) of the language. To single out one example, Germ. Fried-hof ‘peaceyard (cemetery)’ appears in Old High Germ. as frı¯t-hof, in Middle High Germ. as vrı¯thof ‘fenced-in-yard (forecourt [of house or church])’. “Undisturbed” development would have resulted in Modern Germ. Freit-hof, which indeed is the form in certain dialects. Association of the meaning ‘church-yard’ with ‘Friede’ (‘peace’), however, was responsible for the remotivated form Fried-hof, which is used in most varieties including Standard German. For detailed information concerning the development of the individual examples the reader is referred to the etymological dictionaries (Gamillscheg 21969; Kluge 221989; Onions 1966). 2.2.1. Total remotivation of foreign words This sub-section and the next contain the most frequently quoted examples of folk-etymology. Typically, in German and English, a polysyllabic word of Romance origin undergoes secondary motivation (other origins, however, are also possible); alternately, an obsolete element of an indigenous word is replaced by a still existing morpheme (see 2.2.2). ⫺ Eng. beak-iron < Old Fr. bicorne ‘twohorn’; ⫺ (American) Eng. carry-all < Fr. carriole ‘small cart’; ⫺ Eng. cause-way < Fr. chausse´e < Lat. calciata (via) ‘lime-road’; ⫺ Eng. Charter-house < Fr. chartreuse < Anglonorm. chartrouse ‘Carthusian monastery’; ⫺ Eng. cut-lass < Fr. coutelas < Lat. cultellus ‘knife’; ⫺ Eng. goose-berry (with obscured pronunciation) < Fr. groseille; ⫺ Eng. lance-knight < Germ. Lands-knecht (with remotivation of Lands- as lance, and
149. Remotivation and reinterpretation
⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺
etymologically, but not semantically correct interpretation of -knecht as -knight) ; Eng. leg-horn (hen) < It. Legorno (now Livorno); Eng. man-drake < Medieval Lat. mandragola; Eng. pent-house < Old Fr. appentis < Medieval Lat. appenditium ‘that which is attached’; Eng. prim-rose < Fr. primerole < Lat. prim(er)ula ‘firstling’; Eng. rose-mary < Lat. ros marinus ‘seadew’; Eng. sparrow-grass (dialectal) < Lat. asparagus (see also 2.2.8); Fr. chou-croute < Alsatian sur-krut (⫽ Germ. Sauerkraut); Fr. (jeu de) l’aˆne sale´ < Eng. (game of) Aunt Sally; Germ. Arm-brust < Lat. arcuballista < arcus ‘bow’ ⫹ballista ‘catapult’ (< Gk. ballein ‘[to] throw’); Germ. Fell-eisen < Fr. valise; Germ. Hänge-matte < Du. hang-mat (here the remotivation occurred) < Sp. hamaca < Taino (Haiti) hamaka; Germ. Murmel-tier < murmuntin < Lat. murem montis ‘mountain-mouse’; Germ. Viel-fraß < Norw. fjell-fross ‘mountain-cat’; Germ. Wetter-leuchten < Wetter-leichen ‘weather-dancing’.
The following examples show only partial remotivation of foreign words, which has resulted in formations containing a nonce-morpheme: ⫺ Eng. cray-fish < Fr. e´crevisse < Middle High Germ. krebis; ⫺ Eng. cater-pillar (with possible analogy to caterwaul) < Fr. chatepelose ‘hairy cat’; ⫺ Eng. fe-male < Fr. femelle; ⫺ Eng. mush-room < Fr. mousseron. 2.2.2. Partial re-motivation of native words Partial re-motivation is found in the following examples: ⫺ Eng. black-mail < black-ma¯l ‘black agreement’; ⫺ Eng. bride-groom < brı¯de-guma ‘brideman’; ⫺ Eng. mistle-toe < mistle-ta¯n ‘mistle-twig’; ⫺ Eng. sand-blind < sam-blind ‘half-blind’ ⫺ Eng. shame-faced < scam-faest ‘shamebound’; ⫺ Eng. tit-mouse < tit-ma¯se ‘small bird’;
1617 ⫺ Eng. wed-lock < wed-la¯c ‘wedding gift, pledge’; ⫺ Germ. Fast-nacht < fase-nacht ‘rovingnight’; ⫺ Germ. Fried-hof < vrı¯t-hof ‘fenced-in yard’; ⫺ Germ. Hage-stolz < hagu-stalt ‘owner of (fenced-off) yard’; ⫺ Germ. Maul-wurf < molt-werf ‘earththrower’ < mu¯werf ‘heap-thrower’ with re-motivation (tertiary motivation); ⫺ Germ. Lein-wand < 17th c. lein-wa¯t ‘linnen-garment’; ⫺ Germ. Sünd-flut (dialectal) < sint-vluot ‘great flood’. The following example shows remotivation of a foreign element in a hybrid word: ⫺ Germ. Elfen-bein < helfent-pein ‘elephant-bone’. 2.2.3. Remotivation without change of form (“reinterpretation”) The examples in 2.2.1 and 2.2.2 all involve phonetic and/or graphemic change of the original word. However, examples of merely semantic reinterpretation without formal change also occur (see also the examples in Mayer 1962: 140⫺233): ⫺ Eng. auburn < Lat. alburnus ‘whitish’, now associated with burn; ⫺ Fr. ouvrable < Lat. operari ‘[to] work’, now associated with ouvrir; ⫺ Germ. irritieren ‘[to] confuse, mislead in an annoying way’ < Lat. ir-ritare ‘[to] excite’, now associated with irre-(machen etc.). This type may also involve homonymic elements, as in the association of Eng. breakfast with fast (adj.). The following words are examples of remotivation through graphemic change (with no change in pronunciation): ⫺ Eng. island (with learned introduction of <s> from Lat. insula (Old Eng. ¯ıgland, Middle Eng. ¯ıland); ⫺ Eng. sovereign (with introduction ofand association with reign); ⫺ Eng. posthumous (with introduction of and association with Lat. humus); ⫺ Eng. hiccough (with association with cough). Change from “obscured” to “transparent” pronunciation (with no change in spelling) is now common with a number of English
1618
XVII. Morphologischer Wandel I: Theoretische Probleme
words whose spelling is still transparent. This phenomenon of “spelling pronunciation” is a type of re-motivation. Examples are forehead, waist-coat, and boat-swain. 2.2.4. Remotivation with change in morphemic structure Examples of the type Eng. beak-iron and Germ. Fell-eisen (2.2.1) show acquisition or change of morpho-semantic structure (at least as far as the remotivating language is concerned), while those of the type Eng. black-mail and Germ. Fast-nacht (2.2.2) retain an already existing structure. Examples of change of morphemic structure in indigenous words are the following: ⫺ (occasional) Eng. cow’s lip < cow-slip < cu¯-slyppe ‘cow- dung’; ⫺ Eng. worm-wood < wermod (monomorphemic Old English word); ⫺ Germ. Ein-öde < ein-oete (suffix reinterpreted as root). Remotivation of compound as derivative or vice versa occurs in: ⫺ Germ. weis-sagen < wı¯zagon ‘act as prophet’ (derivative of wı¯zag ‘knowing’, now interpreted as compound); ⫺ Eng. brid-al < bride-ale < Old English bry¯d-ealo; ⫺ Germ. Junk-er (with doubtful transparency) < jungk-herr; ⫺ Eng. fore-most, inner-most, utter-most with reinterpretation of superlative suffix -mest as most. 2.2.5. Remotivation as simplex Most types of secondary motivation result in complex words (more frequently compounds than derivatives). The following are examples of monomorphemic remotivation:
⫺ Fr. corne < Lat. cornu-a; ⫺ Fr. levre < Lat. labr-a; ⫺ Fr. voile < Lat. vel-a. Conversely, the Old Northern French singular form cherise was reinterpreted in English as the plural form cherries. 2.2.7. Remotivation involving syntax In some examples remotivation transcends the word boundary, as when the combination article ⫹ noun is reanalysed “wrongly” through association with another word (“agglutination” and “deglutination”), e.g.: ⫺ Eng. an aitch-bone < a nache-bone < Fr. nache ‘buttock’; ⫺ Eng. a nick-name < an e¯ke-name ‘an also-name’. Secondary motivation may also affect syntagmas, which in the process undergo changes in syntactic structure or are reinterpreted as complex or simple words: ⫺ Eng. atone < (maken) at o¯n ‘make whole, reconcile’; ⫺ Eng. help-mate < help meet (for him), (from Genesis 2: 18,20, with misunderstanding of meet ⫽ ‘suitable’); ⫺ Eng. upside down < up so down; ⫺ Germ. während des Sommers < währendes Sommers. Occasionally, an opaque word is reanalysed as a phrase, as the name of a coarse material, mungo, which in Yorkshire is associated with “it mun go” (it must go), allegedly said by the inventor of the machine for it (Weekley 5 1961: 161).
⫺ Eng. jerusalem (type of artichoke) < It. girasole; ⫺ Germ. Kater (‘hangover’) < Katarrh.
2.2.8. Other examples The following examples show remotivation of a simplex (of either native of foreign origin) with expansion by addition of explanatory determinatum:
Monomorphemic remotivation must be kept apart from word confusion and resulting semantic convergence (rave/rove, sheen/shine, etc.), where two words of different origins, but similar form and already related meanings approach each other semantically even further.
⫺ Eng. rein-deer, Germ. Renn-tier < Icel. hreinn; ⫺ Germ. Lind-wurm < lint ‘winding’; ⫺ Germ. Maul-beere < Lat. mu¯rus; ⫺ Germ. Maul-tier < Lat. mu¯lus; ⫺ Germ. Wind-hund < wint ‘Vendic dog, greyhound’.
2.2.6. Remotivation involving inflexion The reinterpretation of inflexional endings occurred with a number of Latin neuter plurals, which appear in French as feminine singulars, e.g.:
In some cases, remotivation operates simultaneously with other processes, particularly blending (cf. Art. 91). Eng. run-a-gate is both a remotivated word (from Fr. rene´gat) and a blend of this with run-away.
149. Remotivation and reinterpretation
Formal analogizing, rather than remotivation, occurred in Eng. lar-board < lade-board ‘loading side’, which was brought in line with its opposite, starboard ‘steering side’, whereby partial loss of transparency was compensated with formal similarity to a frequent collocate. A quite different type of analogizing was the source of the semitransparent word-creation mono-kini, which owes its origin to the remotivated word bi-kini (originally from Bikini Atoll). Initialization, i.e. the formation of acronyms, may lead to secondary motivation if the sequence of letters/phonemes results in a pronounceable word, particularly if this has some semantic relation with the initialized phrase. Thus such acronyms as AIDS (⫽ acquired immune deficiency syndrome), CALL (⫽ computer assisted language learning), CARE (⫽ Cooperative for American Relief to Europe), and WASP (⫽ White AngloSaxon Protestant) are primarily and secondarily motivated for some speakers (cf. Art. 92; for further discussion see Ungerer 1991). Secondary motivation may operate in loan-translation, as is the case with Germ. Tausend-gülden-kraut. For the formation of this compound, the Latin word centaureum (< Gk. kentauros) was reanalyzed into the elements centum ‘a hundred’ and aurum ‘gold’ (see also 5.2). Secondary motivation is sometimes restricted to a short period in the history of the word. Thus the remotivation of Eng. coward as cow-heart or cow-herd and abundance as hab-undance was only transitory, and the process was reversed. Remotivated forms may also persist, or exist only, in non-standard varieties: “the corruption of asparagus to sparrow-grass, which is now regarded as vulgar […] was in good use in the eighteenth and early nineteenth centuries” (Greenough & Kittredge 1900: 334).
3.
Causes and hindering factors of secondary motivation
3.1. Causes Secondary motivation is commonly attributed to the need of the language user for transparency and association with familiar elements (Koziol 21974: 117), “the desire to motivate what is, or has become, opaque in language” (Ullmann 1972: 102) or an “Urbedürfnis des Menschen nach Deutung und Stützung der Wörter, die sich durch feh-
1619 lende Motivation seinem Verständnis entziehen,” (Scheler 1977: 116, cf. also Leisi 7 1985: 76⫺79, where remotivation is discussed as a reaction against the “dissociation” of “hard words”). This explanation, however, is not entirely satisfactory in view of the existence of the reverse process of obscuration, which is almost as frequent as that of remotivation. The assumption of “two opposite tendencies […] at work all the time in the development of language,” (Ullmann 1972: 93), or of “two rival factors, conventionality and motivation” (Ullmann 21959: 92) is not convincing if it is not further explained. Somewhat more plausible is the assumption of two attitudes towards language, a naive, “non-reflective” one associated with the speaker’s interest in an economical use of language, and a “reflective” one resulting from the speaker’s need to make him/herself understood and from the hearer’s search for meaningful elements (Fill 1980: 25⫺27); these attitudes are thought of as present in both hearer and speaker, with one of them predominating depending on the situation ⫺ but each of them can also be thought of as having been prevalent in different periods, as for instance the reflective one, which favours secondary motivation, in the Renaissance period. The opposite process of obscuration may even have favoured remotivation in some cases. Obscured and transparent forms of compounds frequently existed side by side, so that with words showing some formal relation to the opaque form hypercorrect formations began to appear. Thus the existence of the forms gozzard and goose-herd, shepherd and sheep-herd may have been an important factor in the remotivation of coward as cowherd (Götz 1971: 22 f.). Other explanations seek the need for transparency in the facilitation of language acquisition and word retention. These explanations are supported by evidence from the language of aphasics and from the linguistic associations of children. Examples: “breakfast is called breakfast because you have to eat it fast when you rush to school”, “Friday is a day when you have fried fish” (Berko 1958: 170), chaotisch ⫺ wenn jemand am Boden liegt und zusammengeschlagen ist < k. o. (Karpf 1990: 129), chicken pox ⫺ chicken pots, ice cream cone ⫺ ice cream comb (Bolinger 1968: 105), forec’sle reinterpreted as foxhole by two children in Richard Hughes’ novel A High Wind in Jamaica (cf. Fill
1620
XVII. Morphologischer Wandel I: Theoretische Probleme
1980: 26). Ad-hoc folk-etymologies of this kind are also provided by parents teaching new words to their children (Germ. Bagger explained as Pack-er, cf. Fill 1980: 26). Frequently certain specific remotivations first appear in certain varieties of a language, e.g. in a specific dialect, or even in the idiolect of individual speakers. The association of Eng. ballistic with ball, of mysticism with mist, of standard with stand (cf. Waldron 1967: 124), of hurricane with hurry (cf. Mayer 1962: 197) and of Germ. Gastronomie (< Gk. gaster ‘stomach’) with Gast, Dienstag with Dienst and Freitag with frei (Weisgerber 1950: 60 f.) is certainly still specific, not universal. Idiolectal remotivations may, however, spread and become “accepted”. Another cause of remotivation concerns place names which conquering nations or new settlers fail to understand and try to adapt to their own language. “The first American hearing a French trapper in Colorado refer to the Purgatoire River judged that what he heard was Picket Wire with a foreign accent” (Bolinger 1968: 104 f.). Other American examples of this type are Smackover < Fr. Chemin Couvert and Low Freight < L’Eau Frais. The Old Eng. place name Eofor-wı¯c (now York), ‘boar-town’ is remotivated from Celtic Eburacum, Bear Park in Durhamshire from Fr. Beau Repas (Görlach 1974: 84). Examples of remotivated place-names in German (Mai-land < Mediolanum, Merse-burg < Slavic -bor ‘wood’, etc.) are already given by Förstemann (1852: 15 f.). For a discussion of folk-etymology in place-names see Vennemann (1999). Until the beginning of the 19th century, the attempts of etymologists to explain the origin of words were frequently unscholarly guesses wide off the mark. Medieval writers, in particular, supplied examples of “learned folk-etymology” (Weekley 51961: 156) or “guesstymology” (Bolinger 1968: 105) aimed at edification rather than presentation of the facts. Philippe de Thaun (12th century), for instance, explained Fr. lundi as ‘day of light’ (lumie`re), mercredy as ‘day of market’ (marche´), jeudi as ‘day of joy’ (joie), vendredi as ‘day of truth’ (ve´rite´), etc. (Weekley 5 1961: 156). Wild etymologizing of the type “lucus a non lucendo, canis a non canendo” (⫽ ‘grove from not being light, dog from not singing’) was also frequent in Classical Antiquity and in the Renaissance period. However, very few of these learned mis-etymolo-
gies left any traces in actual language use, so that as a cause of folk-etymology they can be disregarded. 3.2. Hindering factors There are only few factors which hinder or prevent secondary motivation. Remarkably, it is not prevented or hindered by initial lack of semantic relation between the parts and the whole of the word. Analysis into familiar elements seems to be more important at first than the semantic connexion of these elements with the whole. With compounds, frequently one root element shows a semantic relation with the whole, the other one being completely arbitrary (cf. Eng. mush-room < mousseron, Eng. cut-lass < coutelas). Here, morphological transparency seems to be more important than constructional transparency, i.e. rule-dependent semantic relation between the elements and the whole (Plank 1981: 195 f.). With non-compounds, a semantic relation is sometimes re-established through semantic change (at least on the connotative level) of the remotivated word (see 2.2.3). The laws of sound-change likewise do not prevent or hinder secondary motivation. Baudouin de Courtenay already characterizes “outer” folk etymology as one of the changes which do not follow these laws (cf. Mugdan 1984: 113). Indeed, remotivation involving phonetic change almost by definition defies sound-laws: the need for analysis into existing elements seems to be stronger than the rules for semantic derivation or phonetic change. Phonotactics, however, may sometimes play a role as a cause of secondary motivation (Mayer 1962: 54⫺61). A factor which has sometimes hindered or even reversed folk-etymological processes can be seen in the influence of (scholarly) etymologists, lexicographers, and orthoepists, particularly of the 19th century. Especially in German, a number of words which had undergone folk-etymological changes in previous centuries (e.g. Sprichwort, which in the 16th century had adopted the form Sprüchwort from Spruch) returned to their original form under this influence (Andresen 4 1886: 17 f.).
4.
Effects of secondary motivation
Those effects of secondary motivation which are connected with its causes are self-evident: morphological and/or semantic transparency
1621
149. Remotivation and reinterpretation
are maintained or re-established, and problems of pronunciation are removed. However, the new morphological structure (or connexion) naturally influences the semantic structure of the word and may result in semantic changes on the denotative or the connotative level. Frequently, certain semantic elements are given more prominence than before the remotivation. This is the case, for instance, with ‘Maul’ in Germ. Maulwurf and Maulesel, with ‘Sünde’ in Germ. Sündflut or with ‘bell’ in Eng. belfry (< Old Fr. berfray < Middle High Germ. ber(c)-vrit). Other words develop an occasional meaning strongly influenced by a secondary element (Eng. pundit, also ‘sophist’ with influence from pun). The usual meaning of a word, too, can be affected by secondary motivation (e.g. Germ. irritieren). Folk-etymology has frequently influenced popular beliefs concerning saints, sects, plants, and animals, and has given rise to anecdotes about the origin of customs and to legends about the origin of place-names (for examples see Weisgerber 1950: 56⫺66 and Kainz 1972: 366⫺374, who also discusses prejudices and superstitions originating in secondary motivation). A well-known English example of an anecdote inspired by remotivation is the story (told by Jonathan Swift of King James I) of the knighted loin of beef (sir-loin, really from Fr. surlonge ‘above the loin’, cf. Weekley 51961: 162; Greenough & Kittredge 1900: 331; Mayer 1962: 95). Child-etymology has been shown to be an important phenomenon in child language acquisition, since it facilitates the acquisition of a more comprehensive vocabulary by establishing semantic links between words already stored in the brain and words to be newly acquired (Leopold 1949: 114). To what extent these early attempts at motivation persist and have an influence on the thinking of the grown-up still remains to be investigated. Other effects concern intentional remotivation by writers and journalists. These effects can be either humorous or poetic. A humorous example is the formation Vi-queens to Vikings (Weekley 51961: 142; quoted from Punch). Comic effects are also intended by Shakespeare in The Merchant of Venice (“more than sand-blind, high-gravel blind” Act II, scene ii), and Friedrich von Schiller in Wallensteins Lager (the “Kapuzinerpredigt”, scene 8). For didactic or poetic purposes remotivation has been used by many writers
from Bede (Angli ⫺ Angeli) to James Joyce (particularly in Finnegans Wake; see also the examples in Mayer 1962: 288⫺310). Effects of both comic relief and characterization are achieved by “malapropisms” in The Rivals by R. B. Sheridan and Der Zauberberg by Thomas Mann.
5.
Different approaches to secondary motivation
Secondary motivation has been investigated with different approaches depending on the different levels of language involved, and from various theoretical positions adopted by the different schools of linguistics. The different linguistic levels from which it has been approached are the phonological (Ullmann 1972: 101; Mayer 1962: 54⫺79, 234⫺287), the morphological (Erben 1964, Koziol 21972: 317⫺320, Ullmann 1972: 101⫺105), the semantic (Waldron 1967: 124 f., 153 f.; Koziol 2 1974: 117⫺121; Mayer 1962: 127⫺233), and the psycho-pragmatic (Kainz 1972: 366⫺ 374). There are, however, no comprehensive treatments of remotivation which concentrate specifically on any one of these levels. The different approaches depending on linguistic school and theoretical background will be sketched out in the following sub-sections. 5.1. The historical approach Secondary motivation, under the term folketymology, was already investigated in some detail in the second half of the 19th century. The interest of scholars was, however, chiefly historical. Three types of etymology were distinguished, the popular, the learned and the scholarly, of which the popular was held to be the oldest (Förstemann 1852: 2). In accordance with the aims of Comparative Philology, examples of folk-etymology were investigated in many ancient languages (Ancient Greek, Latin, Gothic, Anglo-Saxon, Old Norse, etc.) and a few modern ones, particularly German, English, and the Scandinavian languages (Andresen 1876; Förstemann 1877). Folk-etymology was judged as “distortion of words” (“Verunstaltungen der Wörter”, Förstemann 1852: 7; “volksetymologische Entstellung”, Andresen 41886: 16), as “false derivation” and “verbal corruption” (Palmer 1882: subtitle) or as “the tendency of the uneducated […] to distort an unfamiliar or un-
1622
XVII. Morphologischer Wandel I: Theoretische Probleme
intelligible word into some form which suggests a meaning” (Weekley 51961: 105 f.). Evaluative judgments of this kind have occasionally persisted into the 20th century (in Saussure 1916 folk-etymology is discussed in the diachronic section as a rare and pathological phenomenon). Since folk-etymology defies any claims concerning the regularity of sound-change (claims which were made in the 19th century by the Neogrammarians), it was frequently regarded from the onomasiological point-ofview, and remotivated words were categorized according to object groups or conceptual categories (proper names, place-names, words denoting plants, animals, etc., Andresen 1876). There was also a tendency to see folk-etymology largely as a thing of the past, at least as far as the standard language was concerned (Förstemann 1852: 14). In a number of publications, folk-etymology is discussed as one of the strange and curious adventures which words may undergo inthe course of their romantic history (Greenough & Kittredge 1900; Weekley 51961). 5.2. The synchronic approach Since the advent of Structuralism and the possibility of a synchronic approach, the majority of scholars have taken a more tolerant and even positive attitude towards the phenomena under discussion. This change in attitude has also affected judgments concerning the frequency and statistical importance of the phenomena. While scholars adopting the historical approach tended to view folk-etymology as an exceptional and rare phenomenon restricted to a small number of examples of the type listed in 2.2.1 and 2.2.2, secondary motivation was later and is now regarded as frequent, natural, and indispensable for a living language: “Sekundäre Motivation offenbart sich als ein durchgängiges Prinzip der Sprache” (Mayer 1962: 348). Consequently, scholars interested in secondary motivation no longer restrict their research to previous stages of linguistic development (e.g. the period between the 14th and the 17th century), but increasingly take into account recent and contemporary instances. There is a growing awareness that remotivation and the phenomena related to it (e.g. rhyme association) are extremely widespread, particularly in colloquial speech and substandard varieties, and that the instances brought to the notice of researchers represent only a fraction of the actually existing cases (Miettinen
1965: 29). Interest also no longer focuses on the word alone, but scholars now consider syntagmatic and syntactic reinterpretation to be at least closely related to folk-etymology (Mayer 1962: 310⫺321). With the synchronic approach and the introduction of the term “motivation”, the phenomenon under discussion began to be seen as a “force” (Mayer 1962: 349) and an expression of the creativity of a language (“schöpferisches Mißverständnis”, Knobloch 1968). This view is most strongly manifested in the so-called “inhaltsbezogene” and “leistungsbezogene Sprachbetrachtung” (Weisgerber 1950: 56⫺66), where folk-etymology is discussed as one of the powers which create and determine folk-customs and types of behaviour, and from which popular beliefs about animals (Vielfraß), plants (Tausendgüldenkraut), patron saints (St. Vincent ⫺ vin), etc. are derived. Folk-etymology is now seen as “acquisition of morphological and semantic motivation” (Ullmann 1972: 101⫺105). In an extension of this approach, the following types of secondary motivation have been distinguished (Mayer 1962; see also the criteria of subclassification in 2.1): ⫺ phonological with mere adjustment on the level of sounds (e.g. Eng. shamrock < Ir. seamro´g ‘small clover’); ⫺ phono-semantic, roughly corresponding to Ullmann’s “morphological motivation” (e.g. Eng. beak-iron < Fr. bi-corne); ⫺ semantic motivation, i.e. semantic association without phonological change (e.g. Engl. hurricane, with association to hurry); ⫺ onomatopoeic motivation, i.e. acquisition of sound-symbolism (e.g. Eng. rush < Fr. reuser); ⫺ intentional secondary motivation (e.g. Viqueen, her-story; see also the examples discussed in 4); ⫺ parallel syntactic developments (Eng. help-mate < help meet (noun ⫹ adjective)). Particularly with the inclusion of “semantic motivation” (in Mayer 1962 about 140 examples from English are discussed), secondary motivation becomes a frequent every-day phenomenon with idiolectal differences, or rather a dynamic process based on the “synchronic etymological competence” of the speaker (Augst 1975: 156⫺230; see also 5.3).
1623
149. Remotivation and reinterpretation
The generative-transformational approach to language has contributed to viewing secondary motivation as a process, which goes on in the individual, rather than as a historical fact. However, this approach has not played an important role in the further investigation of the phenomena. Folk-etymology is too “irregular” to admit the setting-up of rules. The products of secondary motivation do not normally show constructional transparency, since frequently only one element contributes to semantic motivation (Plank 1981: 195 f.). Besides, remotivated words do not follow one particular pattern of wordformation; rather, they show a variety of structures, from endocentric noun ⫹ noun structures (e.g. Fried-hof ) to exocentric noun ⫹ noun (e.g. Arm-brust) or verb ⫹ noun (e.g. cut-lass) structures. In addition, most of them must be regarded as highly idiomatized complex words. As far as their generation is concerned, they are fairly isolated and lexicalized formations, for whose production no generalization is possible (Dressler 1976: 157). 5.3. Psychological and sociological aspects Psychological aspects concerning the role of memory were first drawn attention to around the turn of the century (Paul 41909: 221 f.), when folk-etymology was also explained as a phenomenon related to a specific type of slip of the tongue, viz. substitution (Meringer & Mayer 1895: 76), slips being regarded as nonpathological phenomena stemming from constantly operating psychic forces which are subject to rules (Meringer & Mayer 1895: 9). In contrast to this, more recent psychological explanations of folk-etymology seek its causes in the “interpretative correction” of unfamiliar words and phrases (Bolinger 1968: 104). The psychological aspects of secondary motivation have received special attention with the investigation of child language acquisition. Many children have private meanings for compounds, which they analyse in their own way, without being aware of or considering the history of the word (Berko 1958: 169). It is now generally accepted that child etymology is an important factor in language acquisition and concept-formation (Leopold 1949: 114 f.; Porzig 41967: 203; Fill 1976: 14 f.; Karpf 1990: 129). It has even been suggested that secondary motivation, which is thought to be based on synchronic etymological competence and to precede generative competence, can be practised to facilitate fur-
ther language acquisition (Augst 1975: 178). In child etymology, as in folk-etymology, the formal similarity need not be very close, as in pretty coat for pettycoat (Leopold 1949: 115). Likewise, constructional transparency (i.e. rule-based relation between form and meaning) is probably aimed at, but rarely achieved, as in the example “it is called mushroom, because it has rooms in it” (Fill 1976: 14). Social aspects of remotivation, which concern for instance its particular frequency in non-standard social and regional varieties of a language, have received attention from the very beginning (Förstemann 1852: 20⫺22; Andresen 1876). More recently, it has been shown that remotivation and the processes related to it are subject to variation and fluctuation in the different dialects and sociolects, and that no dividing line can be drawn between accidental formations due to carelessness or ignorance and intentional word associations created for comic and other effects (Miettinen 1965: 29 f.). Although dialects abound in remotivated formations, there is as yet no comprehensive study of the phenomena under discussion in specific dialects, let alone in larger groups of non-standard varieties.
6.
References
Andresen, Karl Gustaf (1876), Über deutsche Volksetymologie. Heilbronn: Henninger Andresen, Karl Gustaf (41886), Sprachgebrauch und Sprachrichtigkeit im Deutschen. Heilbronn: Henninger [11880] Augst, Gerhard (1975), Untersuchungen zum Morpheminventar der deutschen Gegenwartssprache. Tübingen: Narr (Forschungsberichte des Instituts für deutsche Sprache 25) Bergenholtz, Henning (1975), “Volksetymologie oder synchrone Etymologie”. Muttersprache 85, 89⫺94 Berko, Jean (1958), “The Child’s Learning of English Morphology”. Word 14, 150⫺177 Blank, Andreas (1997), Prinzipien des lexikalischen Bedeutungswandels am Beispiel der romanischen Sprachen. Tübingen: Niemeyer Bolinger, Dwight (1968), Aspects of Language. New York etc.: Harcourt, Brace & World Dressler, Wolfgang (1976), “Das Zusammenspiel verschiedener Ebenen und Prozesse in der diachronen Wortbildung: Thesen”. In: Panagl, Oswald (ed.), Wortbildung diachron ⫺ synchron. Innsbruck:
1624
XVII. Morphologischer Wandel I: Theoretische Probleme
Inst. für Sprachwissenschaft (Innsbrucker Beiträge zur Sprachwissenschaft 18), 155⫺157
Leisi, Ernst (71985), Das heutige Englisch. Heidelberg: Winter [11955]
Erben, Johannes (1964), “Deutsche Wortbildung in synchronischer und diachronischer Sicht”. Wirkendes Wort 14, 83⫺93
Leopold, Werner F. (1949), Speech Development of a Bilingual Child: A Linguist’s Record, Vol. III: Grammar and General Problems in the First Two Years. New York: AMS Press
Fill, Alwin (1976), “Synchrone oder diachrone etymologische Kompetenz”. Klagenfurter Beiträge zur Sprachwissenschaft 2.2⫺3 (Etymologie I, ed. Oswald Panagl & Hans-Dieter Pohl), 3⫺16 Fill, Alwin (1980), Wortdurchsichtigkeit im Englischen. Innsbruck: Institut für Sprachwissenschaft (Innsbrucker Beiträge zur Sprachwissenschaft 24) Förstemann, Ernst (1852), “Über deutsche Volksetymologie”. Zeitschrift für Vergleichende Sprachforschung auf dem Gebiete der Indogermanischen Sprachen (⫽ Kuhns Zeitschrift) 1, 1⫺25 Förstemann, Ernst (1877), “Über deutsche Volksetymologie”. Zeitschrift für Vergleichende Sprachforschung auf dem Gebiete der Indogermanischen Sprachen (⫽ Kuhns Zeitschrift) 23, 375⫺385 Gamillscheg, Ernst (21969), Etymologisches Wörterbuch der französischen Sprache. Heidelberg: Carl Winter Universitätsverlag [11928] Gillie´ron, Jules (1922), Les e´tymologies des e´tymologistes et celles du peuple. Paris: Champion
Mayer, Erwin (1962), Sekundäre Motivation: Untersuchungen zur Volksetymologie und verwandten Erscheinungen im Englischen. Köln: Universität Köln Meringer, Rudolf & Mayer, Karl (1895), Versprechen und Verlesen: Eine psychologisch-linguistische Studie. Stuttgart: Göschen Miettinen, Erkki (1965), “Beiträge zur deutschen Volksetymologie: Assoziative Umbildungen und Umdeutungen romanischer und lateinischer Entlehnungen”. Neuphilologische Mitteilungen 66, 28⫺91 Mugdan, Joachim (1984), Jan Baudouin de Courtenay (1845⫺1929): Leben und Werk. München: Fink Olschansky, Heike (1996), Volksetymologie. Tübingen: Niemeyer Onions, C[harles] T[albut] (1966), The Oxford Dictionary of English Etymology. Oxford: Clarendon Press
Görlach, Manfred (1974), Einführung in die englische Sprachgeschichte. Heidelberg: Quelle & Meyer
Palmer, Abraham Smythe (1882), Folk-etymology: A Dictionary of Verbal Corruption or Words Perverted in Form or Meaning, by False Derivation or Mistaken Analogy. London: Bell
Götz, Dieter (1971), Studien zu den verdunkelten Komposita im Englischen. Nürnberg: Carl
Paul, Hermann (41909), Prinzipien der Sprachgeschichte. Halle: Niemeyer [11880]
Greenough, James Bradstreet & Kittredge, George Lyman (1900), Words and Their Ways in English Speech. New York: Macmillan [paperback reprint 1961]
Plank, Frans (1981), Morphologische (Ir-)Regularitäten: Aspekte der Wortstrukturtheorie. Tübingen: Narr (Studien zur deutschen Grammatik 13)
Kainz, Friedrich (1972), Über die Sprachverführung des Denkens. Berlin: Duncker & Humblot Karpf, Annemarie (1990), Selbstorganisationsprozesse in der sprachlichen Ontogenese: Erst- und Fremdsprache(n). Tübingen: Narr (Tübinger Beiträge zur Linguistik 352) Kluge, Friedrich (221989), Etymologisches Wörterbuch der deutschen Sprache, ed. by Elmar Seebold. Berlin, New York: Walter de Gruyter [11883]
Porzig, Walter (41967), Das Wunder der Sprache. Bern, München: Francke [11950] Saussure, Ferdinand de (1916), Cours de linguistique ge´ne´rale. Lausanne: Payot Saussure, Ferdinand de (1967), Grundfragen der allgemeinen Sprachwissenschaft, transl. by Herman Lommel. Berlin: de Gruyter Scheler, Manfred (1977), Der englische Wortschatz. Berlin: Schmidt (Grundlagen der Anglistik und Amerikanistik 9)
Knobloch, Johann (1968), “Das schöpferische Mißverständnis”. Lingua 21, 237⫺249
Schreiner, Markus (1987), Bibliographie zur Volksetymologie. Münster: Inst. für Allgemeine Sprachwissenschaft (Studium Sprachwissenschaft 11)
Koziol, Herbert (21972), Handbuch der englischen Wortbildungslehre. Heidelberg: Winter [11937]
Ullmann, Stephen (21959), The Principles of Semantics. Oxford: Blackwell [11951]
Koziol, Herbert (21974), Grundzüge der englischen Semantik. Wien, Stuttgart: Braumüller [11967]
Ullmann, Stephen (21966), Language and Style. Oxford: Blackwell [11964]
150. Lexicalization and demotivation Ullmann, Stephen (1972), Semantics: An Introduction to the Science of Meaning. Oxford: Blackwell [11962] Ungerer, Friedrich (1991), “Acronyms, Trade Names and Motivation”. Arbeiten aus Anglistik und Amerikanistik 16, 131⫺158 Vennemann, Theo (1999), “Volksetymologie und Ortsnamenforschung”. Beiträge zur Namensforschung N. F. 34, 269⫺322
1625 Waldron, R[oland] A. (1967), Sense and Sense Development. London: Deutsch Weekley, Ernest (51961), The Romance of Words. London: Murray [11912] Weisgerber, Leo (1950), Die Muttersprache im Aufbau unserer Kultur. Düsseldorf: Schwann
Alwin Fill, Graz (Austria)
150. Lexicalization and demotivation 1. 2. 3. 4.
Terminology Lexicalization Demotivation References
1.
Terminology
1.1. Lexicalization The term lexicalization (Germ. Lexikalisierung) is of relatively recent origin; according to the OED (21989) (see verb lexicalize, from which lexicalization has probably been derived) it was first attested in 1949. It has been used more widely from the 1960s onwards, but it is apparently still confined to the language of linguists. On its history see, e.g., Grimm (1991: 57⫺70) and Lipka (1992: 2⫺ 6). I shall briefly present some of the more important definitions that have emerged in the discussion. 1.1.1. Accepting into the lexicon The OED (21989) defines lexicalize (1) as ‘to accept into the lexicon, or vocabulary, of a language’, and lexicalization as ‘the action or process of lexicalizing’. In this sense simple and complex words, native as well as loanwords can be lexicalized. Thus, Lyons (1968: 352) says “that the relationship of the transitive to the intransitive is ‘lexicalized’” in pairs like kill and die (Bill died ⫺ John killed Bill), i.e. the transitive (and causative) concept of ‘to cause someone to die’ is expressed by a separate word, to kill (someone). Quirk et al. (1985: 1525 f.) restrict lexicalization to words formed by word-formation processes, explaining it as the process of creating a new word (a complex lexical item) for a (new) thing or notion instead of describing this thing or notion in a sentence or with
a paraphrase. The use of words is more economical because they are shorter than the corresponding (underlying) sentences or paraphrases, and because they can be more easily used as elements of sentences. Thus one does not say ‘someone who writes a book [...] for someone else, who then often pretends it is their own work’, one says ghostwriter instead (see the definition in DCE 2 1987). Quirk et al. (1985) also point out that the precise form of the lexicalization often cannot be predicted. Sometimes there are competing patterns: ‘to make national’ has been lexicalized as to nationalize, but ‘to make beautiful’ as to beautify, ‘to make broad(er)’ as to broaden, ‘to make narrow(er)’ as to narrow; sometimes different formations (synonyms) are used for the same concept, often with regional or stylistic differentiation: ‘a track round which horses race’ is a race-track in American English, but a race-course in British English (according to the DCE 21987); ‘Saturday’ is Samstag in Southern German, but Sonnabend in Northern German. 1.1.2. Replacement of a semantic configuration by a lexeme Within the framework of generative semantics, lexicalization (or lexical insertion) refers to “a process in which a configuration of semantic elements in an abstract representation is replaced by a lexeme” (Lipka 2002: 111), e.g. [cause become not alive] is lexicalized as kill, and in German [bewirken werden sauber(er)] is lexicalized as säubern, but I shall not pursue this aspect here. 1.1.3. Formal and/or semantic changes The third sense of lexicalization is now probably the most usual one; it is also the one which will be discussed in greater detail in
1626
XVII. Morphologischer Wandel I: Theoretische Probleme
this article. It presupposes the first sense (inclusion of a word into the lexicon), however. Broadly speaking, lexicalization refers to the phenomenon that complex words and expressions are often not identical with the sum of their parts; instead they show idiosyncratic properties, i.e. formal (phonological, morphological) and/or semantic peculiarities which cannot be explained from their constituents or from the word-formation pattern. As Quirk et al. (1985: 1526) put it, the meaning of lexicalized words “is not recoverable from the form”, i.e. it cannot be deduced from the constituents, or at least not entirely. Thus holiday originally referred to a ‘holy day’, i.e. a church festival, but in Modern English it refers to ‘a day free from work (or school)’; Germ. Hochzeit originally referred to ‘any kind of festival’ but in Modern German it refers specifically to a ‘wedding’. Lexicalization in this sense is a complex phenomenon, and accordingly much more detailed definitions are possible (see Lipka 2002: 113; 1992: 7 f.). Lexicalization overlays regular word-formation processes; Bauer says that a lexicalized lexeme has taken on a form “which it could not have if it had arisen by the application of productive rules” (Bauer 1983: 48). Lexicalization thus is a diachronic process. Even though they did not use the term lexicalization, the phenomenon referred to here was known to and at least partly described by earlier scholars, in particular by Paul (1880) and Brugmann (1900). Lexicalization stands in relation to a number of terms from which it is not always easy to distinguish. Therefore idiomatization and institutionalization must also be briefly discussed. 1.2. Idiomatization Idiomatization (or its result, idiomaticity) also refers to semantic peculiarities of a complex word or a phrase which cannot or only partly be explained from its constituents, but whereas lexicalization can include phonological and morphological differences, too, idiomatization refers to semantic differences only. Moreover, lexicalization implies a diachronic process, e.g. change of meaning (cf. the examples of holiday and Hochzeit given in 1.1.3) or the addition of meanings, leading to polysemy, e.g. reader (a) ‘a person who reads (something)’, (b) ‘a person who reads books to put mistakes right before printing [i.e. a proofreader], or to decide whether to print them’; (c) ‘a senior British university
teacher just below the rank of professor’ (cf. DCE 21987), whereas idiomatization can refer to a synchronic process, whereby complex words or expressions are coined so as to present a specific meaning. Middle English loveday, for example, was probably coined to mean ‘day for the settlement of quarrels out of court’ (presumably as a loan translation of Lat. dies amoris) and did not mean something else first; the same applies to ghostwriter (see 1.1.1), also to phrasal idioms such as golden wedding/Germ. goldene Hochzeit ‘50th wedding anniversary’, ins Gras beißen ‘to die’. It is, of course, not always easy to ascertain whether a word or phrase was coined with a specific meaning or whether it developed this meaning only later; accordingly the distinction between lexicalization as a diachronic change of meaning and idiomatization as a synchronic specialization of meaning is sometimes difficult. Scholars are therefore not always consistent in their use of these terms. Consistency is especially difficult because idiomatized formations can then also be lexicalized in other ways: chairman /tseemen/, for example, was probably coined to mean ‘a person who is in charge of a meeting’ or ‘a person who is the head of a large organization’ and did not simply have ‘man who sits in a chair’ as its original meaning, but rather ‘man who sits in a chair of authority’ (see OED 21989), but its second element man was, probably subsequently, weakened from /mæn/ to /men/. But it is perhaps going too far to doubt the value of the distinction altogether, as Kastovsky (1982: 164 f.) seems to do. Formations which are similar on the surface can nevertheless be idiomatized quite differently: callgirl ‘prostitute’ has the meaning ‘girl who is called (by men on the phone asking for paid sex)’, whereas the original meaning of callboy was ‘boy who calls (actors onto the stage)’ see, e.g., Kastovsky (1982: 165). 1.3. Institutionalization This term was apparently made popular in morphology by Bauer (1983: 48). It also relates to the history of vocabulary items, but refers to an earlier stage than lexicalization. Bauer distinguishes three such stages, namely formation, institutionalization, and lexicalization. 1.3.1. Formation Words are usually first coined in a particular context and in order to name a specific (new) thing or concept. Many are never taken over
1627
150. Lexicalization and demotivation
by the speech community, they remain nonce-formations (ad-hoc formations, hapax legomena, Germ. Augenblicksbildungen, Einzelprägungen). 1.3.2. Institutionalization Words are institutionalized when they enter the common vocabulary or at least the vocabulary of a certain group of speakers. The term institutionalization is thus roughly equivalent to (synonymous with) the definition of lexicalization as ‘accepting into the lexicon’ (see 1.1.1). Institutionalization can also be connected with the change from typefamiliarity (recognizing a particular pattern of word-formation) to item-familiarity (knowing a particular word). In institutionalized words, possible ambiguity (polysemy) is often ignored and their use is restricted to one of their possible meanings. Almsman, for example, usually meant ‘man who receives alms’ and not ‘man who gives alms’, Germ. Schüler ‘pupil’ refers to ‘someone who is taught at school’ and not to ‘someone who teaches at school’ (the latter meaning is institutionalized as Lehrer in German). On the other hand, formations with similar or partly identical elements can be institutionalized (and idiomatized) in different ways, as shown in 1.2 for callboy and callgirl, cf. also pairs such as pork butcher and family butcher, Germ. Brillenträger ⫺ Hosenträger, etc. But even in institutionalized formations, polysemy is sometimes retained (or re-introduced), e.g.: fireman can refer to ‘a person whose job is putting out fires’ or to ‘a person who looks after the fire in a steam engine or furnace’ (DCE 21987); Germ. Holzschuppen can refer to a ‘shed made of wood’ or a ‘shed where wood is stored’ (or both at the same time). 1.3.3. Lexicalization The third stage in the development of a lexical item is then lexicalization (in the sense of developing idiosyncratic formal and/or semantic features, i.e. change of meaning and/ or form), as in bonfire ‘fire made for pleasure etc.’ (originally bone-fire), or Germ. Drittel (< Dritteil (der dritte Teil)). Not all institutionalized complex words become lexicalized, however: words such as goldsmith, stone wall, etc. have existed since Old English without any apparent change of meaning, and have undergone regular sound changes only, which affected the compounds as well as their ele-
ments. To cover both institutionalized and lexicalized formations Bauer (1983) also uses the term established formations. 1.4. Obscuration This term refers primarily to the formal side of lexicalization. In obscured compounds, the elements show a different phonological (and consequently often also morphological) development in the compounds and in independent use. Holiday /hAled=/, /hAl=de=/ is, for example, slightly obscured from its constituents holy /hewl=/ and day /de=/, whereas gospel (from Old English go¯d > Modern English good /gwd/ and Old English spell > Modern English spell /spel/) is totally obscured (/gAspl/) and has become monomorphemic. Obscured compounds often show semantic changes, too. On obscured compounds in English, see, e.g., Götz (1971), Faiß (1978), and Sauer (1992: 345⫺358). 1.5. Other terms Some authors use terminologies quite different from the one favoured here. Lyons (1977: 535 ff.) calls the stage of the formation of words lexicalization (see 1.1.1). Syntactic compounds is his term for completely regular, non-lexicalized compounds which can be explained according to productive patterns of compounding. Compound lexemes is his term for semantically lexicalized compounds (i.e. with idiosyncratic semantic features) which he also terms petrified (1977: 547 f.) ⫺ for a similar use of petrification see Leech (21981: 225⫺227). From these Lyons distinguishes fossilized compounds; they are derived according to rules that are no longer productive in the language system (such as the so-called imperative compounds like pickpocket, turnkey) ⫺ these are called morphologically lexicalized in our terminology (see 2.3.3). Lyons’s terminology does not seem very clear, however; for criticism see also Bauer (1983: 49 f.).
2.
Lexicalization
2.1. Causes of lexicalization Lipka (1981: 131) mentions several possible causes of lexicalization, which often work hand in hand. Lexicalization stresses the unified character of complex words and at the same time diminishes their character as syntagms. It should be borne in mind, however, that phenomena such as word-character,
1628
XVII. Morphologischer Wandel I: Theoretische Probleme
naming function, and frequency of use can lead to lexicalization, but do not necessarily do so: Old English formations such as grassgreen (Old English gærsgrene), Sunday (Old English sunnandæg) are still used in their original meaning and show only regular sound-changes which also affected their constituents; therefore they can be regarded as institutionalized, but not as lexicalized. 2.1.1. Word-character In compounds unification of the constituents starts with the imposition of compound stress: syntactic group bla´ck bo´ard vs. compound bla´ckbo`ard, cf. Germ. bla´ue Be´ere vs. Bla´ube`ere; this is, however, a fairly regular process and therefore not yet a case of lexicalization in the stricter sense. But unification can end in the coalescing of originally bimorphemic compounds into monomorphemic words, i.e. in total obscuration, e.g. Old English scı¯rgere¯fa > Modern English sheriff (but Old English scı¯r > shire; gere¯fa > reeve). 2.1.2. Naming function A consequence of the naming function is hypostasation (Germ. Hypostasierung), i.e. the phenomenon that the existence of a (simple or complex) word makes people assume that there is a referent (thing, person, etc.) in reality (in the world) which the word refers to. Often this assumption is, of course, correct; there exist things or persons to which words like bookcase, bookseller, etc. refer. But this is not always the case ⫺ authors of science fiction stories, for example, invent new worlds with things and actions that do not exist in our world, such as to uncreate, to unmurder (cf. Kastovsky 1982: 152). 2.1.3. Frequent use It can be assumed that frequent use was the reason for the lexicalization and eventual obscuration of a number of original compounds such as daisy, gospel, lord, lady, sheriff, woman, etc. It seems, however, impossible to decide precisely how often a formation must be used in order to become lexicalized, and, as stated in 1.3.3, there are a number of Old English compounds which are still commonly used in Modern English without having been lexicalized. 2.2. The scope of lexicalization The majority of the examples given so far were compounds, but lexicalization can also occur with other types of word-formation,
for example: prefix-formations such as Germ. abkratzen in the sense ‘to die’; suffix-formations such as reader (see 1.2) or sleeper in the sense ‘heavy piece of wood, metal, etc., supporting a railway track’, Germ. Zeitung ‘newspaper’ (originally ‘message, news’); zero-derivations (conversions), e.g., to corner someone ‘to put someone into a corner so that he cannot escape’ (Leech 21981: 226), onomatopoeic formations such as Lat. cuculus ‘cuckoo’ > Fr. cocu ‘cuckold’, clippings such as vamp ‘seductive woman’ (from vampire). Syntactic groups (phrases) can be lexicalized, too; in this case they are usually called idioms. Examples are (cf. also 1.2): adjective ⫹ noun groups (public school, in British English ‘private school’; Germ. geheimer Rat [as a title]); groups of noun ⫹ preposition ⫹ noun (man in the street/Germ. der Mann auf der Straße ‘average person’); verb ⫹ object groups (kick the bucket/Germ. ins Gras beißen ‘to die’); verb ⫹ particle constructions (phrasal verbs: to bottle up ‘keep (one’s feelings) under control’; to do in ‘to kill’). The borderline between syntactic groups (phrases) and complex words (generated through word-formation processes) is not always easy to draw, anyway: syntactic groups can, for example, coalesce into complex words (e.g. Germ. geheimer Rat to Geheimrat; Old English do¯mes dæg to Modern English doomsday); combinations with the first element in the genitive are regarded as compounds by some scholars but as syntactic groups by others (craftsman, driver’s seat, bull’s-eye, etc.); phrasal verbs are certainly lexical units, but since they are separable (they done the old woman in [Pygmalion]), they cannot be regarded as complex words. 2.3. Types of lexicalization Lexicalization can affect all linguistic levels, see, e.g., Bauer (1983: 50⫺61), Grimm (1991: 71⫺121). Often, several of these types co-occur in a formation. 2.3.1. Graphemic lexicalization Graphemic lexicalization usually reflects phonological changes, e.g. tuppence /1tvp(e)ns/ for twopence. 2.3.2. Phonological lexicalization This can be divided into segmental and suprasegmental. (a) Segmental lexicalization: Lexicalization of segmental features occurs when the elements of a complex word develop phoneti-
150. Lexicalization and demotivation
cally different in independent use. This is especially frequent in compounds; here the causes may be loss of stress, shortening of an element, assimilation, etc.; often two or more of these factors are combined and lead to obscuration (opacity). Woman (from Old English wı¯f > Modern English wife and man /man, mon/ > man /mæn/), for example, shows shortening of the first element (before two consonants), loss of stress on the second element, and assimilation of the end of the first element to the beginning of the second element (regressive assimilation): Old English wı¯fman > wı˘fman > wimman > Modern English woman /wwmen/; Germ. hohe Zeit vs. Hochzeit. As stated in 1.4, there is a scale from slight obscuration (e.g. cupboard /kvbed/ vs. cup /kvp/ and board /bc6d/) to total obscuration; totally obscured compounds such as barn (< Old English bere-ern ‘barleyhouse’), lord, lady (< Old English hla¯fweard, hlæfdige), lammas, gospel, gossip, etc. have become monomorphemic, whereas partly or slightly obscured compounds can still be regarded as bimorphemic, e.g. holiday. Sometimes the pronunciation is obscured, but not the spelling, e.g. in breakfast (/brekfest/ vs. / bre=k/ and /fa6st/). Phonological lexicalization can also occur with suffixes; in most of the formations with the (no longer productive) suffix -th, the underlying simplex has been changed in the derivation: broad – breadth, deep ⫺ depth, etc. (which can be explained historically as i-mutation caused by the suffix -ı¯Ìo¯ > -th). For words to take on a different form as the basis of derivations was relatively frequent in Old English, while it is rare in Modern English, apart from derivative alternants due to loan-influence, e.g. consume /ken1sju6m/ ⫺ consumption / ken1svm(p)sn/, decide /d=1said/ ⫺ decision / de1s=zen/. (b) Suprasegmental (prosodic) lexicalization: Compounds often have a different stress pattern from corresponding syntactic groups; in English and German, for example, syntactic groups typically have double stress (i.e. two main stresses), whereas compounds typically have a main stress on their first element and a secondary stress on their second element, e.g. bla´ck bo´ard vs. bla´ckbo`ard, Germ. schö´nes We´tter vs. Schö´nwe`tter. Some loan suffixes retract the accent, e.g. -a´tion (decla´re /d=1klee/ vs. declara´tion /dekle1reisn/), -e´e (train vs. traine´e), Germ. -ieren (Lack ⫺ lackı´eren, Marsch ⫺ marschı´eren). Nouns de-
1629 rived from phrasal verbs by zero-derivation (conversion), on the other hand, change the main accent from the second to the first element: to pick u´p > a pı´cku`p; to le`t do´wn > a le´tdo`wn. But all these phenomena are apparently synchronically productive processes and thus do not yet lead to lexicalization in the stricter sense. The stage of phonological lexicalization is reached, however, if, for example, the second element of a compound loses its secondary stress and is also reduced phonetically as a consequence, e.g. in some institutionalized compounds with man, such as chairman /tseemen/, where the second constituent is pronounced as /men/ instead of /mæn/ and thus approaches the character of a suffix. But -man is not a genuine suffix yet, because the weakening of /mæn/ to /men/ probably took place after the formation of these compounds. In some cases, however, suffixes originated from the weakened second elements of compounds, e.g. the suffix -ful /f(e)l/ from the word full /fwl/ (careful, grateful, thankful); frequently lexicalized elements thus became productive in a different function. In other cases weakening of the second element due to loss of the (secondary) stress apparently was the first step to obscuration, especially in a number of older compounds. 2.3.3. Morphological lexicalization This type of lexicalization can also be classified into a number of subtypes. (a) Linking elements, e.g. Germ. Arbeit but Arbeitsvertrag. Bauer (1983: 53 f.), who regards the insertion of linking elements as one kind of morphological lexicalization, points out that often their addition follows synchronically productive rules and thus cannot be regarded as lexicalization in the strict sense, but there can also be diachronic changes resulting in genuine morphological lexicalization: Germ. Schwanenhals, for example, retains the old form of the genitive, whereas in independent use Schwan now has the s-genitive (des Schwans); something similar applies to Lat. pater familias, which also retains the old genitive, whereas in classical Latin, the genitive of familia is familiae. While linking elements are fairly frequent in German, they are rare in English (-in- in nightingale < Old English nihtegale, cf. Germ. Nachtigall), apart from the -s, which can, however, usually be regarded as the genitive -s, sometimes combined or merged with the plural -s, cf. ladies’ room, salesman, townsman, etc.
1630
XVII. Morphologischer Wandel I: Theoretische Probleme
(b) Unproductive word-formation patterns or elements: Words formed according to unproductive patterns or containing unproductive affixes are also morphologically lexicalized. In English this applies to the socalled imperative compounds (pickpocket, dolittle, speakeasy), cf. (Lyons 1977: 547 f.), or to derivations with the suffix -th (warmth, growth) ⫺ formations like length, strength are thus lexicalized phonologically as well as morphologically. English moreover has a large number of loan-words where the suffix then also became productive in English. Words with these suffixes fall into two groups: (i) loans where the base does not occur independently in English, e.g. atheism, atheist (but no *athe-), baptism, baptist (but no *bapt-), ⫺ these could be regarded as lexicalized, but usually they are just regarded as loan-words; they show, however, that the distinction between loanwords and English word-formation is also sometimes blurred. (ii) Formations where the base also occurs independently in English, e.g. escape > escapism, escapist, but these are not lexicalized (at least not morphologically). (c) Blocked morphemes: Blocked (unique) morphemes are elements of compounds which occur just in one compound and which do not or no longer occur as independent words, e.g. many first elements of the days of the week such as Tues- in Tuesday/Germ. Dienstag, or rasp- in raspberry, Germ. Bromin Brombeere, -gale/-gall in nightingale/Germ. Nachtigall (cf. also Sauer 1992: 340⫺345). Blocked morphemes are frequent in English place-names, e.g. the first elements of Manchester, Gloucester, Worcester, Salisbury, etc. (d) Obscured compounds: Morphological lexicalization has, of course, also taken place in obscured compounds which have become monomorphemic, e.g. nostril (< Old English nos-Ìyrel, lit. ‘nose-hole’), see 2.3.2 (a). (e) With partly obscured compounds the question arises of whether the obscured elements should still be regarded as allomorphs of the corresponding independent words, e.g. whether Mon- /mvn/ in Monday should be seen as an allomorph of moon /mu:n/. The answer probably depends on whether the speakers still feel a semantic connection between the forms, which is probably not the case in moon and Mon(day), while it is probably true (due to the spelling) in cup(board) /kv-/ and cup /kvp/.
2.3.4. Semantic lexicalization Semantic lexicalization entails change of meaning. This is often subdivided into addition of semantic features, which results in restriction (narrowing) of meaning, loss of semantic features, which conversely results in extension (widening) of meaning, and a mixture of both. These changes also happen in simplex words, e.g. fowl (Old English fugol ‘bird’ > Modern English fowl ‘farmyard bird’) and conversely bird (Old English bridd ‘young bird’ > Modern English bird ‘bird’), but in connection with lexicalization we are only concerned with semantic changes in complex words. Examples are: (a) Addition of semantic features, e.g. Germ. Hochzeit (see 1.1.3). Here belong probably also formations containing implied features in addition to the stated features (in the terminology of Warren 1978): Middle English hand-whil means ‘time needed to turn one’s hand’, i.e. ‘very short time, moment’; Germ. Augenblick has the same implied feature ‘very short time’. Strictly speaking, hand-whil/ Germ. Augenblick are probably rather cases of idiomatization than of lexicalization. Addition of features also occurs if one element of a phrase is lost through ellipsis and the remaining element in addition takes on the meaning of the lost element, e.g. private (soldier), Germ. Süddeutsche (Zeitung). For the distinction between general (systematic) features such as in sleepwalker/Germ. Schlafwandler [⫹habitual] and idiosyncratic features such as in pushchair [⫹for children], see also 2.4. (b) Loss of semantic features, e.g. arrive (< Fr. arriver < Lat. *adripare, *arripare) ‘come to a place’, originally ‘land at the shore’. Cases where a constituent loses its meaning entirely are treated under demotivation. (c) Mixture (combination) of addition and loss of features, e.g. holiday (has lost the feature ‘religious’ and added the feature ‘free from work or school’), blackboard (has lost the feature ‘black’, because blackboards are nowadays usually green or white, and added the feature ‘made to write upon’). This phenomenon seems to be more frequent than simple loss or addition of features. Here belong probably also instances of metaphoric transfer, e.g. the use of bottleneck ‘neck of a bottle’ for ‘narrow part of a road which slows down traffic’ (DCE 21987). Demotivation can also result in change of semantic features:
150. Lexicalization and demotivation
a Junggeselle (Germ.) usually is no longer young, but he is unmarried. (d) As pointed out in connection with reader, the same formation can be semantically lexicalized in several ways. Examples are numerous among the phrasal verbs, e.g. to hold out ‘to offer’, ‘to continue to exist’, ‘to continue in spite of difficulties, endure’. The examples discussed so far seem fairly clear, but semantic description in connection with the question of lexicalization and idiomatization also has to face a number of problems (cf. Bauer 1983: 56⫺59); I shall mention two of them. (e) In connection with added semantic features it is sometimes unclear where the borderline between a semantic (linguistic) description of a word and an encyclopedic description of the thing or being concerned lies. A typewriter can be linguistically explained as an ‘instrument that writes with types’ and as such is not lexicalized (but rather institutionalized), but of course a typewriter has many more elements to it than just types. The same problem often arises with plant and animal names: There are many animals that hop around in the grass (e.g. hares), but only a certain kind of insect is called grasshopper/ Germ. Grashüpfer. (f) In other cases the distinction between semantic processes (lexicalization or idiomatization) and word-formation processes is not easy to make. The so-called exocentric compounds, with the subgroups of bahuvrihi (possessive) compounds like redcoat ‘person who has/a red coat’, paleface/Germ. Bleichgesicht, etc. and of imperative compounds like dolittle, pickpocket, Germ. Vergißmeinnicht, Fürchtegott ‘person who performs the action indicated by the verb’ all contain the element ‘person’ (or ‘animal’ as in stickleback, or ‘thing’ as in greenback, scarecrow), which is, however, not expressed on the surface. There are two possible explanations for this, one in semantic terms and one in terms of word-formation: These formations could be seen as a kind of metonymy with a part standing for the whole, or the action standing for the agent, and the feature ‘person’ (or ‘animal’ or ‘thing’) could then be regarded as an additional semantic feature, which would make them lexicalized (or idiomatized) compounds. But since this added feature can also be regarded as typical of the entire patterns of word-formation concerned, it could be ar-
1631 gued with Kastovsky (1982: 167 f.), who follows Marchand (21969), that all formations based on these types (paleface, pickpocket) are regular formations with a complex determinant and a zero-morpheme as a determinatum, e.g. redskin/Ø ‘someone who has a red skin’ pickpocket/Ø ‘someone who picks (other people’s) pockets’. Marchand (21969) and Kastovsky (1982) classify the zero-morpheme as a zero-suffix with the systematic meaning ‘person’ (or ‘animal’ or ‘thing’), apparently as a parallel to explicit suffixes such as -er (with the systematic meaning ‘person’, as in baker, or ‘instrument’, as in bottleopener). This would, moreover, classify the exocentric formations as (zero) derivations rather than as compounds. 2.3.5. Syntactic lexicalization This is the most problematic of the types. At least in English and German, there are fewer clear cases of syntactic lexicalization than of semantic, phonological and morphological lexicalization. (a) Bauer (1983: 59) says that syntactic lexicalization occurs when a syntactic pattern according to which complex lexemes have been formed falls into disuse. An example of this would be Germ. Vergißmeinnicht ‘forget-menot’, derived from the sentence vergiß mein(er) nicht, whereas in Modern German the construction is vergiß mich nicht. In this connection Bauer also mentions the so-called imperative compounds (pickpocket, scarecrow, to which Vergißmeinnicht belongs, too). As stated in 1.5, this pattern seems to be unproductive in present day English and is thus morphologically lexicalized, but it is not syntactically lexicalized (Vergißmeinnicht being an exception) because the underlying sentence pattern verb ⫹ object, e.g. ‘(to) scare crows (away)’ is still usual in Modern English. (b) Many compounds show a change of word order as compared with the underlying (or paraphrase) sentence or phrase, e.g. as green as grass J grassgreen, the door at the front J the front door. But since this change of word order is a fairly regular phenomenon in compounding, it should not be regarded as indicative of lexicalization. 2.3.6. Mixed lexicalization Often two or more types of lexicalization are combined in one word. This has become clear from some of the examples given above: holi-
1632
XVII. Morphologischer Wandel I: Theoretische Probleme
day and Germ. Hochzeit (see 1.1.3, 1.2, 2.3.4) show phonological as well as semantic lexicalization (different pronunciation of at least one constituent and change of meaning). Bauer points out “that once a form is lexicalized in one way it is easier for it to become lexicalized in others” (Bauer 1983: 61). This is certainly true for many of the totally obscured English compounds (barn, lord, lady, lammas, gospel, sheriff ), which show phonological changes with resulting morphological changes (originally bimorphemic, now usually monomorphemic), and semantic changes (idiomatization or semantic lexicalization). 2.4. Degrees of lexicalization There is a gliding scale from slight lexicalization to total obscuration (or demotivation). Examples of phonological and morphological obscuration have been given in 1.4 and 2.3.2 (relatively slight: cupboard, holiday; medium: bonfire; total: lord, lady). Here I want to concentrate on degrees of semantic lexicalization (idiomatization). Usually, a distinction is made in the literature between the addition of general and systematic semantic features and the addition of specific and idiosyncratic features (see, e.g. Lipka 1981: 125⫺ 129; Kastovsky 1982: 195 ff.; Sauer 1992: 46⫺ 50, 384⫺388). (a) General (systematic) additional features: Here belong, e.g., [⫹habitual], e.g. crybaby, [⫹professional], e.g. baker/Germ. Bäcker. Instruments as a rule have the feature [⫹purpose] and they can be paraphrased as ‘B is (made) for A’, e.g. bottle-opener/ Germ. Flaschenöffner, candlestick/Germ. Kerzenständer. This kind of analysis is not without its problems, however. Features like [⫹habitual] and [⫹professional] are certainly inherent in some formations, e.g. [⫹professional] in baker, miller/Germ. Bäcker, Müller, whereas in others they seem to depend on the context. Writer/Germ. Schreiber in the writer of this letter/der Schreiber dieses Briefes is a pure agent nominalization (‘the person who writes or has written this letter’) without any additional semantic features; in she is a regular letter-writer (or writer of letters) the feature [⫹habitual] is indicated by regular. Authors like G. B. Shaw and Hemingway were professional writers, but Caesar and Chaucer were writers in their pastime. (b) Specific (idiosyncratic) additional features: The elements of the historical term
Germ. Morgengabe/Old English morgengifu just state that it is ‘a gift given in the morning’, but the precise meaning is ‘a gift (present) given by the husband to his wife in the morning after the wedding night’. Cf. also examples discussed in 1.2, such as ghostwriter, reader, etc. Leech (21981: 225⫺227) calls the addition of such specific and unpredictable semantic features petrification. As far as I can see they are impossible to classify. (c) Additional semantic features can, of course, be combined: Compounds like crybaby and sobsister have [⫹habitual, ⫹derogatory]. There are also combinations of added general features and added specific features: In formations like playboy, callgirl the general features [⫹habitual] or [⫹professional] are combined with added (or deleted) specific features: a playboy is usually not a boy, but a man; he plays in a very special way, he is rich and does not work. A crybaby, moreover, usually is no longer a baby, a sobsister is not necessarily a sister, etc. (d) As has been shown before (cf. reader), the same formation can also be non-lexicalized (have no additional semantic features) in one sense and be heavily lexicalized (have specific features) in another sense. (e) It is not always clear where the general additional features end and the specific (idiosyncratic) features begin. Crybaby, sobsister/ Germ. Heulsuse also have the feature [⫹derogatory], which, according to Marchand (21969: 73 f.), occurs relatively frequently in formations of the type verb ⫹ noun. [⫹derogatory] could therefore be regarded as a fairly general semantic feature, too, but it is not usually mentioned as such in the literature, although it is usually given as a stylistic label in dictionaries. (f) It is not even always clear where exactly lexicalization (or idiomatization) begins. If one follows the distinction made by Bauer (1983), the use of a formation in just one of its possible meanings to the exclusion of others is a case of institutionalization but not yet of lexicalization. Institutionalization entails that in the paraphrase only one verb (or a group of synonymous verbs) can be used, e.g. for stone wall ‘the wall consists/is made of stone’, and not, e.g. *‘the wall produces stone’ or *‘the wall resembles stone’. Lipka (2002: 114) mentions the historical highwayman among the unpredictable idioms with rather specific semantic features. It can, how-
150. Lexicalization and demotivation
ever, be paraphrased as ‘man who robs (others) on the highway’ (cf. DCE 21987), and according to this analysis it would be a simple case of institutionalization, excluding meanings such as *‘man who meets others on the highway’ or *‘man who helps others on the highway’. On the other hand, while one author listed country house among the compounds without additional features, Lyons (1977: 535⫺544) spends several pages to point out that, at least in British English, it means not just ‘house in the country’, but rather ‘large house in the country owned by a member of the aristocracy, as opposed to his town house’ (cf. Sauer 1992: 119). (g) Features which are additional in some formations can already be part of an element in others. [⫹derogatory] or [⫹negatively evaluated] is additional in crybaby, playboy, etc., but in Middle English flescheslust/Germ. Fleischeslust it is probably inherent because the flesh was (is) regarded as inherently weak and sinful by the church. Similarly, in formations such as Germ. Langweiler, Trödler, the feature [⫹derogatory] is not added but inherent in one of the constituents (Langeweile, trödeln).
3.
Demotivation
Demotivation has been described as the loss of the sign-character of one or both constituents of a complex lexeme. Seen from the semantic aspect it indicates that one or both elements have lost their original meaning (and in some cases have acquired a new meaning). Thus demotivation partly overlaps with the phenomenon of semantic lexicalization. But before discussing demotivation in some more detail, the underlying and antonymous term motivation should be briefly explained. 3.1. Motivation Since F. de Saussure it has been fairly generally accepted that the linguistic sign is on principle arbitrary, i.e. unmotivated, because as a rule there is no natural connection between the signifiant (expression side, phonetic form) and the signifie´ (content side, meaning). This is clear from several facts: Both sides can change, the expression side through sound-changes (e.g. Old English sta¯n > Middle English sto¯n > Modern English stone / stewn/), and the content side through semantic change, i.e. change of meaning (e.g.
1633 Germ. Gift, originally ‘gift, present’ [as still in Mitgift ‘dowry’] J ‘poison’). Moreover, the same meanings or denotata can be expressed by different signifie´s in different languages, e.g. Lat. arbor, Eng. tree, Germ. Baum, as well as in the same language; in the latter case they are called synonyms and are often dialectally or stylistically differentiated, e.g. begin, commence, start or British English railway vs. American English railroad. But there are a number of exceptions to this general principle; words can be phonetically, morphologically or semantically motivated (see Ullmann 1962: 82⫺93; Fill 1980: 14⫺16). (a) Phonetic motivation: Onomatopoeic words, such as miaow/Germ. miauen, are imitations of animal cries and thus phonologically motivated; phonological motivation also plays a role in words formed with the help of sound-symbolic elements, such as Eng. -ash (indicating quick movement (ending suddenly)) in clash, crash, dash, flash, lash, etc.; for details see Marchand (21969: 397⫺428). (b) Morphological motivation: Complex words, especially compounds, prefix- and suffix-formations are morphologically motivated, i.e. if one knows the meaning of the constituents (which are unmotivated in themselves), one can predict or at least make an informed guess at the meaning of the entire combination. If the speaker/hearer knows for example, that re- means ‘again, once more’ and tell ‘narrate’ he will also know that retell means ‘to tell again’; this aspect has been called type-familiarity. (c) Semantic motivation: This applies in cases of semantic transfer by metaphor or metonymy; if one speaks, for example, of the foot of a mountain/Germ. Fuß des Berges, or if one calls a narrow part of a road a bottleneck. Compounds used metaphorically are thus morphologically and semantically motivated; the same applies to metaphorically used syntactic groups like wet blanket (cf. Lyons 1977: 548). As stated earlier, exocentric nouns (bahuvrihi nouns such as paleface and imperative nouns such as dolittle) can generally be seen as metonymic formations (‘person with a pale face’, ‘person who does little’, etc.). Semantic transfer, of course, also occurs in simplexes, e.g. in fox/Germ. Fuchs ‘clever person full of tricks’ (in Germ. also ‘horse with a reddish colour’, in Eng. also ‘sexually attractive woman’).
1634
XVII. Morphologischer Wandel I: Theoretische Probleme
3.2. Loss of motivation Motivation on all the levels mentioned can also be lost; I shall concentrate on loss of morphological motivation. There are several ways of classifying this phenomenon, e.g. according to extralinguistic/intralinguistic changes or according to the number of constituents affected (cf. Lipka 1981: 127 f.; Sauer 1992: 49 f.). (a) Extralinguistic changes are changes in the cultural background or in our knowledge of the world. Shoemakers/Germ. Schuhmacher, Schuster and watchmakers/Germ. Uhrmacher usually no longer make shoes or watches, but repair them; blackboards were originally black, but are nowadays usually green or white; a Schreibfeder (Germ.) is now usually made of steel and no longer of feathers (quills). Strictly speaking, in all these cases it is not the meaning of the compounds or derivations that has changed, but the activity or structure of the referent; shoemaker can still be paraphrased as ‘someone who makes shoes’, as opposed to Germ. Junggeselle, which cannot be paraphrased as *‘der Geselle ist jung’ or *‘junger Geselle’. (b) Intralinguistic changes are due to a change of meaning in one or both constituents. In mincemeat ‘mixture of fruit and spices ...’ meat does not have its modern meaning ‘edible flesh’, but retains the older meaning ‘food in general’. Germ. Mitgift ‘dowry’ retains the old meaning of Gift, namely ‘gift, present’, whereas the modern meaning of Gift is ‘poison’. (c) Loss of meaning in one constituent: The first constituent has lost its original motivation and therefore its original meaning in, e.g. blackboard, Whitsunday (originally, white garments were worn on this Sunday), Germ. Bockbier, Bockwurst, the second constituent in bluebottle (a kind of blue fly), mincemeat, sweetmeat, Germ. Handschuh, Bahnhof, Lehrkörper. The original motivation has often been lost in plant and place names: speakers usually do not know what, for example, horses have to do with horse-chestnuts or horsethistles; many towns are still called Newtown, Newton/Germ. Neustadt, although many of them were founded in the Middle Ages. A routine formula such as how do you do has also partly lost its motivation. Here belong also blocked morphemes, which do not (or no longer) carry any meaning in themselves, but serve to differentiate
one compound from another; Tues-/Germ. Diens- differentiate Tuesday/Dienstag from other days of the week, rasp-/Germ. Himdifferentiate raspberry/Germ. Himbeere from other berries (blueberry, blackberry, Germ. Blaubeere, Stachelbeere). (d) Loss of meaning in both constituents: Here belong, e.g., animal names such as butterfly (which is neither a fly nor is it known what its connection with butter should be), understand (connected neither with under nor with stand), Germ. Junggeselle (who is neither young nor a Geselle ‘companion’, but unmarried), Schornstein (with a blocked morpheme as its first element) ⫺ cf. phrases like alter Junggeselle. A French example is beaucoup ‘a lot’ (no longer ‘a nice blow’). It has been said that formations such as butterfly, Junggeselle, etc. have a misleading motivation (cf. Fill 1980: 70). A number of place names and surnames have also lost their original motivation: Oxford/Germ. Ochsenfurt were originally settlements at a ford where oxen could cross; the towns are still there, but the fords for oxen have gone. People called Fowler, Forester, Germ. Bartenschlager, Strohschneider have quite different occupations or professions than their ancestors had. Obscured compounds which have become monomorphemic (barn, gospel, lord, sheriff ) have, of course, also lost their motivation completely. Loss of motivation in both elements is furthermore frequent in phrasal idioms, e.g. white elephant, red tape, including phrasal verbs such as to do in ‘to kill’, etc. 3.3. Degrees in the loss of motivation As with lexicalization in general, there are also degrees in the loss of motivation. Formations such as shoemaker/Germ. Schuhmacher, are still analyzable and probably represent only slight cases of demotivation, since we still know about the cultural background that existed during their formation. Formations with blocked morphemes or where one constituent has lost its meaning (raspberry, Germ. Brombeere, Handschuh) show a middle degree of demotivation, and formations where both elements have lost their meaning as well as entirely obscured compounds show total loss of motivation (butterfly; Germ. Junggeselle; lord ), cf. Fill (1980: 69⫺73). The loss of motivation and accordingly the degree of idiomatization is stronger in Großmutter than in Großstadt: a Großstadt is a large city (although the minimum size of a Großstadt
1635
150. Lexicalization and demotivation
has been officially defined), whereas the groß in Großmutter has lost the meaning ‘large’ (see Fleischer 51982: 13). There will probably always be difficult or disputed cases. Some dictionaries say that the reason for (i.e. the motivation of) the plant name strawberry is not known, which would make it a partly demotivated compound (with loss of meaning in straw). I believe, however, that strawberry fruits were (as they still are) protected from decay by surrounding them with straw or fine shavings, which would make strawberry a fully motivated compound (‘berry which is protected with straw’). 3.4. Remotivation Native as well as loan words that have partly or entirely lost their motivation are sometimes remotivated. (a) If the spelling is still intact this can happen through spelling pronunciation, e.g. in waistcoat /1wesket/ > /1we=stkewt/; this seems to be frequent in place-names, e.g. Cirencester /1s=s=te/ > /1sa=erenseste/. (b) In other cases it can happen through popular etymology (folk etymology, secondary motivation; cf. Olschansky 1996). Here an unmotivated word or the unmotivated part of it (e.g. a blocked morpheme) is changed so as to coincide with existing morphemes. This can be done with loan-words which were never motivated in the receiving language: Lat. levisticum became Liebstöckel in German and lufestice ‘love-stitch’ in Old English, cf. Modern English lovage; asparagus (monomorphemic) was turned into sparrowgrass (bimorphemic). It can also happen with native words which lost their motivation partly or entirely: The second element of Old English brydguma lit. ‘bride-man’ was exchanged so as to yield bridegroom; the first element of Germ. Sintflut (‘big flood’ with blocked Sint-) was exchanged so as to yield Sündflut (‘flood [as a punishment] for the sins’). But the new morphological motivation is often just superficial, once more a kind of misleading motivation: aspargus/sparrowgrass is neither a kind of grass nor have sparrows anything to do with it. (c) Sometimes the same combination survives in different stages of transparency or lexicalization: the Old English syntactic
group ha¯lig dæg ‘holy day’ survives as syntactic group holy day, as lexicalized compound holiday and as still more lexicalized (obscured) compound Halliday (only used as a surname).
4.
References
Bauer, Laurie (1983), English Word-Formation. Cambridge: Cambridge University Press Brugmann, Karl (1900), “Über das Wesen der sogenannten Wortzusammensetzung” [reprinted in: Lipka & Günther (1981, eds.), 135⫺178] DCE (21987) ⫽ Longman Dictionary of Contemporary English, Della Summers (editorial director). Harlow: Longman [11978] [now 4 2003] Faiß, Klaus (1978), Verdunkelte Compounds im Englischen. Tübingen: Narr Fill, Alwin (1980), Wortdurchsichtigkeit im Englischen. Innsbruck: Institut für vergleichende Sprachwissenschaft Fleischer, Wolfgang (51982), Wortbildung der deutschen Gegenwartssprache. Leipzig, Tübingen: Niemeyer [11969] Götz, Dieter (1971), Studien zu den verdunkelten Komposita im Englischen. Nürnberg: Hans Carl Grimm, Ursula (1991), Lexikalisierung im heutigen Englisch am Beispiel der -er Ableitungen. Tübingen: Narr Kastovsky, Dieter (1982), Wortbildung und Semantik. Düsseldorf: Schwann-Bagel Leech, Geoffrey (21981), Semantics: The Study of Meaning. Harmondsworth: Penguin Lipka, Leonhard (1981), “Zur Lexikalisierung im Deutschen und Englischen”. In: Lipka & Günther (eds.), 119⫺132 Lipka, Leonhard (1992), “Lexicalization and Institutionalization in English and German”, Linguistica Pragensia 1, 1⫺13 Lipka, Leonhard (2002), English Lexicology. Tübingen: Narr [ 11990 under the title An Outline of English Lexicology] Lipka, Leonhard & Günther, Hartmut (1981, eds.), Wortbildung. Darmstadt: Wissenschaftliche Buchgesellschaft Lyons, John (1968), Introduction to Theoretical Linguistics. Cambridge: Cambridge University Press Lyons, John (1977), Semantics, Vol. 1⫺2. Cambridge: Cambridge University Press
1636
XVII. Morphologischer Wandel I: Theoretische Probleme
Marchand, Hans (21969), The Categories and Types of Present-Day English Word-Formation. Munich: Beck [11960] OED (21989) ⫽ The Oxford English Dictionary, ed. by Sir James Augustus Henry Murray et al.; 2nd ed. by John Andrew Simpson & Edmund S. C. Weiner. Oxford: Clarendon Press [11933]
Grammar of the English Language. London: Longman Sauer, Hans (1992), Nominalkomposita im Frühmittelenglischen. Tübingen: Niemeyer Ullmann, Stephen (1962), Semantics: An Introduction to the Science of Meaning. Oxford: Blackwell
Paul, Hermann (1880), Prinzipien der Sprachgeschichte. Tübingen: Niemeyer [51920]
Warren, Beatrice (1978), Semantic Patterns of Noun-Noun Compounds. Göteborg: Acta Universitatis Gothoburgensis (Gothenburg Studies in English 41)
Quirk, Randolph & Greenbaum, Sidney & Leech, Geoffrey & Svartvik, Jan (1985), A Comprehensive
Hans Sauer, Munich (Germany)
Olschansky, Heike (1996), Volksetymologie. Tübingen: Niemeyer (Germanistische Linguistik 175)
151. Change in productivity 1. 2. 3. 4. 5.
Introduction Types of changes in productivity Synchronic vs. diachronic productivity Linguistic factors affecting the productivity of bases References
1.
Introduction
Most languages reveal fluctuations in the productivity of word formation patterns over time. Despite this, the diachronic aspect of productivity has been little investigated in a systematic fashion. A discussion of changes in productivity depends, first of all, however, on a coherent definition of the notion of productivity (cf. Art. 33). Different scholars have used the term in different ways (cf. Botha 1968: 149), although common to most definitions is an implicit notion of frequency of occurrence, as in Bauer (1983): “Basically, any process [...] is said to be productive if it can be used synchronically in the production of new forms, and non-productive if it cannot be used synchronically in this way.” (Bauer 1983: 18)
Thus, in order for some process to be productive, new words must be generated. It follows that a greater degree of productivity should be reflected in a greater number of new forms. Thus, in German, for instance, the suffix -bar ‘able’ is both productive and frequent, e.g. trinkbar. Forms that are nonproductive or lexically idiosyncratic must be listed in the lexicon, and appear before the level of lexical insertion.
2.
Types of changes in productivity
Within the group of productive processes, some are more productive than others, while non-productive processes can be accounted for in terms of a list of the bases to which they have applied. Changes in productivity may affect roots as well as affixes. It is well known that in English, for instance, a number of new affixes not existing in Old English were borrowed, e.g. -ment as in involvement, -ity as in purity, mal- as in malformed, etc., while other native ones disappeared or declined in use. Although words containing the suffix -ment are still analyzable, the suffix is no longer productive as can be seen from the fact that new words are not being coined. Other nominalizing suffixes are preferred instead and existing forms with -ment have become lexicalized. Similarly, all nominalizations from adjectives which end in -th are now lexicalized, e.g. warmth, depth, length, etc., and most also have unproductive root forms. No new forms can be created with -th so that *realth, for instance, is not synchronically possible. Moreover, new suffixes are created all the time by analysis of existing lexemes, e.g. forecast led to funcast (i.e. the weather outlook for sporting activities), quickcast (i.e. a brief summary of the weather), travelcast (i.e. forecast for travellers), etc. Productive processes may become less productive over time, while non- or semi-productive processes can become more productive. The upper limits on productivity are
1637
151. Change in productivity
vague. Gruber (1976: 322) has suggested that to be fully productive, an affix must be usable with all [bases] definable by some semantic, syntactic or phonological property. Few, if any, affixes would meet this criterion. Many scholars have, however, suggested that the suffix -ness is fully productive (cf. Jespersen 1924). Its productivity extends to both major and minor categories: ⫺ major categories: goodness, jackassness, wellness; ⫺ minor categories: muchness, threeness, betweeness, whyness; ⫺ phrasal/semi-phrasal: broken-heartedness, fed-up-ness, up-to-dateness, know-nothingness. Some of these formations provide counterexamples to generalizations about word formation, namely that only major categories can be the product of word formation rules and that only form classes can be used as bases in the formation of derivatives (cf. Aronoff’s principles of word formation; cf. Aronoff 1976: 21). In addition, however, -ness managed to attach itself to native as well as nonnative roots and therefore was able to compete with French suffixes such as -ity, cf. e.g. stableness vs. stability; oddity vs. oddness, etc. Looking at the cognate suffix -nis in the history of German, we see that the opposite state of affairs prevails in the modern language, where many of the Old High German formations have disappeared, and most of the modern formations were established by 1800. As in English, there is a competing French suffix -ität, which applied mainly to foreign bases. Unlike English, however, -nis never really extended itself to foreign bases and most of the formations are deverbal. There are a total of 428 derivations in Mater’s (1971) reverse dictionary, more than 3/4 of which are compounds or derivatives of only 29 bases. Word forms ending in -ität account for nearly as many of the total abstract noun formations (48 percent in -ität as compared to 52 percent in -nis), as shown in Table 151.1.
English German
-ness/-nis/?
-ity/-ität/-(i)te´
33 percent 52 percent
67 percent 48 percent
Tab. 151.1: Formations in -ness/-nis and -ity/-ität/ -(i)te´ in English and German
From a comparative historical perspective it is clear that not only has -ness met a more productive locus in English, but the same is true of English -ity compared with the productivity of its cognate in French. In French there is a total of 797 words ending in the suffix -ite´/te´. In English there is a total of 1926. Nouns ending in -ity in English account for 33 percent of the total number of -ness and -ity formations, while those ending in -ness account for 67 percent (cf. Lehnert 1971; Juilland 1965). The difference between English and French on the one hand and English and German on the other is quite striking. Scholars such as Thompson (1975: 347) have drawn a distinction between productivity and analogy. An analogical formation is a new formation clearly modelled on one already existing lexeme and does not give rise to a productive series. The problem in drawing an absolute distinction between productivity and analogy is that an analogical formation may provide the impetus for a series of formations. Thus, at one extreme would be cases which are more clearly analogical formations resulting in the creation of only one new form which cannot be accounted for by rule, e.g. wargasm, from war ⫹ orgasm (see Bauer 1983: 96, 100). This would be the limiting case of productivity. At the other end, however, are cases such as sit-in providing the model for teach-in, swim-in, and hamburger providing the model for cheeseburger, fishburger, etc. (see Art. 148). A somewhat different example of a change in word formation patterns which involves an alteration in productivity can be seen in German. In Middle High German compounds with first elements ending in -keit were formed without any linking element by adding one stem to another, e.g. Gerechtigkeitliebe ‘love of justice’ consists of Gerechtigkeit ‘justice’ ⫹ Liebe ‘love’. In modern German, however, all words ending in -keit add a linking element -s (i.e. Kompositionsfugenelement) when they enter into compounds. Lexemes which were institutionalized before the change in linking element either have now added -s in modern German, e.g. Gerechtigkeitsliebe, or have become lexicalized in their old form since they can no longer be generated by synchronic productive rules. It is also possible for the same two lexemes to form compounds at two different time periods with different linking elements. In such cases the compounds may constitute minimal pairs
1638
XVII. Morphologischer Wandel I: Theoretische Probleme
which differ in meaning, e.g. Wassernot ‘drought’ vs. Wassersnot ‘flood’ (cf. Bauer 1983: 53 f.).
3.
Synchronic vs. diachronic productivity
The extent to which it is possible to distinguish between synchronic and diachronic productivity depends at least partly on maintenance of a distinction between synchrony and diachrony which has become increasingly less tenable. Most scholars recognize that a rigid boundary between the two is difficult to defend and that synchronic speech forms reflect diachronic processes. Matthews (1972: 191, n.2) notes that the “introduction of diachronic evidence ... in no way implies that it is a diachronic and not a synchronic fact that it is evidence for.” If this view is accepted, then there can be no clear-cut boundary between diachronic and synchronic productivity. Words occur all the time with varying degrees of probability in both speech and writing. A large proportion of complex word forms heard everyday are nonce formations, and do not become established in the language. It is only with hindsight that we can determine if a word has become established and thus contributed to the productivity of a morphological process over the long term rather than just momentarily. The rules allowing speakers to create new words by applying patterns of word formation can be thought of as a matter of synchrony while the subsequent fate of established lexemes is a matter for diachrony. Although some have said that processes such as back formation are strictly diachronic (cf. Art. 60), there must be some synchronic rule which allows analogies to be made between apparently complex forms with other cases where the suffixed and non-suffixed forms are both lexemes, e.g. laze is derived from the earlier form lazy, possibly by analogy with forms such as craze and crazy (cf. Bauer 1983: 64 f.). It is possible to gain some retrospective idea of productivity, and hence some insight into its diachronic dimension by looking at a dictionary, as in the comparisons made above among German, French and English with respect to -ität, -ite´/te´ and -ity (cf. 2). In providing a list of the established lexemes of a given language, a dictionary includes the existing products of various word formation rules which have applied in the past to produce conventionalized and accepted words.
One way to study productivity diachronically is to look at the output of various word formation rules as recorded in dictionaries at different points in time. It is particularly instructive to compare word formation processes which compete for the same bases. In such cases the factors constraining productivity become clearer, and it is evident that synchronic restrictions on productivity are essentially the result of diachronic changes. Romaine (1983: 186 f.) examined a list of 100 adjectives in the contemporary Oxford English Dictionary (OED) and compared these with lists obtained from two earlier dictionaries, Samuel Johnson’s Dictionary of the English Language (1755) and Thomas Sheridan’s A General Dictionary of the English Language (1780) to see how many -ity and -ness formations were listed, e.g. pompous: pomposity/pompousness. Over time there has been an increase in the number of words taking both suffixes. For example, the OED records 60 words out of the list of 100 which may take either -ity or -ness, compared to only 22 for Johnson’s dictionary and 18 for Sheridan’s. Similarly, more words take neither suffix in Sheridan’s and Johnson’s dictionaries (15 and 21, respectively) than in the OED (1). There are only minor differences among the dictionaries with regard to those words which are listed as taking -ness only. The OED records 32, Samuel Johnson’s dictionary, 35, and Thomas Sheridan’s, 42. The differences are greater for -ity. The OED lists only seven words which can take -ity only, while Sheridan has 25 and Johnson, 20. This method could be used with other affixes, and to look at individual authors writing in earlier time periods. This procedure, however, gives only a partial glimpse of productivity because dictionaries list only actual, attested and not possible words (cf. Art. 171). It must not be forgotten that dictionaries are in some sense arbitrary collections of the words in a language at a given time and will not necessarily reflect the actual written or spoken usage of a speech community (cf. Mugdan 1984: 240⫺ 259). Thiel (1973: 379) found that 62 percent of the compound words in an issue of the German popular news magazine, Die Zeit, were not listed in dictionaries. Morphological processes may be exceedingly productive and never result in established lexemes which are recognized by a dictionary as being legitimate words of the language. One well known example is found in the English use of infixed
1639
151. Change in productivity
Base Type
Formations in -ness Number
-ous -ive -able -al -ible -ile -ic
generousness transitiveness reasonableness musicalness fallibleness fragileness domesticness
Formations in -ity Example
Number
Example
514 391 346
generosity transitivity reasonability
94 96 200
169 99 17 15
musicality fallibility fragility domesticity
315 156 75 63
Tab. 151.2: Formations in -ness and -ity in relation to base type
forms such as absobloodylutely. Here the infixes are potential word forms rather than bound morphs like most other English affixes. Although infixed forms such as these have been in colloquial use for some time, it is unlikely any of them will ever appear in a dictionary due to their taboo status. Just because such words do not appear in the dictionary, this does not mean that infixation is non-productive, or that the forms resulting from them are not established lexemes in the mental lexicon.
4.
Linguistic factors affecting the productivity of bases
Just as it is impossible to talk about productivity in a meaningful sense synchronically without taking into account other factors, e.g. phonological, syntactic, it is impossible to do so diachronically. Although some have assumed that the morphological form of the base is the most important factor affecting productivity (cf. 4.1), for others, semantic considerations are primary (cf. 4.4). The successful spread and productivity of certain suffixes such as -ness can be accounted for in terms of the interaction of these factors (see also Baayen 1989). 4.1. Morphological form of the base as a factor restricting productivity It is obvious that the morphological form of the base plays an important role in determining productivity (cf. Aronoff 1976: 53 f.). Compare the difference in the number of -ness and -ity formations which obtain when the suffixes are attached to two distinct classes of adjectives, those ending in -ive, and
those ending in -ile. There are approximately five times as many words ending in -iveness as there were ending in -ivity. The second base, however, yields different results. The number of words ending in -ility overwhelmingly exceeded those ending in -ileness. Table 151.2 shows some results for different base types from Romaine (1983: 182) using Lehnert’s (1971) dictionary. The boxes indicate where the locus of productivity lies for each suffix. Nouns ending in -ness are most productively formed with bases ending in -ous, -ive, and -able, while -ity occurs most frequently with -al, -ible, -ile, and -ic. Although a consideration of the morphological type of base imposes a considerable refinement on the analysis of productivity, it still does not deal with the issue of recurrent vs. possible lexemes, and whether the productivity of existent types bears a regular relationship to that of possible types. Aronoff (1976: 37) claims that speakers of a language have intuitions about productivity which entail the notion ‘likelihood of being a word of the speaker’s active vocabulary’. He furthermore speculated that if pairs such as perceptiveness and perceptivity are presented to native speakers of English, they will say that although both are possible, perceptiveness is preferable. Romaine (1983), however, did not find this to be the case in an experiment conducted to test speakers’ intuitions about the productivity of -ness and -ity formations. Comparing the results with those from Table 151.1 based on the dictionary, Romaine found that in all cases where we would expect -ness to be more productive than -ity, it was not; nor was -ity more productive than -ness in all those cases where we would expect it
1640
XVII. Morphologischer Wandel I: Theoretische Probleme
to be. This indicates a significant discrepancy between productivity measured in terms of existent lexemes as recorded by dictionaries and possible lexemes as indicated by speaker judgements. The suffix -ity turned out to be more productive than -ness over all the base types. These results suggest that the factor of morphological type of base is itself a dependent and not an independent variable in productivity. There is also wide variation among the words belonging to a particular type, which indicates a greater degree of lexical idiosyncrasy than Aronoff allows. 4.2. Lexical restrictions on productivity Productivity may also be affected by idiosyncratic lexical restrictions. For example, the suffix -ric occurs only with bishop, e.g. bishopric (cf. Bauer 1983: 93). In some -ity nominalizations from adjectival bases ending in -ous, the -ous becomes -os-, e.g. curiosity, while in others it is deleted, e.g. voracity from voracious (cf. Aronoff 1976: 40). Because these processes are not predictable, these exceptions must be dealt with on an ad hoc basis in the lexicon. Another type of conditioning can be seen in the preference of suffixes such as -ee to apply more productively to transitive verb bases, e.g. employee (cf. Bauer 1983: 98). 4.3. Phonological restrictions on productivity Phonological factors affecting the productivity of bases include stress shift and vowel and consonant alternations. Speakers tend not to use words whose pronunciation they are unsure of. In order to add the suffix -ity to the base electric to produce the derived form electricity, a speaker must apply rules of stress shift and velar softening. In experiments testing the alleged psychological reality of these kinds of processes as formulated by Chomsky & Halle (1968) in their generative model of English phonology, it was found that when speakers were asked to apply such rules to bases which normally did not undergo them, e.g. methanity (methane ⫹ -ity), they did not use them (cf. Steinberg & Krohn 1975; Baldi et al. 1985). In other cases the phonological shape of the base determines whether it may be used as the input to a word formation rule. In French the diminutive suffix -ette is not normally added to bases ending in /t/ or /d/, while in English the adverbial suffix -ly is not productively added to adjective bases already
ending in -ly, e.g. *sillily. The suffix -eer shows a preference for bases ending in /t/, e.g. profiteer, racketeer, etc. In still other cases productivity is sensitive to syllable structure. Thus, the English prefix mal- may apply only to polysyllabic bases such as adapted, e.g. maladapted. Another case where syllable structure impedes productivity can be found in the infixed forms mentioned earlier. In most cases the infixes must be inserted immediately before the syllable of the base which bears the lexical stress so that forms such as *abbloodysolutely are ruled out. Stress is also a relevant conditioning factor since it is difficult to insert such infixes in words with stress falling on the first syllable with no subsequent subsidiary stress, e.g. solid. Most commonly infixes occur in words of three or more syllables (cf. Bauer 1983: 89⫺91). 4.4. Semantic restrictions on productivity There is one general semantic factor which limits the productivity of word formation rules, which can be called hypostatization. This means that words will not be formed to denote non-existent entities (cf. Bauer 1983: 85; Art. 80). The use of a word presupposes the existence of a relevant category. Some have suggested there is a simple relationship between productivity and meaning: a morpheme whose meaning is more semantically coherent will be more productive (cf. Aronoff 1976: 39). Comparing -ness and -ity, it can be easily shown that the semantics of -ness nominalizations has been highly coherent throughout history. The meaning of X-ness can be stated in terms of a choice among three operations on the base X: (a) ‘the fact that someone/something is X’, e.g. His kindness (‘the fact that he is kind’) amazes me. (b) ‘the extent to which someone/something is X’, e.g. His kindness (‘the extent to which he is kind’) amazes me. (c) ‘the quality or state of being X’, e.g. Kindness is a virtue. By contrast, the semantics of -ity derivatives is not nearly so coherent. Many of the established lexemes ending in -ity have strayed from their original abstract sense to denote concrete objects and now have a collective or technical sense (cf. Romaine 1985). The root sense of many -ity formations is abstract, i.e. a noun denoting a state, quality or condition,
151. Change in productivity
e.g. nobility in the sense of ‘nobleness’, the quality or state of being noble. Some of these formations then became used as nouns denoting occupations, offices, positions or titles signifying the abstract qualities of the occupations, e.g. nobility in the sense ‘class of nobles’. Many of the formations ending in -ty (from French -e´te´/te´) denote offices or titles of occupations, e.g. royalty, admiralty, mayoralty, etc. A number of these denote occupations or offices associated with feudalism and are no longer in use, e.g. squiralty, vassality, etc. Some -ity formations developed a sense of collectivity or totality of people belonging to an occupation or who have the quality of being associated with it, e.g. humanity, Christianity, polity, society, nationality, fraternity, etc. Some also came to be used as scientific or technical terms, e.g. gravity, velocity, and others to denote things having the abstract quality of the adjective base associated with them, e.g. oddity, antiquity, curiosity, etc. Thus, nouns ending in -ity have drifted away from the semantics of their derivational base. This semantic itinerary is established diachronically for -ness too, although the number of -ness formations which have strayed from their original sense is far smaller. In Old English there were some cases where a noun ending in -ness denoted something concrete, e.g. smirines ‘oil’, or the still surviving wilderness or witness used to refer to both the person offering evidence and the evidence or testimony itself. There is an inverse relationship between productivity and lexicalization: the most productive patterns are not lexicalized and fully lexicalized processes are not productive (cf. Bauer 1983: 88). Formations such as witness, business, wilderness and likeness are examples of this. There were also cases where -ness formations became used as titles connected with occupations, e.g. highness, holiness. The meaning of -ness in this sense is apparently still available to speakers synchronically since it is used humorously in Agatha Christie’s novel Death on the Nile, when an Indian character addresses the famous detective as “Your Sleuthness”. Finally, the word business refers to a collectivity. Where competing -ness and -ity nominalizations exist, they occupy different semantic spaces. Thus, nobleness refers only to the capacity or state of being noble, while nobility also refers to the capacity or state of being noble. Humanness refers to the quality of being human, while humanity refers to the col-
1641 lectivity of humans; antiqueness refers to the state of being antique, while antiquity refers to an old object or time period, etc. Lexicalized business referring to a collectivity has the counterpart busyness denoting the state of being busy. The semantic lexicalization of business is accompanied by phonological lexicalization in that the final vowel of busy is lost. Many words spread themselves over more than one slot in semantic space. For example, curiosity can denote both the condition or state of being curious and also an object having the quality or state associated with the adjective but curiousness can be used only to denote the quality or state of being curious. Oddity and oddness contrast in the same way. The fact that the semantics of -ity formations is less predictable and less regular is implicated in productivity since speakers will tend not to use words whose meaning is unclear to them. Van Marle (1988: 141 f.) has discussed a similar case in Dutch, where the greater semantic heterogeneity of deverbal adjectives ending in the suffix -lijk has led to its decline in productivity, while -baar has increased its domain of application. He also suggests that the loss of productivity of one of two competing word formation rules must be explained by internal factors, i.e. in terms of factors pertaining to the properties of the category of formations produced by that process. In Dutch it seems that there are non-standard dialects where -lijk has declined in productivity, but has not been replaced by -baar. Thus, -baar has not been responsible for the decline in productivity of -lijk. An important factor in this case is the large-scale occurrence of non-literal, metaphorical readings of words ending in -lijk, e.g. onberispelijk ‘perfect’ from the verb berispen ‘to rebuke, reprehend’. The meaning of the underlying verb has completely disappeared and the derivative occurs only in combination with the negative prefix on-. This represents a drift away from the core or inherent properties of the verbs to which the suffix is attached (cf. van Marle 1988: 144 f.). Once this happens, the semantics of the derivatives becomes less transparent. Semantic considerations are also probably responsible for the recent increased productivity of -free over -less (cf. Slotkin 1990). Originally, -free and -less were semantically equivalent, and in some formations, still are, e.g. sugarfree and sugarless. However, in other formations they are not, e.g. childless
1642
XVII. Morphologischer Wandel I: Theoretische Probleme
vs. childfree. The adjective ending in -less has negative overtones, while the one in -free is positive. Thus, to describe someone as childless implies that having children is desirable, but to be childfree imputes a positive value of being free from the burden of having children. The negative connotations of -less may have been acquired through the conventionalization of compounds such as feckless, hopeless and ruthless, which denote undesirable qualities. The original meaning of careless to indicate someone free from care is now expressed by carefree and careless has undergone pejoration. Thus, careless means ‘not taking care’ and the result of carelessness is generally negative. Now -less seems to be attached primarily to nouns to indicate undesirable connotations and has declined in productivity. The suffix -free, however, has increased its productivity and is used increasingly in a variety of nonce denominal adjective formations. Blocking conditions on productivity have been formulated in terms of meaning so they too may be seen as a type of semantic constraint. Aronoff (1976: 56) predicts that it will be impossible for there to be two words with the same meaning and the same root in one person’s lexicon at the same time. Others have also suggested that some possible formations will be blocked if there is already an existing lexeme with that meaning (e.g. Regel der besetzten Stelle, or preemption by synonymy; cf. Art. 85). Thus, we would predict that stealer is not likely to become conventionalized due to the existence of thief. There is no nominalization ending in -eur from voler ‘to fly’ in French due to the prior existence of a homophonous form voleur meaning ‘thief’. Similarly, the presence of nominals such as glory and fury blocks the application of a word formation rule suffixing -ity to the adjectives glorious and furious. Despite this, however, nominalizations ending in -ness are possible, i.e. furiousness, gloriousness. Blocking does not in any case prevent coining. It acts only as a brake on conventionalization. Formations in -ness often appear in the semantic slots of nouns formed with other abstract deadjectival noun forming suffixes. The individual limits and differences in semantic distinguishability between pairs such as sincereness/sincerity, chasteness/chastity, stableness/stability, etc. remain to be investigated.
4.5. Pragmatic factors restricting productivity Closely related to such meaning considerations affecting changes in productivity are pragmatic factors having to do with non-referential associations such as the stylistic and social connotations associated with a particular pattern of word formation. Since they are optional, word formation rules may be sensitive to the same kinds of constraints which affect other sociolinguistic variables, e.g. age, sex, social class, region, etc. From a diachronic perspective it is possible to establish that certain patterns of word formation vary with time-depth and style. Nonce formations are often glossed as such in written texts. In modern writings new terms are often put in inverted commas. When new formations are introduced, they may be glossed, e.g. cherite Ìet is luve (Ancrene Riwle) or used with a synonym, e.g. sikernesse and surete. The repetition of two synonymous forms, one native and one foreign, is still used in legal registers, where there is a high proportion of Latin and French vocabulary, e.g. life and limb, break and enter, legal and lawful, etc. Whether a new term gains in currency will depend on many factors, one of which is the status of the person who uses it. Where new concepts are concerned, e.g. upon the invention of the television or internet, new words are required, but prestige is often a factor in motivating new formations. Nowadays popular media, in particular, the world wide web, are an important influence. Historically speaking, almost as soon as French and Latin words were introduced into English, native prefixes and suffixes were added to them, e.g. stableness. As soon as sufficient lexical material was borrowed for foreign word formation patterns to be transparent and isolable, they could be used productively with both native and newly borrowed words, e.g. oddity. For the educated in society, knowing the derivational systems of both French and English made mixing foreign and native systems of derivational morphology possible. Authors such as Chaucer and Wyclif experimented with these types of formation very early. Wyclif, for example, uses doublets such as feersness (cf. Modern English: fierceness) and feerste (cf. Modern English: fiercity and ferocity) and bareyness (cf. Modern English: barrenness) and bareynte (cf. Modern English: *barrenty). Once massive borrowing in the Middle English period had disrupted the
1643
151. Change in productivity
prevailing native patterns of word formation, there were numerous derivatives which are closely related semantically and possibly also etymologically, but which have different roots, e.g. regal and royal, eat and edible, etc. In many cases only one member of these sets of roots remains productive in modern English. It does not seem possible to create new lexemes in English by adding suffixes to the root ed-, but eat has been productive in new forms such as eatery. English is remarkable in its ability to accommodate and integrate the derivational morphological systems of Romance with its own native Germanic one. Neither Czech nor German normally mixes native bases with foreign affixes or foreign bases with native affixes. Thus, German *Sterbation is ungrammatical, while English starvation is not. In so far as there are some restrictions on the compatibility of foreign and native elements, these can be considered to be of the same type as others which are due to the morphological form of the base (cf. 4.1). Aronoff (1976: 51 f.) proposes that bases have to be marked as [⫹latinate] or [⫺latinate]. Some foreign suffixes have never combined with native roots, but over time such restrictions have been relaxed. It is not clear what factors are responsible for this diachronic change in status. Speakers’ reactions to novel expressions may also affect productivity. When speakers notice a new form, they often oppose it and write letters to the newspaper complaining about the use of the new form on etymological, grammatical, semantic or aesthetic grounds. One example is Time magazine’s article (cf. Time 1962) entitled “The Nesselrode to ruin” complaining about the productivity of -ness. The article regarded the prolifness of -ness formations such as inwardness, thereness, matter-of-factness, etc. as a dangerous tendency which was bringing about a decline in standards of English usage. English speakers have also complained about the extension of -wise meaning ‘with respect to’ or ‘in regard to’ in formations such as moneywise, I’m o.k. which means ‘I have enough money’ (cf. Houghton 1968). This increase occurred sometime around the middle part of this century. Earlier in the century, it looked as if this suffix was unproductive since it existed primarily in a few lexicalized forms such as likewise, otherwise and clockwise.
4.6. Other factors restricting productivity Among the other possible factors limiting the productivity of word formation processes is simply time itself. The longer a process of word formation has been used productively, the more likely it is that a sizable number of the bases to which the rules can be applied will become institutionalized lexemes. This means that over time there will be fewer bases available for the process to apply to productively (see Bauer 1983: 98). The relationship between productivity and lexicalization is still not well understood. Nor is the interaction of all the factors already discussed. Bauer (1983: 99) speculates that potential formation of a particular word is blocked by a cumulation of factors whose precise weightings we are not in a position to assess at the moment. Van Marle (1988: 147) however, argues that semantic factors are primary and are in effect the root cause of productivity. Semantic coherence is a requisite for productivity. This implies that all word formation processes which are productive are semantically coherent, while those which display semantic coherence are not necessarily productive. This is supported by the existence of non-productive processes such as the denominal female personal suffix -in in Dutch. All existing formations in Dutch are regular, e.g. leeuw ‘lion’ and leeuwin ‘lioness’, etc., but newly coined formations are decidedly jocular in character.
5.
References
Aronoff, Mark (1976), Word Formation in a Generative Grammar. Cambridge/MA: MIT Press (Linguistic Inquiry Monograph 1) Baayen, Harald (1989), A Corpus-based Approach to Morphological Productivity. Ph. D. dissertation, Vrije Universiteit Amsterdam Baldi, Philip & Broderick, V. & Palermo, David S. (1985), “Prefixal Negation of English Adjectives: Psycholinguistic Dimensions of Productivity”. In: Fisiak, Jacek (ed.), Historical Semantics; Historical Word-Formation. Berlin: Mouton, 33⫺57 Bauer, Laurie (1983), English Word Formation. Cambridge: Cambridge University Press Botha, Rudolf P. (1968), The Function of the Lexicon in Transformational Generative Grammar. The Hague: Mouton Chomsky, Noam & Halle, Morris (1968), The Sound Pattern of English. New York: Harper & Row
1644
XVII. Morphologischer Wandel I: Theoretische Probleme
Gruber, Jeffrey S. (1976), Lexical Structures in Syntax and Semantics. Amsterdam: North Holland
ity in the Lexicon”. Australian Journal of Linguistics 3, 177⫺200
Houghton, Donald E. (1968), “The Suffix -wise”. American Speech 43, 209⫺215
Romaine, Suzanne (1985), “Variability in Word Formation Patterns and Productivity in the History of English”. In: Fisiak, Jacek (ed.), Papers from the 6th International Conference on Historical Linguistics [Poznan 1983]. Amsterdam: Benjamins; Poznan: Adam Mickiewicz Univ. Press (Current Issues in Linguistic Theory 34), 451⫺465
Jespersen, Otto (1924), The Philosophy of Grammar. London: Allen and Unwin Juilland, Alphonse (1965), Dictionaire inverse de la language franc¸aise. The Hague: Mouton Lehnert, Martin (1971), Rückläufiges Wörterbuch der englischen Gegenwartssprache. Leipzig: VEB Marle, Jaap van (1988), “On the Role of Semantics in Productivity Change”. Yearbook of Morphology 1, 139⫺154 Mater, Erich (1971), Rückläufiges Wörterbuch der deutschen Gegenwartssprache. Leipzig: VEB Matthews, Peter H. (1972), Inflectional Morphology: A Theoretical Study Based on Aspects of Latin Verb Conjugation. Cambridge: Cambridge University Press Mugdan, Joachim (1984), “Grammatik im Wörterbuch: Wortbildung”. In: Wiegand, Herbert Ernst (ed.), Studien zur neuhochdeutschen Lexikographie IV. Hildesheim etc.: Olms (Germanistische Linguistik 1⫺3/83), 237⫺308 Romaine, Suzanne (1983), “On the Productivity of Word Formation Rules and the Limits of Variabil-
Slotkin, Alan R. (1990), “Adjectival -less and -free: A Case of Shifting Institutional Currency”. American Speech 65, 33⫺49 Steinberg, Danny & Krohn, Robert (1975), “The Psychological Reality of Chomsky and Halle’s Vowel Shift Rule”. In: Koerner, Konrad (ed.), The Transformational Generative Paradigm and Modern Linguistic Theory. Amsterdam: Benjamins, 233⫺ 259 Thiel, Gisela (1973), “Die semantischen Beziehungen in den Substantivkomposita der deutschen Gegenwartssprache”. Muttersprache 83, 377⫺404 Thompson, Sandra A. (1975), “On the Issue of Productivity in the Lexicon”. Kritikon Litterarum 4, 332⫺349 Time (1962) ⫽ “The Nesselrode to Ruin”. Volume 79, May 11, 70
Suzanne Romaine, Oxford (United Kingdom)
152. Morphologische Entlehnung und Lehnübersetzung 1. 2. 3. 4. 5.
Zum Gegenstandsbereich Die außersprachlichen Bedingungen für Interferenz Grundtypen der Interferenz Gliederung der Phänomene Zitierte Literatur
1.
Zum Gegenstandsbereich
Die Veränderung natürlicher Sprachen ist nicht nur eine Angelegenheit der betreffenden Sprachen selbst, sie geht zu erheblichen Teilen auf den Kontakt mit anderen Sprachen und die dadurch ausgelösten Interferenzen zurück (einen allgemeinen Überblick geben Boretzky & Igla 1994). Im folgenden sollen einige Begriffe eingeführt werden. Leider muß man sagen, daß es keine allgemein akzeptierte Terminologie gibt und teilweise auch verschiedene Ansichten über den Cha-
rakter der im Kontakt auftretenden Prozesse bestehen, bedingt auch dadurch, daß manches nicht mehr empirisch zugänglich ist. Wir wollen uns deshalb auf das Notwendigste beschränken. Viele Linguisten meinen mit Interferenz (interference) die in der parole zweisprachiger Individuen zustandegekommenen Abweichungen von der intendierten Sprache. Wir wollen im folgenden darunter die Gesamtheit der durch Sprachkontakt bewirkten Phänomene verstehen. Die aktuelle Übertragung von Elementen aus einer Sprache in die andere, also den Prozeß selbst, nennen wir Transfer (transfer), den Übergang von transferierten Elementen und Relationen aus der parole in die langue aber Entlehnung (borrowing). Im folgenden wird in der Sache fast nur von Entlehnung die Rede sein. Wie die empirische Forschung gezeigt hat, bleibt
1645
152. Morphologische Entlehnung und Lehnübersetzung
grundsätzlich keine Ebene oder Unterebene der Sprache von fremden Einflüssen ganz verschont. Es gibt also kaum absolute universale Restriktionen für Interferenz; Versuche, solche Beschränkungen zu formulieren (vgl. Moravcsik 1978: 110⫺113), haben die außersprachlichen Umstände der jeweiligen Situation zu wenig berücksichtigt. Interferenzphänomene fallen uns vor allem im Bereich der Lexik und in etwas geringerem Umfang im Bereich der Wortbildung (s. 4.1) auf, aber auch die Übertragung von Lauten (mit oder ohne Konstituierung neuer Phoneme) und von syntaktischen Regeln ist nicht ungewöhnlich. Viel seltener begegnet uns die Übertragung von Flexionsmorphologie (s. 4.2.3), weshalb auch schon früh die Behauptung ausgesprochen wurde, diese Art von Interferenz existiere so gut wie nicht (Meillet 1921: 82; auch noch Oksaar 1972: 492). Die Erfahrung zeigt jedoch, daß auch Flexionsmorpheme übertragbar sind, wie bereits Schuchardt (s. Spitzer 1928: 195) betont hat. Häufiger finden wir jedoch im engeren grammatischen Bereich, also dem Bereich der grammatischen Kategorien und ihrer Ausdrücke, andere Interferenzphänomene: Entlehnung grammatischer Partikeln oder (flektierter) Auxiliare (s. 4.2.1); Lehnübersetzungen (loan translations, calques) von Partikeln, Auxiliaren (s. 4.2.2) bzw. umfangreicheren periphrastischen grammatischen Konstruktionen, d. h. sinngemäße Wiedergabe, sofern die genannten Elemente eine eigenständige Bedeutung erkennen lassen. Auch Entlehnungen aus dem Bereich der kleinen Wortklassen beinhalten mehr als syntaktische Phänomene, ohne doch direkt die Morphologie zu berühren. Sie können hier jedoch nicht behandelt werden. Dies alles sind interferenzbedingte Aufbauprozesse. Morphologie kann aber auch unter fremdem Einfluß abgebaut werden (s. 4.2.2). Es sei noch angemerkt, daß mit Entlehnung gebundener Morphologie am ehesten in solchen Sprachen zu rechnen ist, die selbst schon vergleichbare Morphologie haben. Von Transfer und Entlehnung gebundener Morphologie ist zu unterscheiden die morphologische Adaptation lexikalischer Elemente. Im ersten Fall werden gebundene Flexions- und Wortbildungsmorpheme entlehnt und mit Erbwörtern kombiniert, im zweiten Fall werden Wörter entlehnt und in der Regel mit ererbter Flexionsmorphologie versehen bzw. mit ererbten Wortbildungsmorphemen kombiniert. Hierbei ist z. B. zu untersuchen:
bis zu welchem Grad morphologische Adaptation erfolgt; nach welchen Kriterien Nomina in Genussprachen ein Genus zugeordnet wird; in welche Flexionsklassen die Nomina eingeordnet werden; welche Form der Verben (Wurzel bzw. Stamm, Präsensstamm, Präteritalstamm usw.) als Grundlage der morphologischen Adaptation dient; ob für Lehnwörter besondere Flexionsklassen entwickelt werden (wie z. B. im europäischen Romani; vgl. Boretzky 1989), u. a. m. Auch diese Thematik kann aus Raumgründen hier leider nicht weiter verfolgt werden.
2.
Die außersprachlichen Bedingungen für Interferenz
Interferenz zwischen zwei oder mehreren Sprachen verläuft normalerweise nicht ungesteuert, sondern in einer bestimmten Richtung, die durch das Verhältnis der am Kontakt beteiligten Gruppen diktiert wird. Die Qualität sprachlicher Strukturen ist dabei von geringer Bedeutung, entscheidend sind vielmehr außersprachliche Faktoren, die man grob mit dem Stichwort ‘Prestige’ umreißen kann ⫺ jedenfalls, soweit Entlehnung im engeren Sinne gemeint ist. Entlehnungen können nützlich sein, indem sie im Sinne einer Sprachökonomie (s. Martinet 1955) dazu beitragen, Lücken im grammatischen System zu füllen, Strukturen zu parallelisieren u. a. m., als notwendig wird man sie jedoch kaum je bewerten können. Zwar trifft es sehr oft zu, daß Wörter aus einer Kultur entlehnt werden, die neue Begriffe zu bieten hat, die in gewissem Sinne also notwendig sind; für die anderen Ebenen, insbesondere für die Morphologie, kann man dies jedoch schwerlich behaupten. Als Voraussetzung für die Entlehnung von Wortbildungsmorphemen und noch mehr von Flexionsmorphemen gilt, daß vorher bereits Interferenz auf anderen Ebenen stattgefunden hat, zunächst auf der lexikalischen, dann auf der syntaktischen und phonologischen (vgl. Moravcsik (1978: 110); auch von Thomason & Kaufman (1988: 20 f.) für normale Entlehnung, nicht für Substratwirkung akzeptiert). Wie mir scheint, darf man dies jedoch weniger als eine linguistisch-strukturelle Voraussetzung verstehen. Eher sind hier Faktoren wie Kulturgefälle, Intensität und Dauer des Kontakts und wohl auch die Frequenz der verschiedenen Morphemtypen entscheidend, die den Boden für weitergehende
1646
XVII. Morphologischer Wandel I: Theoretische Probleme
Entlehnungen psychologisch vorbereiten. Der durch solche Faktoren begünstigte hohe Grad an Zweisprachigkeit nimmt den fremden Strukturen viel von ihrer Fremdartigkeit. Es sei eingeräumt, daß es auch rein linguistische Vorbedingungen für Entlehnung von Morphologie gibt: Wenn Wortbildungsmorpheme übernommen werden, dann ist diesem Prozeß gewöhnlich die Entlehnung von Wortmaterial, in dem solche Morpheme vorkommen, vorausgegangen. Für reine Flexionsmorphologie gilt dieser Gesichtspunkt in viel geringerem Maße, denn der Transfer von Wörtern mit ihrer Flexion ist nur selten zu beobachten, am ehesten im Stadium des Sprachverfalls (language decay; s. Sasse (1990: 45); Art. 153; vgl. auch Problem “Mischsprachen” 3.3).
3.
Grundtypen der Interferenz
Verschiedene Arten von Interferenz haben unterschiedliche Konsequenzen für die beeinflußten Sprachen (s. Thomason & Kaufman 1988). Entlehnung als Übergang einzelner Elemente und Relationen aus einer Sprache in eine andere ist verständlicherweise überall beteiligt, wir können aber je nach Art und Ausmaß vier Typen von Interferenz unterscheiden: 3.1. Entlehnung Der bekannteste Typ ist die Entlehnung oder, wie wir in Ermangelung eines besseren Terminus sagen wollen, die normale Entlehnung, bei der keine der beteiligten Sprachen ihre Identität verliert. Die beeinflußte Sprache, die Kopiesprache (replica language) übernimmt Elemente verschiedener Ebenen aus der Modellsprache (source language, model language). Wenn die Interferenz größeren Umfang annimmt, kann sich der Charakter der Sprache wesentlich ändern; es können Strukturen akzeptiert werden, die den ererbten Strukturen zunächst durchaus fremd waren. Die Sprache bleibt jedoch “dieselbe”, sie wird in ununterbrochener Folge im natürlichen Spracherwerbsprozeß tradiert. 3.2. Sprachwechsel Interferenz kann auch durch Sprachwechsel (language shift) zustandekommen. Eine Sprechergruppe wechselt ihre Sprache, meist innerhalb eines relativ kurzen Zeitraumes, und nimmt aus ihrer Muttersprache oder ihrer ersten Sprache verschiedene Phänomene mit in die neu angenommene Sprache (Zielsprache,
target language) hinüber. Die aufgegebene Sprache wird dabei gewöhnlich als Substrat (substratum) bezeichnet. Gibt eine Gruppe ihre Sprache auf, ohne die Chance gehabt zu haben, die Zielsprache ausreichend zu erlernen, entstehen Kreolsprachen (v. a. in der Kolonialzeit entstandene Sprachen mit gebrochener Entwicklungslinie und Zügen aus verschiedenen Sprachen, deren genaue Entstehung umstritten ist; s. auch Art. 153). Die Unterscheidung zwischen normaler Entlehnung und Interferenz durch Sprachwechsel ist deshalb wichtig, weil sich die Ergebnisse deutlich unterscheiden: Nur im ersten Fall sind viele Lehnwörter aus der Modellsprache zu erwarten, im zweiten Fall findet man eher Lehnübersetzungen nach dem Muster des Substrats, gerade auch für den grammatischen Bereich. Dies hat vielleicht damit zu tun, daß Lehnwörter für die sprachwechselnden Sprecher leichter als fremde Elemente zu erkennen sind als die viel weniger als solche ins Bewußtsein tretenden Lehnübersetzungen und deshalb besser eliminiert werden können. Bezüglich der Morphologie und der grammatischen Kategorien kann man konstatieren, daß bei normaler Interferenz sowohl Morpheme als auch abstrakte Strukturen von grammatischen Kategorien übertragen werden, während beim Sprachwechsel zwar mit der Nachahmung von grammatischen Kategorien, aber kaum mit dem Transfer von morphologischem Material zu rechnen ist. Die Gründe scheinen dieselben wie bei der Vermeidung von Lehnwörtern zu sein. 3.3. Mischsprachen Ein weitreichender Fall von Interferenz, der früher in Abrede gestellt wurde, hat zu Mischsprachen (mixed languages; s. auch Bakker & Mous 1994, eds.) geführt. Im Unterschied zu normaler Entlehnung handelt es sich hier um Sprachen, die genetisch verschiedenen Wurzeln entspringen, also keine eindeutige Abstammungslinie haben. Sie gleichen darin den Kreolsprachen, zeichnen sich zusätzlich aber dadurch aus, daß einzelne Sprachebenen, ja selbst Unterebenen der Morphologie, verschiedener Herkunft sein können. Im Gegensatz dazu beziehen Kreolsprachen ihre Lexik im wesentlichen aus einer Sprache, ihre Kategorien bauen sie jedoch selbst (zum Teil unter Substrateinfluß) neu auf. Beispiele für Mischsprachen: (a) In Europa gehören dazu verschiedene, von Zigeunern gesprochene Idiome wie Angloromani und Calo´ (in Spanien). Hier
1647
152. Morphologische Entlehnung und Lehnübersetzung
ist die Lexik zum überwiegenden Teil ererbt, fast die gesamte Grammatik, Morphologie wie Syntax, entstammt aber dem Englischen bzw. dem Spanischen. Hier ein Beispielsatz aus dem Angloromani, der keine Abweichungen von der englischen Morphosyntax zeigt (englische Elemente fett; zu diesem Typ s. Hancock 1984; Boretzky 1985): (1) once apre a chairus a einmal auf indef Zeit indef Rommany chal chor-ed a Roma Kerl stehl-pr‰t indef rani chillico. Frau Vogel ‘Es war einmal ein Rom, der einen Truthahn stahl (vgl. engl. once upon a time a Romani guy stole a turkey)’ (b) In Kanada wird das Mitchif oder Me´tif von Indianern und Mischlingen gesprochen. Das Nominalsystem stammt oberflächenstrukturell aus dem Französischen, das Verbsystem aber oberflächlich wie tiefenstrukturell aus dem Cree (s. Bakker 1989). (c) Die Media Lengua aus Ecuador kombiniert spanische Lexik mit gebundener Quechua-Morphologie und -Syntax (s. Muysken 1981). (d) Im Ma’a oder Mbugu aus Kenia und Tansania ist ein südkuschitischer Grundwortschatz mit einer durchweg bantuisierten Grammatik vereint (s. Thomason 1983). 3.4. Diffusion Die Verbreitung bestimmter grammatischer Phänomene über große geographische Regionen, in denen miteinander nicht verwandte Sprachen gesprochen werden und die auch kaum durch lexikalische Entlehnung verbunden sind, läßt die Existenz einer weiteren Art von Interferenz vermuten. Wir stellen z. B. fest, daß die meisten afrikanischen Sprachen den Komparativ semantisch einheitlich mithilfe eines Verbs ‘passieren; übertreffen’ ausdrücken. Hier muß es eine Art Diffusion (Sapir 1921: 203, 206) abstrakter grammatischer Strukturen geben, die möglicherweise ohne lexikalische Entlehnung abläuft. Vielleicht ist eine Bedingung für diese Art von Interferenz, daß es zwischen den Sprechergruppen kein nennenswertes Prestige- und Kulturgefälle gibt. Dieser Typ kann als eine Unterart von Entlehnung (s. 3.1) angesehen werden.
4.
Gliederung der Phänomene
Zunächst werden die Phänomene behandelt, die häufiger vorkommen und weniger gravierende Einbrüche fremder Strukturen darstellen, es folgen dann die selteneren und ungewöhnlicheren. Die Entlehnung von Wortbildungsaffixen ist insgesamt häufiger anzutreffen als die von Flexionsmorphemen, offenbar weil erstere eher konkrete Bedeutungen haben und kaum in festgefügte Paradigmata eingebunden sind. Wenn ganz neue Affigierungstypen entlehnt werden, stellt dies bereits einen stärkeren Eingriff in die vorhandene Sprachstruktur dar. Beispiele für die Entlehnung von Flexionsmorphologie finden sich eher in kleinen Sprachen oder in Randdialekten von geringer sozialer Bedeutung, die uns besser bekannten westeuropäischen Sprachen sind fast frei davon. 4.1. Wortbildung Die Übertragung von Wortbildungsaffixen auf muttersprachliches Material wird meist als mittelbare oder vermittelte Entlehnung rekonstruiert: Die Affixe gelangen mit Lehnwörtern in die Replika, werden dort abstrahiert und dann mit Erbwörtern kombiniert (Weinreich 1970: 31). Dies mag in der Mehrzahl der Fälle so sein; man sollte die Vorstellung jedoch nicht auf alle Fälle verallgemeinern, denn es gibt Kontaktsituationen, für die dieses Szenario eher unwahrscheinlich ist. Das Albanische hat z. B. eine große Zahl von Suffixen aus dem Slawischen entlehnt, z. B. -ac, -ac¸, -ak, -ice¨, -ik, -ine¨, -ishte¨ u. a. (s. Xhuvani & C ¸ abej 1962), ohne daß sich slawische Lehnwörter in nennenswertem Umfang finden würden. Meines Erachtens reicht weitgehende Zweisprachigkeit oder Bilinguismus (bilingualism) aus, um solche Phänomene zu erklären. Die Entlehnung von Wörtern mit solchen Suffixen und die Übertragung von diesen Suffixen auf ererbtes Wortmaterial können parallel ablaufen, die Sprecher brauchen nicht unbedingt auf Lehnwörter zurückzugreifen, wenn ihnen die Kontaktsprache voll zugänglich ist. Wir sollten uns nicht nur den strukturellen, sondern auch den psycholinguistischen Aspekt solcher Situationen klarmachen. Wenn es bereits eine große Menge von Lehnwörtern mit den entsprechenden Morphemen gibt, die wichtige Begriffe besetzen und frequent sind, dann ist die Annahme gerechtfertigt, die Morpheme seien von diesen Lehnwörtern auf ererbtes Wortmaterial übertragen worden. Wo es jedoch nur eine
1648
XVII. Morphologischer Wandel I: Theoretische Probleme
Handvoll von Lehnwörtern des entsprechenden Typs gegeben hat, die u. U. eher ad hoc gebraucht werden, ist es unwahrscheinlich, daß sich die Sprecher an ihnen ausgerichtet haben. Bei Lehnübersetzungen oder semantischen Nachbildungen von Wortbildungs- und Flexionsmorphemen (s. 4.2.2) wäre die hier in Zweifel gezogene Position ohnehin schlecht haltbar; wir müßten ja dann mit innersprachlichen Übersetzungen rechnen. Beispiele für Präfixe: englisch re- < französisch/lateinisch re-: re-write, re-draw, reedify, evtl. über Fälle wie re-turn < französisch retourner; schwedisch be- < deutsch be: be-vittna ‘bezeugen’, be-gagna ‘gebrauchen’, be-teende ‘Verhalten’, evtl. über Lehnwörter wie be-fria, be-döva ‘betäuben’ vermittelt. Rumänische und slawische Verbpräfixe sind in verschiedene Romanidialekte eingedrungen; vgl. do-lav ‘nehmen’ < slawisch do‘zu-’, za-bistrava ‘vergessen’ < slawisch za-. Hierbei ist das Romaniverb nach serbisch zaboraviti geformt worden. Es handelt sich um eine Verstärkung (reinforcement), da bistrava bereits ‘vergessen’ bedeutet. Weiter des-mekljol ‘auftauen’ < rumänisch des- ‘ent-’ u. ä. (eigene Erhebungen). In diesen Fällen haben wir es eher mit Innovationen, Bereicherungen zu tun, da vor der Entlehnung keine direkten Entsprechungen vorhanden waren. Bei den neuen Präfixen im Romani ist sogar eine neue Wortbildungsklasse mit einem neuen Morphemtyp entstanden, denn Romani kannte keine Präfigierung. Anders verhält es sich in Fällen, in denen ein neues Suffix ein altes substituiert, z. B. bei deutsch -er < lateinisch -arius, in Bäck-er, Töpf-er usw., wofür es ältere Bildungen wie mittelhochdeutsch becke, neuhochdeutsch Beck gab; oder bei dem maskulinen Deminutivsuffix -ici im Romani < griechisch -ı´tsi, z. B. in bobici (eine Bohnensorte) zu bobo ‘Bohne’. Hier gibt es ein ererbtes Äquivalent -orˇo, das allerdings meist nur an Erbwörter tritt (eigene Erhebungen). Eine Besonderheit bietet der folgende Fall: das Istrorumänische, eine dem Rumänischen nahestehende Sprachform auf Istrien, hat aus dem Kroatischen eine Reihe von Verbaffixen übernommen. Dieser Vorgang geht jedoch über bloße Wortbildung hinaus und betrifft grammatische Kategorien, denn die Affixe dienen dem Aufbau eines Aspektsystems, also der Schaffung von imperfektiv-perfektiven Verbpaaren nach slawischem Muster, (zu den Details und weiterer Literatur s. Sala 1988).
Zwischen Wortbildung und Flexionsmorphologie stehen die folgenden, in der Geschichte des Romani zu beobachtenden Phänomene. Lehnwörter werden im Romani anders als die voreuropäischen Erbwörter, und zwar meist nach griechischen Mustern, flektiert; so gehen ererbte Maskulina mehrheitlich in der Grundform Nominativ Singular auf -o´ aus, z. B. cˇhav-o´ ‘Junge’, entlehnte Maskulina aber mehrheitlich auf unbetontes -os, das aber als -o´s- auch in den obliquen Kasus erhalten bleibt, weshalb -os nicht einfach als Morphem ‘nom.sg.m’ bezeichnet werden kann, sondern auch Wortbildungsfunktion hat. Als Beispiel sei du´b-os < slawisch dub ‘Eiche’ genannt. Dieses -os geht auf griechisch -os und -o´s zurück, die im Griechischen gleich häufig sein mögen. Aus diesem Fall ersieht man, daß die Replika hier eine aktive Auswahl getroffen hat, indem sie ein Allomorph gewählt hat, das von dem ererbten -o´ möglichst verschieden ist. Interessanterweise finden sich unter den Gräzismen des Romani außerhalb Griechenlands so gut wie keine Wörter, die im Griechischen selbst auf -os oder -o´s ausgehen. Das deutet darauf hin, daß sich die Roma bei solchen Entlehnungen am ehesten nach Frequenzverhältnissen des Griechischen und anderen Faktoren gerichtet haben. Auch bei den Affixen gibt es neben der direkten Entlehnung von Morphemen als lautliche Gebilde den Fall der Übersetzung bzw. semantischen Nachformung. So ist schwedisch genom- ‘durch-’ nach deutsch durch-, z. B. in genom-föra ‘durch-führen’, geprägt worden; genom war im Schwedischen zunächst nur Präposition. Das Sorbische hat Verbalpartikeln (Präverbien) nach deutschem Muster entwickelt (s. Schuster-Sˇewc 1977: 466), obwohl es als slawische Sprache über eine große Zahl von Verbpräfixen verfügt. So finden wir das Element horje als Entsprechung von deutsch ‘hinauf’ in horje-nosic´ ‘hinauftragen’. Wie im Deutschen kommt es aber auch in der Position am Ende des Prädikats vor: (2) ja nosˇu te meˇchi horje ich trage die Säcke hinauf 4.2. Flexionsmorphologie und Syntax 4.2.1. Entlehnung grammatischer Partikeln und Auxiliare Solche Prozesse betreffen zunächst die Morphosyntax, teilweise aber auch die Morphologie. Auch wenn diese Elemente unmittelbar
152. Morphologische Entlehnung und Lehnübersetzung
bei dem Wort stehen, das sie determinieren, sind sie leichter isolierbar als echte Flexionsmorpheme, denn sie haben oft eine gut greifbare Bedeutung und werden wohl deshalb leichter entlehnt. Zunächst ein Fall aus dem Anglo-Irischen (Hibernian English), einer vom irischen Substrat beeinflußten Variante des Englischen. Hier ist ein Element be [bi] < irisch bı´- übernommen worden, das zum Ausdruck eines habituellen Präsens dient (s. (3)). (3) we be often wondering where he gets the money ‘wir pflegen uns (oft) zu wundern, woher er das Geld nimmt’ (Henry 1957: 170) Wie man sieht, steht be nicht unmittelbar beim Verb, hat also nicht den Status eines Flexionsmorphems angenommen. In Romanidialekten des Kosovo ist tuj/tyj u. ä. < albanisch tue/tye u. ä. für den Ausdruck des Progressivs übernommen worden. Anders als im Albanischen verbindet sich tuj mit dem finiten Verb, nicht mit einem Partizip oder sonstigem Infinitum. (4) avel tuj rov-ol er.komm part wein-3.sg ‘er kommt weinend’ (Boretzky 1989: 367) Hier ist der Marker bereits zu einem Verbpräfix und zu einem Flexionsmorphem geworden, obwohl er dies im Albanischen nicht war. Da ältere indogermanische Verbflexion normalerweise suffigierend ist, werden solche Fälle oft falsch kategorisiert. Auch bereits bestehende Kategorien können auf diese Weise eine Ausdruckserneuerung erfahren. Nicht selten tritt dann die Partikel an die Stelle älterer suffigierter Flexion. So hat das Albanische lateinisch magis als maˆ/me¨ zur Bildung des Komparativs übernommen (Meyer 1881); vgl. maˆ (i) mire¨ ‘besser’ (bzw. ‘der bessere’). Zwischen Partikel und Adjektiv können Artikelformen treten, so daß hier kein Flexionsmorphem entstanden zu sein scheint. Es ist aber auch möglich, Komparativmarker und Artikel als präfigierte Flexionskette zu interpretieren. Vor dem lateinischen Einfluß muß es synthetische Bildungen für die Komparation gegeben haben. Hier dürfte nicht nur das Komparationselement, sondern das ganze Verfahren aus einer spätlateinischen Vorlage stammen (vgl. auch rumänisch mai für den Komparativ).
1649 4.2.2. Entlehnung grammatischer Strukturen durch Lehnübersetzung Möglicherweise noch häufiger als die Entlehnung von Partikeln, Auxiliaren u. a. in ihrer lautlichen Form ist die Lehnübersetzung (loan translation, calque), die Nachahmung fremder Kategorien mit eigenen Mitteln. Naturgemäß läßt sich hier der Beweis für Interferenz weniger leicht führen als bei direkter Entlehnung von Morphemen, weshalb solche Fälle oft von Verfechtern eines (nur) innersprachlich bedingten Wandels in Zweifel gezogen werden. Wo aber eine innersprachliche Erklärung mit einer kaum wahrscheinlichen Häufung von Zufällen rechnen müßte, sollten wir doch Entlehnung annehmen. Kategorien, die auf diese Weise eingeführt werden, haben eher periphrastische als morphologische Struktur. Im Gefolge einer innersprachlichen Entwicklung kann der Ausdruck aber morphologisiert werden. Als Beispiel für letzteres kann z. B. der bulgarische postponierte Artikel -aˇt, -ta, -to, -te dienen, der von einem Demonstrativpronomen stammt. Das Verfahren dürfte aus balkanromanischer (und evtl. albanischer) Quelle stammen; vgl. bulgarisch zˇena-ta, rumänisch femei-a, albanisch grua-ja ‘die Frau’ (s. Sandfeld 1930: 165⫺173). Wir wissen nicht genau, wie und wo die Artikelbildung im Bulgarischen eingesetzt hat, fest steht jedoch, daß das Altbulgarische keinen Artikel hatte. Der Artikel ist heute im Bulgarischen eine echte Flexionskategorie. Im Neupersischen finden sich PossessivSuffixe (neben selbständigen Elementen) statt -pronomina wie sonst im Indogermanischen; vgl. pidar-am, -at, -asˇ usw. ‘mein, dein, sein Vater’ usw. wie türkisch baba-m, -n, -sı usw. (s. Doerfer 1967: 59). Die persischen Suffixe stammen von enklitischen Pronomina. Auch hier ist echte Flexion entstanden. Entlehnung ist sehr wahrscheinlich, weil das Persische noch viele andere türkische Einflüsse aufweist (s. Doerfer 1967). Diese Art von Interferenz reicht verschieden weit, je nachdem ob sie nur das morphologische Muster betrifft oder die Übernahme einer Kategorie zusammen mit dem morphologischen Muster. Im Prinzip ist es nicht einmal ausgeschlossen, daß fremder Druck zur Entwicklung einer Kategorie führt, ohne daß im Ergebnis eine strukturelle Ähnlichkeit zwischen Modell und Replika zu erkennen ist. Der Beweis für Interferenz ist hier besonders schwer zu erbringen. Fremder Einfluß wird dann wahrscheinlicher, wenn es sich um eine in den Sprachen der Welt selten anzu-
1650
XVII. Morphologischer Wandel I: Theoretische Probleme
treffende Kategorie handelt, die in zwei nicht näher verwandten Nachbarsprachen vorkommt. So muß beim bulgarischen Narrativsystem (System der indirekten, wiedergegebenen Rede) mit türkischer Herkunft gerechnet werden, obwohl die Ausdrücke nur schwache strukturelle Ähnlichkeit miteinander aufweisen (zu den Details s. Roth 1978: 55 ff.) Bisher sind nur Fälle angeführt worden, in denen durch Interferenz neue Strukturen geschaffen wurden. Interferenz kann sich aber auch negativ auswirken und zur Beseitigung ererbter Strukturen führen. Naturgemäß ist auch hier der Interferenzbeweis nicht leicht zu führen, weil Kategorien auch im Zuge von internem Sprachwandel schwinden können. Bei einem Fall aus dem baltisch-finnischen Sprachkontakt dürfte aber Interferenz die Ursache sein: in den livonischen (tamischen) Dialekten des Lettischen, die auf livischem, also finnischem Substrat erwachsen sind, gibt es starke Tendenzen, das Femininum und damit das ganze 2-Genus-System zu beseitigen. Dies wirkt sich nicht nur dahingehend aus, daß die Kongruenz maskulin wird, sondern daß die ursprünglichen Feminina direkt maskuline Kasusformen annehmen; vgl. saimniec laimig-s (statt fem. laimig-a) ‘die Wirtin ist glücklich’; vist-s ‘Huhn’ (statt vist-a); im Dativ Plural ma¯t-am statt ma¯t-em ‘Mutter’ (s. Endzelin 1922: 342). Da dieser Prozeß nur in Mundarten auf finnischem Substrat zu beobachten ist, kann bloßer Zufall als Erklärung kaum herhalten. 4.2.3. Entlehnung von Flexionsmorphemen Flexionsmorpheme werden zugegebenermaßen recht selten übernommen. Sprachen unterziehen sich diesem Einfluß wohl am ehesten, wenn sie unter besonders starkem fremdsprachigen Einfluß stehen. Solche Fälle werden auch deshalb so selten entdeckt, weil man eher in den Standards und zu selten in der parole nach ihnen sucht (Weinreich 1970: 33). Es hat den Anschein, als ob bei dieser Art von Interferenz meist fremde an die Stelle eigener Morphologie tritt, also die Zahl der Allomorphe eines Flexionsmorphems vermehrt wird, ohne daß irgendwelche zusätzlichen Funktionen ausgedrückt oder gar neue grammatische Kategorien eingeführt würden. Strukturelle Faktoren können allenfalls da eine Rolle spielen, wo in einer Kategorie ein schwächer markiertes Allomorph durch ein stärker markiertes ersetzt wird (vgl. Weinreich 1970: 33) und dadurch im Sinne der Natürlichkeitstheorie (Prinzip des Konstruk-
tionellen Ikonismus; vgl. Mayerthaler 1981: 23) innerhalb einer Kategorienopposition die stärker markierte Kategorie auch den merkmalhaltigeren Ausdruck erhält. Recht häufig finden wir die Entlehnung fremder Pluralmorphe. Der folgende Fall ist wahrscheinlich anders, nämlich als ein Wandel in der Flexionsklasse, zu verstehen. Im Jiddischen wird der maskuline Plural -im, der hebräischer Herkunft ist, auf einige Nomina deutscher Herkunft übertragen; z. B. pojer-im ‘Bauern’, doktojr-im ‘Doktoren’ (s. Weinreich 1970: 31). Da es im Jiddischen aber hebräische Wörter in großer Zahl gibt und diese dahin tendieren, die hebräische Flexion zu behalten, muß man -im als ein zum System des Jiddischen gehöriges Allomorph betrachten und den Fall als innersprachliche Expansion eines Allomorphs interpretieren. Auch der nächste Fall könnte eine ähnliche Geschichte haben. In einem türkisch beeinflußten Romanidialekt, der heute in Athen gesprochen wird, sind die 1. und 2. Person Plural durch türkisch -muz bzw. -nuz verstärkt worden; z. B. gel-am ‘wir gingen’ zu gela-mus, gel-en ‘ihr gingt’ zu gela-nus. Dieser Dialekt weist aber türkische Lehnverben in großer Zahl auf, die durchweg türkische Präsens- und Präteritalflexion beibehalten. Es ist daher wahrscheinlich, daß es sich hauptsächlich, wenn auch nicht ausschließlich, um einen innersprachlichen Übertragungsprozeß handelt. Die Übertragung ist wohl auch durch die Ähnlichkeit der Morpheme begünstigt worden (ausführlicher s. Igla 1996). Die weiteren Fälle lassen sich eher nicht in der Weise deuten. Das ambigene Plural-Allomorph des Rumänischen -uri wird in Romanidialekten an Lehnwörter verschiedener Herkunft angefügt, und zwar in der Form -ur(j) oder -ur(j)-a, d. h. um das Plural-Allomorph -a´ des Romani erweitert; vgl. taljanurja ‘Italiener’. Wie das Beispiel zeigt, tritt das Allomorph anders als im Rumänischen auch an Belebte, was wieder eine innersprachliche Expansion ist. Es ist aber weniger sicher, daß -urj(a) an sich aus rumänischen Lehnwörtern isoliert worden ist, denn die Roma folgen im allgemeinen konsequent der Regel, Lehnwörter grammatisch voll zu adaptieren und fremde Flexionsmorpheme nur dort beizubehalten, wo sie ganze Syntagmen aus der Kontaktsprache ad hoc übernehmen. Solche Prozesse fallen unter Interferenz, sind aber nicht als Entlehnung zu werten. Es wäre wohl eine verkürzte strukturalistische Sicht, wenn man bei allen diesen Fäl-
1651
152. Morphologische Entlehnung und Lehnübersetzung
len den Faktor Wortentlehnung allein in Betracht zöge, ohne den Faktor der geläufigen Zweisprachigkeit zu berücksichtigen. Das deutsche Plural-Allomorph -en ist in rätoromanische Dialekte eingeführt worden; vgl. vacch-en ‘Kühe’ (Sala 1988: 8). Im Albanischen tritt türkisch -lar in begrenztem Umfang an albanische Wörter, besonders an Familiennamen; vgl. Belko-llare¨ ‘die Belkos’ (s. Boretzky 1975: 238). Die Regeln der Modellsprache wurden also nur selektiv kopiert. Das Baskische bildet regelmäßig eine infinite Verbform auf -tu, für die allgemein lateinischromanische Herkunft (< -tus, -ta, -tum) angenommen wird (s. Hurch 1989: 12, mit weiterer Literatur). Der genaue Hergang des Prozesses ist leider nicht bekannt. Im Fall des Meglenorumänischen aus Nordgriechenland, das vor allem slawische Sprachkontakte hatte, sind an die 1. und 2. Person Singular Präsens -u, -i einer Reihe von Verben die slawischen Morpheme -m, -sˇ anfügt worden, z. B. aflu-m ‘ich finde’, afli-sˇ ‘du findest’ (s. Capidan 1925: 94; Weinreich 1970: 33). Da die ursprünglichen Morpheme nicht wegfallen, ist dies ein Fall von Verstärkung. Die eher seltenen slawischen Lehnverben werden in diesem Dialekt nicht in ihrer slawischen Form flektiert, weshalb die Verstärker-Elemente direkt aus dem Slawischen übernommen sein dürften. Das Albanische hat vom Türkischen das Adverbmorphem -c¸e übernommen; vgl. vendc¸e ‘nach heimischer Art’, malcor-c¸e ‘wie die Bergbewohner’ (s. Boretzky 1975: 268). Im Albanischen finden wir zwar -as/azi als Adverb-Marker, aber sie werden vornehmlich nicht mit Nomina kombiniert. Daher ist auch eine grammatische Bereicherung mit der Entlehnung verbunden. Nordtadzˇikische (d. h. persische) Dialekte haben u. a. aus dem Özbekischen, einer Türksprache, verschiedene Kasussuffixe übernommen, z. B. den Ablativ -dan, vgl. yakom klasasˇ-dan ‘aus der ersten Klasse’. Daneben besteht aber weiter die Möglichkeit, Präpositionalphrasen nach iranischem Muster zu bilden (s. Doerfer 1967: 54). Diese Dialekte nähern sich dem Typ der Mischsprachen (s. 3.3). Im folgenden Fall ist die fremde Kategorie nur unvollkommen adaptiert worden; gleichwohl handelt es sich um eine funktionelle Neuerung: In 4.1 war die Entlehnung slawischer Verbalpräfixe in Romanidialekte erwähnt worden. Im Slawischen sind diese Präfixe nicht nur wortbildend, sondern liefern u. a. den Ausdruck für den perfektiven
Aspekt des Verbs. Nach diesem Muster ist in russischen Romanidialekten nun in Ansätzen ein Verbalaspekt, also eine neue Kategorie, entstanden. Es fehlt allerdings ein Imperfektivierungsverfahren (zu den Details s. Boretzky 1989: 367 f.). Noch stärker als die Möglichkeit der Entlehnung einzelner grammatischer Morpheme ist die ganzer Paradigmata in Abrede gestellt worden. Indessen gibt es selbst für solch weitreichende Interferenz Belege. Mednyj Aleut, eine ursprünglich aleutische, mit dem Eskimoischen verwandte Sprache, hat seine komplizierte Verbalflexion durch die des Russischen ersetzt; vgl. für das Präsens unucˇ-im, -isˇ, -jat ‘sitz-pr‰s1.sg, -pr‰s.2.sg, -pr‰s.3.sg’, und für das Präteritum unucˇi-l ‘sitz-pr‰t.m’ und unucˇi-la ‘sitzpr‰t.f’, d. h. es sind die Formantien der sog. i-Konjugation angetreten. In derselben Weise ist auch die Bildung des Infinitivs und des Futurs (mit budu ⫹ Infinitiv) auf das Aleut übertragen worden. Morphologisch haben hier russische Suffixe aleutische substituiert; man kann jedoch nicht sagen, daß die Strukturen des Aleut dabei wirklich erhalten geblieben sind (zu den Details s. Menovsˇcˇikov 1969: 132 und Thomason 1986: 277 f.). Im Ergebnis hat sich hier eine Übergangsstufe zu den Mischsprachen eingestellt (s. 3.3).
5.
Zitierte Literatur
Bakker, Peter (1989), “Relexification: The Case of Me´tif (French Cree)”. In: Boretzky et al., 119⫺137 Bakker, Peter & Mous, Maarten (1994, eds.), Mixed Languages: 15 Case Studies in Language Intertwining. Amsterdam: IFOTT Boretzky, Norbert (1975), Der türkische Einfluß auf das Albanische, Teil 1: Phonologie und Morphologie der albanischen Turzismen. Wiesbaden: Harrassowitz Boretzky, Norbert (1985), “Sind Zigeunersprachen Kreols?” In: Boretzky, Norbert & Enninger, Werner & Stolz, Thomas (Hrsg.), Akten des 1. Essener Kolloquiums über “Kreolsprachen und Sprachkontakte” vom 26. 1. 1985 an der Universität Essen. Bochum: Brockmeyer, 43⫺70 Boretzky, Norbert (1989), “Zum Interferenzverhalten des Romani”. In: Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung 42, 357⫺374 Boretzky, Norbert & Enninger, Werner & Stolz, Thomas (1989, Hrsg.), Vielfalt der Kontakte. Beiträge zum 5. Essener Kolloquium über “Grammati-
1652
XVII. Morphologischer Wandel I: Theoretische Probleme
kalisierung: Natürlichkeit und System-Ökonomie” vom 6. 10.⫺8. 10. 1988 an der Universität Essen, Bd. I. Bochum: Brockmeyer
In: Highfield, Arnold & Valdman, Albert (Hrsg.), Historicity and Variation in Creole Studies. Ann Arbor: Karoma, 52⫺78
Boretzky, Norbert & Igla, Birgit (1994), Kontaktinitiierter Wandel, 1. Interferenz (= B. Jeßing (Hrsg.), Sprachdynamik. Auf dem Wege zu einer Typologie sprachlichen Wandels, Bd. III. Bochum: Brockmeyer)
Oksaar, Els (1972), “Bilingualism”. In: Sebeok, Thomas A. (Hrsg.), Current Trends in Linguistics, Vol. IX: Linguistics in Western Europe. The Hague: Mouton, 476⫺511
Capidan, Theodor (1925), Meglenoroma˘nii I. Bucures¸ti: Cultura Nat¸ionala˘ Doerfer, Gerhard (1967), Türkische Lehnwörter im Tadschikischen. Wiesbaden: Steiner (Abhandlungen für die Kunde des Morgenlandes 37.3)
Roth, Juliana (1978), Die indirekten Erlebnisformen im Bulgarischen. München: Sagner Sala, Marius (1988), El problema de las lenguas en contacto. Me´xico: Universidad de Me´xico Sandfeld, Kristian (1930), Linguistique balkanique. Paris: Champion
Endzelin, Jan (1922), Lettische Grammatik. Riga: Gulbis
Sapir, Edward (1921), Language. New York: Harcourt, Brace & World
Hancock, Ian (1984), “Romani and Angloromani”. In: Trudgill, Peter (Hrsg.), Language in the British Isles. Cambridge: Cambridge University Press, 367⫺383
Sasse, Hans-Jürgen (1990), Theory of Language Decay und Contact-induced Change: Similarities and Differences. Institut für Sprachwissenschaft ⫺ Universität zu Köln. Arbeitspapiere Nr. 12 (Neue Folge) Schuster-Sˇewc, Heinz (1977), “Die deutschen Lehnübersetzungen im Obersorbischen und ihre Stellung im System der obersorbischen Lexik und Wortbildung”. Zeitschrift für Slavistik 7, 455⫺468
Henry, Patrick Leo (1957), An Anglo-Irish Dialect of North-Roscommon. Zürich: Aschmann & Scheller Hurch, Bernhard (1989), “Hispanisierung im Baskischen”. In: Boretzky et al. (Hrsg.), 11⫺35 Igla, Birgit (1996), Das Romani von Ajia Varvara: Deskriptive und historisch-vergleichende Darstellung eines Romanidialekts. Wiesbaden: Harrassowitz Martinet, Andre´ (1955), Economie des changements phone´tiques: Traite´ de phonologie diachronique. Bern: Francke Mayerthaler, Willi (1981), Morphologische Natürlichkeit. Wiesbaden: Athenaion Meillet, Antoine (1921), Linguistique historique et linguistique ge´ne´rale. Paris: Champion Menovsˇcˇikov, Georgij Alekseevicˇ (1969), “O nekotorych social’nych aspektach evoljucii jazyka”. Voprosy social’noj lingvistiki. Leningrad: Nauka, 110⫺134 Meyer, Gustav (1881), Etymologisches Wörterbuch der albanesischen Sprache. Strassburg: Trübner Moravcsik, Edith (1978), “Language Contact”. In: Greenberg, Josef H. (Hrsg.), Universals of Human Language, Vol. I. Stanford: Stanford University Press, 93⫺122 Muysken, Pieter (1981), “Halfway between Quechua and Spanish: The Case for Relexification”.
Spitzer, Leo (1928), Hugo Schuchardt-Brevier: Ein Vademecum der allgemeinen Sprachwissenschaft. Halle/Saale: Niemeyer Thomason, Sarah Grey (1983), “Genetic Relationship and the Case of Ma’a (Mbugu)”. Studies in African Linguistics 14, 195⫺231 Thomason, Sarah Grey (1986), “Contact-induced Language Change: Possibilities and Probabilities”. In: Boretzky, Norbert & Enninger, Werner & Stolz, Thomas (Hrsg.), Akten des 2. Essener Kolloquiums über “Kreolsprachen und Sprachkontakte” vom 29. und 30. 11. 1985 an der Universität Essen. Bochum: Brockmeyer, 261⫺284 Thomason, Sarah Grey & Kaufman, Terence (1988), Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press Weinreich, Uriel (1970), Languages in Contact. The Hague: Mouton Xhuvani, Aleksande¨r & C ¸ abej, Ekrem (1962), Prapashtesat e gjuhe¨s shqipe. Tirane¨: Universiteti Shtete¨ror
Norbert Boretzky, Bochum (Deutschland)
1653
153. Pidginization, creolization, and language death
153. Pidginization, creolization, and language death 1. 2. 3. 4.
Introduction Reduction processes: pidginization and language death Expansion processes: creolization References
1.
Introduction
In the field of pidgin and creole studies the main question is how exactly a new language can come into existence, and how the particular grammatical properties of the newly formed languages, pidgins and creoles, are related to the way in which they have emerged. A pidgin language is generally defined as a strongly reduced linguistic system that is used for incidental contacts between speakers of different languages, and that is the native language of nobody (DeCamp 1971). A creole language is a language that has emerged when a pidgin has acquired native speakers. The social circumstances that lead to the emergence of a jargon (a very primitive contact system) and consequently of a more stable pidgin generally involve the migration of a socially dominated group. This can be in the context of slavery or of contract labour in a colonial setting. Often trade carried out on an unequal footing is involved. A group of people is forced by the circumstances to develop a new communication system, to be used with foreigners who do not speak their own language. Even though various languages are involved when a jargon or pidgin emerges, the vocabulary of a pidgin generally derives from one language: the language that is socially or politically dominant in the original contact situation. Because most pidgins have resulted from the European (and later more generally Western) colonial expansion, starting in the 15th century, the vocabulary of most pidgins and creoles is derived from a European language (Portuguese, English, Spanish, French, Dutch). Most pidgins still in existence are spoken in Africa and in the Pacific Ocean. Creole languages are found predominantly in the Caribbean, in West Africa, in the Indian Ocean, and in the Far East. The most wellknown of the creoles are spoken, of course, by the descendants of the black slaves who were brought to the Caribbean to work in the sugar plantations.
Language death is obviously the falling out of use of a language, either because its original speech community is dissolved or because the community shifts to another language. Language death may (but need not, this is a point of debate) involve gradual reduction of the resources of the language in question, lexical, morpho-syntactic, stylistic. Language death occurs in every part of the world; at present it is particularly frequent in the Americas and in Australia, where many native languages are in the process of disappearing. When looking at pidginization, creolization, and language death from the perspective of morphology, it is necessary to consider both reduction and expansion.
2.
Reduction processes: pidginization and language death
From the point of view of morphology, the most important question concerning pidginization and language death is: can morphological elements survive the processes of morphological reduction inherent in these developments, and if so, which ones? This question will first be discussed in relation to pidginization, and then to language death, before a more general, comparative perspective is introduced. 2.1. Pidginization Most Caribbean creoles appear to have lost the largest part of the suffixes of their lexical source languages. This is particularly true of inflection. A case in point is the Haitian Creole verb, which is invariant in form, whereas French, the lexical source language, has an extensive paradigm with tense, mood, and person inflections: invariant reme ‘love’ versus e:me, e:mo˜, e:me, etc. The Mediterranean contact language Lingua Franca (Schuchardt 1909) exemplifies a pidgin Romance with invariant verb forms: (1) mi star contento mirar per ti I be glad see for you “I am glad to see you” Here the source language infinitive endings are generally used on the invariant verb forms. There is one exception to the invariance in verbal morphology:
1654
XVII. Morphologischer Wandel I: Theoretische Probleme
(2) mi mira-to per ti I see-pf for you “I have seen you” Here the past participle is used to mark perfective aspect. The loss of morphology in pidgins could result from a principle such as “Express each separately intuited element of meaning by a separate form” (Naro 1978) if speakers of the dominant languages are responsible for the reduction (in a kind of foreigner talk strategy), or from properties of early second language learning processes (in which words are acquired first in an invariant form), if the speakers of the subordinate languages are (Ferguson 1975; Andersen 1983, ed.). Nonetheless, before attempting to explain this supposed absence of inherited morphology, one should ask whether it is real. It turns out that in the Caribbean creoles, many derivational affixes from the European lexical source languages did survive (Holm 1989; see also DeGraff 2001). An example is Papiamentu, spoken on some of the Netherlands Antilles and on Aruba, where the action nominalizer -shon (< Ibero-Romance -cio´n) and the agentive marker -do` (< Ibero-Romance -dor) are used productively. Non-European-based pidgins and creoles (Heine 1978) provide further evidence against the supposed loss of morphology and show that it is risky to make generalizations on the basis of pidgin and creole derived from European colonial languages. Thus Nubi Arabic has lost the Arabic binyan system, but preserves a number of Arabic compounds, prefixes, and suffixes (Heine 1982). Similarly, in Kituba there has been loss of subject and object clitic pronouns and partial loss of tense/aspect morphology, but preservation of derivational morphology. What explains this partial loss and partial retention? Consider first some data from Amazon Quechua (Ecuador). Later a tentative general answer in terms of morphological typology will be given. A simple example can be used to illustrate some of the properties of Amazon Quechua, a resultant of ethnolinguistic restructuring concomitant with colonization and drastic population decline in the Amazon basin. Many different small groups of Indians are forced to disband and they form new, more stable, communities in which the peasant language of the Andean highlands is adopted.
(3) riku-u-ni yura-ma pishku-una see-pres-1 tree-to bird-pl tiya-nau-nga-chu ni-sha be-pl-3.fut-int say-sr ‘I am looking at the trees to find out if there are any birds (lit. saying: will there be any birds).’ Amazon Quechua presents many features that may be attributed to pidginization and creolization: (a) the loss of morphophonemic adjustment rules such as the vowel lowering rule (Adelaar 1977); (b) loss of nominal person marking; (c) reduction from 17 to 7 modal suffixes on the verb; (d) formation of a periphrastic future/modal; (e) reinterpretation of the subordinator -ptias -k-pi ‘agentive-locative’, and extension to other nominalizers, yielding -shka-pi and -na-pi; (f) development of a periphrastic construction involving the complementizer ni ⫹ sha ‘saying’ (this may be due to Jivaroan substrate influence). Nonetheless, morpheme counts (for 100 words of text) suggest that Amazon Quechua is morphologically less complex than e.g. Cuzco Quechua (Peru), but certainly has not lost all of its morphology:
Cuzco Quechua: Amazon Quechua:
traditional narrative
syntactic field notes
2.01 1.74
2.54 2.20
Tab. 153.1: Morpheme/word proportions in two Quechua varieties
To comprehend that the pattern of reduction is not random, consider the set of Quechua person markers in Cuzco Quechua and Amazon Quechua (see Table 153.2). A perusal of the forms that have survived shows that: (a) the first person plural inclusive/exclusive distinction has been lost for subjects; (b) the only object forms that have survived are those that can be fit into the scheme {1.sg object} ⫺ {subject}. The syntagmatically irregular forms have disappeared. The fact that so many affixes have survived at all in pidgin/creole varieties of languages such as Quechua as well as in Swa-
153. Pidginization, creolization, and language death
Cuzco
Amazon
-ni -nki -n -y-ku -n-chis
-ni -ngi -n
-nki-chis -n-ku -yki -wa-nki -wa-n -su-nki -wa-nchis -yki-ku -yki-chis -wa-nki-ku -wa-nki-chis -wa-nchis-ku -su-nki-chis -su-nki-ku -wa-n-ku
-nchi -ngichi (-nau..)-n -wa-ngi -wa-n
-wa-ngichi
1.sg 2.sg 3.sg 1.pe 1.pi 1.pl 2.pl 3.pl sbj.1.sg&obj.2.sg sbj.2.sg&obj.1.sg sbj.3.sg&obj.1.sg sbj.3.sg&obj.2.sg sbj.3.sg&obj.1.pi sbj.1.pe&obj.2.sg sbj.1.sg&obj.2.pl sbj.2.sg&obj.1.pe sbj.2.pl&obj.1.sg sbj.3.pl&obj.1.pi sbj.3.sg&obj.2.pl sbj.3.pl&obj.2.sg sbj.3.sg&obj.1.pe
Tab. 153.2: Quechua person markers
hili, for instance, no doubt has to do with their agglutinative character. This leads to the issue of morphological typology (cf. Anderson 1985). At least five dimensions need to be distinguished in a morphological typology, with respect to the survival of specific morphological patterns: (a) The type of concept that can be expressed by morphemes (Sapir 1921). In most of the pidgin and creole languages only derivational concepts are expressed morphologically, but it is not clear whether these are inherited morphemes. (b) The extent to which words are complex. Complexity by itself does not determine whether a form is maintained or lost in pidginization processes. (c) The extent to which there are morphophonological relations between the components of a word. Consider in this respect allomorphy, the type of fusion found in Semitic systems, etc. It will be clear that this dimension plays an important role. Only morphological patterns that rate low on this dimension can survive a process of reduction. (d) The extent to which there are non-syntactic selectional restrictions between the
1655 components of a word. Relevant in this respect are elements sensitive to the declension class of their base, or subject to lexical specification. Again, elements low on this dimension have a better chance to survive a process of reduction, and agglutination can be defined as the absence of morphological selectivity. (e) The extent to which the meanings of complex forms are paradigmatically rather than syntagmatically defined. The case of Amazon Quechua person marking suggests that paradigmaticity could play a role: only those elements that are syntagmatically interpretable survive. The loss of Indo-European inflectional morphology in the Caribbean creoles can be interpreted in the same light: involved are highly paradigmatized morphological subsystems. It is beyond the scope of this article to further develop the notions sketched here. What is important is the need for a much more differentiated view of the process of morphological reduction (see Kusters 2003). 2.2. Language death Although highly frequent, the processes of language attrition, language obsolescence, and language death have been little studied. In part this is because it is a subject fraught with methodological problems. To mention but two: (a) what is the reference point for the process of loss or decay investigated; and (b) how to account for the discrepancy that many speakers show between active and passive competence in the language undergoing loss? What follows is based on two of the bestknown studies in this field: one on Scottish Gaelic (Dorian 1981) and one on the Australian aboriginal language Dyirbal (Schmidt 1985). The studies compare the speech of the older, more traditional speakers with that of the younger speakers, primarily using oral sentence translation tasks. Although the differences between East Sutherland fishing communities and the Jambun communities in northeast Queensland, Australia, could not be larger, there are some striking similarities in the results of the two studies. Language death is not directly apparent in a decreasing fluency, but rather in morphological and syntactic changes, which could also have been the result of ordinary linguistic change, but are more numerous
1656
XVII. Morphologischer Wandel I: Theoretische Probleme
and proceed at a faster rate. Both studies document a gradual decrease in use of some traditional morphological elements of the language, though not of others (Dorian 1981: 146⫺151; Schmidt 1985: 76⫺78). For both languages, case endings are highly unstable; it may well be that the alternative signalling function of word order and borrowed prepositions (both languages are giving way to varieties of English) is a factor here. For the rest, typological differences make a detailed comparison impossible. In Gaelic verbal number is also unstable as a category, but nominal number and gender are stable. Tense and passive morphology are maintained by younger fluent speakers but decline in frequency in the speech of semi-speakers of Gaelic. In young people’s Dyirbal, almost all derivational affixes survive, as well as some mood and aspect markers. Consider an example such as (Schmidt 1985: 217): (4) oh she baji-baji-yarra-nyu down oh she fall-rdp-aspect-nfut down “Oh! she started to fall down!” In addition to case markers, some tense markers and other verbal inflectional affixes tend to disappear. Both studies stress the role of borrowing in language attrition and death, but borrowing is not a reliable diagnostic for it. What makes a fragment of spoken Quechua, for instance, show signs of attrition is not the amount of Spanish borrowing: much more fluent speakers also use many Spanish words. Rather it is the lack of variation in the syntax and morphology: only a few suffixes are used, even if they are used frequently. What is urgently needed is a comparative and controlled study of morphological loss and reduction in the context of language death, where the relative weight of formal linguistic factors (e.g. the dimensions referred to in 2.1), frequency, phonological saliency, functional load, etc. is studied. 2.3. A comparison between reduction due to pidginization and to language death Both studies of language death briefly discussed stress the differences between language death and pidginization (Dorian 1981: 152⫺156; Schmidt 1985: 214⫺218). In addition to the numerous functional differences, the main difference mentioned is the amount of morphology retained from the source language.
As seen in 2.1, morphological simplicity per se is not a necessary concomitant of pidginization, when the source language is agglutinative, for instance. Indeed, one could imagine a morphologically complex pidginDyirbal, given the transparent suffixal nature of its morphology. For Gaelic, the situation is quite different, however. Here the morphological distinctions are highly fusional and allomorphic, involving initial lenition and nasalization, and final consonant mutation as well. It is hard to conceive of a pidgin that would have maintained even a small portion of Gaelic morphology. We may conclude that the evidence points to a qualitative difference between the two processes: the type of disruption in the transmission of linguistic forms is much more radical in the case of pidginization.
3.
Expansion processes: creolization
The creole languages that emerged out of pidgins with various degrees of elaboration and complexity reconstituted the lexical richness characteristic of the native language of a speech community in various ways. Here several of the lexical expansion strategies used will be surveyed: compounding, phrasal compounding, affixation, reduplication, toneor stress-shift, conversion. The survey can only be preliminary because creole morphology is a sorely neglected area of study (due to the wide-spread belief that creoles have no morphology to speak off). Three preliminary general remarks are in order. First, the reconstituted system is neither in its organization nor in its formal aspects necessarily a replica of what is found in the European colonial lexical source language. A case in point is Papiamentu, where no less than seven different forms replaced the Ibero-Romance “reflexive” clitic se (Muysken 1993; Muysken & Smith 1994: 285 f.): (5)
pan˜a (< Port. pano, Span. pan˜o ‘cloth’) mi ta bisti pan˜a (< Span. vestirse) ‘I dress (myself)’
(6)
kurpa (< Port. corpo, Span. cuerpo ‘body’) lanta kurpa (< Span. levantarse) ‘get up’
(7)
null reflexive diskulpa´ (< Span. disculparse) ‘excuse oneself’
1657
153. Pidginization, creolization, and language death
(8)
possessive ⫹ kurpa skonde su kurpa (< Span. esconderse) ‘hide’
(9)
object pronoun sinti e tristi (< Span. sentirse) ‘feel sad’ han˜a e (< Span. hallarse) ‘find oneself’
(10) object pronoun ⫹ mes (< Port. mesmo ‘self’) e ta komport’e (mes) (< Span. comportarse) ‘he behaves himself’ (11) possessive pronoun ⫹ mes hasi su mes malu (< Span. hacerse mal) ‘hurt oneself’ Examples like these illustrate the morpholexical autonomy of the creoles with respect to their lexical source languages, even in a case such as Papiamentu, which would seem to be rather close to its lexical source languages Spanish and Portuguese. Second, particularly in the domain of tense/mood/aspect, European verbal affixes have been replaced by separate, grammaticalized, pre-verbal particles (Holm 1989). Thus the progressive marker ta, as in mi ta´ ho´ndi pı´ngo ‘I hunt boar’ in the Surinam maroon language Saramaccan derives from Eng. stand (Smith 1987). These particles have a fixed position, but may not be considered verbal prefixes. When subject to sandhi rules, they tend to form a unit with the preceding subject rather than with the following verb. Third, lexical expansion also occurs in pidgins, to some extent. Consider the question words of Chinese Pidgin English:
who what when where why how
forms
analysis
who (-man) wat ting wat-time wat-side wat-for how (-fashion) wat-fashion
‘who (-man)’ ‘int-thing’ ‘int-time’ ‘int-side’ ‘int-for’ ‘how (-fashion)’ ‘int-fashion’
Tab. 153.3: The question words of Chinese Pidgin English (Bisang 1985, as analyzed in Muysken & Smith 1990)
This system is highly transparent in that there is a stable question particle derived from English what. 3.1. Compounding The primary means of lexical expansion in almost all, if not all, creoles is compounding, which tends to be highly productive. Com-
pounding is used for all word classes. The creole language of coastal Surinam Sranan has mofoneti ‘mouth night (midnight)’ and bobimofo ‘breast mouth (nipple)’, Haitian has bouch kabrit ‘mouth goat (Cassandra)’ (Hall 1953: 41). Where Papiamentu has bula bay ‘fly go (fly away)’, a compound derived from a serial verb construction, Haitian has magne manje ‘touch eat (just eat a little)’ (Hall 1953: 42). Serial verbs are verbs that can be concatenated within a single clause. Compounding is also used to form new function words. Some examples from Papiamentu and Chinese Pidgin English were already given in 3. Compounding is also illustrated in an example from the reflexives in Saramaccan and Sranan (cf. Muysken & Smith 1994: 276):
myself yourself himself ourselves yourselves themselves
Saramaccan
Sranan
mi-se´ei ju/i-se´ei en-se´ei wi/u-se´ei unu-se´ei den-se´ei
mi-srefi ju-srefi en-srefi wi-srefi unu-srefi den-srefi
Tab. 153.4: Reflexives in Saramaccan and Sranan
At first sight, it would seem that these forms are direct reflexes of the English forms. Notice, however, that they are directly based on Saramaccan and Sranan pronouns, respectively. Only for the third person forms a direct inheritance could be postulated. We find that quantifiers, different kinds of pronouns, question words, complex prepositions, etc. are all formed through compounding in many creole languages. 3.2. Phrasal compounding Very common in some creole languages is phrasal compounding. It involves (a) cases where one of the members of the compound (in fact, the non-head) is a phrase rather than a word, and (b) cases where the structure of the compound itself reflects a syntactic rather than a morphological pattern. An example of the first type is Saramaccan agentive compound formation, involving the noun ma ‘man’. It is attached to verbs or verbal complexes: (12) paı´-ma bear-man ‘pregnant woman’
1658
XVII. Morphologischer Wandel I: Theoretische Probleme
ho´ndi-ma hunt-man ‘huntsman’ tja´-bu´ka-ma carry-mouth-man ‘messenger’ paı´-ku-muje´e-ma bear-with-woman-man ‘midwife’ Notice first of all that the meaning of ma is no longer exclusively masculine but rather ‘person that ...’, i.e. the form is somewhat grammaticalized (cf. 3.3). Second, the lefthand member can contain nouns (bu´ka ‘mouth’), prepositional phrases (ku muje´e ‘with woman’), and even clauses and serial constructions. Similar perhaps are cases in Haitian where verb ⫹ object noun combinations function as nouns (Hall 1953: 41) (cf. 3.6): (13) pase raj pass rage ‘exotic dance’ la˜vi mouri want die ‘imprudent person’ pote mak bear mark ‘he who is scarred’ An example of the second type is Papiamentu, where most compounds include a linking morpheme di/i ‘of’ (Dijkhoff 1987): (14) palu di garganta stick of neck ‘neck bone’ kabes di boto head of boat ‘lift’ barba di yo`nkuman beard of young.man ‘herb’ It can be shown that these forms, in spite of their syntactic appearance, behave as lexical islands for pluralization, adjectival modification, extraction, etc. (Dijkhoff 1987). 3.3. Affixation Affixation is more limited in the morphological systems of creoles. It is not easy for a language to develop new affixes. There are two ways in which this can be done: (a) by borrowing; in 2.1 the case of Papiamentu nominalizers was mentioned as an example of retention from the source language; (b) reanalyzing part of a compound; it is possible that the righthand member of a compound is semantically bleached and phonologically weakened so that it can develop into a suffix.
A possible example may be Saramaccan ma mentioned in 3.2. Similar Saramaccan examples are fa´si compounds: (15) tjalı´ fa´si sorrow fashion ‘sorrowful’ tooka´ fa´si change fashion ‘change in appearance’ bu´nu fa´si good fashion ‘generous’ Other Saramaccan nouns which may be turning into affixes in a similar way are se ‘side’, kamia ‘place’, libi ‘life’, sembe ‘somebody’, soni ‘something’, and te ‘time’. Another way to introduce affixation is through borrowing. Two nominal affixes in Berbice Dutch creole are borrowed from the African language Eastern Ijo: -je ‘nr’ and -apu ‘pl’ (Kouwenberg 1991; Smith et al. 1987). The form -je occurs in forms such as: (16) di kalije ‘the small one’ en gu feteje ‘a big fat one’ di eshtije ‘the first one’ Notice that the category plural, when attached to terms for individuals (‘Hilda’ below) can have the meaning of an associative plural (see Art. 100), something common in creoles: (17) matj-ap gugu-j-apu hild-apo
‘friends ‘the big ones’ ‘Hilda and her family’
The nominalizing element je in Berbice Dutch may be an enclitic noun, or a phrasal affix with nominal status. It is possible to add this element more than once, as can be seen from: (18) kali-je-apu-je-apu [kalijapjapu] small-nr-pl-nr-pl ‘the ones that belong to the small ones’ A more complex form of borrowing involves cases where the lexical source language and a substrate donor language have a morpheme similar in form in both languages. The adoption in the creole has been analyzed as the result of conflation (Kihm 1988) or convergence (Kouwenberg 1992). In Guinea-Bissau Portuguese Creole pairs occur such as: (19) (a) firma ‘stand’ firma-nta ‘raise, put on its base’ (b) sibi ‘go up’ sibi-nti ‘make go up’ (c) yentra ‘go in’ yentra-nda ‘make go in’
1659
153. Pidginization, creolization, and language death
The causative ending could have two sources (Kihm 1988): (i) the rather unproductive Portuguese derivational suffix -nt- ‘causative, inchoative’ that occurs in:
(c) iterative-repetitive mu-mu ‘to keep going (step by step)’ (d) iterative-distributive findi-fint ‘to open one by one’
(20) (a) ferver ‘boil/be boiling’ aferve-nt-ar ‘boil/bring to boil’ (b) quebrar ‘break’ quebra-nt-ar ‘break something’
Reduplication is recursive, as can be seen from examples such as:
(ii) a causative suffix in the neighbouring West-African languages: (21) (a) Manjaku: -lenp ‘work’ -lenpandan ‘make work’ (b) Mandinka sonka ‘quarrel’ sonkandi ‘cause to quarrel’ The process of conflation would involve a mutual reinforcement of these two. The Berbice Dutch Creole perfective marker -te (Kouwenberg 1992) can be linked both to a similar Ijo form and to the Dutch past tense suffix -te. Again, presumably its presence in both contributing languages favored its adoption in the Creole. 3.4. Reduplication Many Europeans will immediately associate reduplication with pidgins and creoles, because it is part of stereotypical foreigner talk, “primitive” language use. However, the idea of reduplication as an iconic means of intensification or pluralization, still adopted by Schuchardt (1914), for instance, does not necessarily correspond to the way this device is used in most creoles. Reduplication occurs frequently with adjectives and verbs in Berbice Dutch (Kouwenberg 1991), but it has different meanings with each category. With adjectives it can be:
wer (24) (a) eni bu-bu-bu-bu-te they drink-rdp-rdp-rdp-pf again ‘they kept drinking on and on again’ (b) ek hafu I have.to bjonto-bjonto-bjonto-bjonto remember-rdp-rdp-rdp ‘I have to keep on remembering’ In this case reduplication does have its iconic intensifying meaning. It may be the case that reduplication has two sources: as an iconic device, not fully grammaticalized, and as a special type of affixation, in which an unspecified affix is added which is then spelled out as a copy of the base form. In Saramaccan there are many kinds of reduplication (Bakker 1987). Productive, grammaticalized reduplication can have a meaning completely different from the iconic one. With verbs it is used to form resultative adjectives: (25) di lai-lai goni ‘the loaded gun’ (lai ‘to load’) di dee-dee koosu ‘the dried cloth’ (de ‘to dry’) Adjectives and nouns can also be reduplicated, yielding the meaning ‘X-ish’ or ‘X-like’: (26) geligeli ‘yellowish’ (geli ‘yellow’) baafubaafu ‘soup-like’ (baafu ‘soup’)
With verbs it is iterative, which has a different nuance of interpretation with different examples of verbal usage:
3.5. Stress and tone shift Though not as common as other morphological devices, there are instances of stress- and tone-shift in creoles. One illustration is the verb duna ‘give’ in Papiamentu. Its citation form has penultimate stress and a low-high tone pattern. In the imperative, however, there is high-low stress with bisyllabic verbs of the duna class (Römer 1983). The past participle form has shift of stress to the last syllable, but maintains low-high tone. The same shift pattern occurs with some deverbal nouns:
(23) (a) iterative-habitual futel-futel ‘converse regularly’ (b) iterative-aimlessness kap-kap ‘cut here and there’
(27) (a) piska ‘to fish’ piska´ ‘a fish’ (b) pika ‘to stab, to burn pika´ ‘a stab, a burn’
(22) (a) intensifying kal-kali ‘tiny’ (b) emphatic eshti-eshti ‘the very first’ (c) distributive de pote-pote kenap ‘the old people (one by one)’
1660
XVII. Morphologischer Wandel I: Theoretische Probleme
3.6. Multifunctionality In all creoles lexical items are multifunctional in that they can belong to several word classes at a time. Assuming there to be a base category for each item, a morphological process of conversion or zero-derivation can be postulated (cf. Art. 90). Voorhoeve (1981) has studied this problem in some detail for Sranan. He posits rules of conversion, on the basis of regular patterns of correspondence. Thus the word siki can be the adjective ‘ill’, the noun ‘illness’, the transitive verb ‘make (someone) ill’ and the intransitive verb ‘be ill’. The same pattern holds for hebi ‘heavy’, blaka ‘black’, etc. while dede ‘dead’, bigi ‘big’, and bun ‘good’ cannot have the interpretation of being a transitive verb, while sharing the other three options. Wasi ‘wash’ can only be a transitive and an intransitive verb, but not a noun. The major problems posed by forms such as these are (a) how to establish a base form; (b) how to formulate rules that account for the regularities in what is possible and what is not? A complex case of multifunctionality, where the category of the element involved is contextually determined, involves adjectives /stative verbs in Sranan. These elements function as stative verbs when used predicatively and cannot have a preceding copula (de in Sranan): (28) (a) a liba bradi ‘the river is wide’ (b) * a liba de bradi Bradi is a verb, and hence a copula is not allowed, presumably. Consider now a case where bradi is preceded by the adjectival modifier so: (29) (a) * a liba so bradi (b) a liba de so bradi ‘the river is so wide’ Here the copula suddenly is obligatory. The same holds when bradi is question with the particle o ‘how’: (30) (a) * o bradi a liba (b) o bradi a liba de ‘How wide is the river?’ When bradi has a pre-head specifier, it is an adjective. Similar, it must be viewed as an adjective in attributive position: (31) a bradi liba ‘the wide river’
3.7. Concluding remarks Almost the full spectrum of morphological techniques or strategies available has been called upon in the process of morphological expansion, even if compounding, reduplication, and multifunctionality play a more central role than affixation or stress-shift, for instance.
4.
References
Adelaar, Willem (1977), Tarma Quechua. Leiden: Brill Andersen, Roger (1983, ed.), Pidginization and Creolization as Language Acquisition. Rowley/MA: Newbury House Anderson, Stephen A. (1985), “Typological Distinctions in Word Formation”. In: Shopen, Timothy (ed.), Language Typology and Language Description, Vol. III: Grammatical Categories and the Lexicon. Cambridge: Cambridge University Press, 3⫺50 Bakker, Peter (1987), “Reduplications in Saramaccan”. In: Alleyne, Mervyn C. (ed.), Studies in Saramaccan Clause Structure. Amsterdam: University of Amsterdam; Mona/Jamaica: University of the West Indies (Caribbean Culture Studies 2), 17⫺40 Bisang, Walter (1985), Das Chinesische Pidgin Englisch. Amsterdam: University of Amsterdam DeCamp, David (1971), “Introduction: The Study of Pidgin and Creole Languages”. In: Hymes, Dell (ed.), The Pidginization and Creolization of Languages. Cambridge: Cambridge University Press, 13⫺39 DeGraff, Michel (2001), “Morphology in Creole Genesis: Linguistics and Ideology”. In: Kenstowicz, Michael (ed.), Ken Hale: A Life in Language. Cambridge, Mass.: MIT Press, 53⫺121 Dorian, Nancy (1981), Language Death: The Life Cycle of a Scottish Gaelic Dialect. Philadelphia: University of Pennsylvania Press Dijkhoff, Marta (1987), “Complex Nominals and Composite Nouns in Papiamentu”. In: Maurer, Philip & Stolz, Thomas (eds.), Varia Creolica. Bochum: Brockmeyer, 1⫺10 Ferguson, Charles J. (1975), “Towards a Characterization of English Foreigner Talk”. Anthropological Linguistics 17, 1⫺14 Hall, Robert A. Jr. (1953), Haitian Creole: Grammar, Texts, Vocabulary. American Anthropological Association Memoir 74) Heine, Bernd (1978), “Some Generalizations on African-Based Pidgin-Languages”. Afrika und Übersee 61, 219⫺229
154. Morphological reconstruction
1661
Heine, Bernd (1982), The Nubi Language of Kibera: an Arabic Creole. Berlin: Dietrich Riemer
Pidgins and Creoles: An Introduction. Amsterdam: Benjamins, 271⫺288
Holm, John (1989), Pidgins and Creoles, Vol. I: Theory and Structure. Cambridge: Cambridge University Press
Naro, Anthony J. (1978), “A Study on the Origins of Pidginization”. Language 54, 314⫺347
Kihm, Alain (1988), “Conflation as a Directive Process in Creolization”. In: Boretzky, Norbert & Enninger, Werner & Stolz, Thomas (eds.), Beiträge zum 4. Essener Kolloquium über “Sprachkontakt, Sprachwandel, Sprachwechsel, Sprachtod”. Bochum: Brockmeyer, 111⫺38 Kouwenberg, Silvia (1991), Berbice Dutch Creole. Ph.D. dissertation, University of Amsterdam Kouwenberg, Silvia (1992), “The Convergence Hypothesis in the Genesis of Berbice Dutch Creole”. Lingua 88, 263⫺300 Kusters, Wouter (2003), Linguistic Complexity. The Influence of Social Change on Verbal Inflection. Ph. D. dissertation, Leiden University Muysken, Pieter (1993), “Reflexes of the Ibero-Romance Reflexive Clitic in Papiamentu”. In: Byrne, Francis & Winford Donald (eds.), Focus and Grammatical Relations. Amsterdam: Benjamins, 285⫺ 302 Muysken, Pieter & Smith, Norval (1990), “Question Words in Pidgin and Creole Languages”. Linguistics 28, 883⫺903 Muysken, Pieter & Smith, Norval (1994), “Reflexives in Pidgin and Creole Languages”. In: Arends, Jacques & Muysken, Pieter & Smith, Norval (eds.)
Römer, Raul G. (1983), “Papiamentu Tones”. Papiamentu: Problems & Possibilities. Zutphen: Walburg, 85⫺96 Sapir, Edward (1921), Language. New York: Houghton Schmidt, Annette (1985), Young People’s Dyirbal: An Example of Language Death from Australia. Cambridge: Cambridge University Press Schuchardt, Hugo (1909), “Die Lingua Franca”. Zeitschrift für romanische Philologie 33, 441⫺461 Schuchardt, Hugo (1914), Die Sprache der Saramakkaneger in Surinam. Amsterdam: Johannes Muller (Verhandelingen der Koninklijke Akademie van Wetenschappen te Amsterdam, Nieuwe Reeks, Deel 14, No. 6) Smith, Norval (1987), The Genesis of the Creole Languages of Surinam. Ph.D. dissertation, University of Amsterdam Smith, Norval & Robertson, Ian & Williamson, Kay (1987), “The Ijo Element in Berbice Dutch”. Language in Society 16, 49⫺90 Voorhoeve, Jan (1981), “Multifunctionality as a Derivational Problem”. In: Muysken, Pieter (ed.), Generative Studies on Creole Languages. Dordrecht: Foris, 25⫺34
Pieter Muysken, Nijmegen (The Netherlands)
154. Morphological reconstruction 1. 2. 3. 4. 5. 6. 7.
1.
Introduction: basic questions and methods Reconstruction: The first steps Further guiding principles Going beyond simple reconstruction of forms Pushing the limits: reconstructed states as real languages Conclusion References
Introduction: basic questions and methods
Morphology is above all concerned with the forms (morphemes and words) of a language. These forms include roots and affixes, as the
basic building blocks of words, as well as the patterns of combination ⫺ both derivational and inflectional ⫺ by which words are built up with these elements, and the phonological adjustments that apply within words as part of the overall word-formation process. Moreover, of interest too are the categories that these forms express and mark, e.g. gender, number, person, animacy, tense, mood, etc. All of these aspects of morphology can be reconstructed for earlier stages of a language and/or proto-language, once one has appropriate data and methods to work with and some guiding principles. The basic methods are those that serve reconstruction in other domains (especially
1662
XVII. Morphologischer Wandel I: Theoretische Probleme
phonological). That is, they include primarily the Comparative Method and Internal Reconstruction, as well as whatever other means one can employ to reach a reasonable set of assumptions about the paths of development from reconstructed elements to attested outcomes (e.g., observation of attested changes and models derived from them). These methods are illustrated, and expanded upon with other useful principles, in the sections that follow.
2.
Reconstruction: The first steps
For the most part, the reconstruction of the formal side of morphological units can be fairly straightforward, drawing on the successes linguists have had with the reconstruction of sounds and sound systems. That is, once a set of valid sound correspondences holding among (related) languages have been worked out, and phonological units for a proto-language for those languages have been reconstructed based on those correspondences, one can exploit those reconstructions in the reconstruction of the form of morphemes. Essentially, all one has to do is observe the sound correspondences in a comparison of actual morphemes ⫺ as opposed to simply matching up sounds that are, so to speak, disembodied ⫺ and string together the reconstructible segments to give whole reconstructed forms (a practice Anttila 1972: 351 has called doing morphological reconstruction by “applied phonology”). This is a reasonable step to take, since the correspondences themselves emerge out of a comparison of morphemes in the first place. For example, the comparison sets p/p/p/f, i/a/a/a, t/t/t/d, etc. across Sanskrit, Greek, Latin, and Gothic, respectively, in pitar-/pate´r-/pater/fadar, and the recognition that each set is a regular, i.e. valid and well-established, sound correspondence that licenses the reconstruction of a Proto-Indo-European phonological unit, together allow one to put the reconstructed segments together sequentially to yield a Proto-Indo-European form *peter-. Similarly, putting a/e/e/i and s/s/s/s together across these same languages, along with t/t/t/t (in the environment s__) and i/i/Ø/ Ø (in the environment __#), in asti/e´sti/est/ ist warrants a reconstruction for Proto-IndoEuropean of word *esti (actually, more properly *H1 esti in laryngealistic terms). With a less obvious but no less compelling set of cor-
respondences, the equation of Skt. ch, Hittite sk, Lat. sc, and Alb. h in verbal forms permits one to reconstruct a Proto-Indo-European morpheme *-sk’-, given that the sound correspondences in evidence here are impeccable (cf. Skt. cha¯ya- and Alb. hije ‘shadow’, where Gk. skı´a gives a clear indication of *sk’ as the starting point for this set). Within the individual languages, several of the forms that these reconstructions are based on fall into paradigmatic sets with other forms that are equatable across the languages, e.g. along with asti in Skt. there is asmi, along with Gk. esti, there is emmi (in the Aeolic dialect), and so on, and these together license a reconstruction *esmi (actually *H1 esmi). To the extent that these reconstructed forms *esmi/*esti constitute members of the same paradigm ⫺ as their outcomes in the various languages do ⫺ segmentation of the forms into constitutent morphemes *es- and *-ti/*mi is possible, and thus inferences about the order of morphemes in the reconstructed words become possible. What has not been addressed here is the meaning to be assigned to these reconstructed forms, yet that is a crucial step in morphological reconstruction, to the extent that morphemes are seen to be pairings of both form and meaning. In the simplest case, involving lexical items with a clear meaning, as with pitar-, etc. above, all of which mean ‘father’, or as-, etc., all of which mean ‘be’, the decision is straightforward; thus a formmeaning nexus *peter- ‘father’ or *es- ‘be’ can safely be assumed for Proto-Indo-European. Similar considerations also hold for non-root elements, so that the observation that the reflexes of *-ti all mark third person singular and those of *-mi all mark first person singular forms in their respective languages licenses one to assign those values to the proto-morphemes in question. It is also therefore safe to assume that the person markings were suffixes, since they are suffixal in all of the languages represented. In somewhat trickier cases, the meanings of the elements in the comparison set do not match perfectly, and thus more difficult decisions need to be made. To some extent, reconstruction in such cases becomes a matter of historical semantics and semantic reconstruction, not morphological reconstruction per se, but the ramifications for reconstructed proto-morphemes should be clear. Thus, Skt. ma¯tar-/Gk. ma¯ter-/Lat. ma¯ter, etc. mean ‘mother’ but reconstructing
1663
154. Morphological reconstruction
that meaning for the Proto-Indo-European word must take the Alb. cognate mote¨r, meaning ‘sister’, into account; conceivably the original of the word was broader (e.g. ‘female in nuclear family’) or else Albanian alone innovated (perhaps through the involvement of an original dvandva (coordinative) compound ‘sister-and-mother’ as a merism (defined by Watkins 1995: 9 as “a twopart figure which makes reference to the totality of a single higher concept”) for female kin). At the grammatical level, semantic/functional mismatches for cognate markers can pose problems for reconstruction, especially in the absence of clear bases for understanding how category markings can themselves mutate and be altered. Thus, in the *-sk’case mentioned above, the Hittite morpheme marks iterative action, the Latin morpheme marks inchoative (beginning) action, the Sanskrit morpheme is simply one of several markers for present tense system stems, and the Albanian morpheme marks non-active voice (with passive, reflexive, reciprocal, and stative functions) but only in the present tense system. Thus there is an element of presentiality in most of the reflexes of *-sk’-, but one needs to ask whether the original function of this morpheme was just to mark present, or instead was more specific, with one of the attested values as the starting point in Proto-Indo-European. Any answer here runs the risk of seeming to be arbitrary, but without an answer, the reconstruction for this morpheme is incomplete; reconstructing the form alone without the meaning or function is satisfying only half of the burden. The above cases draw on the Comparative Method, the mainstay of historical reconstruction for nearly two centuries, but other methods can be employed as well in reconstructing morphology. In particular, Internal Reconstruction can be used, where, in essence, alternations within a single synchronic stage of a language are “undone” as it were, and an earlier state without the alternation is reconstructed. Morphological reconstruction in such a case consists in the positing of a unique form for a given alternation found in a later stage, generally taking the later allomorphy to be the result of conditioned sound changes. Thus the allomorphy seen in the Modern English plural marker -s/-z/-ez allows for a hypothesis of an earlier marker with a single undifferentiated form, e.g. -ez, with the variants having arisen, as is so often the case, via sound changes, e.g. syncope and
voicing assimilation. But even this method can break down; there is no way to reconcile the plural marker -en (restricted in Modern English just to oxen and brethren, though children contains it too) with the -s/-z/-ez forms, and that is as it should be, given that its distribution is lexically idiosyncratic and not rule-governed and not tied to phonological conditioning in any way. But even cases involving clear phonological conditioning of the allomorphy can be problematic for reconstruction; for instance, the Korean nominative markers are in complementary distribution: -i occurring after a consonant-final noun, -ka after a vowel-final stem, but the phonetic distance between the two forms makes a reconstruction of a single nominative marker that gave rise to both of these alternants most unlikely. The third method, or rather a principle (cf. 1), namely ensuring that there is a reasonable path of development from the earlier reconstructed stage to the attested ones, must be invoked here, so that one does not take -i and -ka back to the same proto-form, internally reconstructing by brute force, as it were. Presumably they each have an independent origin and have come, by various developments, to stand as phonologically distributed functionally equivalent alternants.
3.
Further guiding principles
More can be said about these methods and principles. Critical to positing a reconstruction that requires only a reasonable set of changes for the development of any given element is an understanding of what the expected processes of language change are. What is thus especially important here, besides recognizing the regular nature of sound change (which gives the regular sound correspondences drawn on in doing morphological reconstruction as “applied phonology”, as above, cf. 2), is allowing for the workings of other processes of change, such as analogy. For instance, the consonantal matchings between the root nouns (where the root with nothing added constitutes the stem) seen in Gk. pod- and Lat. ped-, both meaning ‘foot’, are perfectly regular, but the vowels do not match up as expected, as Gk. o usually corresponds to Lat. o (reflecting PIE *o) and Lat. e usually corresponds to Gk. e (reflecting PIE *e). Armed with the knowledge that there are Indo-European languages in which
1664
XVII. Morphologischer Wandel I: Theoretische Probleme
reflexes of *e and reflexes of *o alternate in grammatically determined environments (e.g. English present tense sing from *seng wh-, versus past tense sang from *song wh-), rather than reconstructing a different vowel for the o ~ e correspondence, a step which would ultimately lead one to reconstruct a different vowel for every such correspondence (and there are others, especially when languages other than Greek and Latin are included), one can instead reconstruct a process by which *o and *e alternate within the grammatical forms of ‘foot’ in the proto-language. One can then treat the Greek and Latin forms with fixed vowels as the result of analogical change within the paradigm as the individual languages took shape. The exact conditions for the *e /*o alternation may be unclear, but the method of allowing for the “undoing” of the potential effects of analogy leads one to the reconstruction of a morphological process (of vowel gradation) for the proto-language, that is, an aspect of a ProtoIndo-European word-formation process. It can be noted as well that the inferences about morpheme order derivable from the comparison of *esti and *esmi similarly reflect reconstructive assumptions made about word-formation processes of the proto-language, i.e. suffixing (at least for person/number marking). Moreover, another useful principle can be invoked to illuminate the nature of the reconstructed *e/*o alternation. In particular, it is known that the relics of processes or earlier states that were once productive can be found embedded in compounds; for instance, English with- retains its original adversive meaning (‘against’) only in composite forms such as withstand (‘stand against’, not ‘stand alongside of’ with the synchronically regular comitative meaning of with). Therefore, one can look to compounds for some insights into aspects of earlier formations. The occurrence of *e in the genitive case of an old root noun found in the Greek composite form despo´te¯s ‘master, owner’, from *dems-pote¯s, literally ‘dwelling:gen-master’ as opposed to the consistent *o found in derivatives related to the root nouns, such as Lat. domus ‘house’, invites the suggestion that *e was proper, originally, to the oblique cases (e.g. genitive) whereas *o was found in the direct cases (e.g. nominative) of root nouns in the proto-language. Therefore, one can infer an original paradigm with nominative stem *pod- and genitive (oblique) stem *ped-.
A corollary of the use of relic forms for guidance in reconstructing earlier states is what can be called the Meillet Principle (after Antoine Meillet, based on work discussed in Arlotto 21981: 144 f.), namely to reconstruct from synchronic irregularities and isolated forms, that is any sort of unproductive material present in a language at a given time. The rationale is that synchronically unproductive material is exactly what demands a historical explanation, whereas synchronically productive forms could in principle have been created at any time by means of the productive and regular processes, and thus are not an indicator of the presence of some element in the proto-language. For example, when confronted with Skt. bhrø-ta- and Old Irish breth as past passive participles of *bher- ‘carry’, one can mechanically reconstruct *bhrø-to- as a Proto-Indo-European pre-form underlying them historically, even though the *-to- participial formation is clearly the productive one for Proto-IndoEuropean and on into the offspring languages, to judge from its widespread occurrence in all of the branches of the family. Adding in Lat. la¯-tus and Gk. ois-to´s as the participles to the outcomes of *bher- in these languages (present tense fero¯/phe´ro¯, respectively) complicates the picture somewhat, though. The Latin form seems clearly to be an import from the regular participle of a semantically related verb tollo¯ ‘pick up; take on’ (la¯tus from *tla¯tus, formed regularly with *-to- from a root *telH2-) and thus presumably simply shows the substitution of a participial form from a different paradigm; such a substitution is understandable, as tollo¯ supplied the form tulı¯ that functions as the perfect tense associated with fero¯. However, which form did la¯tus replace, an outcome of *bhrø-to- or something else again? Gk. oisto´s provides the answer, since it is synchronically isolated within Greek (the root ois- occurs elsewhere only in a suppletive future of phe´ro¯) and thus can only be explained historically, as a relic of an earlier irregular (i.e., suppletive) state. We thus reconstruct the Proto-Indo-European participle as *oistos and explain the Sanskrit and Irish forms as simply creations within those branches using the productive pattern. Meillet’s Principle thus licenses the reconstruction of suppletion, that is, a proto-language morphological irregularity, and in that sense can lead to a different solution than one might reach with internal reconstruction
154. Morphological reconstruction
where irregularities are taken back to earlier regular states. Similarly, allowing for the reconstruction of suppletion means that not every paradigm that one might want to reconstruct has to be fully articulated, with all members intact (i.e., all cells filled), in keeping with the observations from attested language states that there can be defective paradigms and that paradigms can be built up piece-meal (see Watkins 1962 for discussion of how paradigms can be reconstituted). The overriding factor here, as always, is to give the best account of the facts. In a sense, then, Meillet’s Principle is akin to the Accountability Principle, known from sociolinguistics (cf. Labov 1982: 30; Winford 1990: 227), in that it requires that all the forms in a set of comparanda be accounted for ⫺ in this case it is not enough to account for just the Sanskrit and Old Irish forms with a Proto-Indo-European form that does not allow for a straightforward account of the Greek suppletion. That is, an explanation is possible for why bhrøta- is to be found in Sanskrit or Celtic: at any time due to productivity of *-to- participial formation, such a form could have come into being; the Gk. ois-, on the other hand, demands a different kind of account, e.g. continuing an inherited suppletive paradigm, as suggested above.
4.
Going beyond simple reconstruction of forms
What else can be reconstructed of the morphology in a proto-language? Clearly, once one makes assumptions about the functions of particular reconstructed items, one has also made inferences about the relevant grammatical categories for the proto-language, e.g. number, person, tense, etc., based on *-ti and *-mi in Proto-Indo-European. However, it is also possible to go beyond these categories, and engage in some internal reconstruction on the value of categories reconstructed for a proto-language. For example, the suppletion that was reconstructed in the paradigm of *bher-, as discussed above (cf. 3), is found with other verbs in Proto-Indo-European, most often focusing on present versus past tense forms (as with Lat. fero¯/tulı¯, noted above (cf. 3), or Alb. ha ‘eats’, from a root *(H1)ed- vs. (he¨n)gra ‘ate’, from the root *gwro¯- (with prefixes *Ho-en-). One way to make sense of the relatively widespread occurrence of apparent present/past
1665 suppletion is to assume that it has to do with some inherent properties of the roots in question, derivable from their respective basic meanings. However, present versus past seems to be an unlikely distinction to be associated with the lexical meanings of particular roots (after all, how could an action be inherently associated with past time). Thus, it is more likely that the original distinction was aspectual in nature, e.g. imperfective (durative) versus perfective (completive), since that is a lexically encodable property (what is sometimes referred to as “Aktionsart”). Presumably, these aspectual distinctions were rearranged and altered somewhat on the way to the attested languages. Internal reconstruction, therefore, attempts to rationalize the irregularity of reconstructed suppletion by projecting it back to a previous state where it is not unmotivated but rather follows from some other property, in this case, aspect as determined by the semantics of a given root. This exercise thus leads to some reasonable inferences about the pre-ProtoIndo-European state of affairs regarding verbal categories. Internal reconstruction can also be carried out on the reconstructed morphemes themselves. For instance, the relation between the reconstructed third person singular present ending *-ti for Proto-Indo-European (as above, 2) and the comparable past ending *-t (cf. Skt. -t# ⫽ Lat. -d#), as well as *-mi vs. *-m in the first person, invites the analysis that the *-i itself marks present time, and that moreover the present ending was diachronically derived from the endings *-t and *-m through the addition of the *-i. It would follow, therefore, that the original value for *-t and *-m was just to mark person (and number), not to mark past time reference directly; only once the opposition of *-ti with *-t emerged was a present/nonpresent distinction relevant. Hypotheses such as these made by internally reconstructing from reconstructed proto-language elements are not subject to confirmation the way that internal reconstruction on an attested state is (e.g. reconstructing /-ez/ for the English plural allomorphy is verified by examining earlier English s-plurals). However, they are compelling scenarios to the extent that they are based on what is known about language in general. In a similar vein, and as noted above, to the extent that one can be sure of various trends in language change in general, they
1666
XVII. Morphologischer Wandel I: Theoretische Probleme
can be employed in reconstruction. Thus, although it is not impossible for affixes to turn into full-fledged words (see Joseph & Janda 1988, Campbell 1991, and Janda 2001 for discussion), it is nonetheless true that the opposite development, in which words develop into affixes, is by far the more common direction for changes involving words and bound morphemes. That is, examples such as Old English ha¯d ‘condition/state’ giving the Modern English derivational suffix -hood (cf. child/child-hood) from original compounds (‘the state of being X’) or the English negative adverb not giving, at least in part via phonological reduction in an unstressed position, the inflectional affix -n’t (see Zwicky & Pullum 1983), are relatively common, while examples such as the Old English bound genitive case-suffix -(e)s giving, via a reanalysis and aligning with the pronoun his, the lessbound and somewhat word-like possessive marker ’s in Modern English (less bound in that it attaches at the end of a phrase, as in the King of England’s hat) are significantly rarer, though definitely attested. This observation means that in some cases, we can reasonably locate the history of a given affix in a word, and this holds whether the basic methodology being employed is the Comparative Method or Internal Reconstruction. For instance, an internal comparison of the English adjectival and adverbial suffix -ly with the similarly used and phonologically similar free word like, as in quickly/ quick-like, friendly/friend-like, might suggest an historical derivation of the suffix from a reduction of the free word, perhaps under conditions of low accentual prominence, and a cross-linguistic comparison of the French adverbial suffix -ment with the more wordlike element in Spanish, -mente (word-like in that a single instance is distributable across two conjoined adverbs, e.g. clara y rapidamente ‘clearly and rapidly’), permits the reasonable inference that the suffix derives from a once less-bound element. In the case of -ly/ like, earlier English evidence confirms the reconstruction (cf. the use of -lic in Old English), as does Latin evidence in the case of -ment/mente (regarding the use of the ablative mente of the noun ‘mind’ in adverbial phrases, e.g. obstinata¯ mente ‘with an obstinate mind’, i.e. ‘obstinately’). Still, such hypotheses are not iron-clad, and synchronic resemblances can be misleading ⫺ the -less of friendless, for instance, has nothing to do
with the independent word less (from Old English læ¯s(sa)) deriving instead from the Old English preposition le¯as ‘without’.
5.
Pushing the limits: reconstructed states as real languages
Some of the successes of morphological reconstruction discussed in the previous sections involve simply applying analytic techniques that are well-known in linguistics, and in some instances, applying them to a reconstructed proto-language, essentially treating it as just another language, a synchronic state that can be analyzed and, among other things, subjected to internal reconstruction. As a final example of morphological reconstruction, an example is presented of how one can reconstruct morphophonemic rules for a proto-language, based on the Proto-IndoEuropean paradigm for the present of the verb ‘be’. This example serves as a suitable conclusion, as it draws on comparative methodology, internal reconstruction, and Meillet’s Principle, as well as general principles of morphological analysis. Besides the first and third person forms *esmi and *esti, it is possible to reconstruct a second person form as well in the singular. The comparison can be made of Skt. asi ‘you are’ with Gk. ei ‘you are’, both of which are irregular within their respective languages (e.g., the non-occurring **assi might be expected in Sanskrit, all things being equal). The “applied phonology” methodology (cf. 2) leads to a reconstruction for Proto-IndoEuropean of *esi, and the shared irregular status of the Sanskrit and Greek forms ensure that this form is to be posited for the proto-language; that is, the paradigm was not a defective one in Proto-Indo-European lacking a second person singular form. Moreover, this form itself is an irregularity within the Proto-Indo-European verbal system, since **essi would be expected, based on the secure reconstruction of a root *es- (cf. 2)) plus the second person singular ending *-si (seen, for instance, in Skt. bhara-si ‘you carry’, Gothic bairi-s, etc.). Given the difference between the form expected on morphological grounds, *essi, and the reconstructed form based on the comparative evidence, *esi, it is reasonable to reconcile the two forms by treating *essi as the (morphologically motivated) underlying form for ProtoIndo-European and *esi as the surface form,
154. Morphological reconstruction
also for Proto-Indo-European; this step means that a morphophonemic rule converting an underlying /-ss-/ into a surface [-s-] must be posited for the proto-language. Moreover, it licenses the inference that the underlying form was the actual pre-ProtoIndo-European form, and that a (sibilant) degemination sound change operated between pre-Proto-Indo-European and ProtoIndo-European to give the surface form *esi, and create the reconstructed irregular singular paradigm with *esmi/esi/esti.
6.
Conclusion
In a sense, then, morphological reconstruction is not significantly different in its goals, methods, and guiding principles from phonological reconstruction. It is thus not surprising that the topic (often under the rubric of “grammatical reconstruction”) is not discussed at great length in any standard textbooks, except insofar as internal reconstruction leads to results that have consequences of a morphological (most usually morphophonological) nature; still, interested readers should consult Anttila (1972; 1989), Fox (1995), Hock (21991), and Trask (1996) for additional general discussion and examples.
7.
References
Anttila, Raimo (1972), An Introduction to Historical and Comparative Linguistics. New York: MacMillan Anttila, Raimo (1989), Historical and Comparative Linguistics. Amsterdam: John Benjamins [revised edition of Anttila 1972] Arlotto, Anthony (21981), Introduction to Historical Linguistics. Washington, D.C.: University Press of America [11972, Boston: Houghton Mifflin Co.]
1667 Campbell, Lyle (1991), “Some Grammaticalization Changes in Estonian and Their Implications”. In: Traugott, Elizabeth & Heine, Bernd (eds.), Approaches to Grammaticalization, Vol. I. Amsterdam: John Benjamins, 285⫺299 Fox, Anthony (1995), Linguistic Reconstruction: An Introduction to Theory and Method. Oxford: Oxford University Press Hock, Hans Henrich (21991), Principles of Historical Linguistics. Berlin: Mouton de Gruyter [11986] Janda, Richard D. (2001), “Beyond ‘Pathways’ and ‘Unidirectionality’: On the Discontinuity of Language Transmission and the Counterability of Grammaticalization”. Language Sciences 23 (Special Issue: Campbell, Lyle (ed.), Grammaticalization: A Critical Assessment), 265⫺340 Joseph, Brian D. & Janda, Richard D. (1988), “The How and Why of Diachronic Morphologization and Demorphologization”. In: Hammond, Michael & Noonan, Michael (eds.), Theoretical Morphology: Approaches in Modern Linguistics. San Diego: Academic Press, 193⫺210 Labov, William (1982), “Building on Empirical Foundations”. In: Lehmann, Winifred & Malkiel, Yakov (eds.), Perspectives on Historical Linguistics. Amsterdam: John Benjamins, 17⫺92 Trask, R. L[awrence] (1996), Historical Linguistics. London: Arnold Watkins, Calvert (1962), Indo-European Origins of the Celtic Verb. Dublin: Institute for Advanced Studies Watkins, Calvert (1995), How to Kill a Dragon: Aspects of Indo-European Poetics. Oxford: Oxford University Press Winford, Donald (1990), “Copula Variability, Accountability, and the Concept of ‘Polylectal’ Grammars”. Journal of Pidgin and Creole Languages 5, 223⫺252 Zwicky, Arnold M. & Pullum, Geoffrey K. (1983), “Cliticization vs. Inflection: English n’t”. Language 59, 502⫺513
Brian D. Joseph, Ohio (U.S.A.)
XVIII. Morphologischer Wandel II: Fallstudien Morphological change II: Case studies 155. From Old English to Modern English 1. 2. 3. 4.
Introductory remarks Inflection Word-formation References
1.
Introductory remarks
1.1. Language periods The linguistic history of English may be divided into the following periods: Old English from the 7th century to about 1100, Middle English from ca. 1100 to ca. 1500, Early Modern English from ca. 1500 to ca. 1700, Late Modern English covering the 18th and 19th centuries, Modern English leading to Present-Day English. 1.2. Dialects Four Old English dialects can be distinguished: West Saxon, Kentish, Mercian, Northumbrian. The last two are called Anglian. In the Middle English period, they roughly correspond to Southern (or Southwestern), Kentish (or Southeastern), Midland and Northern respectively. Northern is also referred to as Northumbrian, the others as Southumbrian. The Midland area may be further subdivided into West and East, the East itself into North and South. Additional subdivisions, like Northwest Midland, are applied when necessary. The variety of English spoken in Lothian and Lowland Scotland developed from Northumbrian, being originally called Inglis, renamed Scottis in the 15th century. The dialect which came to be accepted as standard in Early Modern English was that of the London area, where political, military and economic factors contributed to its acquisition of this status. But most important was William Caxton’s publication in 1477 of the first book to be printed in England. He
based his spelling on that of manuscripts which had previously appeared there, particularly official documents, thus helping to spread the use of this dialect into other areas of England.
2.
Inflection
2.1. From inflection to non-inflection Old English is highly inflected, whereas Middle English shows the transition to a nearly non-inflected language, which was gradually reached in Early Modern English, Middle English being the time-span when the language changed from a (relatively) synthetic to a (relatively) analytic one. This will be illustrated by outlining the history of the inflectional systems of the noun and the verb. For adjectives and pronouns cf. Faiß (1989: 128⫺214; 1992: 21⫺38). 2.2. Nouns 2.2.1. Gender, Number, Case, Declension Like the German noun, the Old English one has three grammatical genders, i.e. masculine, feminine, neuter. The grammatical gender disappeared in the 14th century. Again, like in German, Old English nouns have the number-distinction singular – plural and four cases: nominative, genitive, dative, accusative. The instrumental, which differed occasionally from the dative in the earliest texts only, may be disregarded. Both languages, German and Old English, share two main declensions, strong (or vocalic) and weak (or consonantal), according to whether the stem originally ended in a vowel or -n respectively. Stems originally ending in other consonants than -n are usually subsumed under minor declensions (cf. Wright & Wright 1908: 160).
1669
155. From Old English to Modern English
Map 155.1: Old English Dialects
2.2.2. Strong declensions An example of the strong masculine declension is OE. eorl ‘nobleman’, one of the strong feminine declension OE. a¯r ‘honour’ (see Table 155.1). The strong neuter declension is different from the strong masculine one in that it has -u/-o in the nominative/accusative plural, if the noun has a short stem-syllable like hof ‘dwelling’, scip ‘ship’, but -Ø, if it is long as in ba¯n ‘bone’, hors ‘horse’, nett ‘net’ (cf. Faiß 1992: 16f.).
nom. gen. dat. acc.
2.2.3. Weak declensions An example of the weak masculine declension is OE. lida ‘sailor’ (see Table 155.2).
Weak feminine and neuter nouns like tunge ‘tongue’ or e¯age ‘eye’ are inflected the same way.
masculine
feminine
sing.
plural
sing.
plural
eorl eorles eorle eorl
eorlas eorla eorlum eorlas
a¯r a¯re a¯re a¯re
a¯re, -a a¯ra, -(e)na a¯rum a¯re, -a
Tab. 155.1: Strong masculine and feminine declensions
1670
XVIII. Morphologischer Wandel II: Fallstudien
Map 155.2: Middle English Dialects
nom. gen. dat. acc.
singular
plural
lida lidan lidan lidan
lidan lidena lidum lidan
nom. gen. dat. acc.
singular
plural
fo¯t fo¯tes fe¯t fo¯t
fe¯t fo¯ta fo¯tum fe¯t
Tab. 155.2: Weak declensions
Tab. 155.3: Minor declensions
2.2.4. Minor declensions An example of one of the minor declensions is OE. masculine fo¯t ‘foot’ (see Table 155.3). The morphonemic alternation /fo:t/~/fe:t/ is due to i- (or palatal) mutation, cf. the original Germanic forms of the dative singular,
*fo¯ti, and of the nominative plural, *fo¯tiz, whence the i-mutated vowel spread into the accusative plural, whose original form was *fo¯tunz (cf. Brunner 31965: 226; Wright & Wright 1908: 192).
155. From Old English to Modern English
2.2.5. Plural and genitive The Old English inflectional system of the noun was significantly reduced on its way to Present-Day English (cf. Faiß 1989: 109⫺ 127). The only markers to have been retained for inflection are in fact -(e)s for plural and genitive as well as -(e)n and vowel change (or /replacive/) for plural as, e.g., in oxen and feet (cf. Faiß 1992: 15f.). The other markers had been dropped by the 14th century. Since then the noun has had a two-case system with a common (or unmarked) case and a genitive (or possessive or marked) case together with the important number-distinction singular ⫺ plural. Among the inflectional suffixes retained, -(e)s has become productive. As a plural indicator it goes back to -as of the Old English strong masculine declension. -as yielded ME. -es, which was common by the end of the 14th century. It was gradually replaced by -s from about 1500 onwards except after sibilants. The modern distribution of the phonemically conditioned allomorphs of {plural}, /=z, z, s/, established itself during the 17th century. Genitival -(e)s stems from the genitive singular of the Old English strong masculine and neuter declensions and became generalized in the 14th century. The ousting of the Middle English successors to the Old English genitive plural markers came to an end in the 14th century too, so that, apart from a few exceptions, -(e)s expressed both singular and plural genitive. The tendency towards generalisation proved stronger than the resistance of individual nouns. As to the spelling, <’s> was not fixed until the 1690s. Obviously, <s’> is not attested before the 2nd half of the 18th century. The pronunciation and the allomorphic distribution of the s-genitive are identical to the s-plural. (For the history of mixed plurals like loaf-loaves, house-houses, mouth-mouths, which show voicing of the final consonant of the singular, and for plurals of nouns from Greek and Latin such as alumnus, alga, velum, crisis, oxymoron and many others cf. Faiß 1992.) 2.2.6. of-Genitive, to-Dative Clear signs of the language’s adoption of an analytic character are the of-genitive and the to-dative. They originated in Old English from constructions consisting of the prepositions of and to¯ plus noun phrase, which were preferably used to denote origin and direction/goal respectively. Yet it was not before
1671 the Middle English period that they took over the syntactic-semantic functions of the inflected (or synthetic) genitive and dative. The question remains whether French constructions with the prepositions de and a`, which had replaced the Latin synthetic genitive and dative, encouraged the expansion of the corresponding English ones. 2.3. Verbs Old English verbal inflection may be compared to that of German, Greek, Latin, Italian, Spanish, to a lesser extent to that of French, with respect to the number of inflectional signs denoting person, number, tense and mood. The verb in Present-Day Standard English, on the contrary, has but one such sign at its disposal, i.e. -(e)s for 3rd person singular present indicative as in he pull-s /he push-es. Old English strong verbs comprise eight classes. The two last ones are generally put together in one with a subdivision into a) and b) or 1. and 2. Old English weak verbs are generally grouped in three classes. 2.3.1. Old English strong verbs A characteristic feature of the Old English strong verbs is ablaut, representing some kind of internal inflection which has been retained in the corresponding modern stem-forms. They were reduced from four to three in the course of time, with but one exception: the verb be has still four, namely be-was-werebeen. The Old English forms which are usually adduced are: infinitive ⫺ 1st/3rd person singular preterite indicative ⫺ plural preterite indicative ⫺ past participle (cf. Tab. 155.4) 2.3.2. Preterite and past participle The nowadays common distribution of the phonemically conditioned allomorphs of {preterite} and {past participle}, /id, d, t/, gradually established itself in Early Modern English. The usual Middle English ending was -ed. It is the merger of OE. -ed(-), which indicates preterite and past participle of certain members of the 1st weak class, and of West Saxon -od(-), non-West Saxon -ad(-), as typical indicators of the 2nd weak class. The Old English strong verbs were marked by -en for past participle and also by -an/-on/ -un in the non-West Saxon dialects. These suffixes merged into ME. -en. -n denoted the past participles of be¯on ‘be’, do¯n ‘do’, ga¯n ‘go’. Middle English is characterized by a
1672
XVIII. Morphologischer Wandel II: Fallstudien
infinitive
preterite singular
1st class PrE. 2nd class PrE. 3rd class a) PrE. 3rd class b) PrE. (weak) 3rd class c) PrE. (weak) 3rd class c) PrE. 4th class PrE. 5th class PrE. 6th class PrE. 7th class a) PrE. (weak) 7th class b) PrE.
rı¯dan ride sce¯otan shoot bindan bind helpan help ceorfan carve feohtan fight brecan break giefan give sc(e)acan shake læ¯tan let cna¯wan know
plural
ra¯d
ridon rode
sce¯at
sc(e)uton shot
band/bond
bundon bound
healp
hulpon helped
cearf
curfon carved
feaht
fuhton fought
bræc
bræ¯con broke
geaf
ge¯afon gave
sc(e)o¯c
sc(e)o¯con shook
le¯t
le¯ton let
cne¯ow
cne¯owon knew
past participle riden ridden sc(e)oten shot bunden bound holpen helped corfen carved fohten fought brocen broken gefen given sc(e)acen shaken læ¯ten let cna¯wen known
Tab. 155.4: Strong verbs
high degree of variation between past participles with and without -n, cf. iwurÎen, ihalde, ofslage, idon. Variation between -(e)n and -Ø still existed in Early Modern English. From the 17th century onwards, the present distribution has been established, but not obligatorily, as is proved by modern variants like beaten/beat, bidden/bid and others. The prefix i- is descended from OE. ge-, cf. Germ. ge-, which was optionally attached to past participles. Also writtenin Middle English, it disappeared towards the end of that period, lingering on, however, as an archaism in the following periods, above all in poetry. 2.3.3. Present participle The common marker for present participle in Old English was -ende, varying with -ande in Northumbrian, which mainly occurred with verbs of the 2nd weak class. -ende continued to exist in Middle English, especially in the Southeast Midland. -ande appeared as -and in Northumbrian and Scottish after the loss of unstressed final -e in Northern Early Middle English. A third form arose in the 12th century, -inde, showing [=] through the raising
of unstressed [e] before [nd]. This suffix was widely used, yet most frequently in the West Midland and the South. PrE. -ing, though, does not go back to any one of them. In Old English the predecessor of this suffix, -ing, served as a means of deriving mostly feminine abstract nouns from verbs belonging to the 1st weak class such as nering ‘protection, defence’ from nerian ‘save’. The first instances of present participles with -ing can be detected about 1200 in some southern regions. It was rapidly spreading becoming generalized in the whole Southumbrian area at the end of the 14th century, with some sporadic -inde relics surviving in 15th century texts. It was in this century that -ing entered into Northumbrian and Scottish, there competing with -and, which was eventually ousted except for some dialectal traces in Southern Scots. 2.3.4. Infinitive Old English {infinitive} has the allomorphs /an, ian, n/. /an/ occurs with weak verbs of the 1st and 3rd classes and with strong verbs, cf. de¯man ‘deem’, libban ‘live’, rı¯dan ‘ride’, /ian/ with weak verbs of the 1st class having stem-final -r and with those of the 2nd class,
1673
155. From Old English to Modern English
cf. nerian ‘save’, endian ‘end’, /n/ with contracted verbs such as te¯on ‘draw’ (strong), Ìe¯on ‘press’ (weak) and with be¯on ‘be’, do¯n ‘do’, ga¯n ‘go’. /an/ and /ian/ generally merged into /en/ <en> in the 12th century, while /n/ remained as such. There was a great deal of variation between n- and n-less infinitives in Early Middle English. Owing to the loss of final -n already in Late Old Northumbrian, followed by that of final -e in the 13th century in its dialectal successors, the inflectionally unmarked infinitive was the rule in 14th century Northumbrian and Scottish, cf. accord, disclar, succeid. In Southumbrian the infinitive grew unmarked in the 15th century, cf. do, reduce, translate. 2.3.5. Person, number, tense, mood Old English verbs are inflected for person, number, tense, and mood, often showing a combination of internal and external inflection, as do German verbs. In keeping with Verner’s Law stem-final consonants can alternate, and so can the stem-vowels in the 2nd and 3rd persons singular present indicative on account of -i- in the original inflection of these forms. An example of this kind of verbs is ce¯osan ‘choose’ (cf. Tab. 155.5; imperative singular ce¯os, plural ce¯osaÌ). Some peculiarities need explaining: -t in c=¯ est and -Ì in c=¯ esÌ came into existence through vowel deletion in -est and -eÌ respectively, cf. de¯m(e)st, de¯m(e)Ì (1st weak class). -est, Northumbrian -es, is also present in the 2nd person singular preterite indicative of weak verbs as in Ìu¯ de¯mdest. The verbs of the 2nd weak class have -ie in the 1st, -ast in the 2nd, -aÌ in the 3rd person singular and -iaÌ in the plural present indicative, cf. lo¯cie, lo¯cast, lo¯caÌ, lo¯ciaÌ of lo¯cian ‘look’.
During the Middle English period, the rather sophisticated inflectional system, which, by the way, was still more complex on account of marked dialectal variation, became radically simplified (cf. Faiß 1989: 215⫺297). The only overt person-denoting inflectional signs typical of Early Modern English were in fact -(e)st and -(e)Ì/-(e)s for the 2nd person singular present and preterite and 3rd person singular present indicative respectively. In the standard language -(e)st dropped out in the 18th century, when thou became replaced with you. -(e)th yielded to -(e)s around the middle of the 17th century, being distributionally equal to plural and genitive -s. With the exception of the were-subjunctive in the 1st and 3rd persons singular, -Ø in the 3rd person singular and be for all persons in the present subjunctive, this mood is no longer recognizable formally, cf. if I/he were, he goes vs. he go, I … be. The history of -s for 3rd person singular present indicative has been long debated. Yet the following may be rather safely concluded. The starting-point is the Late Northumbrian dialect (2nd half 10th century). Here the 2nd person present indicative is mainly marked by -s, which is a general characteristic of early texts but became superseded by -st in other dialects. -s first spread into the plural present indicative gradually replacing -Ì, before it appeared in the 3rd person singular present indicative (cf. Berndt 1956: 184⫺196, 202⫺213). The linguistic fact itself is best explained by a vivid tendency towards the generalization of one single inflectional sign, which is not only obvious in this case, but is also illustrated by, e.g., the s-plural and the s-genitive. In Middle English, -s dominated in Northumbrian and Scottish, whence it made its way slowly via the Midlands to the
present
preterite
singular
plural
singular
plural
indicative 1st person 2nd person 3rd person
¯˘ıc ce¯ose Ìuu˘¯ c=¯ est hee¯˘ c=¯ esÌ
wee¯˘ ce¯osaÌ gee¯˘ ce¯osaÌ h=¯ e ce¯osaÌ
¯˘ıc ce¯as Ìuu˘¯ cure hee¯˘ ce¯as
wee¯˘ curon gee¯˘ curon h=¯ e curon
subjunctive 1st/2nd/3rd persons
¯˘ıc/Ìuu˘¯ / hee¯˘ ce¯ose
wee¯˘ /gee˘¯ / h=¯ e ce¯osen
¯˘ıc/Ìuu˘¯ / hee¯˘ cure
wee¯˘ /gee˘¯ / h=¯ e curen
Tab. 155.5: Inflection
1674
XVIII. Morphologischer Wandel II: Fallstudien
South. There, however, it had to fight it out against -th, even in London, where the latter persisted to the middle of the 17th century. After being ousted from the standard language by -s, -th has survived in literary style and dialectally, as has -st for the 2nd person singular. In contrast to -st or -s as the successor to Northern -s of the 2nd person singular present indicative as well as to Early Modern English reduced -st, -th is practically restricted to Eastern Cornwall and Southern Devonshire. 2.4. Conclusion Unlike that of German and to a lesser extent those of the Romance languages the inflectional history of English is characterized by a clear-cut tendency towards generalization. This necessarily means a radical minimizing of inflectional signs of the parts of speech discussed. Thus -s for plural, genitive and 3rd person singular present indicative, -ing for present participle/gerund, -ed for preterite and past participle have become generalized. As a consequence new devices had to be invented to express syntactic relations. This was, on the one hand, the fixing of word-order and on the other the use of prepositions, as is proved by the of-genitive and the to-dative. The developments described set in mostly in Middle English. Some of them can first be detected in the northern dialects making their way more or less rapidly to the South and into the standard language.
3.
Word-formation
The instances below have been mainly selected from BT (1976), BTS (1973) (Old English), MED (1954 ff.) (Middle English), Marchand (21969), OED (21989), more recent ones from Ayto (1989), Barnhart et al. (1990) as well as from newspapers and the like (cf. also Faiß 1992: 54⫺111).
‘very hard’, though no longer productive, has been retained, as in forbid, forget. un-1 with negative meaning as in unfæger ‘not fair, ugly’, unfriÌ ‘unpeace, war’ is highly frequent in Old English (cf. Marchand 21969: 201), yet only few coinages have survived, like unclean, unripe. un-2 with reversative and privative meaning as in unfealdan ‘unfold’ and ungeocian ‘unyoke’ respectively formed a lot of verbs in Old English, of which but a few still exist such as undo, untie. Old English suffixes have been preserved to a greater extent. -ed ‘provided with, having’ as in hilted ‘hilted’, hringed ‘ringed’ has gained nearly unlimited productivity with bases consisting of, e.g., adjective ⫹ noun: beautiful-gowned, high-heeled, and noun ⫹ noun: diesel-engined, eagle-eared (cf. Faiß 1987: 124⫺129). Many were coined in Early Modern English. Shakespeare has, among others, light-winged, nimblefooted and honey-tongued, iron-witted (cf. Scheler 1982: 118). -er
‘one having to do with’ used to form agent nouns (cf. Kastovsky 1971) such as bæcere ‘baker’ from bacan ‘bake’ (deverbal) and bo¯cere ‘scribe’ from bo¯c ‘book’ (denominal). Predominantly from the 16th century onwards, -er has come to derive instrument nouns like blotter, nutcracker, sailer. The suffix is extraordinarily productive in PresentDay English owing to its capability of being appended to practically any base (cf. Faiß 1987: 118⫺121).
3.1.1. Affixation Some Old English prefixes can still be observed in Present-Day English.
-isc, PrE. -ish, ‘belonging to’ originally derived ethnic adjectives with i-mutated stem-vowels, such as frencisc ‘French’, englisc ‘English’, scyttisc ‘Scottish’. The meaning of the suffix changed to ‘nature of, condition’ as in OE. ceorlisc ‘churlish’, cı¯ldisc ‘childish’. In Middle English, it developed a disparaging tinge as in foolish (before 1300), swinish (ca. 1200). Nowadays -ish is widely used in colloquial and journalistic language (cf. Faiß 1987: 124).
for- ‘forward, away’ (often intensifying) as in forbærnan ‘burn (away)’, forheard
-nes, PrE. -ness, ‘state, condition’ mainly derived adjectives and participial adjec-
3.1. Old English Old English basically shares the word-formation types of the Germanic languages: affixation (prefixation and suffixation), compounding and zero-derivation (cf. Krahe & Meid 1967; Art. 37, 87, 89, 90).
155. From Old English to Modern English
tives as æÌelnes ‘nobility’ from æÌele ‘noble’, cy¯Ìnes ‘testament’ from cy¯Ìed, past participle of cy¯Ìan ‘make known’. In Present-Day English, the suffix may be appended to almost any base except verbs, thus to adjectival composites as in selfconceitedness, square-tonguedness, adjectival compounds as in airsickness, homesickness, and adjectival phrases as in get-at-ableness, little-boyishness (cf. Faiß 1987: 123). 3.1.2. Compounding Compounding is a productive process not only in Old English but in all Germanic languages. Old English compounds are predominantly nouns. The majority of them follow the patterns noun ⫹ noun: niht-waco ‘night-watch’, scip-ra¯p ‘ship-rope’; adjective ⫹ noun: eald-fæder ‘grandfather, ancestor’, he¯ah-clif ‘high cliff’; verb(-stem) ⫹ noun: etelond ‘pasture land’, hwet-sta¯n ‘whet-stone’; verbal noun in -ung/-ing ⫹ noun: langunghwı¯l ‘time of longing, weariness’, stemping¯ısern ‘stamping-iron’. Old English has its share of compound adjectives as well. Productive patterns are: noun ⫹ adjective: ælmes-georn ‘alms-eager, generous’, mere-we¯rig ‘sea-weary’; adjective/adverb ⫹ adjective: wı¯d-cu¯Ì ‘widely known’, wı¯shycgende ‘wise-thinking’. Additive copula compounds are rare in Germanic in general, being still rarer in Old English. The numerals thirteen to nineteen, e.g., belong here: Ìre¯oty¯ne, fe¯owerty¯ne, fı¯fty¯ne, siexty¯ne, siofonty¯ne, eahtaty¯ne, nigonty¯ne. Productivity has considerably increased in recent years: poet-critic-humanist, saxophonist-composer-bandleader, soldier-statesman-author-orator. The development of Ìre¯oty¯ne to thirteen and fı¯fty¯ne to fifteen leads to the problem of obscured compounds (cf. Faiß 1978), meaning that original compounds can no longer be analyzed as such from the point of view of modern English. 3.1.3. Zero-derivation Derivation by means of a zero-morpheme is a productive process in Old English. It occurs most often with verbs yielding nouns (cf. Kastovsky 1968) and with nouns yielding weak verbs of classes 1 and 2. Examples of deverbal nouns are: from strong verbs: drinc ‘drink, draught’ from drincan ‘drink’, stand ‘stand, pause, delay’ from standan ‘stand’; from weak verbs: hete ‘hatred’ from hatian
1675 ‘hate’, ræ¯d ‘council, advice’ from ræ¯dan ‘advise’. Denominal verbs are: class 1 (with imutation): ble¯dan ‘bleed’ from blo¯d ‘blood’, de¯man ‘deem’ from do¯m ‘doom’, fe¯dan ‘feed’ from fo¯da ‘food’; class 2: beddian ‘prepare/ make a bed’ from bed(d) ‘bed’, cealfian ‘calve’ from cealf ‘calf’, horsian ‘provide with (a) horse(s)’ from hors ‘horse’. Verbs from adjectives are: class 1 (with imutation): fy¯lan ‘befoul’ from fu¯l ‘foul’, fyllan ‘fill’ from full ‘full’; class 2: deorcian ‘grow dark’ from deorc ‘dark’, ealdian ‘grow old’ from eald ‘old’. Class 2 verbs from particles are: forÌian ‘send forth’ from forÌ ‘forth’, innian ‘get within’ from in ‘in, within’; from verbs: fandian ‘examine’ from singular preterite fand of findan ‘find’, lifian ‘live’ from preterite stem lif- of libban ‘live’. The productivity of zero-derivation has been on the constant increase throughout the history of English. This was favoured by the fact that, with no grammatical suffixes to denote infinitive since around 1500 at the latest, words easily shifted from one lexical category to the other. Mention has to be made of exocentric or pseudo-compounds, i.e. zero-derived composites (cf. Marchand 21969: 386⫺389). Most Old English formations of this type are adjectival bahuvrihi compounds with the patterns adjective ⫹ noun, numeral ⫹ noun and noun ⫹ noun like bær-fo¯t ‘barefoot(ed)’; a¯n-e¯age ‘one-eyed’; wulf-heort ‘wolf-hearted’. Already in Old English -ed coinages were strong competitors. They succeeded in replacing those patterns in Early Middle English, generally speaking, with the exception of barefoot. Concerning nominal bahuvrihis, Old English has some loan-translations of Latin animal and plant names: a¯n-horn for unicornis, fı¯fle¯af for quinquefolium, genitival hundes tunge for cynoglossum, to which may be added dæges-e¯ge ‘daisy’, oxan-slyppe ‘oxlip’ (plants) and coinages such as lady’s bedstraw, lady’s finger (plants), bull’s-eye, cat’s-eye (objects), which are parallelled by genuine compounds like doomsday and driver’s seat (cf. Marchand 2 1969: 65⫺69). They are equally attested in Old English: Tı¯wes-dæg ‘Tuesday’ and Wo¯dnes-dæg ‘Wednesday’ are but two instances. 3.2. Middle English 3.2.1. Affixation Many prefixes which are still productive like de-, dis-, in-, non- came over from French during the Middle English period. Among
1676 them non- ‘not’, which represents Lat. no¯n, Anglo-Norm. noun-, Central Fr. non-, has developed a combinatory flexibility which is unequalled by other prefixes. Although having achieved its full productivity after the middle of the 17th century, it can be regarded as fairly productive in the 15th and 16th centuries. The first instances are direct borrowings from French: noumper ‘umpire’ (ca. 1350), nonpower ‘non-power’ (before 1387) (nouns), noncerteyn ‘non-certain’ (probably before 1400), nounparalle Fr. ‘non-pareil’ (ca. 1450) (adjectives), and from Latin: noun laudabile (before 1398). As to Eng. non-, the following may be observed. In the 15th century, it predominantly combines with deverbal nouns denoting action, activity, condition, quality. The pattern has been documented by numerous examples to the present day, with the exception of nouns in -ing, whose connection with nonobviously dropped out of use in the 19th century. Some instances are: non-Punyshynge (1422), Non-finding (1525), non-syllogizing (1651), non-repealing (1791), non-understanding (1833), non-labelling (1901); noun obeisaunce (1447), non-distribucioun (1540), noncommitance (1650), non-submission (1763), non-pronunciation (1842), nontransitivity (1974). Combinations with agent nouns have become frequent since the 19th century, some instances being: non housholder (1439), nonsubscribers (1599), Non-contributors (1643), Nonjuresses (1723), non-smoker (1846), nonachiever (1972). Coinages with other nouns are not as frequent: noncredibility (before 1450), non-religion (1600), non-freemen (1760), non-form (1886), nonfriends (1971). If used attributively, though, the pattern appears to be rather free from restrictions, which is corroborated by lots of examples from the 19th and 20th centuries like nonfactory (1835), non-market (1884), Nonkernel (1963), nontitle (1973). These centuries also abound with adjectival coinages: non-adherent (1870), non-analytic (1890), nonearthly (1970), nonpermissive (1972). non- has easily combined with adjectival present and past participles since the 17th century: Nonpreaching (1622), non-reading (1797), non-enquiring (1864), non-fattening (1971); Non-incarnated (1671), non-perforated (1776), nongalvanized (1849), non-palatalized (1956). Since the 19th century, it has been attached to adjectival compounds whose determinata are present participles, such as non-officeholding (1888), non-English-speaking (1949).
XVIII. Morphologischer Wandel II: Fallstudien
It has been used with adverbs since the 17th century: non-attendingly (1678), non-vortically (1882), nonbiologically (1974). Though reluctantly, it has been prefixed to verbs since the 17th century as well: Non-act (1645), nonconcur (1703), non-apply (1846), non-intervene (1969); with infinitives it has been employed since the 1920s to form an attributive phrase as in non-crush (1924), non-shrink (1962). Since the late 1960s it has acquired the additional meaning ‘not being true, genuine; pseudo-’ as in the nouns non-actor (1970), non-newspaper (1973). The impact of French on English suffixes is about the same as that on prefixes (cf. Kastovsky 1985: 227). Among those arising in Middle English can be found -able and -ess. -able ‘fit for doing, fit for being done’, with the variant -ible, forms adjectives from verbs and nouns like determinable, understandable; honourable, pesible ‘peaceable’. Deverbal adjectives have always been more frequent than denominal ones, in Middle English as well as in later periods. Instances are: bombable (1930), demountable (1909), flyable (1936); carriageable (1702), fissionable (1945). -esse, PrE. -ess, forms [⫹female] nouns parallelly to [⫹male] personal nouns such as charmeresse, daunceresse, frendesse. The suffix is most productive in the 16th and 17th centuries slowly declining afterwards. The 19th century has squiress, visitress, waitress; rebeless dates from 1912. 3.2.2. Compounding Early Middle English continues the compounding traditions of Old English, with new types and patterns emerging (cf. Sauer 1992). In inversion compounds the usual sequence determinant ⫺ determinatum is permuted under French influence: knight erraunt (ca. 1400) from chevalier errant ‘knight errant’, herbe Ive, herbe Robert (plant names; before 1300). Further instances are: courtmartial, cousin german, heir apparent. Gender-denoting attributive copula compounds with he- and she- as determinants came into being about 1300, originally designating but animals: hee-catte ‘he-cat’, he lambe ‘he-lamb’; she-asse ‘she-ass’, shee ape ‘she-ape’. Words for persons have been recorded since the 16th century, those with shebeing more frequent: he-friend, he-lover; she-
155. From Old English to Modern English
baker, she-favorite, she-friend, 20th-century being: he-male (1909); She Guardian (1937), she-robot [⫹human] (1991). Compounds with particles like over, OE. ofer, as in overlord or overbysy ‘overbusy’, and out, OE. u¯t, as in outcrie ‘outcry’, outlep ‘outleap’ remain common in Middle English. 3.2.3. Zero-derivation In Middle English, the number of nominal bahuvrihi-compounds increases considerably. The paraphrase of such a combination is usually ‘X has/is characterised by what is expressed by the composite’. Instances include Curtmauntel ‘C(o)urtmantle’ (surname of Henry II), long-here ‘person with long hair’, redbrest ‘redbreast’ (bird), stykylbak ‘stickleback’ (fish). Person names and nicknames are most frequent: Henheued ‘hen-head’, Hertblod ‘heart-blood’, Irnefot ‘iron-foot’ (cf. Jönsjö 1979: 111, 115; Erlebach 1979: 71⫺87, 93⫺99 for names of French origin such as Bonquer ‘bon cœur’ and Cordelion ‘cœur de lion’, i.e. ‘Lionheart’). In the following centuries, many [⫹human] as well as [⫺human] nouns have been coined. The meaning of the former is mainly pejorative, that of the latter only occasionally. As in Middle English, the patterns adjective ⫹ noun and noun ⫹ noun occur most frequently, cf. wetback ‘illegally immigrated Mexican’, whitethorn (plant), whitethroat (bird); bricktop ‘red-haired person’, cottontail ‘rabbit’, paperbelly ‘person unable to drink much’. This highly restricted selection may be completed by the weak patterns numeral ⫹ noun as in four-eyes ‘person wearing glasses’, four-wheel (vehicle) and reduced adjectival past participle ⫹ noun as in crackbrain, scatterbrain ‘stupid person’. The latter obviously did not come into existence before the 16th century. A different pattern originating in Middle English consists of verb and complement. It arose under the influence of French imperative phrases and is hence sometimes termed imperative compound (but cf. Giurescu 1975: 69⫺71). Early Middle English coinages follow French conventions in that they mostly denote personal agents: pyke-purs ‘pickpurse’, trayle-bastoun ‘s. o. who carries a club or a cudgel’ (a criminal), wesche-disch ‘dishwasher’. Numerous person names and nicknames, a few attested in the 2nd half of the 11th century already (cf. Reaney 21976: xl), are formed this way, like Schaketaille ‘shake-tail’, Schaketrot ‘shake-trout’, Schaksper ‘shake-spear’ (cf. Jönsjö 1979: 156). All
1677 of them have a pejorative tinge, which is not inherent in the rare [⫺animate] nouns dating from the 14th and 15th centuries such as breakfast and turnpike. 3.3. Early Modern English and Later The transition from Middle to Early Modern English and then to Late Modern English was not as drastic as that from Old to Middle English. No significant breaks occurred in word-formation processes, which nonetheless were enriched by new concepts. 3.3.1. Affixation Some modern affixes of note became productive in Early Modern English. Among the prefixes are counter-, de-, inter-, pre-, re-; suffixes include -ation, -ee, -ive, -ment, as in counterevidence (1665), depopulate (1545), intermarry (1574), pre-elect (1570), reconsider (1571); derivation (1530), debtee (1531), trustee (1647) (patient nouns, agent nouns such as escapee (1875⫺76) are later and particularly of American origin), persuasive (1589), retirement (1596). Besides, Early Modern English experienced the rise of combining forms, which have become highly productive. They combine not only with free morphemes, but also with other combining forms. Early instances are: -cracy ‘rule, power’ from aristocracy (1561), -logy ‘doctrine, study of’ from theology (as early as 1362), poly- ‘many’ from poly-tragicke (1605), some current formations being: Eurocracy, mediacracy, meritocracy; escapology, mafiology, urbanology, with reference made to the study of political situations or events as in Kremlinology, Pekin(g)ology, Sovietology; polydrug, polymorphous, polynuclear. Further combining forms being productive nowadays are: bio- ‘life’ as in bio-ethics, biometric; eco- ‘ecology, ecological’ as in ecotourist, ecotoxicology; Euro- ‘Europe, European’ as in Europerson, Euroslit; mega- ‘extremely; very large; one million’ as in megaboring, mega-busy; megastore; megatanker. 3.3.2. Compounding Fowel of flight (ca. 1300) is the first instance of a new compound pattern the determinant of the inversion compound being a prepositional phrase with of. Productivity is achieved in Early Modern English with instances like Bill of Exchange (1579), House of Commons (1621), House of Lords (1672).
1678
XVIII. Morphologischer Wandel II: Fallstudien
3.3.3. Zero-derivation The pattern verb ⫹ complement has become frequent since Early Modern English (cf. Uhrström 1918), although, in our century, it has lost much of its productive force. Person nouns are: lack-love, mumble-news, pleaseman from Shakespeare, further instances being killjoy (1776), turnkey (1654), turn-penny (1824). Animal nouns are: shearwater (ca. 1671), shufflewing (1829), wagtail (1510), plant names catch-fly (1597), cover-shame (1693), heal-all (1853), others (tools, devices, garments, etc.) breakwater (1721), covershame (1629), stopgap (1533). A different kind of zero-derivation is represented by the patterns showoff and blackout. Both gave rise to numerous coinages especially in the 19th century and afterwards, the frequency of non-animate blackout nouns being higher than that of mainly [⫹human] showoff nouns, which did not come into existence till the 2nd half of the 16th century, while members of the other group, though rarely, appeared a century earlier. Some instances are: go-between, look-alike, runabout, runaway; build-up, cutback, cut-down (cf. Marchand 21969: 380⫺386).
blatterature (ca. 1512) from blatter and literature. From various spheres are: with the first free constituent clipped: compucenter, docudrama, helihop; with the second one clipped: beefalo, happenstance, steelionaire. Blending does not seem to be frequent until the 19th century: lemeren ‘glisten, glimmer’ (probably before 1400) from lemen ‘glow, shine’ and glimeren ‘gleam’, blurt (1573) from blow and spurt, Barsolistor (1888) from barrister and solicitor, clantastical (1803) from clandestine and fantastical are some instances (cf. Bergström 1906: 51⫺70). In Present-Day English, though, blending is a productive means of expressing the mixture of two concepts, such as languages in contact: Engdish from English and Yiddish, Frenglish from French and English, Japlish from Japanese and English. Acronymics is the latest addition to shortening in the English language. It obtained some productivity in the 19th-century special language of chemistry (cf. Marchand 21969: 453), but the first real surge of acronyms did not come before World War I: ANZAC ‘Australian New-Zealand Army Corps’, DORA ‘Defence of the Realm Act’ are but two examples.
3.3.4. Shortening The shortening processes clipping and blending (cf. Art. 91; 92), developed to a certain degree in Middle English, intensified during the centuries to follow. Together with acronymics (cf. Art. 92) they have become outstanding features of the modern language. Clipping got firmly established in Early Modern English. Examples comprise: mere ‘siren, mermaid’ (1225) from mere-min, gent (1564) from gentleman, mas (1575) from master (back-clippings); lepi ‘single’ (ca. 1303) from onlepi, wig (1675) from periwig, winkle (1585) from periwinkle (fore-clippings). Backclipping has turned out to be the most productive among clipping processes. Some instances are: celeb from celebrity, cig from cigarette, fax from facsimile machine (backclippings); burger from hamburger, liner from airliner (fore-clippings); fridge from refrigerator, Soo from Lake Superior (fore- and backclippings, rare); Ten-Tom from TennesseeTombigby-Waterway, Tex-Mex from TexasMexican (multiple back-clippings). Clipping also affects compounds. Consisting of one clipped and one non-clipped free constituent (either the first or the last), they are called “clipping compounds”. An early instance is
3.3.5. Back-derivation Often considered a sub-class of shortening, back-derivation (or back-formation) is the coining of new words by the deletion of actual or supposed suffixes in longer ones (cf. Marchand 21969: 391⫺395). Most back-derivations are verbs, as laze (before 1592) from lazy, dizz (1632) from dizzy; infract (1798) from infraction, book-keep (1886) from bookkeeping, caretake (1893) from caretaker; team-teach (1976) from team-teaching, backcalculate (1987) from back-calculation. 3.3.6. Reduplication Being sparsely attested in Middle English, reduplication (cf. Art. 57) has become more common since Early Modern English and is highly productive in the modern language. Rhyming combinations are: handy-dandy ‘covert bribe or present’ (1362); hocus-pocus ‘jugglery, trickery’ (1624), hurly-burly ‘tumult, confusion’ (1539); hubble-bubble ‘confused talk’ (1720), mumble-jumble ‘speak indistinctly’ (1833), quavery-wavery ‘undecided’ (1749). The 20th century has clankie-tankies ‘members of the Royal Tank Regiment in the Gulf War’, heebie-jeebies ‘state of being scared’, hotsy-totsy ‘fine’.
155. From Old English to Modern English
Ablaut combinations have quickly gained ground: pitter-patter ‘rapid repetition of words’ (ca. 1425); bibble-babble ‘idle talk’ (1532), clitter-clatter ‘chatter’ (1535), dillydally ‘waste time’ (before 1610); twiddle-twaddle ‘foolish chatter’ (1798). From the 19th century onwards have been coined drizzledrazzle ‘continuous rain’, fizz-fuzz ‘CocaCola’, rickety-rackety ‘unsteady’. Reduplicative compounds (or repetitions), infrequent in Middle English, remained rare to the 18th century. Since the 19th, however, their frequency has been increasing considerably (cf. Thun 1963: 49⫺200 passim). Some sound-imitating coinages are: cuccu ‘cuckoo’ (ca. 1240); diddle-diddle (fiddle) (1523), jinglejingle (jingling sound) (1664); tick-tick (1774) (clock), too-too (flute) (1812), woo-woo (wind) (1841), zip-zip (light sharp sound) (1875). choo-choo (train), chuff-chuff (engine), clop-clop (hooves/wooden sandals) are 20th century. 3.4. Conclusion The basic Old English word-formation types affixation, compounding and zero-derivation have generally persisted to the present day. As in the case of inflection, Middle English takes a crucial position. Not only did it experience the influx of French prefixes and suffixes and the coming out of use of Old English ones, but it also laid the foundation for new patterns such as the zero-derivations cutpurse and blackout, new compound types and the shortening processes clipping and blending. The most important contribution of Early Modern English to word-formation is doubtless the combining form, which has gained high productivity during the following centuries showing an ever-increasing frequency in Present-Day English. Of Early Modern English origin is furthermore the zero-derivational pattern showoff, whereas acronymics was not en vogue until World War I. Since then, however, its productivity has been growing steadily, as is illustrated by numberless coinages.
1679 Bergström, Gustav A. (1906), On Blendings of Synonymous or Cognate Expressions in English: A Contribution to the Study of Contamination. Ph.D. dissertation. Lund: Ohlsson Berndt, Rolf (1956), Form und Funktion des Verbums im nördlichen Spätaltenglischen. Halle/S.: Niemeyer Brunner, Karl (31965), Altenglische Grammatik. Tübingen: Niemeyer [11942] BT (1976) ⫽ Bosworth, Joseph, An Anglo-Saxon Dictionary, ed. and enl. by T. Northcote Toller. London etc.: Oxford Univ. Press [repr., 11898] BTS (1973) ⫽ Bosworth, Joseph, An Anglo- Saxon Dictionary. Supplement by T. Northcote Toller with rev. and enl. addenda by Alistair Campbell. London etc.: Oxford Univ. Press [repr., 11921] Erlebach, Peter (1979), Die zusammengesetzten englischen Zunamen französischer Herkunft. Heidelberg: Winter (Anglistische Forschungen 137) Faiß, Klaus (1978), Verdunkelte Compounds im Englischen. Tübingen: Narr (Tübinger Beiträge zur Linguistik 104) Faiß, Klaus (1987), “Words, Words, Words: Compound or Suffixation in German and English”. In: Lörscher, Wolfgang & Schulze, Rainer (eds.), Perspectives on Language in Performance: Studies in Linguistics, Literary Criticism, and Language Teaching and Learning. To Honour Werner Hüllen on the Occasion of His Sixtieth Birthday, Vol. I. Tübingen: Narr, 110⫺132 Faiß, Klaus (1989), Englische Sprachgeschichte. Tübingen: Francke Faiß, Klaus (1992), English Historical Morphology and Word-Formation. Loss versus Enrichment. Trier: wvt (Fokus Bd. 8) Giurescu, Anca (1975), Les Mots Compose´s dans les Langues Romanes. The Hague, Paris: Mouton (Janua Linguarum Series Practica 228) Jönsjö, Jan (1979), Studies in Middle English Nicknames. I: Compounds. Lund: Gleerup (Lund Studies in Eng. 55) Kastovsky, Dieter (1968), Old English Deverbal Substantives Derived by Means of a Zero Morpheme. Ph.D. dissertation, Univ. Tübingen Kastovsky, Dieter (1971), “The Old English Suffix -ER(E)”. Anglia 89, 285⫺325
4.
References
Ayto, John (1988), The Longman Register of New Words. Harlow: Longman Barnhart, Robert K. & Steinmetz, Sol & Barnhart, Clarence L. (1990), Third Barnhart Dictionary of New English. n.p.: Wilson
Kastovsky, Dieter (1985), “Deverbal Nouns in Old and Modern English: From Stem-Formation to Word-Formation”. In: Fisiak, Jacek (ed.), Historical Semantics. Historical Word-Formation. Berlin etc.: Mouton, 221⫺261 Krahe, Hans & Meid, Wolfgang (1967), Germanische Sprachwissenschaft, Vol. III: Wortbil-
1680
XVIII. Morphologischer Wandel II: Fallstudien
dungslehre. Berlin: de Gruyter (Sammlung Göschen 1218/1218a/1218b)
englischen Nominalkomposition. Tübingen: Niemeyer (Buchreihe der Anglia 30)
Marchand, Hans (21969), The Categories and Types of Present-Day English Word-Formation: A Synchronic-Diachronic Approach. München: Beck [11960 Wiesbaden: Harrassowitz]
Scheler, Manfred (1982), Shakespeares Englisch: Eine sprachwissenschaftliche Einführung. Berlin: Schmidt (Grundlagen der Anglistik und Amerikanistik 12)
MED (1954 ff.) ⫽ Middle English Dictionary, Vol. 1 ff. Ann Arbor/MI: Univ. of Michigan Press [predominantly] [various editors]
Thun, Nils (1963), Reduplicative Words in English: A Study of Formations of the Types Tick-tick, Hurly-burly and Shilly-shally. Uppsala: Bloms
OED (21989) ⫽ The Oxford English Dictionary, Vol. 1⫺20, prepared by J. A. Simpson & E. S. C. Weiner. Oxford: Clarendon Press [11933]
Uhrström, Wilhelm (1918), Pickpocket, Turnkey, Wrap-Rascal, and Similar Formations in English: A Semasiological Study. Stockholm: Bergvall
Reaney, P. H. (21976), A Dictionary of British Surnames. London, Boston: Routledge & Kegan Paul [11958]
Wright, Joseph & Wright, Elizabeth Mary (1908), Old English Grammar. London etc.: Frowde, Oxford Univ. Press
Sauer, Hans (1992), Nominalkomposita im Frühmittelenglischen: Mit Ausblicken auf die Geschichte der
Klaus Faiß, Mainz (Germany)
156. Vom Althochdeutschen zum Neuhochdeutschen 1. 2. 3. 4. 5. 6. 7.
1.
Periodisierung der deutschen Sprachgeschichte Synchrone Heterogenität und historische Homogenität Forschungsgeschichte Entwicklungsgeschichtliche Grundzüge Flexion Wortbildung Zitierte Literatur
Periodisierung der deutschen Sprachgeschichte
Der theoretischen Vielfalt der Kriterien für eine sprachgeschichtliche Periodisierung entspricht die tatsächliche Vielfalt der Periodisierungsversuche der deutschen Sprachgeschichte (vgl. die Zusammenstellung bei Roelcke 1995; 2001). Oberhalb aller Feindifferenzierungen gilt seit Wilhelm Scherer vor allem eine Unterteilung in die vier Sprachstufen Althochdeutsch (Ahd.), Mittelhochdeutsch (Mhd.), Frühneuhochdeutsch (Frnhd.) und Neuhochdeutsch (Nhd.) (vgl. Scherer 2 1878: 13⫺15). Soweit die Unterteilung erststellig durch innersprachliche Kriterien begründet wird, handelt es sich dabei zumeist um Kriterien phonetisch-phonologischer Art; in nur einem Fall wird eine (flexions)morpho-
logische Entwicklung als für die Periodisierung wesentlich benannt: die den Übergang des Frühneuhochdeutschen in das Neuhochdeutsche markierende Ausgleichung des präteritalen Numerusablauts (Ablautausgleich) bei den starken Verben (ahd./mhd./frnhd. fand ⫺ funden > frnhd./nhd. fand ⫺ fanden; vgl. Behaghel 51928: 149; Moser 1929: 1). Diese primäre Orientierung an phonologischen Prozessen zeigt die weiterwirkende junggrammatische Bindung der Sprachgeschichtsschreibung: Aufgrund vorgängig physiologisch-lautgesetzlicher Veränderungen werden auch grammatische Strukturen zerstört; erst in der Folge werden die defekten Strukturen aufgrund von Assoziationen der Sprecher (als Analogie) wieder in eine sinnvolle, neue Ordnung gebracht (vgl. Kohrt 1984: 392; Werner 21998: 575). Damit ist eine erste Ursache auch des morphologischen Wandels primär physiologisch-lautgesetzlich erklärt, die im Bereich der Morphologie wirksame Analogie bekommt “eine nachgeordnete komplementäre Erklärungsfunktion” (Putschke 21998: 476). Es folgt somit der junggrammatischen Logik, die wesentlichen Periodengrenzen über die als Ursache der grammatischen Veränderungen gedachten Lautveränderungen zu markieren.
1681
156. Vom Althochdeutschen zum Neuhochdeutschen Im Althochdeutschen (frühes 7. Jahrhundert bis Mitte 11. Jahrhundert) beginnt die geschichtliche und d. h. schriftsprachliche Zeit des Deutschen (vgl. Wolf 1981: 66). Da als sprachgeschichtlich wesentliche Entwicklung die “Transponierung und Bewältigung der Latinität in die Volkssprache” gilt, meint “Althochdeutsch” in erster Linie “die Sprachverwendung und erst in zweiter Linie das Sprachsystem” (Wolf 1981: 72). Als innersprachliches Abgrenzungskriterium gegenüber dem chronologisch vorausliegenden Südgermanischen gilt die 2. oder althochdeutsche Lautverschiebung, bei der u. a. aus germanischen Tenues neue Spiranten/ Affrikaten entstanden. Da die Lautverschiebung nicht alle (süd)germanischen Stammesdialekte des späteren deutschen Sprachraumes ergriffen hat (Ausnahme besonders das Sächsische), liefert sie zugleich auch das bis in die Gegenwart hinein gültige Kriterium zur dialektgeographischen Abgrenzung eines hochdeutschen (hd.) und eines niederdeutschen (nd.) Sprachraums. Als wesentliches Abgrenzungskriterium des Mittelhochdeutschen (Mitte 11. bis Mitte 14. Jahrhundert) gegenüber dem vorausliegenden Althochdeutschen gilt die in einheitlicher Graphie -e- aufscheinende Abschwächung der vollen Endsilbenvokale (ahd. ich bitt-u > mhd. ich bitt-e, ahd. er bitt-it > mhd. er bitt-et), daneben auch die Durchführung des Sekundärumlauts (ahd. wahsit > mhd. wähset; vgl. Wolf 1981: 64); die jüngere Grenze zum Frühneuhochdeutschen hin markieren weitere Lautwandel, so besonders die sogenannte (früh)neuhochdeutsche Diphthongierung/Monophthongierung (z. B. mhd. wıˆn > nhd. Wein; mhd. bruoder > nhd. Bruder). Zum ersten Mal in der deutschen Sprachgeschichte wird im schriftsprachlichen Bereich eine funktional (literarisch) wie diastratisch (höfisch-ritterlich) bestimmte Vorbild-Varietät herausgebildet. Diese als höfisches ⫺ klassisches ⫺ und auch als Normalmittelhochdeutsch bezeichnete Sprachform (vgl. Sonderegger 1979: 167) stellt jedoch nur einen regional wie funktional begrenzten Ausschnitt der Sprachwirklichkeit des Zeitraumes dar, so daß in neuerer Forschung der Sprachbegriff Mittelhochdeutsch teilweise bewußt vermieden wird (vgl. Wells 1990: 106). Im Frühneuhochdeutschen (Mitte 14. bis Übergang 18. Jahrhundert; vgl. besonders Reichmann 1988) tritt die Varietätenvielfalt des Deutschen erstmals deutlich in die Überlieferung ein. Die schriftdialektale Bindung und d. h. eine in unterschiedlicher Ausprägung bestehende landschaftliche Variation auch in der Schriftlichkeit wird zum charakteristischen Merkmal der gesamten Periode (vgl. Penzl 1984: 12 f.). Im Neben- und Miteinander der verschiedenen Varietäten findet im Verlauf des 14. bis 18. Jahrhunderts auf der Ebene der Schriftlichkeit ein in Stufen verlaufender Aus- und Angleichungsprozeß statt: Klar bestimmbare sonderlandschaftliche Schreibsysteme werden in eine überregionale und oberhalb der gesprochenen Mundarten liegende Schriftsprache überführt; dabei sind im Verlauf des 15./16. Jahrhunderts mindestens 12
verschiedene Schreib-/Kanzlei- und Druckersprachen unterscheidbar (vgl. Besch 1980: 590⫺592). Diese Schriftsprache wird zur Grundlage der in der Folge einsetzenden Kodifizierung und Normierung: Das Frühneuhochdeutsche wird zur entscheidenden Phase für die Herausbildung der neuhochdeutschen Schriftsprache. Da insbesondere im 16. Jahrhundert die eigenständige niederdeutsche Schriftsprache des norddeutschen Raumes durch die hochdeutsche Schriftlichkeit verdrängt wurde (vgl. Sodmann 22000), ist damit erstmals in der deutschen Sprachgeschichte eine oberhalb aller dialektalen Differenzierungen liegende schriftsprachliche, überdachende Varietät des gesamten deutschsprachigen Raumes herausgebildet: Die die historische Einzelsprache Deutsch ausmachende Existenzform eines differenzierten Diasystems ist in seinem Grundgerüst herausgebildet.
2.
Synchrone Heterogenität und historische Homogenität
Die sprachgeschichtliche Skizze zeigt, daß eine geschichtliche Morphologie des Deutschen bereits hinsichtlich des Sprachbegriffs ambivalent ist: So meint “Deutsch” eine konkrete Sprachform, die innerhalb eines historisch-sozial bestimmten Kommunikationsraumes eine allgemeine Geltung besitzt, konkret die (neu)hochdeutsche Schrift- oder Hochsprache; diese Sprachform ist erst seit dem 18. Jahrhundert entwickelt. Zugleich eint “Deutsch” eine abstrakte und inner- wie außerlinguistisch bestimmte Zusammengehörigkeit einer Reihe konkret vorhandener Sprachformen innerhalb des Gültigkeitsbereichs der Standardsprache (z. B. Dialekte), die eine jede nur über eine eingeschränkte Geltung verfügt. Innerhalb der unterscheidbaren Varietäten/Subsysteme ist die Standardform (das “Hochdeutsche”) selbst nur eine Varietät neben anderen. Ihre Bedeutung liegt jedoch u. a. darin, daß erst ihre (geschichtliche) Herausbildung und Stabilisierung die zu ihr komplementäre Bestimmung der Varietäten ermöglicht. Der Zustand diasystemischer Gegliedertheit sowie das konkrete Varietätengefüge sind Ergebnis einer insbesondere im Frühneuhochdeutschen vollzogenen Entwicklung, die in besonderer Weise mit dem neuzeitlichen Prozeß der Herausbildung nationalstaatlicher Verkehrs- und Kommunikationsräume verbunden ist. Dabei zeigt sich, daß die mundartlichen Grundlagen der überdachenden Standardvarietät allein innerhalb des südlichen, hochdeutschen Sprachraums liegen (Teilhabe an der 2. oder ahd. Lautverschiebung), daher die in den Epochen-
1682 bezeichnungen (z. B. “Althochdeutsch”) vorliegende Eingrenzung. Somit ist der Begriff “Hochdeutsch” innerhalb des Varietätensystems ebenfalls doppeldeutig: normativ im Sinne der Normvarietät, historisch und sprachgeographisch im Sinne der durch die Lautverschiebung definierten Sprachstufen und Mundarten des Deutschen (vgl. Sonderegger 1979: 166). Gemäß der Varietätendifferenzierung stellt auch das historisch entwikkelte System der gegenwärtigen standardsprachlichen Flexions- und Wortbildungsmorphologie nur eines unter anderen deutschen Systemen dar. Dabei hat die Morphologie jeder der unterscheidbaren Varietäten ebenso ihre eigene Geschichte wie auch die besondere diasystemische Gegliedertheit der deutschen Morphologie insgesamt. Damit ist das grundlegende Problem der diachronen Betrachtung des Deutschen angesprochen: Es geht um ihre Identität in Geschichte und Raum (vgl. Mattheier 21998: 825). Tatsächlich kann von einem “Wandel” sprachlicher Einheiten/Strukturen nur innerhalb eines als identisch zu bestimmenden funktionalen, sozialen und regionalen Gültigkeitsbereichs gesprochen werden. Das System der Flexionswie Wortbildungsmorphologie der im späten 18. Jahrhundert schon weitgehend erreichten “Normalvarietät” des Neuhochdeutschen wäre somit zu vergleichen mit dem System entsprechender Varietäten der vorausliegenden Sprachstufen: Eine “Normalvarietät” ist aber weder für das Althochdeutsche, noch für das Mittelhochdeutsche oder das Frühneuhochdeutsche anzunehmen; entsprechende Paradigmen der historischen Grammatiken, auf denen der sprachgeschichtliche Überblick gründen muß, haben ausschließlich heuristischen Wert. Die relativierende Formulierung erweist sich besonders bezüglich der geschichtlichen Genese der neuhochdeutschen Standardvarietät als bedeutsam. Denn im Unterschied zu anderen europäischen Kultursprachen weist das Deutsche nur begrenzt eine historische Homogenität auf: Im deutschen Reich fehlte ein klar bestimmbares Zentrum, dessen Varietät sich im sprachgeschichtlichen Prozeß zur allgemeinen und d. h. Standardvarietät hätte qualifizieren können. Somit ist auch das im Neuhochdeutschen entwickelte morphologische System in wesentlichen Teilen nurmehr ein Produkt der Überlagerung und eklektizistischen Addition verschiedener und varietätendifferenziert herausgebildeter Teilsysteme/-strukturen/-prinzipien. Auch und gerade außersprachliche Ur-
XVIII. Morphologischer Wandel II: Fallstudien
sachen haben in der Sprachgeschichte des Deutschen eine wesentliche Rolle gespielt (zur strukturellen Disposition sowie den weiteren Prinzipien des im Frühneuhochdeutschen erfolgten Sprachausgleichs vgl. Besch 2 2003). Die Berücksichtigung der Varietätendifferenzierung hat Folgen für die sprachgeschichtliche Argumentation: Die gemeinhin als “Wandel” oder “Fortentwicklung des Systems” klassifizierte Veränderung z. B. mhd. gibe > nhd. gebe (1. Sg. Ind. Präs.) ergibt sich nur aus dem unmittelbaren Vergleich der neuhochdeutschen Standardvarietät mit der in den entsprechenden Grammatiken ausgewiesenen mittelhochdeutschen “Normalform”. Tatsächlich kann die Entwicklung adäquater beschrieben werden als Perpetuierung und anschließende Verallgemeinerung einer schon in mittelhochdeutscher Zeit neben gibe im mitteldeutschen Sprachraum parallel existierenden Form gebe. Damit ist gibe nicht durch einen “Wandel” verändert, sondern aufgrund einer sprachlandschaftlich veränderten Fokussierung nurmehr innerhalb der Schriftlichkeit verdrängt; die Form gibe ist in oberdeutschen Mundarten unverändert bis in die Gegenwart hinein gebraucht.
Die deutsche Sprachgeschichte und damit auch die Geschichte des morphologischen Systems erweist sich zu einem nicht unbedeutenden Teil als die chronologisch aufeinanderfolgende Fokussierung mal der einen, mal der anderen Varietät; insofern liegt nicht zwingend eine Entwicklung des Systems, sondern die unterschiedliche Wirksamkeit varietärer Ausprägungen vor.
3.
Forschungsgeschichte
Der wesentliche Aspekt der Forschungsgeschichte liegt in der Ausdifferenzierung der Wortlehre (i. e. Morphologie) in eine jeweils eigenständige Flexions- und Wortbildungslehre (Solms 1998). Für die Frühphase einer Grammatikographie des Deutschen, die der geschichtlichen Herausbildung einer Standardvarietät im Übergang des Frühneuhochdeutschen zum Neuhochdeutschen parallel geht (im 17. Jahrhundert insbesondere Justus Georg Schottel, im 18. Jahrhundert besonders Johann Christoph Gottsched und Johann Christoph Adelung), gilt eine noch weitgehend undifferenzierte Thematisierung der Wortlehre. Eine solche Differenzierung gelang erst durch eine adäquate Berücksichtigung der Funktion und d. h. der durch Flexion und Wortbildung jeweils realisierten spezifischen “Bedeutung”: Während die Fle-
156. Vom Althochdeutschen zum Neuhochdeutschen
xionslehre “es der Syntax überlässt, die Bedeutung und Verwendung” der Flexive zu bestimmen, ist jene der Wortbildungsmittel allein durch eine eigene, die Wortbildungslehre zugleich unmittelbar konstituierende “Funktionslehre” zu beschreiben (Paul 1981: 17 f.). Den Weg zu einer solchen Unterscheidung wies zuerst Jacob Grimm, indem er die Differenzierung von Wortbildung und Flexion als eine im Fortgang der sprachgeschichtlichen Entwicklung erst herausgebildete erkannte: Grimm erkannte innere Wortbildung (trinken, der Trank) und Wurzelflexion (trinken ⫺ trank) als ursprünglich ungeschieden. Im darin wirksamen Ablaut sah Grimm das wesentliche und d. h. historisch wie genetisch grundlegende Form- und Bauprinzip der deutschen Sprache: er erkannte es als ihr ältestes Prinzip, wobei die Wurzeln aller Wörter mit dem Grundsatz des Ablaut sowie “dieser mit der natur der wurzeln wesentlich in gemeinschaft” stehen (Grimm 1826: 4⫺7). So wenig somit innere Wortbildung und Wurzelflexion zu unterscheiden sind, so wenig kann man “die flexion, d. h. die dem verbum anwachsende pronominalform, die dem nomen anwachsenden geschlechtszeichen und partikeln strenge nicht vom begriffe der wortbildung ausschließen” (Grimm 1826: 86). Erst in der geschichtlichen Entwicklung werden die formalen Ausdrucksmittel und -muster funktional profiliert und als Kategorie faßbar, so z. B. die von Grimm so bezeichneten “eigentlichen” Komposita (z. B. Berg-werk ohne Fugenelement versus Krieg-s-geschrei mit Fugenelement): “Zur charakteristischen Compositionsform wurden diese altertümlichen Gefüge [erst] durch die Ausbildung der Flexion. Denn eben dadurch, dass man sich gewöhnte, die Beziehungen zwischen Wörtern durch Flexionen zu bezeichnen, erschienen die Verbindungen mit der reinen Stammform als Composita.” (Wilmanns 1906⫺1909: 4)
Eine konsequente Berücksichtigung des damit formulierten relationalen Charakters der Unterscheidung von Wortbildung und Flexion und ihrer historischen Entwicklung ist bisher nicht geleistet. Für Grimm gehört die Morphologie neben der Lautlehre zum Kern von Sprachvergleichung wie Sprachgeschichte. Diese Gegenstandskonstitution bleibt auch im späteren junggrammatischen Sprachwandelkonzept bestehen (vgl. Eichinger 1984: 436). Indem nun jedoch die bei Grimm vorherrschende Überzeugung, Sprachgeschichte erweise in ihren historischen Stu-
1683 fen/Epochen den fortschreitenden “Verfall” der ursprünglichen Klarheit und Systematizität, überwunden wird, werden nun durch den junggrammatischen Ansatz eines zu allen Zeiten “nach denselben strengen phonetischen und psychologischen Regeln” (Werner 2 1998: 575) erfolgenden Sprachwandels alle Sprachstufen als prinzipiell gleichwertig erkannt; auf der Grundlage eines jeweils reichen empirischen Materials eröffnet dies grundsätzlich die Möglichkeit zum Systemansatz und d. h. für eine Beschreibung der jeweiligen Sprachstufen als eigenständiger und hinsichtlich übereinstimmender Merkmale homogener Synchronien (vgl. Eichinger 1984: 436). Der junggrammatische Ansatz führt zu einer Reihe historischer Grammatiken (nicht nur zum Deutschen), die bis heute ihren Wert als Standardwerke nicht eingebüßt haben (vgl. zusammenfassend Werner 1998: 575 f.); einige Grammatiken beschreiben den Gesamtzusammenhang der geschichtlichen Entwicklung des deutschen Flexionssystems vom vorgermanischen Ausgangspunkt bis in das Neuhochdeutsche hinein (Wilmanns 1906⫺ 1909; Paul 1916⫺1920), andere bleiben auf jeweils eine der definierten Sprachstufen begrenzt (Braune 141987; Weinhold 21883; Paul 24 1998; Michels 51979). Mit Ausnahme der periodenübergreifenden Darstellungen gerät die Wortbildung nicht in den Blick, unter der Formenlehre wird nun ausschließlich Flexionslehre betrieben. In eben dieser Begrenzung hatte auch Virgil Moser (1929: XII) seine “frühneuhochdeutsche Formenlehre” geplant; die Gleichsetzung von Morphologie und Flexion gilt auch für neuere Darstellungen als eine gegebene (vgl. Kern & Zutt 1977: 1; vgl. auch Kienle 21969 und Russ 1978). Der Schwerpunkt der epochenbezogenen Grammatiken liegt in der Dokumentation der jeweiligen Zustände; auf systematische Umstrukturierungen, auf mögliche Ursachen oder funktionale Verschiebungen wird kaum eingegangen. Damit geht für das Altund Mittelhochdeutsche eine mehr oder weniger konsequente Konzentration auf ein hypostasiertes “Normalsystem” und damit eine nur ansatzweise Berücksichtigung u. a. der diatopischen Varietäten einher. In der jüngeren und einem nicht mehr junggrammatischen, sondern strukturalistischem Ansatz verpflichteten Frühneuhochdeutschen Grammatik ist der Versuch einer Korpusgrammatik unternommen, bei der die Varietätendifferenzierung zum Prinzip der Darstellung geworden ist; ein Normalsystem wird nicht
1684
XVIII. Morphologischer Wandel II: Fallstudien
mehr konstruiert. Eine nur annähernd entsprechende Darstellung der Wortbildung, die zudem die funktionalen Verschiebungen innerhalb des Inventars der Wortbildungsmittel erfaßt, steht jedoch auch für das Frühneuhochdeutsche noch aus. Hier muß man bis auf weiteres auf schon ältere und zumeist etymologisch orientierte Darstellungen von Wilmanns (1906⫺1909), Paul (51920), Kluge (21925) und Henzen (31965) zurückgreifen. Aus der Vielzahl einzelner Untersuchungen und epochenbezogener Überblicksdarstellungen sei zusätzlich zu den historischen Grammatiken verwiesen auf die einschlägigen Beiträge in Besch et al. (21998; 22000, Hrsg.), Schwarz (1986), Bergmann et al. (1987, Hrsg.), Müller (1993), Habermann (1994), Winkler (1995), Joeres (1995), Prell & Schebben-Schmidt (1996); zur bibliographischen Erschließung vgl. Ronneberger-Sibold (1989).
4.
Entwicklungsgeschichtliche Grundzüge
Der geschichtliche Zeit- und Varietätenraum des Deutschen zeigt eine Vielzahl morphologischer Einzelveränderungen: Einige davon profilieren wortartübergreifende und die Wortbildungs- und/oder Flexionsmorphologie insgesamt prägende Prozesse; andere prägen nurmehr die Flexion/Wortbildung einer Wortart; wiederum andere zeigen sich als wesentlich in der Profilierung klarer(er) subsystemischer Flexionsstrukturen innerhalb einer Wortart, z. B. nur der sogenannten starken Verben oder nur der sogenannten schwachen Substantive; andere schließlich fügen sich nur schwer in einen überordnenden Zusammenhang und bleiben als Einzelveränderung singulär. Oberhalb der Vielzahl einzelner Veränderungen ordnet die sprachgeschichtliche Perspektive die Einzelerscheinungen zu Entwicklungstendenzen, die sich als historisch konstante (in allen Sprachstufen und bis in die Gegenwart wirkende) oder als inkonstante (historisch nur partiell aufscheinende) zeigen (vgl. Sonderegger 1979: 241⫺255, 340⫺345). Vom neuhochdeutschen Ausgangspunkt geurteilt werden die Einzelereignisse zu Indizien einer auf das Neuhochdeutsche zielenden Teleologie, sie werden entsprechend als “entwicklungsgeschichtliche Grundzüge” bestimmt (Hotzenköcherle 1962). Solche erweisen sich im Ab- sowie dem paradigmatisch verbleibenden Aus- und Umbau des flexivischen Materials:
(a) Als eine konstante und in allen Sprachstufen beobachtbare Entwicklungstendenz erweist die Flexionsgeschichte eine “kontinuierliche Reduktion und Umstrukturierung der Flexionsmorpheme” (Sonderegger 1979: 243); die Umstrukturierung erfolgt paradigmatisch sowie durch Ausweichen auf die syntagmatische Ebene. Indem eine Konzentration der flexivischen Ausdrucksmittel auf bestimmte morphosyntaktische Kategorien erfolgt, werden diese flexivisch profiliert. Wortartbezogen ergibt sich somit bei den Verben eine “Profilierung der Tempuskategorie” (Besch 1980: 594; vgl. Hotzenköcherle 1962: 331), eine “Numeruskristallisierung beim Substantiv” (Hotzenköcherle 1962: 331), eine “Genusprofilierung beim Adjektiv” (Solms & Wegera 1991: 317 f.). Bei der Flexion der Adjektive zeigt sich zudem die Durchsetzung einer syntaktisch vom bezogenen Substantiv und vom sonstigen Substantivbegleiter aus organisierten “Formregel” (Hotzenköcherle 1968: 3). (b) Strukturelle Konsequenz einer flexivischen Profilierung ist eine entsprechende Nivellierung bestimmter Kategorien: insbesondere die des Kasus bei den Substantiven (vgl. Sonderegger 1979: 248 f.), insbesondere die des Modus bei den Verben. (c) Eine gänzliche oder teilweise Ersetzung von Formativen durch markanter ausgezeichnete Formen tritt bei “funktionaler Untauglichkeit” ein (vgl. Wegera & Prell 22000: 1594). Eine solche ist primär phonetisch-phonologisch verursacht: vom Althochdeutschen zum Mittelhochdeutschen die in einheitlich -e- aufscheinende Abschwächung der vollen Nebensilbenvokale, vom Mittelhochdeutschen zum Frühneuhochdeutschen eine Fortsetzung der Nebensilbenabschwächung in der Apokope. So geht z. B. in der Wortbildung die althochdeutsche Möglichkeit zur ausdrucksseitig eindeutigen Markierung der Nomina-agentis-Bildung durch das Derivationssuffix -o (z. B. kebo, ‘der, der gibt’) verloren, es wird durch das lat. Lehnsuffix -ari (nhd. -er) ersetzt; so wird in der Flexion die vokalisch volle Endungsflexion des Althochdeutschen zerstört, Oppositionen werden aufgehoben (z. B. ahd. Nom. Sg. hirti, Nom. Pl. hirta > mhd. Sg./Pl. hirte). Dadurch kommt es zum Mittelhochdeutschen hin zu eingreifenden morphemischen Strukturveränderungen (vgl. Wolf 1981: 202), die sich in einer paradigmatischen Umstrukturierung des Systems oder auch einer verstärkten Nutzung entsprechender Periphrasen zeigen. Die über das Mittelhochdeutsche hinaus weitergehende Apokope erzeugt bei den Substantiven schließlich den gänzlichen Zusammenbruch des alten Flexionssystems, das im Frühneuhochdeutschen neu aufgebaut wird. (d) Die Funktionsverlagerung in die Periphrase (Wandel vom synthetischen zum analytischen
156. Vom Althochdeutschen zum Neuhochdeutschen Prinzip) erscheint in Flexion wie Wortbildung. So werden die semantisch-syntaktischen Wortbildungsfunktionen der germanischen Ableitungssuffixe schwacher Verben sowohl durch die Präfigierung als auch durch entsprechende Periphrasen übernommen: Es erscheinen z. B. statt der im Germanischen möglichen Wortbildung mittels eines Suffixes *-jan, das ein kausatives Verb deriviert (z. B. gotisch diupjan ‘vertiefen’ aus Adjektiv diups ‘tief’ ⫹ Wortbildungssuffix -jan), seit dem Althochdeutschen sowohl Präfixbildungen z. B. mit ir- (irtrockanen ‘trocken machen’) als auch Periphrasen des Basiswortes mit tuon oder machen. Als dann weitergehende Tendenz zur ‘Periphrase in der Morphemstruktur’ zeigt sich die Funktionsablösung von Simplicia durch Präfixverben seit dem Mittelhochdeutschen (z. B. jmdn. weinen > jmdn. beweinen). Bei der Flexion weist die Tendenz zu periphrastischen Umschreibungen auf die kontinuierliche Veränderung innerhalb der im Deutschen genutzten drei Flexionsprinzipien: die aus dem Vordeutschen übernommene “regressive” (von der Endung her erfolgende Steuerung mittels Relationsmorphen, z. B. 3. Sg. Ind. Präs. suoch-it vs. 3. Sg. Konj. suoch-e im Althochdeutschen), die ebenfalls vordeutsche Stamm- oder Wurzelflexion (untrennbarer Einheiten lexikalischer und grammatischer Bedeutung, besonders der Vokalwechsel “Umlaut” und “Ablaut”; Ablaut als morpholexische Erscheinung, vgl. Durrell 1980: 23; ähnlich Simmler 1987: 121 f.; z. B. 1. Pl. Präs. geben vs. 1. Pl. Prät. gaˆben im Mittelhochdeutschen), die “progressive” Steuerung durch vorangestellten Begleiter (z. B. 1. Pl. Ind. Prät. wir gaˆbun vs. 3. Pl. Ind. Prät. sie gaˆbun im Althochdeutschen). Ein jedes dieser Prinzipien erscheint nur selten ausschließlich, zumeist kommen sie in Kombination miteinander vor (z. B. ahd. Nom. Sg. gast vs. Pl. gesti; nhd. Nom./Akk. Sg. das Rad vs. Pl. die Räder). Die deutsche Sprachgeschichte zeigt eine jeweils unterschiedliche und die einzelnen Sprachstufen charakterisierende Nutzung dieser genannten drei Prinzipien: vorwiegend “regressive” oder Endungsflexion im Althochdeutschen, vermehrter Ausbau von Wurzelflexion und progressiver Steuerung im Mittelhochdeutschen, entscheidende Verlagerung im Neuhochdeutschen zur progressiven Steuerung (vgl. Sonderegger 1979: 243). Periphrasen werden vor allem bei den Verben für Perfekt, Futur, Passiv und auch Konjunktiv genutzt; sie finden sich besonders seit dem 15. Jahrhundert mit einem Schwerpunkt im oberdeutschen Sprachraum. (e) Der tendenzielle Ausbau der Wurzelflexion mittels Wechsel des Wurzelvokals zeigt die Morphologisierung und dann auch Systematisierung phonetisch-phonologisch bedingter und schon im Voralthochdeutschen eingetretener Alternationen (i. e. regelmäßiger und syn-
1685 chronisch funktionierender Lautwechsel in etymologisch verwandten Wörtern, hier besonders Umlaut) bei den Verben, den Substantiven sowie auch den Adjektiven (Komparation). Der Vokalwechsel wird besonders zum Mittel der Flexionsmorphologie entwickelt, weniger zu einem der Wortbildung. (f) Die phonetisch-phonologische Verursachung überlagernd, verstärkend und ersetzend sind weitere Entwicklungsmotive abzuleiten: so ein Streben nach Ökonomie, Redundanzsteuerung (vgl. Koenraads 1953; Lüdtke 1988: 1633). Dazu gehört, daß isolierte/lexemspezifische Besonderheiten innerhalb einer Gruppe/Klasse von Lexemen oder einer Wortart (nicht zwingend) abgebaut werden: der grammatische Wechsel bei der Flexion der Verben (z. B. Sg./ Pl. Prät. was ⫺ waˆren > war ⫺ waren; vgl. Sonderegger 1979: 253), suppletive Formen bei den Pronomen (dirre > dieser; vgl. Ebert et al. 1993: § M60), die rückumlautende Verbflexion (stellen ⫺ stallte > stellen ⫺ stellte), die starke Verbflexion sowie ein Teil der ehemaligen Präteritopräsentien. Tendenzielles Ergebnis eines Abbaus von “Besonderheiten” ist die weitere Profilierung und Bestätigung einer “Normalität”. Ein markantes Beispiel für den Abbau systemischer und lexemdistribuierter Varianz ist die Herausbildung der ehedem schwachen und somit spezifischen Verbalflexion zur allgemeinen und somit “Normalflexion” des Verbs im Deutschen. (g) In der Profilierung spezifischer Flexionsweisen erweist sich auch der besondere Einfluß, den die Erweiterung des Wortschatzes durch Entlehnung oder Wortbildung auf die Entwicklung des (geschlossenen) Flexionssystems hat. Denn jedem neu in den Wortschatz aufgenommenen Wort wird unmittelbar eine der möglichen Flexionsweisen zugeordnet. In diesem konkreten Sprachverhalten erweist sich die Produktivität (Analogie) bestimmter Muster (Art. 148). Insofern solche Muster somit durch zusätzliche Wortschatzelemente gestützt werden, kann der Sog dieser Muster so weit verstärkt werden, daß variante Muster ihre Wirkung verlieren. Die ehemals schwachen Verben sind auch hier ein Beispiel; mit ganz wenigen Ausnahmen (z. B. schreiben im Althochdeutschen oder preisen im Mittelhochdeutschen) sind alle entlehnten Verben automatisch als schwache Verben in den Wortschatz integriert worden; auch dadurch ist der Wortschatzanteil starker Verben relativ gesunken. (h) Eine “Tendenz zur formalen Trennung von Flexions- und Derivationsmorphemen” (Wegera & Prell 22000: 1594) und damit auch eine Tendenz zur Trennung von Flexion und Wortbildung zeigt sich sowohl hinsichtlich bestimmter Prinzipien morphologischer Informationsbindung als auch hinsichtlich der funktionalen Trennung der Relationsmorphe. Der in der vordeutschen Wortbildung wesent-
1686
XVIII. Morphologischer Wandel II: Fallstudien
liche Vokalwechsel (z. B. Ablautbildungen Trunk oder Trank zu trinken, Fluss oder Flotte zu fließen) wird seit dem Althochdeutschen primär flexivisch genutzt (Ausbau der Umlautnutzung bei den Substantiven). Bis zum Neuhochdeutschen hin verläuft ein Prozeß der deutlichen Trennung einzelner Flexions- und Wortbildungsmorphe (vgl. Erben 21983: 138): z. B. ist das in der Wortbildung ehedem produktive und inzwischen nurmehr aktive Wortbildungssuffix -e (Femininabstrakta z. B. Rinne) nurmehr als Flexiv produktiv (dient insbesondere der Markierung des Plurals bei den Substantiven); z. B. hat sich das Präfix ge-, das bis in das Mittelhochdeutsche hinein aspektuelle Differenzierungen ermöglichte (z. B. 3. Sg. Ind. Präs. er geas mit perfektiver Bedeutung ‘er hatte gegessen’) und daneben als ein die Verbsemantik modifizierendes Präfix (“Aktionsartenausfächerung”) funktionierte (z. B. Präfixverb geheilen ‘gesund werden’ zum Simplexverb heilen ‘gesund machen’), im Verlauf des Frühneuhochdeutschen nurmehr zum Bestandteil des diskontinuierlichen Part. Prät.-Morphems herausgebildet (vgl. Solms 1991). (i) In der morphologischen Profilierung und funktionalen Bindung von Flexiven und Formativen zeigt sich eine “Tendenz zur deutlichen strukturellen Trennung der Wortarten” (Wegera & Prell 22000: 1594), “unverkennbar” zwischen Substantiv und Adjektiv (Erben 2 1983: 138).
5.
Flexion
5.1. Flexion der Verben Die diachron motivierte Klassifizierung der deutschen Verben unterscheidet neben den schon im Germanischen ausdifferenzierten Hauptklassen der “starken” und “schwachen” Verben eine weitere Gruppe flexivisch heterogener besonderer Verben. Kriterium der Unterscheidung der zwei Hauptklassen ist die ohne oder mit einem Dentalsuffix -t- erfolgende Flexion des Präteritums (schwach: ahd. suochen ⫺ suochta; stark: ahd. grıˆfan ⫺ graif/griffen). Daneben werden als ebenfalls eigenständige Gruppe die sogenannten “Präteritopräsentien” (Prät. Präs.; z. B. mhd. wizzen oder gunnen) unterschieden, ursprünglich entstanden aus dem indogermanischen Perfekt, der teils Vergangenheit und teils Gegenwart bezeichnete (aus letzteren Fällen entstehen Präteritopräsentien, aus ersteren “normale” starke Verben; vgl. Hirt 1932: § 115). Entsprechend zeigen die Präteritopräsentien im Präsens (ablautende) Präteritalformen, zu denen dann schwache Präteritalformen neu gebildet sind. In einer synchronen Klassifizierung gilt die Flexion der Präteritopräsentien schon seit dem Althochdeutschen als Besonderheit nur weniger Lexeme. Dies rechtfertigt ihre Gruppierung zusammen mit einer nur kleinen und hinsichtlich ihres Inventars nicht konstanten Restgruppe auffälliger und individuell
flektierter Verben: ahd. uuellen ‘wollen’, tuon ‘tun’, sıˆn ‘sein’, gaˆn/geˆn/gangan ‘gehen’ und staˆn/steˆn/ stantan ‘stehen’. “Restgruppe” wie “Präteritopräsentien” werden in den historischen Grammatiken zumeist in einer nicht für jede Sprachperiode identischen Gruppe “besonderer Verben” gefaßt (vgl. Braune 141987: § 370⫺385; Paul 241998: § 267⫺288): ihre Flexion ist als jeweils einzellexematische Kombination verschiedener Stamm- (teilweise suppletiv) wie auch Relationsmorphe zu beschreiben.
Die vom Alt- bis zum Ausgang des Frühneuhochdeutschen strukturell wirksamen Veränderungen zeigen sich teilweise als Fortführungen schon im Germanischen begonnener Tendenzen, zum Teil als später einsetzende Entwicklungen; die Veränderungen bleiben partiell subsystemisch, z. B. nur auf die starken Verben begrenzt, partiell führen sie in ihrer Gesamtheit zur Veränderung des gesamten verbalen Flexionssystems: zur Herausbildung der schwachen von einer germanisch besonderen zu der neuhochdeutsch regelmäßigen Normalflexion. In diesen Zusammenhang gehören Angleichungen innerhalb der schwachen Flexionsklassen, Angleichungen der Endungsflexivik der starken an die der schwachen Flexion, der Wechsel zahlreicher starker zu den schwachen Verben, die Übernahme der schwachen Flexionsweise auch bei einigen besonderen Verben sowie auch die automatische Übernahme der schwachen Flexionsweise bei neu in den Wortschatz aufgenommenen Verben. Die Entwicklung der schwachen zur neuhochdeutsch “regelmäßigen” oder “Normalflexion” bedeutet für den im Frühneuhochdeutschen vollzogenen Prozeß bei den starken Verben, daß es sie als eigenständige Klasse spätestens seit der Aufklärung nicht mehr gibt (vgl. Augst 1975: 263); sie haben als Ausnahmen zu gelten. Die noch alt-/mittelhochdeutsche Formenvielfalt der besonderen Verben wird im Frühneuhochdeutschen reduziert, in einigen Fällen wird auch die schwache Flexion durchgesetzt (bei mhd. tugen, gunnen, sollen), teilweise werden die Besonderheiten in anderer Weise aufgegeben (beginnen ist nurmehr stark). Besonders die Veränderungen bei den besonderen Verben sowie auch die Reduktion des “Grammatischen Wechsels” und stärkere Systematisierung des Ablauts bei den starken Verben zeigt eine (partielle) Aufhebung von Suppletion zugunsten geregelter Alternationen; hier liegt ein als Stamm- oder morphologisches Prinzip bezeichneter Prozeß zur einheitlichen Wurzel-/Stammstruktur in allen Flexionsformen vor.
156. Vom Althochdeutschen zum Neuhochdeutschen
Die Herausbildung einer schwachen Verbflexion ist eine das Germanische vom Indogermanischen typologisch unterscheidende Neuerung (vgl. Schmidt 21998: 995 f.). Das schwache (endungsflexivische) Prinzip wird zum Merkmal der zumeist derivierten Verben; das starke (wurzelflexivische) Prinzip des Ablauts (seltener auch der Reduplikation) erweist sich als das zunehmend abgebaute Prinzip des zumeist schon dem Indogermanischen angehörenden Grundbestandes an Verben. Die (Wort)Bildung der neuen und schwach flektierten Verben des Germanischen erfolgt suffixal: Die ausdrucksseitig unterschiedenen Stämme (durch die stammbildenden Suffixe *-ja, *-oˆ, *-eˆ, *-na) definieren im Germanischen vier Klassen schwacher Verben. Dabei ist jedoch schon für das Althochdeutsche fraglich, inwieweit dieses “Stammelement” noch “als Kennzeichen einer paradigmatischen Struktur” (Simmler 1987: 135) segmentierbar ist. Wenn für das Althochdeutsche trotzdem noch schwache Stammklassen (vgl. Sonderegger 1979; 22000; Braune 141987) unterschieden werden, dann zeigt sich das deliminative Kriterium (Übereinstimmung in der phonologischen Substanz des Auslauts, d. h. des ursprünglichen Stammelements in allen Wortformen) jedoch nur noch teilweise; die übliche Klasseneinteilung des Althochdeutschen ist somit vorwiegend historisch und nicht synchron motiviert. Es werden unterschieden: die -ien/-en-Verben (< germ. *-jan) als schwache Verben der Klasse I (z. B. deverbal-kausativ: se(z)zen zu si(z)zen), die -oˆn-Verben als Klasse II (z. B. denominal-instrumental/ornativ: salboˆn zu salba ‘Salbe’), die -eˆn-Verben (unter Einschluß der germanischen *-nan-Verben) als Klasse III (z. B. denominal-ingressiv/inchoativ: alteˆn ‘alt werden’). Dabei erweisen einzelne Oppositionen das noch teilweise Funktionieren der alten Stammklassenelemente (vgl. Wolf 1981: 200): ahd. roˆten ‘rot machen’ (< *roˆt-jan) neben roˆteˆn ‘rot werden’ (< *roteˆ-n); darüber hinaus zeigen sich die einzelnen Klassen als mehr oder weniger produktiv: Alle lateinischen Lehnwörter schließen sich der Klasse II an (vgl. Russ 1987: 336). Möglicherweise begründet hier die hohe Type-Anzahl dieser Verben ihre historische Produktivität: Von den ca. 1798 althochdeutsch schwachen Verben (vgl. Russ 1987: 336) gehören zu Klasse I 860 Lexeme, zu Klasse II 660 Lexeme, zu Klasse III nur 278 Lexeme. Formalmorphologisch scheinen die ursprünglichen Ableitungs- oder Stammsuffixe
1687 im Althochdeutschen eindeutig: -i-, -oˆ-, -eˆ gelten als jeweiliger Bindevokal (vgl. Braune 14 1987: § 306) zwischen der nur historisch/etymologisch zu fassenden Wurzel und dem jeweiligen Endungsflexiv. Diese Wertung enthält jedoch ein Problem, da -oˆ/-eˆ in allen Formen des jeweiligen Verbs erscheinen; sie erweisen sich als Bestandteile eines paradigmatisch nicht mehr in Wurzel oder Stamm (bzw. Stammelement oder Flexiv) segmentierbaren Morphs. Gegenüber dem Germanischen wird die Klassifizierung im Althochdeutschen verkompliziert, da aufgrund phonologisch zu fassender Veränderungen bei einem Teil der Verben aus Klasse I der “Bindevokal” (-i-) innerhalb des Präteritums und des flektierten Part. Prät. zum Althochdeutschen hin ausgefallen ist; als Folge daraus (nicht eingetretener Umlaut) hat sich der einige schwache Verben bis heute charakterisierende “Rückumlaut” (brennen vs. brannte) herausgebildet. Aus der (sekundären) Berücksichtigung dieser apophonischen Varianten (mit Vokalwechsel) sowie in erster Linie der historisch/etymologisch bestimmten Stammklassen und der ihnen entsprechenden Bindevokale (im Präteritum und flektierten Part. Prät.) wird für das Althochdeutsche ein Fünfklassensystem der schwachen Verben definiert (vgl. Sonderegger 1985: 1023), die sich jedoch endungsflexivisch, von wenigen Ausnahmen abgesehen, identisch verhalten: I.1: mit Vokalwechsel/ohne Bindevokal (brennen, branta); I.2: ohne Vokalwechsel/ohne Bindevokal (hoˆren, hoˆrta); I.3: ohne Vokalwechsel/mit Bindevokal -i- (leggen, legita); II.: ohne Vokalwechsel/mit Bindevokal -oˆ- (machoˆn, machoˆta); III: ohne Vokalwechsel/mit Bindevokal -eˆ- (tageˆn, tageˆta).
Die zum Mittelhochdeutschen eintretende Abschwächung der vollen Nebensilbenvokale verändert die Klassifizierung; das -e- der Nebensilbe wird als Bestandteil des Flexivs gewertet. Das Vorhandensein/Fehlen dieses Bindevokals im Präteritum wird zum zentralen Klassifizierungskriterium (vgl. Paul 24 1998: § 265): I.1: mit Vokalwechsel/ohne Bindevokal (brennen, branta); I.2: ohne Vokalwechsel/ohne Bindevokal (teilen, teilte); II: ohne Vokalwechsel/mit Bindevokal (legen, legete).
1688 Da ein ausdrucksseitiges Kriterium für den Gebrauch des -e- nicht vorliegt (mit Ausnahme der Kopplung an den Vokalwechsel), ist das mittelhochdeutsche Inventar beider Klassen nur noch partiell durch die althochdeutsche Zugehörigkeit bestimmt. Dies ergibt sich allein schon durch das immense Anwachsen des schwach flektierten Verbwortschatzes (durch Wortbildung, Entlehnung) auf mehr als 5000 Verben (vgl. Sonderegger 1979: 258). Doch auch das jeweils ererbte Flexionsverhalten der einzelnen Lexeme ist aufgrund mehrerer Ursachen nicht klassenstabil. Zwar wird als maßgeblich für das mittelhochdeutsche Vorhandensein des Bindevokals die Quantität der Silbe genannt (vgl. Hempen 1988: 254), doch wird unabhängig davon -e“vielfach” synkopiert (Paul 241998: § 265). Dies führt zur Vermischung der althochdeutsch getrennten Klassen I.2. und II. Auch das Inventar der apophonisch funktionierenden Lexeme stimmt aufgrund der Wirkungen des “Sekundärumlauts” sowie “anorganisch” eintretender Rückumlaute (z. B. keˆren ⫺ karte oder enden ⫺ ante) nicht mehr mit dem der althochdeutschen Gruppe überein (vgl. Paul 241998: § 259⫺266). Obwohl die Gruppe der Verben mit einer apophonischen Wurzelflexion im Mittelhochdeutschen stark anwächst (über 200), nimmt ihr relativer Anteil am Gesamt aller schwachen Verben stark ab. Dieser Prozeß wird im Frühneuhochdeutschen noch wesentlich verstärkt. Es verschmelzen während des Frühneuhochdeutschen auch die im Mittelhochdeutschen noch hinsichtlich des Bindevokals unterschiedenen Klassen. Die Entwicklung verläuft wiederum über eine Vermischung der historischen Gruppen, so z. B. teilete (mittelhochdeutsch ohne Bindevokal), aber legte (mittelhochdeutsch mit Bindevokal). So erweist sich schon zu Beginn des Frühneuhochdeutschen der Gebrauch des Nebensilben-e nurmehr deutlich phonologisch distribuiert. Trotz eines deutlichen Anstiegs in der Verwendung des -e- vor allem im Oberdeutschen des 16. Jahrhunderts ist die neuhochdeutsch gültige Regelung eines beibehaltenen -e- nur nach Dental in der zweiten Hälfte des 17. Jahrhunderts usuell (vgl. Ebert et al. 1993: § M90.2.); es gibt nurmehr eine Klasse schwacher Verben (mit nach Dental eintretendem Bindevokal -e-). Die (neuhochdeutschen) Ausnahmen innerhalb der schwachen Flexion bestimmen sich durch den als Rückumlaut bezeichneten Vokalwechsel z. B. brenn- (Inf./Präs./Part. Präs.) zu brann-
XVIII. Morphologischer Wandel II: Fallstudien
(Prät./Part. Prät.). Im ausgehenden Frühneuhochdeutschen funktionieren gemäß diesem Muster zusätzlich zu den neuhochdeutsch erhaltenen (denken und synchron analog auch bringen, brennen, nennen, rennen, senden, wenden; daneben einige semantisch isolierte und erstarrte Formen, z. B. abgeschmackt zu schmecken, behaftet zu heften, getrost zu trösten) auch noch z. B. legen, trennen, schenken, schätzen, setzen, zerren (vgl. Hoffmann & Solms 1987: 53 f.). Innerhalb des Ausgleichsprozesses zugunsten der stamminvarianten Flexion ist in nur wenigen Fällen ein Ausgleich auch zugunsten der (rückumlautenden) Prät./Part. Prät.-Form eingetreten (z. B. mhd. bestellen ⫺ bestalte > nhd. Inf. bestallen neben bestellen). Das Gesamtinventar der mittelhochdeutsch rückumlautenden Verben umfaßt noch zumindest 221 Lexeme (vgl. Weinhold 21883: § 383). Die Ausgleichung findet dann erst im Frühneuhochdeutschen statt, begünstigt durch die schon im älteren Mittelhochdeutschen und in der ersten Hälfte des Frühneuhochdeutschen gültige sprachlandschaftliche Differenzierung (vgl. Sta˚rck 1912: 308): brennen ⫺ brante ⫺ gebrant im Mitteldeutschen versus prennen ⫺ prant ⫺ geprennet im Oberdeutschen. Insbesondere vom Oberdeutschen ausgehend (“oberdeutscher Präteritumschwund”, d. h. Rückgang des Gebrauchs finiter Prät.-Formen und Zunahme der periphrastischen Tempusform) greift im Verlauf des Frühneuhochdeutschen der Ausgleich zugunsten der Form des Inf./Präs. (und oberdeutsch auch Part. Prät.). Die Entwicklung zeigt sich als Rückbau einer flexivischen “Übercharakterisierung” (durch Dentalsuffix und zusätzlichen Vokalwechsel) des Präteritums (Schirmunski 1962: 501) zugunsten einer als morphologisches Prinzip bezeichneten Stabilität des Stammes in allen Positionen des Paradigmas. Das im Althochdeutschen paradigmatisch wie regional noch differenzierte System der Flexionssuffixe erfährt bis zum ausgehenden Frühneuhochdeutschen eine Vereinfachung, die eine Tilgung teils paradigmatisch und teils auch regional isolierter Flexive darstellt. Die differenzierte Flexivik des Althochdeutschen bestimmt jedoch keine ausgeprägten klassenbezogenen und/oder morphosyntaktisch bezogenen Paradigmen: So liegen einzelne klassenbezogene Unterschiede z. B. bei der 1. Sg. Ind. Präs. vor (starke sowie schwache Verben der Klasse I mit -u, schwache Verben der Klassen II/III mit -m); klassenindifferente Unterschiede hinsichtlich der morphosyntak-
156. Vom Althochdeutschen zum Neuhochdeutschen
tischen Merkmale liegen z. B. bei der Modusdifferenzierung der 1. Sg. Präs. vor (Ind. -u/ -m zu Konj. -e); klassendifferente Unterschiede hinsichtlich der morphosyntaktischen Merkmale liegen z. B. vor bei der 2. Pl. Präs. (Variation -et/-ent bei starken sowie schwachen Verben der Klasse I, -t bei schwachen Verben der Klassen II/III). Dabei zeigt schon das Althochdeutsche klassen- wie merkmalsübergreifenden Ausgleich, so z. B. von der Konjunktivform schwacher Verben ausgehende Verallgemeinerung von -st in der 2. Sg. (beibehaltene Besonderheit: -i im Präteritum starker Verben). Die Entwicklung zum Mittelhochdeutschen ist einerseits geprägt durch den Zusammenfall der Nebensilbenvokale in -e (Aufhebung der ausdrucksseitigen Oppositionen z. B. 1. Sg. Ind. Präs. -u und 1. Sg. Konj. Präs -e); andererseits kommt es zur An- und Ausgleichung innerhalb der wenigen eine Variation aufweisenden morphosyntaktischen Merkmale (z. B. 1. Sg. Ind. Präs.): Soweit es sich um eine auch klassenindifferente Variation handelt (1. Pl. -meˆs, -en), ist die Variation schon zum Ende des Althochdeutschen aufgegeben; dort, wo die Variation zugleich eine Klassendifferenzierung berührt, verläuft die Entwicklung bis in das Frühneuhochdeutsche hinein (z. B. flexivische Besonderheit der 2. Sg. Ind. Prät. bei den starken Verben). Bis auf wenige Positionen (3. Pl. Ind. Präs. -ent zu 3. Pl. Konj. Präs. -en sowie bis in Neuhochdeutsche erhaltener Unterschied bei 1./3. Sg. Ind. Prät. starker/schwacher Verben: z. B. gab- versus sagt-e) erscheint die Endungsflexivik im Normalmittelhochdeutschen sowohl zwischen starken und schwachen Verben als auch hinsichtlich Präs./ Prät. oder Ind./Konj. weitgehend ununterscheidbar (vgl. Paul 241998: § 239, 257). Eine notwendige Berücksichtigung der Varietäten in mittelhochdeutscher Zeit erweist die Verhältnisse jedoch als nicht ganz so eindeutig (z. B. erscheint Flexiv-n der 1. Sg. Ind. Präs. besonders westoberdeutsch und -mitteldeutsch klassenunabhängig und schwindet weitgehend erst im 16. Jahrhundert; vgl. Paul 24 1998: § 367, 395; Ebert et al. 1993: § M88). Im Verlauf des Frühneuhochdeutschen werden durch Ausgleichserscheinungen die schon im Mittelhochdeutschen (paradigmatisch oder regional) isolierten Endungsflexive aufgegeben. Gänzlich erfolgt dies hinsichtlich der Flexion der 2. Sg. Ind. Prät. starker Verben durch einerseits -e und andererseits die (umgelautete) Wortform des Prät. Pl. (gaeb-e > gab-st); nurmehr zeitlich und regional be-
1689 grenzt hinsichtlich der 1./3. Sg. Ind. Prät. der starken Verben, die insbesondere in der zweiten Hälfte des Frühneuhochdeutschen oft mit Relationsmorph -e erscheint (“lutherisches -e”; vgl. Hoffmann & Solms 1987: 42⫺46). Die Flexion der 3. Pl. Ind. Prät. (mittelhochdeutsch -ent) ist überlagert von einer bis in das 17. Jahrhundert reichenden Entwicklung “mit sprachlandschaftsinterner und teilweise erheblicher zwischenlandschaftlicher Variation”, durch die spezifisch landschaftliche Paradigmen herausgebildet werden (Ebert et al. 1993: § M74; Tendenz z. B. zum Einheitsplural -et). Schon für das Althochdeutsche zeigt sich die definierende Eigenschaft des starken Verbs, seine “Tempusstämme nur durch den Vokalwechsel zu unterscheiden” (Braune 14 1987: § 324), als Ergebnis einer vorausliegenden Entwicklung. Während die Unterscheidung im Indogermanischen und partiell noch im Germanischen auch aufgrund konsonantischer Variationen des Stammes, suffixaler und auch präfixaler Elemente erfolgt, sind solche Möglichkeiten im Althochdeutschen weitgehend geschwunden. Entsprechend den vier überkommenen Ablautstufen werden für den Großteil starker Verben vier primäre Tempusstämme (“Stammformen”, bei Berücksichtigung der indogermanischen/ germanischen Sprachgeschichte ist statt vom Stamm besser von “Wurzel” zu sprechen) unterschieden: z. B. rıˆtan (Präs./Inf.) ⫺ reit (1./ 3. Sg. Ind. Prät.) ⫺ riten/riti (1. Pl. und 2. Sg. Ind. Prät.) ⫺ giritan (Part. Prät.). Innerhalb des Präteritums der Klassen I⫺V dient der qualitative/quantitative Ablaut zusätzlich auch als Numerusablaut; Klasse VI und VII (ehemals reduplizierende Verben) haben im Singular und Plural einen jeweils identischen Stamm (z. B. VI: ahd. wahsan ⫺ wuohs ⫺ wuohsun/wuohsi ⫺ giwahsan). Die Divergenz zwischen 4 Ablaut-, aber nur 3 Zeitstufen stellt ein funktionales Mißverhältnis dar (vgl. Augst 1975: 251), das über das Althochdeutsche hinaus bis in das Frühneuhochdeutsche vorliegt. Eine Entwicklung setzt erst im Frühneuhochdeutschen ein, es werden die jeweiligen Numerusablaute weitgehend seit der zweiten Hälfte des 15. Jahrhunderts zumeist zugunsten einer der beiden Tempus-Alternanten ausgeglichen: z. B. ahd./mhd. (Sg./Pl.) reit/riten wird nhd. ritt/ritten, ahd./mhd. fand/ funden wird nhd. fand/fanden. Der Ablaut hat nun erst zum Neuhochdeutschen hin “seine wahrhaft funktionelle Geltung erlangt” (Paul 5 1920: 211): Die morphologische Kategorie
1690 des Tempus erscheint flexivisch profiliert. Eine über den Ablautausgleich bestimmte Periodengrenze des Frühneuhochdeutschen zum Neuhochdeutschen hin ist nicht eindeutig zu ziehen, da der konkrete Ausgleichsprozeß durch mehrere chronologisch versetzt wirksame Prinzipien gesteuert ist. Während der Ablautausgleich bei den starken Verben der Klassen I (z. B. rıˆten) und II (z. B. bieten) schon zum Ende des 17. Jahrhunderts lange abgeschlossen ist, bleibt die Entwicklung bei Klasse III und hier vor allem bei Verben mit Nasalverbindung nach dem Wurzelvokal (Subklasse IIIa, z. B. finden) zum Ende des 17. Jahrhunderts noch unabgeschlossen. Für den raschen Ausgleich in Klasse I und II wird eine strukturelle Disposition darin gesehen, daß der Stammvokal des Partizip Präteritums mit jenem des Singular oder Plural Präteritums übereinstimmte. Hier liegt ein intraparadigmatischer Ausgleich vor, der ohne Einfluß seitens einer anderen Ablautungsmöglichkeit eintrat (vgl. Behaghel 51928: § 436). Eine identische Disposition weisen auch die Verben der Klasse IIIa auf (mhd. finden ⫺ fand ⫺ funden ⫺ gefunden). Ihr Ablautausgleich ist jedoch zum Ende des Frühneuhochdeutschen nicht nur nicht entschieden, sondern die Richtung des im Neuhochdeutschen letztlich erreichten Ausgleichs (fand ⫺ fanden statt *fund ⫺ *funden) widerspricht vielmehr auch der Annahme einer wirksamen und das Part. Prät. einschließenden strukturellen Disposition: Die zugunsten des Sg. Prät.-a ausgleichende Entwicklung ist als eine interparadigmatische und an das Ablautungsmuster der starken Verben der Klasse IV/V (z. B. ne¨men/ge¨ben) angelehnte Ausgleichung zu werten. Das Ergebnis ist jedoch weniger das Ergebnis eines unbewußt vollzogenen Wandels, als vielmehr des bewußten und die Entwicklung beeinflussenden Eingreifens der normierenden und einem Konzept der Analogie verpflichteten Grammatiker des 18. Jahrhunderts (Gottsched und Adelung; vgl. Solms 1993). Im Rahmen dieses schon intentionalen Prozesses werden weitere, in frühneuhochdeutscher Zeit begonnene und das Flexionsund Klassifikationssystem der starken Verben verändernde Entwicklungen weiter forciert. So kommt es inbesondere zum ausgehenden Frühneuhochdeutschen zum internen Klassenwechsel einiger starker Verben; es erscheinen dominante Ablautmuster herausgebildet (interparadigmatischer Ausgleich), wodurch die noch althochdeutsche Systematik der starken Verben endgültig aufgehoben ist.
XVIII. Morphologischer Wandel II: Fallstudien
Über den Ablautausgleich hinaus findet in einigen Fällen ein Ausgleich zugunsten der -oaufweisenden Stammform des Part. Prät. statt, z. B. bei melken, fechten, glimmen. Hier erscheint die Reduzierung und Konzentration auf nurmehr eine Stammform für die Tempusinformation “Vergangenheit” (sei es finit oder periphrastisch mittels des Part. Prät.) konsequent fortgesetzt; vor dem Hintergrund des im Oberdeutschen seit dem 15. Jahrhundert beobachteten “Präteritumschwundes” zugunsten der Periphrase ist die intraparadigmatische Orientierung am Part. Prät. nur die konsequente Fortsetzung der flexivischen Profilierung des Tempus. Darüber hinaus verlieren einige weitere Lexeme ihren ursprünglichen und partiell bis über das 18. Jahrhundert erhaltenen Ablaut und übernehmen in einem interparadigmatischen Ausgleich ebenfalls den Stammvokal -o- (z. B. mhd. heben ⫺ huob ⫺ geheben > nhd. heben ⫺ hob ⫺ gehoben). Zum ausgehenden Frühneuhochdeutschen werden somit dominante Ablautungen herausgebildet, die ihrerseits zum Zielpunkt von Veränderungen werden: Die dominanten Ablautungen mit -i/ie(z. B. schreiben ⫺ schrieb ⫺ geschrieben) oder -o- umfassen zum Neuhochdeutschen hin ca. 51 Prozent aller starken Verben (vergleichsweise mittelhochdeutscher Anteil ca. 31 Prozent; vgl. Solms 1984: 71 f.). Eine Vereinfachung sowohl hinsichtlich der Anzahl flexivischer Muster als auch hinsichtlich der Verallgemeinerung des einfachsten Musters zeigt die Stammalternation in der Präsensflexion der starken Verben. Die aufgrund vorgängig phonologischer Entwicklungen zum Althochdeutschen hin entstandenen drei Muster “sekundärer” und d. h. nicht durch Ablaut erzeugter Stämme (ohne Alternation; Vokalwechsel im Ind. Sg., z. B. Inf. we¨rdan mit 1. Sg. Ind. wirdu, 2. Sg. Ind. wirdis, 3. Sg. Ind. wirdit; Vokalwechsel nur in 2./ 3. Ind. Sg., z. B. varan mit varu, veris, verit) wird im Verlauf des Frühneuhochdeutschen auf nurmehr zwei Muster reduziert (ohne Alternation; Vokalwechsel in 2./3. Ind. Sg, z. B. auch werden mit 1. Sg. Ind. werde, wirdest, wirdet). Während die Anzahl starker Verben vom Althochdeutschen (349) bis zum Mittelhochdeutschen (339) weitgehend gleich bleibt, ist zum Neuhochdeutschen (169) hin ein deutlicher Schwund eingetreten (vgl. Solms 1984: 23, Anm. 33): starke Verben gehen zum Neuhochdeutschen hin unter (z. B. bıˆten) oder nehmen schwache Flexion an (z. B. gıˆgen; vgl.
156. Vom Althochdeutschen zum Neuhochdeutschen
Solms 1984: 27 f.). Die Übernahme der schwachen Flexion zeigt sich strukturell disponiert bei Identität des Stammes von Inf./ Präs. mit dem des Part. Prät., so z. B. bei Verben der Klasse VII wie valten ⫺ vielt ⫺ gevalten (falten ⫺ faltete ⫺ gefaltet) (vgl. Schirmunski 1962: 504; Solms 1984: 323 f.). 5.2. Flexion der Substantive Die diachron motivierte Klassifizierung der deutschen Substantive unterscheidet seit dem Althochdeutschen “starke” (differenzierte flexivische Realisierung der morphosyntaktischen Kategorien) und “schwache” (flexivisch weniger differenziert, im Neuhochdeutschen nur -en in allen obliquen Kasus) Substantive. Die Unterscheidung überlagert die für das Germanische über Stammsuffixe definierte Stammklassengliederung, deren Gültigkeit das Inventar der althochdeutsch vollvokalisch erhaltenen Endungsflexive sowohl am partiell markierten Nom. Sg. (z. B. kann Endung -o nur mask. n-Stamm sein, z. B. hano) als auch über zahlreiche paradigmatische Oppositionen (z. B. Instr. Sg. tagu zu Gen. Pl. tago beim mask. a-Stamm) noch erweist.
Bereits im Althochdeutschen sind genusbezogene Hauptklassen profiliert: die “schwache” n-Deklination (alle Genera), die “starke” a(Maskulinum, Neutrum), oˆ- (Femininum) und i-Deklination (Maskulinum, Femininum); neben den Hauptreihen existieren einige kaum repräsentierte Restklassen, deren Inventar zumeist bereits von den Hauptreihen aufgesogen ist (besonders -iz/-az-Stämme in die a-Deklination; vgl. Wolf 1981: 197). Über die flexivische Angleichung verschiedener Klassen ist im Althochdeutschen gegenüber dem Germanischen zugleich die Tendenz deutlich, flexivische Kasusunterscheidungen einzuebnen, z. B. in der weitgehenden Aufhebung der Distinktion Nom. versus Akk. Sg. bei den germanischen a-Stämmen (gotisch Nom. Sg. dag-s und Akk. Sg. dag, ahd. Nom./Akk. tag; vgl. Hirt 1932: § 38), z. B. auch in der weitgehenden Aufgabe des Instrumentals. In der Folge erweisen die weiteren flexivischen Entwicklungen eine vorwiegende Numerusprofilierung bei gleichzeitiger Kasusnivellierung; dabei wird die Wurzelflexion aus-, die Möglichkeit der regressiven Flexion stark abgebaut. Die Flexion bleibt im Althochdeutschen noch weitgehend regressiv gesteuert, der (Primär)umlaut erscheint lautgesetzlich verteilt und somit stammklassenbezogen sowohl im Singular als auch im Plural (z. B. Dat. Sg. und auch Nom./Akk. Pl. ensti zum fem. i-Stamm anst ‘Gunst’). Die Umlautformen werden teilweise schon im Verlauf
1691 des Althochdeutschen aufgegeben, so daß die für das Deutsche formulierte Tendenz zu umlautlosem Singular und allenfalls umgelautetem Plural bereits im Althochdeutschen greift (vgl. Sonderegger 1979: 309). Die wesentliche Umgestaltung des althochdeutschen Flexionssystems wird dann durch den zum Mittelhochdeutschen eingetretenen Zusammenfall der vollen Nebensilbenvokale in -e bewirkt, wodurch ein strukturell grundsätzlich neues Flexionssystem entsteht (vgl. Stopp 1974): Es werden die meisten paradigmatischen Distinktionen eingeebnet (z. B. Zusammenfall der Plural-Flexivreihe der a- und iDeklination, z. B. Nom/Akk. Pl. tag-/gest-e aus -a/-i), so daß das Inventar der segmentierbaren Flexive von 52 auf 16 im Mittelhochdeutschen reduziert ist (vgl. Sonderegger 1979: 246). Strukturell wirkt sich dieser Vorgang in mehrfacher Weise aus. Einerseits wird nun zusätzlich zur regressiven Endungsflexion auch die Wurzelflexion (Umlaut) genutzt. Der Umlaut erscheint analog übertragen auch bei Lexemen, bei denen er historisch nicht entwickelt ist, so z. B. bei mask. aStämmen (z. B. Nom. Pl. nagele neben negele) oder den angegliederten Restgruppen (z. B. rStamm veter(e) neben vater(e); vgl. Paul 24 1998: § 177; Wolf 1981: 198); damit erhält der Umlaut eine deutlichere flexivische Signifikanz als Kennzeichen des Plurals. Andererseits ist durch die eingetretene Uniformierung der Nebensilbe “automatisch die Grenze zwischen Stamm und Flexionsendung hinter die alte Flexionsendung” (Wegera 1987: § 11) verschoben: Es kommt zu der für das Mittelhochdeutsche wesentlichen Veränderung einer nurmehr zweigliedrigen Wortstruktur von Stamm und Flexionszeichen, durch die der Nom. Sg. zur unmarkierten Grundform wird (vgl. Paul 231989: § 174; Klein 1987: 155). Relational dazu wird das ursprüngliche Stammbildungssuffix -er (germ.*-iz/-az) zum Plural-Kennzeichen profiliert und nun auf eine Reihe weiterer, zum Teil im Mittelhochdeutschen erstmals belegter Lexeme übertragen (z. B. cleid ⫺ cleider). Ebenso wird nun bei allen Substantiven, deren Plural im Althochdeutschen durch vokalisches Flexiv gebildet wird (z. B. Nom. Pl. taga, gebaˆ, listi), das -e zu einem der vorwiegenden mittelhochdeutschen Plural-Flexive, das insbesondere im Mitteldeutschen auch auf bisher nicht markierte Fälle übertragen wird (besonders neutr. a-Stämme, z. B. Nom./Akk. Pl. wort > worte). Damit ist schon im Mittelhochdeutschen die Tendenz zur eindeutigen
1692 Kennzeichnung der Numerusopposition deutlich (vgl. Wolf 1981: 199). Dieser Prozeß wird allerdings durch die seit dem 13. Jahrhundert zuerst im Oberdeutschen eintretende Apokope und die damit eintretende Zerstörung der über -e geleisteten Plural-Kennzeichnung überlagert. Dies prägt die schreibsprachliche Entwicklung bis in das 16. Jahrhundert hinein. Dabei lassen sich die frühneuhochdeutschen Entwicklungen insgesamt als zwei Großprozesse beschreiben: Über eine Umorganisation vorhandener flexivischer Mittel (z. B. -er, -en, Umlaut; vgl. Wegera 1987: § 63) wird der bisher unbezeichnete (mittelhochdeutsche neutr. a-Stämme, z. B. wort ⫺ wort) oder durch Apokope unbezeichnet gewordene Plural “durch deutliche und sichere Flexive” markiert (Wegera 1987: § 11); durch eine (flexivische) Nivellierung der Kasus (z. B. Abbau der -en-Flexive in obliquen Kasus bei Feminina oder gegenläufig auch Übertragung des -en in den Nom. Sg.; vgl. Wegera 1987: § 48), die über andere analytisch-syntaktische Möglichkeiten zum Ausdruck gelangen, erfolgt paradigmatisch eine Profilierung des Numerus (Plural). Da beide Prozesse (zumindest teilweise) genusgesteuert verlaufen, zeigt sich neben Kasusnivellierung und Numerusprofilierung im Verlauf des Frühneuhochdeutschen implizit ebenfalls eine Genusprofilierung (vgl. Wegera 1987: § 11, 18). Schon im 16. Jahrhundert deutet sich aus dem von der Apokope weitgehend unberührt gebliebenen Obersächsischen kommend eine Restituierung des Plural-e an, das im 17. Jahrhundert zunehmend wieder in der Schriftsprache verwendet wird. Hier spielt die Bibelübersetzung Luthers und ihre Verbreitung eine herausragende Rolle (“lutherisches -e”); es wirkt jedoch auch die “strukturelle Überlegenheit des omd. Systems”: Hinsichtlich einer “konsequenteren Numerusmarkierung durch das Plural-e” entsprach dieses schreiblandschaftlich ausgebildete System “der Tendenz zur Numerusprofilierung besser als die nur ansatzweise durchgeführten Numerusunterscheidungen mit Hilfe anderer Flexive” (Wegera 1987: § 95). So wird das Plural-e “vielfach wieder zur bevorzugten Pluralbildung bei Lexemen, die in davorliegenden Zeiträumen andere Pluralbildungen angenommen hatten” (Wegera 1987: § 72). Zum Neuhochdeutschen hin ist -e zum wichtigsten Numerusflexiv innerhalb des Kernwortschatzes herausgebildet (vgl. Augst 1975: 37). Daneben hat die Kombinierbarkeit des Umlauts mit suffixalen Flexiven zum
XVIII. Morphologischer Wandel II: Fallstudien
Neuhochdeutschen deutlich zugenommen (vgl. Sonderegger 1979: 246); mit dem Umlaut nicht kombinierbar sind neuhochdeutsch allein die nicht-fremdsprachlichen Pluralflexive -s und -en. 5.3. Flexion der Adjektive Die Flexionsgeschichte der Adjektive ist insbesondere innerhalb der gesamten Substantivgruppe als Durchsetzung einer Artikel- sowie Gruppenflexion zu beschreiben, die sich sowohl an konkreten flexivischen Veränderungen als auch in der Aufgabe einer Flexion des prädikativen Adjektivs zeigt: Artikelflexion meint einerseits die grammatische Abhängigkeit vom und andererseits die Funktionalisierung auf das bezogene Substantiv hin (das Adjektiv realisiert die morphosyntaktischen Merkmale des Substantivs); dabei wird das Adjektiv in Abhängigkeit von der ausdrucksseitig vorhandenen Markiertheit der Substantivgruppe “determinierend” (genusdifferenzierend im Nom. Sg.: ein lieber Junge, ein liebes Kind) oder “indeterminierend/attribuierend” (genusindifferent im Nom. Sg. -e: der liebe Junge, das liebe Kind) flektiert. “Gruppenflexion” meint die tendenzielle Flexivdistribution in Abhängigkeit besonders von der Markiertheit des Substantivs selbst: “starke” Adjektivflexion bei fehlendem substantivischem Kasusflexiv (groß-er Liebe fähig sein), “schwache” Adjektivflexion bei vorhandenem substantivischem Kasusflexiv (groß-en Mutes fähig sein). Die das Adjektiv innerhalb der Hauptwortarten charakterisierende Artikelflexion ist bereits im Althochdeutschen usuell, die Regelung wird im Frühneuhochdeutschen endgültig fest. Diese “Formregel” (“mechanische Regelung”, Behaghel 1923: § 10) löst eine frühere “Sinnregel” ab, bei der über die Flexion des Adjektivs primär eine ko- oder kontextuelle (Un)Bestimmtheit signalisiert wurde. Dabei stimmen die vordeutschen Flexionsparadigmen der Adjektive weitgehend mit denen der Substantive überein (vgl. Kern & Zutt 1977: 109). Aufgrund jedoch einer den Pronomen identischen Verwendungsweise treten schon im Germanischen zunehmend auch pronominale Formen neben die ursprünglichen nominalen (vgl. Kern & Zutt 1977: 110 f.). Die Durchsetzung dieser pronominalen Flexive ist beim Adjektiv mit dem 16. Jahrhundert abgeschlossen; sie führt zur flexivischen Angleichung von Adjektiv und Pronomen, die sich ähnlich auch bei den Pronomen zeigt, z. B. Dat. Pl./Gen. Pl. des Per-
1693
156. Vom Althochdeutschen zum Neuhochdeutschen
sonalpronomens in > inen, ir > irer oder Gen. Sg./Pl. des > dessen, der > deren, wes > wessen (vgl. Walch & Häckel 1988: § 12.2., Anm. 34). Eine Angleichung der adjektivischen und pronominalen Flexion erweist sich schließlich insbesondere “in der Durchsetzung der Regelung des grammatisch determinierenden/indeterminierenden Flexionsprinzips auch beim Poss. Pron., den Pronominaladj. (und sehr selten auch Pronominalsubst.)” (Ebert et al. 1993: § M60). In diesem Zusammenhang ist die Adjektivflexion als eine Gruppenflexion herausgebildet, für die im Verlauf des 16. Jahrhunderts ein “monoflexivisches Prinzip” usuell wird (vgl. Brinkmann 1964: 96; Solms & Wegera 1991: § 143): Die morphosyntaktischen Kategorien des Substantivs werden tendenziell nur an einer Stelle der gesamten Substantivgruppe ausdrucksseitig signalisiert (der großer Mut > der groß-e Mut; groß-es Mutes fähig sein > groß-en Mutes); eine ähnliche Entwicklung zeigt sich im Frühneuhochdeutschen vereinzelt auch beim Pronomen: meinen Wissens. Parallel der flexivischen Entwicklung wird die Flexionsmöglichkeit bis zum 16. Jahrhundert auf nurmehr die attributive/substantivierte Verwendung begrenzt; die bis in die erste Hälfte des Frühneuhochdeutschen noch mögliche Flexion auch des prädikativen Adjektivs (ein Baum ist großer) wird aufgegeben; damit werden prädikatives und adverbiales Adjektiv formal ununterscheidbar. Markanter noch als beim Substantiv wird das Flexivinventar beim Adjektiv vom Althochdeutschen (47 unterscheidbare Flexive) über das Mittelhochdeutsche (24) zum Neuhochdeutschen hin (6) reduziert (vgl. Sonderegger 1979: 246). Die stufenweise zu verfolgende Reduzierung sowie die beobachtbare Verwendung der jeweiligen Flexive erweist mehrere Tendenzen: (a) Es wird eine tendenzielle 1 : 1-Zuordnung (Form:Funktion) auf der Ebene der “morphosyntaktischen Position” (gemeint ist z. B. ‘Nom. Sg. Mask. nach best. Art.’ oder ‘Dat. Sg. Fem. nach Präposition’; vgl. Solms & Wegera 1991: § 12) herausgebildet. Aufgrund der flexivischen Reduzierung stehen nur noch maximal zwei Flexive je morphosyntaktischem Merkmalbündel zur Verfügung, die auf der Ebene der morphosyntaktischen Position dann eindeutig disambiguiert werden. So sind z. B. für Nom. Sg. Neutr. -(e) und -(e)s möglich, nach dem bestimmten Artikel kann jedoch nur -(e) (das liebe Kind), nach dem unbestimmten Artikel nur -(e)s (ein liebes Kind) erscheinen.
(b) In keinem Fall wird im Frühneuhochdeutschen das flexivische Inventar der SingularKasus so verändert, daß eine bestehende Genusdifferenzierung aufgehoben würde. Vielmehr wird dort, wo flexivisch eine Genuspolysemie vorliegt, im Frühneuhochdeutschen eine Eindeutigkeit erreicht: “Genusprofilierung” (z. B. Profilierung des -0 im Nom./Akk. Sg. zum genusmarkierenden Flexiv des Neutr. bis ins 16. Jahrhundert: Maskulinum wahr-er Mut, Femininum wahr-e Liebe, Neutrum wahr Wort; vgl. Solms & Wegera 1991: § 51⫺55). (c) Die im Nom./Akk. Sg. Neutr. zum späten Frühneuhochdeutschen erfolgte Verdrängung der unflektierten Form zugunsten der Flexion mit pronominalem -es liegt innerhalb des Prozesses der zwischen dem 14. und 16. Jahrhundert erfolgten formalen Ausdifferenzierung der Substantiv- und Verbalgruppe: die flexionslose Form kommt der nicht substantivischen Komponente der Verbalgruppe (als prädikatives Adjektiv), die flektierte Form der nicht substantivischen Komponente der Substantivgruppe zu. (d) Teilweise ergibt sich eine Kasusmarkierung im Zusammenhang der Markierungsfunktion des Adjektivs für die am Substantiv nicht markierten morphosyntaktischen Merkmale: Dal sieht neben der “Kasusnivellierung” beim Substantiv eine gegensteuernde “systemerhaltende Tendenz” beim Adjektiv: Eine solche Kasus(gruppen)markierung (Nom./Akk. Pl. versus Gen./Dat. Pl.) zeigt z. B. die insbesondere im Oberdeutschen bis in das 18. Jahrhundert hinein übliche determinierende Flexion mit -e im Nom./Akk. trotz vorausgehenden bestimmten Artikels (die liebe Kinder) gegenüber Dat./Gen. mit -en (den/der lieben Kinder(n)). Eine generelle Tendenz zur Kasusmarkierung liegt jedoch nicht vor, analog zu den Pronomen (diu, die > die) wird z. B. die formale Kasusunterscheidung Nom./Akk. Sg. Fem. aufgehoben.
6.
Wortbildung
Die geschichtliche Ausdifferenzierung von Flexion und Wortbildung erweist sich auch in der funktionalen Profilierung der formalen Ausdrucksmittel. Zwar werden bereits im Germanischen Komposition, Derivation und auch “innere Wortbildung” genutzt, doch sieht Grimm in letzterer, durch Ablaut erfolgender Wortbildung die älteste Möglichkeit, z. B. erkennbar in ahd. faran ‘fahren’, fuora ‘Fahrt’ (vgl. Grimm 1826: 1). Dieses Prinzip ist im Althochdeutschen schon nicht mehr produktiv, sondern nurmehr “aktiv” (die Wortschatzeinheiten sind morphologisch-relational motiviert); der Lautwechsel ist seitdem primär flexivisch produktiv; ausgebaut werden die Möglichkeiten der Komposition
1694 und affixalen Wortbildung (als explizite Ableitung durch Suffixe oder als Präfigierung). Aufgrund der in den verschiedenen Wortarten unterschiedlichen Nutzung sind die einzelnen Wortbildungsarten schon im Althochdeutschen als wortartcharakteristisch ausgewiesen: Komposition bei den Substantiven, Derivation bei Substantiven und Adjektiven, Präfigierung bei den Verben (eine wortartbestimmte Bevorzugung erweisen mittelhochdeutsche Texte im Verhältnis von Ableitungen zu Komposita mit 3 : 2 bei Substantiven und 3 : 1 bei Adjektiven; vgl. Zutt 1985: 1160). Dabei zeigt sich in der parallelen Existenz einiger Morphe als noch selbständige Lexeme (z. B. starkes Maskulinum/Neutrum tuom ‘Urteil, Gericht, Macht, Fähigkeit’), als erster oder zweiter Teil entsprechender Determinativkomposita (tuomtac ‘Gerichtstag’, cheisertuom ‘Herrschaft des Kaisers’) sowie auch schon nurmehr als Suffix (alttuom ‘Alter’) die ursprüngliche Herkunft affixaler Elemente: Es handelt sich um Lexeme, die aufgrund häufiger und reihenhaft empfundener Komposition ihre Selbständigkeit verlieren und zum Affix werden (z. B. auch Präfix be- aus der Präposition bei). Ähnliche Entwicklungen zeigen sich in allen Sprachstufen des Deutschen, so z. B. auch in -zeug (Nähzeug, Werkzeug). Die Wortschatzerweiterung durch Wortbildung wird in der geschichtlichen Entwicklung des Deutschen zunehmend genutzt, wobei die deutlichste frequentielle Zunahme im Frühneuhochdeutschen und hier insbesondere im 16. und 17. Jahrhundert stattfindet: So liegt z. B. der Anteil derivierter/zusammengesetzter Nomen in exemplarisch ausgewählten mittelhochdeutschen Texten bei unter 10 Prozent, in entsprechenden Texten des Neuhochdeutschen dagegen bei über 50 Prozent (vgl. Zutt 1985: 1160); allein der Anteil von Komposita am Gesamt des jeweiligen Substantivwortschatzes wächst von ca. 7 Prozent in Texten des 12. und 13. Jahrhunderts auf über 18 Prozent in Texten des späten 17. Jahrhunderts. Im Bereich der Derivation wächst der Bestand desubstantivischer Verben zwischen der Früh- und Spätzeit des Frühneuhochdeutschen um ca. 125 Prozent (vgl. Prell & Schebben-Schmidt 1996: 386⫺ 388). In solchen Verhältnissen erweist sich die im Frühneuhochdeutschen konstitutive Offenheit der schriftsprachlichen Entwicklung für varietäre Einflüsse, insbesondere auch die frühneuhochdeutsche Ausbildung und Entwicklung von Fachsprachen.
XVIII. Morphologischer Wandel II: Fallstudien
Das Inventar der Wortbildungsmittel sowie auch die semantisch-funktionalen wie morphologischen Typen der jeweiligen Wortbildungsmuster sind im wesentlichen bereits im Althochdeutschen ausgebildet (vgl. Prell & Schebben-Schmidt 1996: 380) (z. B. denominale Verbableitung des Typs ‘ornatives Verb’ mittels Verbalisierungsmorphs, z. B. salb-oˆn). Allerdings finden wesentliche Veränderungen statt, die sich derzeit (möglicherweise nur aufgrund der Forschungslage) als Entwicklungen insbesondere der frühneuhochdeutschen Zeit zeigen: Entwicklungen des Lexembestandes erkennbarer Wortbildungsmuster/-typen und auch des Inventars der genutzten Affixe, sowie Veränderungen der morphologischen Struktur und Basisbeziehungen. So handelt es sich bei der Mehrzahl mittelhochdeutscher Präfixverben (zu ca. 70⫺90 Prozent) um “untrennbare Präfixverben” (z. B. be-, ent-; “trennbare Präfixverben” z. B. mit an-, aus-), wohingegen diese im Neuhochdeutschen nurmehr ca. 30 Prozent ausmachen (vgl. Solms 1990: 118; Herbers 2002: 107 f.). Dem veränderten Wortbildungsmuster geht eine Verschiebung der Wortbildungsbasis der vorhandenen (un)trennbaren Präfixverben parallel. Der Anteil der sogenannten “Präfix-Suffix-Ableitungen” (mit nominaler Basis, z. B. bebildern) beträgt neuhochdeutsch knapp 40 Prozent; der mittelhochdeutsche Anteil solcher Bildungen liegt bei höchstens 12 Prozent. Der Umschwung zugunsten einer verstärkten Verbableitung auch aus nominalen Basen wird schon im späten Mittelhochdeutschen bemerkbar und setzt vor allem seit dem 16. Jahrhundert ein (vgl. Prell & Schebben-Schmidt 1996: 388). Die fortschreitende Verbalisierung nominaler Basen zeigt für die Verben eine den Nomen identische Tendenz: die Univerbierung umständlicherer Syntagmen in Komposita und Derivaten, nach Wegera ein Moment sprachlicher Ökonomie des Deutschen (Erben 21983: 122; vgl. Wegera & Prell 22000: 1594). Die Verbpräfigierung zeigt vom Mittelzum Neuhochdeutschen eine ähnlich auch im Englischen entwickelte Tendenz zur Iso-Morphie (vgl. Koziol 21972: § 710). Das Mittelhochdeutsche kennt ebenso wie noch das frühe Frühneuhochdeutsche eine “Überproduktion von Präfixverben”, die zu semantischer Äquivalenz von Simplex und Präfixverb führen kann (vgl. Zutt 1985: 1163; Solms 1990: 122; Herbers 2002: 323 f.). Hier zeigt
156. Vom Althochdeutschen zum Neuhochdeutschen
1695
sich in der Folge eine Informationsverlagerung vom Simplex auf das Präfixverb, auf die eine semantische Verengung oder der Untergang des Simplexverbs folgt. Der Wortschatz als Inventar der formalen Ausdrucksmittel wird erweitert, auf der Inhaltsseite aber ist nichts Neues entstanden. Die Präfigierung leistet hier einerseits eine Auflösung der bei Simplicia vorhandenen Polysemie. Zugleich erscheinen die Präfixverben als ausdrucksseitig deutlicher markierte Wortschatzeinheiten für gleichbleibende Inhalte. Es werden einzelne Sprachinhalte “reihenhaft gleichbleibend (‘iso-morph’) ausgeprägt” (Erben 1964), das einzelne Präfix wird zu einem deutlichen Ausdrucksträger profiliert. Da hier eine dem Simplex inhärente Bedeutung durch ein zusätzliches, syntagmatisch gebundenes Element zum Ausdruck gebracht wird, steht auch diese Entwicklung im Zusammenhang einer Veränderung synthetischer zugunsten analytischer Ausdrucksmittel (vgl. Tschirch 31989: 134).
Gruyter (Handbücher zur Sprach- und Kommunikationswissenschaft 2)
7.
Erben, Johannes (1964), “Deutsche Wortbildung in synchronischer und diachronischer Sicht”. Wirkendes Wort 14, 83⫺93
Zitierte Literatur
Augst, Gerhard (1975), Untersuchungen zum Morpheminventar der deutschen Gegenwartssprache. Tübingen: Narr (Forschungsberichte des Instituts für deutsche Sprache 25) Behaghel, Otto (1923), Deutsche Syntax: Eine geschichtliche Darstellung. Bd. I: Die Wortklassen und Wortformen. A. Nomen. Pronomen. Heidelberg: Winter (Germanische Bibliothek I, 10) Behaghel, Otto (51928), Geschichte der deutschen Sprache. Berlin, Leipzig: de Gruyter (Grundriss der Germanischen Philologie 3) Bergmann, Rolf & Tiefenbach, Heinrich & Voetz, Lothar (1987, Hrsg.), Althochdeutsch, Bd I: Grammatik, Glossen und Texte. Heidelberg: Winter (Germanische Bibliothek, Neue Folge 3. Reihe: Untersuchungen) Besch, Werner (1980), “Frühneuhochdeutsch”. In: Althaus, Hans Peter & Henne, Helmut & Wiegand, Herbert Ernst (Hrsg.), Lexikon der Germanistischen Linguistik. Tübingen: Niemeyer, 588⫺597
Braune, Wilhelm (141987), Althochdeutsche Grammatik, bearb. v. Hans Eggers. Tübingen: Niemeyer (Sammlung kurzer Grammatiken Germanischer Dialekte A, 5) [11886] Brinkmann, Hennig (1964), “Das deutsche Adjektiv in synchronischer und diachronischer Sicht”. Wirkendes Wort 14, 94⫺104 Durrell, Martin (1980), “Morphologische Regelmäßigkeiten im deutschen Ablautsystem”. In: Rupp, Hans & Roloff, Hans-Gert (Hrsg.), Akten des VI. Internationalen Germanisten-Kongresses Basel 1980, Teil 2. Bern usw.: Lang, 19⫺28 (Jahrbuch für internationale Germanistik A, 8.2) Ebert, Robert Peter & Reichmann, Oskar & Solms, Hans-Joachim & Wegera, Klaus-Peter (1993), Frühneuhochdeutsche Grammatik. Tübingen: Niemeyer (Sammlung kurzer Grammatiken Germanischer Dialekte A, 12) Eichinger, Ludwig M. (1984), “Die großen historischen Grammatiken des Deutschen”. In: Besch et al. (Hrsg.), 433⫺440
Erben, Johannes (21983), Einführung in die deutsche Wortbildungslehre. Berlin: Schmidt (Grundlagen der Germanistik 17) [11975] Grimm, Jacob (1819⫺1837), Deutsche Grammatik, 4 Teile. Göttingen Habermann, Mechthild (1994), Verbale Wortbildung um 1500: Eine historisch-synchrone Untersuchung anhand von Texten Albrecht Dürers, Heinrich Deichslers und Veit Dietrichs. Berlin usw.: de Gruyter (Wortbildung des Nürnberger Frühneuhochdeutsch 2) Hempen, Ute (1988), Die starken Verben im Deutschen und Niederländischen: Diachrone Morphologie. Tübingen: Niemeyer (Linguistische Arbeiten 214) Henzen, Walter (31965), Deutsche Wortbildung. Tübingen: Niemeyer (Sammlung kurzer Grammatiken germanischer Dialekte B. 5) [11947]
Besch, Werner (22003), “Die Entstehung und Ausformung der neuhochdeutschen Schriftsprache/ Standardsprache”. In: Besch et al. (Hrsg.), 2252⫺ 2296
Herbes, Birgit (2002), Verbale Präfigierung im Mittelhochdeutschen: eine semantisch-funktionale Korpusanalyse. Tübingen: Niemeyer (Studien zur Mittelhochdeutschen Grammatik 1)
Besch, Werner & Betten, Anne & Reichmann, Oskar & Sonderegger, Stefan (21998/2000/2003/2004, Hrsg.), Sprachgeschichte, Bd. 1⫺4 [2., vollst. neubearb. und erw. Aufl.; 11984/1989] Berlin usw.: de
Hirt, Hermann (1932), Handbuch des Urgermanischen, Teil II: Stammbildungs- und Flexionslehre. Heidelberg: Winter (Indogermanische Bibliothek I, 21)
1696 Hoffmann, Walter & Solms, Hans-Joachim (1987), “Zur Flexion der starken und schwachen Verben”. In: Besch, Werner & Klaus-Peter Wegera (Hrsg.), Frühneuhochdeutsch: Zum Stand der sprachwissenschaftlichen Forschung. Zeitschrift für deutsche Philologie 106, Sonderheft, 37⫺59 Hotzenköcherle, Rudolf (1962), “Entwicklungsgeschichtliche Grundzüge des Neuhochdeutschen”. Wirkendes Wort 12, 321⫺331 Hotzenköcherle, Rudolf (1968), “Gegenwartsprobleme im deutschen Adjektivsystem”. Neuphilologische Mitteilungen 69, 1⫺28 Joeres, Rolf (1995), Wortbildungen mit -macher im Althochdeutschen, Mittelhochdeutschen und Neuhochdeutschen. Heidelberg: Winter (Germanische Bibliothek: Reihe 3, Untersuchungen, N.F. 21) Kern, Peter Chr. & Zutt, Herta (1977), Geschichte des deutschen Flexionssystems. Tübingen: Niemeyer (Germanistische Arbeitshefte 22) Kienle, Richard von (21969), Historische Laut- und Formenlehre des Deutschen. Tübingen: Niemeyer (Sammlung kurzer Grammatiken Germanischer Dialekte A. Hauptreihe 11) Klein, Thomas (1987), “Zur althochdeutschen Flexionsmorphologie in synchroner Sicht”. In: Bergmann et al. (Hrsg.), 146⫺168 Kluge, Friedrich (21925), Abriß der deutschen Wortbildungslehre. Halle/S.: Niemeyer Koenraads, Willy Henri August (1953), Studien über sprachökonomische Entwicklungen im Deutschen. Amsterdam: Rapoldi Kohrt, Manfred (1984), “Historische Phonetik und Phonologie”. In: Besch et al. (Hrsg.), 388⫺399 Koziol, Herbert (21972), Handbuch der englischen Wortbildungslehre. Heidelberg: Winter (Germanische Bibliothek, Reihe 1) [11937] Lüdtke, Helmut (1988), “Grammatischer Wandel”. In: Ammon, Ulrich & Dittmar, Norbert & Mattheier, Klaus J. (Hrsg.), Soziolinguistik, Bd. 1⫺2. Berlin usw.: de Gruyter (Handbücher zur Sprachund Kommunikationswissenschaft 3.), 1632⫺1642
XVIII. Morphologischer Wandel II: Fallstudien Gruyter (Wortbildung des Nürnberger Frühneuhochdeutschen 1) Paul, Hermann (1916⫺1920), Deutsche Grammatik, Bd. 1⫺5. Tübingen: Niemeyer Paul, Hermann (51920), Prinzipien der Sprachgeschichte. Halle/S.: Niemeyer [81968 Tübingen: Niemeyer; 11880] Paul, Hermann (1981), “Über die Aufgaben der Wortbildungslehre”. In: Lipka, Leonhard & Günther, Hartmut (Hrsg.), Wortbildung. Darmstadt, 17⫺35 (WDF 564) Paul, Hermann (241998), Mittelhochdeutsche Grammatik, neu bearbeitet von Peter Wiehl & Siegfried Grosse. Tübingen: Niemeyer (Sammlung kurzer Grammatiken Germanischer Dialekte A, 2) [11881] Penzl, Herbert (1984), Frühneuhochdeutsch. Bern usw.: Lang (Germanistische Lehrbuchsammlung 9) Prell, Heinz-Peter & Schebben-Schmidt, Marietheres (1996), Die Verbableitung im Frühneuhochdeutschen. Berlin usw.: de Gruyter (Studia Linguistica Germanica 41) Putschke, Wolfgang (21998), “Die Arbeiten der Junggrammatiker und ihr Beitrag zur Sprachgeschichtsforschung”. In: Besch et al. (Hrsg.), 474⫺ 494 Reichmann, Oskar (1988), “Zur Abgrenzung des Mittelhochdeutschen vom Frühneuhochdeutschen”. In: Bachofer, Wolfgang (Hrsg.), Mittelhochdeutsches Wörterbuch in der Diskussion. Tübingen: Niemeyer (Reihe Germanistische Linguistik 84), 119⫺147 Roelcke, Thorsten (1995), Periodisierung der deutschen Sprachgeschichte. Berlin usw.: de Gruyter (Studia Linguistica Germanica 40) Roelcke, Thorsten (2001), Periodisierung: Die zeitliche Gliederung der deutschen Sprachgeschichte. Frankfurt a. M.: Lang (Dokumentation germanistischer Forschung 4) Ronneberger-Sibold, Elke (1989), Historische Phonologie und Morphologie des Deutschen: Eine kommentierte Bibliographie zur strukturellen Forschung. Tübingen: Niemeyer (GA: Ergänzungsreihe, 3)
Mattheier, Klaus (21998), “Allgemeine Aspekte einer Theorie des Sprachwandels”. In: Besch et al. (Hrsg.), 824⫺836
Russ, Charles V. J. (1978), Historical Germanic Phonology and Morphology. Oxford: Oxford University Press
Michels, Victor (51979), Mittelhochdeutsche Grammatik. Heidelberg: Winter
Russ, Charles (1987), “Die Verben auf -u in der Mundart von Bosco Gurin und die ahd. -oˆn-Verben”. In: Bergmann et al. (Hrsg.), 332⫺344
Moser, Virgil (1929), Frühneuhochdeutsche Grammatik. Bd. I: Lautlehre. 1. Hälfte: Orthographie, Betonung, Stammsilbenvokale. Heidelberg: Winter (Germanische Bibliothek I, 17) Müller, Peter O. (1993), Substantiv-Derivation in den Schriften Albrecht Dürers. Berlin usw.: de
Scherer, Wilhelm (21878), Zur Geschichte der deutschen Sprache. Berlin: Weidmann Schirmunski, Victor M. (1962), Deutsche Mundartkunde: Vergleichende Laut- und Formenlehre der deutschen Mundarten. Berlin: Akademie
156. Vom Althochdeutschen zum Neuhochdeutschen
1697
Schmidt, Karl Horst (21998), “Versuch einer geschichtlichen Sprachtypologie des Deutschen”. In: Besch et al. (Hrsg.), 993⫺1000
Sonderegger, Stefan (22000), “Morphologie des Althochdeutschen”. In: Besch et al. (Hrsg.), 1171⫺ 1196
Schwarz, Hans (1986), Präfixbildungen im deutschen Abrogans: Analyse und Systematik. Göppingen: Kümmerle (Göppinger Arbeiten zur Germanistik 458)
Sta˚rck, John (1912), Studien zur Geschichte des Rückumlauts. Uppsala: Appelbergs Boktryckeri
Simmler, Franz (1987), “Morphologische Grundlagen zur Beschreibung althochdeutscher Schreibdialekte”. In: Bergmann et al. (Hrsg.), 114⫺146. Sodmann, Timothy (22000), “Der Rückgang des Mittelniederdeutschen als Schreib- und Druckersprache”. In: Besch et al. (Hrsg.), 1505⫺1512 Solms, Hans-Joachim (1984), Die morphologischen Veränderungen der Stammvokale der starken Verben im Frühneuhochdeutschen. Ph.D. Dissertation, Universität Bonn Solms, Hans-Joachim (1990), “Das System der Präfixverben in der frühesten Überlieferung des Hartmannschen ‘Gregorius’”. In: Besch, Werner (Hrsg.), Deutsche Sprachgeschichte: Grundlagen, Methoden, Perspektiven. Festschrift für Johannes Erben zum 65. Geburtstag. Frankfurt a. M. usw.: Lang, 115⫺128 Solms, Hans-Joachim (1991), “Die Wortbildung der Verben in Hartmann von Aues ‘Iwein’ (Hs.B) und ‘Gregorius’ (Hs.A): Das Präfix ge- im System der verbalen Präfigierung.” In: Wegera, Klaus-Peter (Hrsg.), Mittelhochdeutsche Grammatik als Aufgabe. Zeitschrift für deutsche Philologie 110, Sonderheft, 110⫺140 Solms, Hans-Joachim (1993), “Ein verspäteter Ablautausgleich im Kontext des regionalen und konfessionellen Meinungsstreits im 18. Jahrhundert. Die starken Verben der Klasse IIIa”. In: Mattheier, Klaus J. & Wegera, Klaus-Peter & Hoffmann, Walter & Macha, Jürgen & Solms, Hans-Joachim (Hrsg.), Vielfalt des Deutschen: Festschrift für Werner Besch. Frankfurt a. M. usw.: Lang, 331⫺351 Solms, Hans-Joachim (21998), “Historische Wortbildung”. In: Besch et al. (Hrsg.), 596⫺610 Solms, Hans-Joachim & Wegera, Klaus-Peter (1991), Flexion der Adjektive. Heidelberg: Winter (Grammatik des Frühneuhochdeutschen: Beiträge zur Laut- und Formenlehre, Band VI) Sonderegger, Stefan (1979), Grundzüge deutscher Sprachgeschichte, Bd. I: Einführung-GenealogieKonstanten. Berlin usw.: de Gruyter
Stopp, Hugo (1974), “Veränderungen im System der Substantivdeklination vom Althochdeutschen bis zum Neuhochdeutschen”. In: Besch, Werner (Hrsg.), Studien zur deutschen Literatur und Sprache des Mittelalters. Festschrift für Hugo Moser zum 65. Geburtstag. Berlin: Schmidt, 324⫺344 Tschirch, Fritz (31989), Geschichte der deutschen Sprache, zweiter Teil, bearbeitet von Werner Besch. Berlin: Schmidt Walch, Maria & Häckel, Susanne (1988), Flexion der Pronomina und Numeralia. Heidelberg: Winter (Grammatik des Frühneuhochdeutschen: Beiträge zur Laut- und Formenlehre, Band VII) Wegera, Klaus-Peter & Prell, Heinz-Peter (22000), “Wortbildung des Frühneuhochdeutschen”. In: Besch et al. (Hrsg.), 1594⫺1604 Wegera, Klaus-Peter (1987), Flexion der Substantive. Heidelberg: Winter (Grammatik des Frühneuhochdeutschen: Beiträge zur Laut- und Formenlehre, Band III) Weinhold, Karl (21883), Mittelhochdeutsche Grammatik. Breslau Wells, Christopher J. (1990), Deutsch: eine Sprachgeschichte bis 1945. Tübingen: Niemeyer (RGL 93) Werner, Otmar (21998), “Historische Morphologie”. In: Besch et al. (Hrsg.), 572⫺596 Wilmanns, Wilhelm (1906⫺1909), Deutsche Grammatik: Gotisch, Alt- Mittel- und Neuhochdeutsch. Straßburg: Trübner Winkler, Gertraud (1995), Die Wortbildung mit -lich im Alt-, Mittel- und Frühneuhochdeutschen. Heidelberg: Winter (Sprache ⫺ Literatur und Geschichte 11) Wolf, Norbert Richard (1981), Althochdeutsch ⫺ Mittelhochdeutsch. Heidelberg: Quelle & Meyer Zutt, Herta (1985), “Wortbildung des Mittelhochdeutschen”. In: Besch et al. (Hrsg.), 1159⫺1164
Hans-Joachim Solms, Halle/S. (Deutschland)
1698
XVIII. Morphologischer Wandel II: Fallstudien
157. From Latin to French 1. 2. 3. 4.
Introduction Inflection Word-formation References
1.
Introduction
There are few languages whose history has been studied in such depth as French. But though there are many excellent studies on single special issues, coverage is by no means complete and, most disappointingly, we still lack a thorough modern synthesis of the evolution of both inflection and word-formation from Latin to Modern French. It goes without saying that this article cannot fill such a gap. For word-formation, it still had to heavily rely on Meyer-Lübke’s elegant monograph of 1921, whose reedition in 1966 by Piel has rightly been the object of severe criticism (see Bork 1968; Höfler 1967; Jänicke 1969). For inflection, the bibliographic situation is more satisfactory, since there exists a series of recent studies that have extended our knowledge of the data or at least reinterpreted wellknown data in the light of modern historical linguistics. But we still lack satisfactory explanations for many changes, though there is no dearth of proposals for any single case. As is well-known, the starting point for the development of the Romance languages is not Classical Latin, but “Vulgar”, i.e. spoken Latin. Nevertheless in this article the Vulgar Latin forms are only used where the difference with respect to Classical Latin is significant for the problem under discussion. The length of Latin vowels is also indicated only where this is crucial for the argument.
2.
Inflection
Ever since Schlegel (1818), the passage from Latin to French has been considered as a paradigmatic case for the evolution from an essentially synthetic language to a predominantly analytic one (cf. Art. 10; Schwegler 1990). Synthetic here means that grammatical categories such as gender, number, case, tense, mood, etc. are encoded in the words themselves via inflections, while analytic refers to situations where these grammatical categories are encoded in separate words or clitics. As we will see below, the analytical tendency has been particularly strong in the
nominal domain, while French verbs have remained closer to the Latin synthetic type, even though the French conjugation is also more analytic than that of other, more conservative Romance languages. 2.1. Nouns and adjectives Latin nouns and adjectives pertained to one of three genders (masculine, feminine, or neuter) and had to be inflected for number (singular, plural) and case (traditionally six: nominative, genitive, dative, accusative, vocative, and ablative). The categories gender, number, and case were represented by one single exponent (portmanteau-morph): thus, e.g., the ending -i of muri solidi, from murus ‘wall’ and solidus ‘solid’, encoded at the same time the categories masculine, singular, and genitive (or, alternatively, masculine, plural, and nominative). This single exponent, furthermore, varied according to the declension class the word pertained to. This Latin nominal system underwent dramatic restructuring at the level of morphological categories on its way to French: while the number categories remained unchanged, the genders were reduced to two (masculine, feminine), and so were the cases in Old French (rectus, oblique), before being totally abandoned during the Middle French period (roughly, the 14th and 15th centuries). The formal expression of the remaining categories has also been heavily changed. Latin, as we have just seen, had three genders, masculine, feminine, and neuter, of which only the first two survived in French. Neuter nouns have joined either the masculine (article: le) or the feminine (article: la) class: Lat. vinum ‘wine’ > Fr. le vin, Lat. gaudium ‘joy’ > Fr. la joie (via the Latin plural gaudia), etc. We can also observe a number of exchanges between masculine and feminine: Lat. porticus (fem.) ‘colonnade’ > Fr. le porche, Lat. dens, -tis (masc.) ‘tooth’ > Fr. la dent, etc. Many of these changes were intended to make the Latin gender system more coherent by establishing a correlation between the final -a and feminine and the final -us (-o in Vulgar Latin) and masculine, a state of affairs still characteristic of most Romance languages. In French, however, the evolution of final vowels has blurred this Vulgar Latin/Romance correlation: on the one hand, due to the deletion of all final vowels
1699
157. From Latin to French
except -a, masculine nouns in -o merged with feminine nouns in -e (cf. V.Lat. flore ‘flower’ > Fr. la fleur, etc.), while on the other hand a final -e resulting from a Latin final -a was also added epenthetically to masculine nouns ending in difficult consonant clusters (cf. V.Lat. comite ‘count’ > conte, etc.). The gender of a non-derived noun is thus highly arbitrary in Modern French ⫺ except for animates, where it tends to coincide with sex ⫺ and only becomes apparent through anaphoric pronouns or agreement with determiners and adjectives. Since number and case were intimately linked both in Latin and in Old French, it is convenient to deal with them together. The six cases of Classical Latin had already been reduced to two, a rectus and an oblique, in the popular speech of most parts of the Roman Empire, among them Gaul. The rectus formally and functionally corresponded to the classical nominative and also served as vocative, while the oblique, which formally goes back to the accusative, took over most functions of the other cases, either in isolation or in combination with a preposition. Prepositional phrases and cases, by the way, coexisted from the earliest times, but the former grew more and more important with the decline in the case system. The cases that resisted best were the most frequent ones, nominative and accusative (cf. Martı´nez Moreno 1993: 112⫺116). Number, on the contrary, was not parallelled by analytic constructions and encoded by inflections throughout the history of Latin. This is essentially the situation that we still find in Old French, where most nouns obey one of the two patterns displayed in Tab. 157.1: singular
plural
rectus oblique rectus oblique m. ‘wall’ f. ‘door’
murs mur porte porte
mur murs portes portes
Tab. 157.1: Case and number in Old French
As one can see, in the most frequent declension classes there was no one-to-one correspondence between form and function, so that listeners had to rely on syntactic context and encyclopedic knowledge in order to correctly determine the syntactic function of a noun. In Middle French, this shaky two-case-
system was abandoned, leading to a system where the final -s was interpreted as a plural marker, along the lines of the feminine: sg. mur/pl. murs, sg. porte/pl. portes. The breakdown of the two-case-system has often been attributed to the deletion of final [s] in the Middle French period, but careful investigations (cf. Schøsler 1984) have shown that the break-down occurred slightly earlier, so that there can be no such causal link. The deletion of the final [s], however, had important consequences for the expression of number in spoken Modern French. While the written language continues to signal plural by attaching -s, most nouns and adjectives have the same form for singular and plural in spoken French: mur and murs are both pronounced [myr], porte and portes both [pcrt]. The only exception to this rule is a closed set of highly frequent irregular nouns and an open set of adjectives in -al, where plural is expressed by alternations ⫺ due to regular sound-change ⫺ as in cheval ‘horse’ [sval]/pl. chevaux [svo], travail ‘work’ [travaj]/pl. travaux [travo], national ‘national’/pl. nationaux, etc. Apart from these cases, plural in spoken Modern French is indicated at the phrasal level, mainly by determiners (cf. la porte [la pcrt]/pl. les portes [le pcrt]), but also by the number of an agreeing verb (cf. leur fils dort ‘their son is sleeping’ [lœr fis dcr] vs. leurs fils dorment ‘their sons are sleeping’ [lœr fis dcrm]) or by “liaison” (cf. leur enfant ‘their child’ [lœr a˜fa˜] vs. leurs enfants ‘their children’ [lœrz a˜fa˜]). The evolution of case and number in adjectives was essentially the same as in nouns. As far as gender is concerned, we also see the reduction to masculine/feminine: Lat. bonus ‘good’ > O.Fr. bons/Lat. bona > O.Fr. bone, etc. In Old French, though, the neuter was still preserved in predicative position after an impersonal pronoun (Lat. bonum > O.Fr. bon). In the main class of adjectives, feminine was signalled by a final -e, and in some cases co-signalled further by a consonantal alternation: V.Lat. sicco > O.Fr. sec/V.Lat. sicca > O.Fr. seche, etc. Furthermore, there were two classes without a gender difference in Old French: the class aigre ‘sour’ (from V.Lat. acro/acra, with an epenthetic -e in the masculine) and the class grant ‘big’ (from V.Lat. grande). During the Old and Middle French period, many members of the third class joined the first by taking -e as a feminine marker (masc./fem. grant > masc. grand/fem. grande, etc.), but at the same time some
1700 members of the first class also joined the second (masc. larc, fem. large > masc./fem. large, etc.). All things considered, however, gender inflection has become more complicated in Old, Middle, and early Modern French, contrary to the general trend towards analyticity. At the written level, Modern French still presents an invariant class and a class characterized by the addition of -e in the feminine. In spoken Modern French, however, the situation has changed dramatically as a consequence of the deletion of the final schwa (-e) and many final consonants: masc. petit ‘small’ [peti]/fem. petite [petit], masc. gris ‘grey’ [gri]/fem. grise [griz], masc. blanc ‘white’ [bla˜]/fem. blanche [bla˜s], etc. This situation has led some linguists to consider the feminine form as the base and derive the masculine forms by deleting the final consonant, a solution that creates more problems than it solves. Semantically, the solution is awkward since the unmarked gender in French is clearly masculine. And formally it makes incorrect predictions with respect to the large group of consonant-final invariable adjectives (cf. masc. clair ‘clear’ [klir]/fem. claire [klir]) and the smaller group of adjectives where the gender distinction is expressed by a consonantal alternation (cf. masc. vif ‘lively’ [vif]/fem. vive [viv], etc.). Popular neologisms like bleuse [bløz] as a feminine form corresponding to invariable bleu ‘blue’ [blø] after the model of masc. creux ‘hollow’ [krø]/fem. creuse [krøz] seem to indicate that speakers take masculine as the unmarked category and consider the feminine to be formed by the addition of a consonant, where there is such a formal difference. Some generative analyses, on the other hand, consider spoken Modern French to be more or less like Middle French at the underlying level, deriving the surface forms by deletion of final schwa and some final consonants. The orthographic final consonant of petit and similar words, by the way, is pronounced not only in the feminine form petite, but also if the following word is vowelinitial (cf. petit enfant ‘small child’ [petit a˜fa˜]). This phenomenon, called “liaison” in French grammar, is dependent in a highly intricate manner on context and register (cf. Encreve´ 1988). 2.2. Determiners In the determiner-system, the most drastic difference between Classical Latin and French is constituted by the introduction of the category (in)definiteness expressed with articles (Heinz 1982).
XVIII. Morphologischer Wandel II: Fallstudien
As in most other Romance languages, the French definitive articles go back to different forms of the Latin demonstrative pronoun ille. Old and Middle French still distinguished rectus and oblique forms in the masculine series (sg. li/le, pl. li/les), while Modern French has generalized the oblique forms (sg. le, pl. les). The feminine counterparts have always been la and les, and before vowel-initial words the singular article has always been simply l’. As far as the function of the definite articles is concerned, it should be noted that in Old French it mainly served to refer to somebody or something previously introduced or given by context or encyclopedic knowledge, while its generic use, its use with proper nouns, appositions and in several other contexts constitute secondary developments. The indefinite articles go back to different forms of the Latin numeral unus ‘one’. Here too, Old French distinguished rectus (uns < Lat. unus) and oblique (un < V.Lat. uno) in the masculine series, of which the oblique form was generalized in Modern French. The feminine has always been une (< Lat. una). In early Old French the indefinite article was restricted to specific reference, but already around 1200 cases of unspecific reference may be found in certain constructions. Before some kinds of nouns, Old and Middle French allowed the plural indefinite articles uns/unes (e.g., uns guanz ‘gloves’), which, however, eventually were replaced by des (cf. Mod.Fr. des gants). This latter form originally had a partitive meaning: des guanz, e.g., meant ‘some of the gloves mentioned’ before it simply came to mean ‘(some) gloves’. Such partitive constructions consisting of the preposition de ‘from’ and the definite article also existed in the singular: manger du pain in Old French meant ‘to eat from the bread in front of us or mentioned before’, while only later, during the Middle French period, it simply came to mean ‘to eat bread’, thus becoming a variant of the indefinite article before mass nouns (du is an amalgamated form of de and le). This so-called “partitive article” is an innovation that sets French apart from most other Romance languages. The development of articles in Romance is often presented as proof of the general trend towards analyticity. This is awkward, since (in)definiteness was not expressed at all in Latin. Furthermore, one should not forget that a Modern French article like la is highly synthetic in itself, since it expresses at the
157. From Latin to French
same time the categories singular, feminine, and definite. The Latin system of demonstratives was characterized by three degrees of distance: near the speaker (hic), near the listener (iste), and further away (ille). While this system is preserved in several Romance languages, in Old French it has been reduced to a two-degree system: cist (< V.Lat. ecce iste) for proximity and cil (< V.Lat. ecce ille) for distance (ecce was an interjection meaning roughly ‘look!’). Both forms could be used as adjectives and pronouns and had special forms for the expression of oblique and plural. The further evolution (cf. Dees 1971) was characterized by a reduction to a one-degree system and by a specialisation of the two formal series in adjectival (cist, Mod.Fr. ce) and pronominal (cil, Mod.Fr. celui) use. In the Middle French period, proximity and distance began to be marked by joining the adverbs ci ‘here’ and la` ‘there’, which have become obligatory in Modern French with non-modified pronouns (celui-ci ‘this one’ vs. celui-la` ‘that one’, etc.), while with demonstrative adjectives these adverbs are still facultative (ce livre ‘this book’, ce livre-ci, ce livre-la`). Latin had the following possessives (only the masculine singular nominative forms are given): 1. meus ‘my’, 2. tuus ‘your’, 3. suus ‘his’ (reflexive), eius ‘his’ (non-reflexive), 4. noster ‘our’, 5. vester ‘your’, 6. suus ‘their’ (reflexive), eorum ‘their’ (non-reflexive). In Vulgar Latin, suus was extended to cover the non-reflexive possessive corresponding to the third person singular and in some languages also to the third person plural, which in the Vulgar Latin of Gaul however was marked with illorum (> Fr. leur). Vester furthermore was changed analogically to voster (cf. Fr. votre, voˆtre) after noster (cf. Fr. notre, noˆtre). All in all, there were few functional changes on the way to Romance. The most important change from Latin to Old French was a formal one, viz. the rise of a stressed and an unstressed series of possessives: unstressed Lat. meus/meum (obl.) became O.Fr. mes/ mon, while under stress it ended up in O.Fr. miens/mien, and similarly for the second and third person singular. In Modern French, the series arising from stressed possessives is only used with the article: mon/ton/son livre ‘my/ your/his book’ vs. le mien/tien/sien ‘mine/ yours/his’, etc. 2.3. Personal, relative, and interrogative pronouns As we have seen in 2.1, with nouns and adjectives the six Latin cases were reduced to two in Old French and have completely disap-
1701 peared in Modern French. Personal pronouns, on the contrary, have preserved three cases ⫺ nominative, dative, and accusative ⫺ up to this day. For the third person feminine, for example, Modern French distinguishes elle (subject, V.Lat. illa), lui (indirect object, < V.Lat. illui), and la (direct object, < V.Lat. la, Cl.Lat. illam). This conservative behaviour is probably due to the high frequency of the personal pronouns that have survived, and shows that for highly frequent units inflexion, especially in the form of suppletion, is the most economical way of achieving brevity (cf. Hunnius 1990: 64). Another conservative trait that distinguishes personal pronouns from nouns and adjectives is the conservation of the neuter in the Old French 3. sg. nom. form el (< V.Lat. illu, Cl.Lat. illud), later on replaced by masculine il (cf. il pleut ‘it is raining’). But the pronoun system also shows some innovative aspects. A French innovation is the indefinite personal pronoun on ‘one’ (cf. on dit ‘one says’), which has still other functions in colloquial French. On goes back to unstressed Lat. homo ‘man’, whose accusative form hominem is the ancestor of French homme ‘man’. Just like in the case of the possessives dealt with at the end of the last paragraph, Latin personal pronouns have also developed differently in Romance in stressed and unstressed positions (cf. Wanner 1987). So, for example, the Lat. 1. sg. acc. pronoun me [me:] remained me [me] in the unstressed preverbal position (cf. vous me regardez ‘You look at me’), while in the stressed final position it was diphthongized to moi (cf. regardez-moi! ‘look at me!’). What distinguishes Modern French ⫺ together with some RhetoRomance and Northern Italian dialects ⫺ from the rest of Romance languages is the fact that it also has a stressed and an unstressed series for subject pronouns. In the other Romance languages ⫺ just like in Latin, and in Old French ⫺ the categories person and number in the verb are expressed by inflectional suffixes, as we are going to see in 2.4, and personal subject pronouns reserved for the expression of emphasis, contrast or disambiguation. But when subject pronouns became obligatory after the Old French period, the stressed direct object pronouns were extended to serve as stressed subject pronouns too. So, O.Fr. chant ‘I sing’ vs. je chant ‘I sing’ corresponds to Mod.Fr. je chante vs. moi, je chante.
1702
XVIII. Morphologischer Wandel II: Fallstudien
Relative and interrogative pronouns also conserved a dative form up to the Old French period (Lat. cui > O.Fr. cui), later fused with the nominative form qui. The rectus/oblique distinction, on the other hand, has been conserved up to the present day in relative pronouns: Lat. qui > Fr. qui, Lat. quem > Fr. que. The Latin neuter interrogative pronoun quid became que in the unstressed position, while under stress it developed into quoi. 2.4. Verbs The Latin conjugation is much better preserved in Romance than the Latin declension (cf. Geckeler 1996: 213⫺220). So its four conjugation classes are all preserved in French, though many verbs changed class in Vulgar Latin or later on. The most spectacular innovation undoubtedly consisted of the replacement of many Latin synthetic categories, especially the less frequent ones, by analytical ones. Synthetic and analytic forms coexisted during many centuries, and if the latter eventually evinced the former this might be due to their greater simplicity, and hence learnability, a big advantage in the Roman Empire, where millions of non-Latin speakers had to learn the official language in precarious conditions (cf. Wuest 1997). Latin verbs, in their finite forms, were inflected for person/number, tense/aspect, mood, and voice. In the following discussion, the fate of these categories in the passage from Latin to French will be outlined. Person and number are realized by a portmanteau-suffix located at the end of the verb in Latin as well as in Romance (see Tab. 157.2). Written Modern French, which is by and large identical to Old French except for the obligatory presence of the personal pronoun, is still very similar to the Latin/PanRomance type, while spoken Modern French has person/number-endings only for the first and second person plural (4 and 5). Latin
1 2 3 4 5 6
cant-o canta-s canta-t canta-mus canta-tis canta-nt
Modern French written
spoken
je chant-e tu chant-es il chant-e nous chant-ons vous chant-ez ils chant-ent
[ze-sa˜t] [ty-sa˜t] [i(l)-sa˜t] [nu-sa˜to˜] [vu-sa˜te] [i(l)-sa˜t]
Tab. 157.2: Person and number in Latin and French verbs
From a functional point of view, it might seem that the obligatory ⫺ if there is no subject noun ⫺ presence of the subject pronoun is due to the ambiguity deriving from the phonetic erosion of the person/number-endings in spoken French, but historians of the French language assure that chronology forbids postulating a relationship of cause and effect between the two. Another interpretation of the facts in Tab. 157.2, very popular in the sixties, also has to be rejected, namely the view that in French the Latin person/ number-suffixes have been replaced by person/number-prefixes: in reality, je, tu, etc. do not have the behaviour of prefixes but of clitics, since they may be separated from the stem by other pronouns, as in je vous la chante ‘I sing it to you’, and have conserved a certain syntactic mobility, as in je dis ‘I say’ vs. dis-je ‘say I’ (cf. Hunnius 1977; Koch 1993). The three-way distinction in past-presentfuture of the Latin tense/aspect-system has been preserved in French. There have been no changes in form (beyond, of course, the effects of regular sound-laws and occasional analogies) or function in the present indicative (see Tab. 157.2) and imperfect indicative (Lat. cantabat > Fr. il chantait, etc.). As far as the perfect indicative (called “passe´ simple” in French grammar) is concerned, the most noteworthy development was its total abandonment in the spoken language. There is no agreement among specialists as to when and why this important change took place. In the spoken language, the passe´ simple has been totally replaced by the passe´ compose´, a Vulgar-Latin and hence pan-Romance innovation. This periphrastic tense has arisen from Latin resultative constructions of the type habeo litteras scriptas ‘I have written letters’, where scriptas ‘written’ originally had an adjectival function before it was reanalysed as part of the verb phrase, with habeo ‘I have’ becoming an auxiliary verb. As a consequence, agreement in gender and number between the original adjective converted into a participle and the following noun was abandoned: j’ai e´crit (and not: e´crites) des lettres (cf. Loporcaro 1998). Of the Latin pluperfect indicative of the type cantaveram some traces can be found in Old French, but it has since been completely replaced by a periphrastic construction of the type j’avais chante´, composed of the auxiliary verb avoir ‘to have’ in the imperfect tense and the past participle. Immediate anteriority in the past is expressed
157. From Latin to French
in literary French by the periphrastic passe´ ante´rieur, a Romance innovation consisting of the passe´ simple of the auxiliary and the past participle: j’eus chante´. If the auxiliary verb is put in the periphrastic past, one gets the passe´ surcompose´: Je suis parti quand j’ai eu termine´ ‘lit. I have gone away when I have had finished’ (cf. Holtus 1984). The Latin future tense of the type cantabo had already been replaced by a periphrastic construction in Vulgar-Latin times, consisting of the infinitive followed by the present tense of habere ‘to have’ (Old French, though, still conserved synthetic future forms for the high-frequency verb ‘to be’: (i)er, (i)ers, etc.). This periphrastic construction was then again amalgamated to a synthetic conjugation, a process already completed in Old French (cf. Kuen 1952): V.Lat. cantare habeo ‘lit. I have to sing’ > V.Lat. cantaraio > O.Fr. chanterai. In Modern French, this synthetic future is again challenged by a new periphrastic future, consisting of the auxiliary verb aller ‘to go’ and the infinitive (je vais chanter ‘lit. I go to sing’). The Latin futurum exactum (past in the future) of the type cantavero has also been abandoned and replaced by a periphrastic construction consisting of the future tense of avoir and the past participle (j’aurai chante´ ‘I will have sung’). A Romance innovation closely linked to the periphrastic future is the conditional, originally a kind of future of the past formed with the infinitive followed by the imperfect tense of habere, but later on amalgamated to a new synthetic conjugation along the lines of the future: V.Lat. cantare habebat > Fr. il chanterait (cf. il a dit qu’il chanterait ‘He said that he was going to sing’). It is probably no coincidence that the only periphrastic forms which amalgamated were the ones where the auxiliary was postponed to the verb. The conditional later also developed a modal use unknown to Latin: il chanterait ‘he would sing (if …)’. As far as mood is concerned, the four Latin subjunctives have been preserved in French as morphological categories, but only the present subjunctive still conserves the Latin forms: Lat. cantem > (que) je chante (O.Fr. chant, the -e is analogical), etc. The imperfect subjunctive in Modern French only survives in very stilted literary language and formally goes back to the Latin pluperfect subjunctive: Lat. cantavissem > Fr. (que) je chantasse, etc. The Latin perfect subjunctive of the type cantaverim has been replaced by a periphrasis composed of the present subjunctive of the
1703 auxiliary and the past participle: (que) j’aie chante´, etc. The pluperfect subjunctive, whose forms, as we have seen, are used as imperfect subjunctive in Romance, is also expressed periphrastically in French, combining the imperfect subjunctive form of the auxiliary with the past participle: (que) j’eusse chante´, etc. It is as marginal in Modern French as the imperfect subjunctive. Of the different Latin imperatives, only the imperative of the singular has conserved the Latin forms: Lat. canta > Fr. chante, etc. The imperative of the second person plural has been replaced by the corresponding indicative ending (cf. Lat. cantate vs. Fr. chantez, etc.), and the negative imperatives are formed with the negative particle ne ... pas: Lat. ne cantaveris, ne cantes, noli cantare vs. Fr. ne chante pas, etc. The Latin future imperative of the type cantato has disappeared completely. Changes have been equally drastic in the domain of infinite verb forms. Latin had a wealth of infinitives: present, perfect, and future, all of them in an active and a passive variant. Of these, only the present active infinitive is conserved both in function and form: Lat. cantare > Fr. chanter, etc. The present passive infinitive has been replaced by a periphrastic form (Lat. cantari vs. Fr. eˆtre chante´), and so was the perfect active infinitive (Lat. cantavisse vs. Fr. avoir chante´). The perfect passive infinitive (cantatus/a/um esse ‘to have been sung’) was already periphrastic in Latin, but has been adapted in French as avoir e´te´ chante´(e), since the formal equivalent eˆtre chante´(e) means ‘to be sung’. The future infinitives both disappeared. Of the three Latin participles, past, present, and future, the latter has disappeared completely. The past participle has been conserved and has even strongly extended its use due to the many periphrastic tenses: Lat. cantatus/a > Fr. chante´(e), etc. The Latin present participle was replaced in Late Latin by the ablative of the gerund, which in French later on was used to translate Latin present participles and so got a participial character (cf. Lausberg 1962: 197): ses mains tremblant de peur ‘his hands trembling with fear’. The Latin gerund is also conserved in the form V.Lat. in cantando > Fr. en chantant. The Latin future participle has disappeared completely, and so have the two forms of the Latin supine. In the category voice Latin distinguished active and passive. Passive was expressed by special inflectional forms in the imperfect,
1704
XVIII. Morphologischer Wandel II: Fallstudien
present, and future tenses, as well as in the infinitive present and perfect, as we have already seen. For the future infinitive as well as the perfect and pluperfect tense, passive was realized periphrastically. In French, as generally in Romance, synthetic passives have disappeared completely, maybe as a consequence of the phonetic erosion of some of the endings (Togeby 1980: 136). Thus, there are also no longer any deponent verbs in French, i.e. active verbs with passive inflection, typical of Latin. Passive now is formed periphrastically with eˆtre ‘to be’ and the past participle, the tense of the auxiliary being identical to the tense of the periphrastic construction: la chanson est/a e´te´/sera/etc. chante´e ‘the song is/was/will be/etc. sung’, etc. Alternatively, in certain contexts, the reflexive construction can be used instead of the periphrastic passive: cette chanson se chante beaucoup ‘lit. this song sings itself very much’, etc.
3.
Word-formation
Word-formation has also undergone deep changes from Latin to French, but it cannot be said that word-formation is more analytic in French than in Latin. Though some patterns have remained more or less unchanged over the centuries, many others have been abandoned or have undergone semantic change or reanalysis, and some French patterns may be considered as innovations (see also Art. 121). From the Middle French period onwards, however, French and Latin word-formation have again become more similar as a consequence of massive borrowing of Latin words on the part of French. Besides Classical Latin, also Neo-Latin, English and, to a lesser extent, other European languages contributed to the renovation of French vocabulary, creating in this way many correspondences with word-formation in Latin and other European languages, to the extent that someone has even spoken of the emergence of a “euromorphology” (Schmitt 1996; Rainer 2002). Another general characteristic that sets Old and Middle French apart from Modern French is the wealth of synonymous derivations in the former stages of the language; this obviously is due to the fact that the language had not yet gained complete geographic homogeneity in those times and that normative pressure was much lower.
3.1. Compounding Latin had relatively few compounds in comparison to Greek or Germanic languages (cf. Lindner 2002). Only very few compounds, furthermore, were passed on to French directly and those that were often lost their transparency on the way, like Lat. undecim ‘eleven’ (unus ‘one’ ⫹ decem ‘ten’) > Fr. onze, etc. Modern French compound types are thus in general later innovations, due in many cases to loan-translations. Only the more important types can be mentioned here. One of the most hotly debated questions in Romance word-formation is the origin of verb-noun compounds of the type essuieglace ‘windshield-wiper’, lit. ‘wipe-windshield’ (cf. Bierbach 1982). Traditional wisdom attributes its origin to a reanalysis, in early Romance, of imperative sentences as proper names, whence the type is supposed to have been extended to common nouns. This developmental sequence, however, is purely speculative, since the first examples, going back to the 11th and 12th centuries in French, already show a great variety of semantic types. It has thus been recently proposed (cf. Bork 1990) to see in this Romance compound type the descendant of the Latin compound type poscinummius ‘money-asking’ (from poscere ‘to ask’ and nummus ‘coin’), which was influenced by the corresponding Greek type. Only sixteen, mostly marginal Latin compound words of this type, however, are documented, and none of them has directly survived in Romance. Noun-noun compounds ⫺ where the first noun is the head, contrary to English ⫺ became more important in French only in the 19th century, but are quite frequent nowadays in the written language. They do not constitute one homogeneous type, but rather a conglomerate of subtypes with varying semantics and productivity. This synchronic heterogeneity reflects the different origin of the subtypes. Hoˆtel-dieu ‘main hospital’, e.g., goes back to an asyndetic Old-French genitive construction meaning ‘lodging (hoˆtel) of god (dieu)’. At least from the 16th and 17th centuries onwards we find copulative compounds of the type imprimeur libraire ‘lit. printer librarian’ or come´die-ballet. The Romantic poets liked compounds with a metaphor as second element: cite´-cadavre ‘a town (cite´) that is like a corpse (cadavre)’, etc. But the great vogue of noun-noun compounds and the diversification of their semantics in the 19th and 20th centuries is attributable to the influ-
157. From Latin to French
ence of English and German, manifest in many loan-translations like e´talon-or ‘gold (or) standard (e´talon)’, etc. A thorough investigation of the history of this compound type is badly lacking. The most frequent nominal compounds in French are the ones formed by noun-preposition-noun or adjective-noun/noun-adjective, like pomme de terre ‘potato, lit. apple-ofearth’, moulin a` vent ‘windmill, lit. mill-bywind’, casque bleu ‘blue helmet, lit. helmet blue’, etc. It seems better, however, to treat such sequences as essentially syntactical constructions. Some subordinative adjective-adjective (or adverb-adjective) compounds may already be found in Old-French (nouveau-ne´ ‘new-born’, clair-voyant ‘clear-sighted’, etc.), but the pattern has remained marginal. The copulative adjective-adjective compound type aigredoux ‘sour-sweet’ was adapted directly from Greek by the Ple´iade poets in the sixteenth century, but never did become very productive in standard French (Hatcher 1951: 11 f.). The copulative type e´thico-moral, on the contrary, is due to adaptations of NeoLatin compounds such as ethico-moralis from the 17th century onwards (Hatcher 1951: 135⫺137). As a consequence of its Neo-Latin origin, the first element of this type of compound is often a learned allomorph, like franco- for franc¸ais, etc. The Vulgar Latin noun-verb compound *manutenere ‘to hold (tenere) with the hand (manu)’ survived in French as maintenir, and so did some other compound verbs of this type (cf. Klingebiel 1989). In Old and Middle French, a few neologisms based on this model like saupoudrer (from sau (Mod.Fr. sel) ‘salt’ and poudrer ‘to strew over’) are attested, but the type never became really productive. Many Latin compounds were borrowed by French in the late Middle Ages and even more Neo-Latin compounds during the Modern Times. Due to the changes in the vernacular language, most such compounds however have remained more or less opaque to the average speaker. While Latin speakers could easily analyse agricola ‘farmer’ into ager ‘field’ ⫹ linking vowel -i- ⫹ colere ‘to till’ ⫹ ending -a, Fr. agricole ‘agricultural’ is much more opaque for speakers of French, where ‘field’ now is champ and ‘to till’ labourer. Where many (Neo-)Latin compounds of the same make-up were borrowed, however, these sometimes have given rise to a
1705 productive pattern in Modern French. Such was the case, e.g., with compounds with a second element mania (Höfler 1972), which entered the French language in loan-translations like de´monomanie, bibliomanie, etc. and then came to be attached even to unequivocally French bases as in bureaumanie, peinturomanie (from peinture ‘painting’), etc. Elements like the -omanie of peinturomanie, however, should better be analysed as suffixes from a synchronic point of view (cf. also Rainer 2003 on the integration of the Latin type aurifer). 3.2. Derivation 3.2.1. Prefixation Latin had a rich system of prefixes, many of them with a clear link to corresponding prepositions or adverbs, which diachronically constitute the most important source for prefixes. It is interesting to note, however, that the opposite evolution, from a prefix to an adverb, is also possible. Lat. trans- ‘across’, e.g., has become the French adverb tre`s ‘very’, and the same process of adverbialisation of a prefix may be observed in Old French par (Lat. per-, cf. permagnus ‘very big’) and re (Lat. re- ‘again, back’, cf. revenire ‘to come back’): par est granz ‘lit. very it-is big’, r-est venuz ‘lit. back he-has come’ (Meyer-Lübke 21966: 139 f.). Due to this close relationship between prefixes on the one side and prepositions and adverbs on the other, it is not always easy to draw a neat borderline between prefixation and compounding. Only relatively few deverbal verb-forming prefixes were passed on directly to French. The most productive ones are re- (< Lat. re-; refaire ‘to do again’, etc.) and reversative de´(s)- (< Lat. dis-; de´faire ‘to undo’, etc.). The others are much rarer: Lat. ad- > Fr. a(apporter ‘to bring; lit. here-carry’, etc.), contra- > contre- (contredire ‘to contradict’, etc.), ex- > e´- (s’e´couler ‘to flow off’, etc.), in- > en- (enfermer ‘to enclose’, etc.), inter- > entre(s’entretuer ‘to kill each other’, etc.), pro- > pour- (poursuivre ‘to persue’, from suivre ‘to follow’, etc.), super- > sur- (surcharger ‘to overload’, etc.). Lat. sub- is only present in some lexicalized forms like (se) souvenir ‘to remember’ (< Lat. subvenire), while in productive formations it has been replaced by sous- (sous-estimer ‘to underestimate’, etc.), the follower of subtus, or by Latinate sub(subdiviser ‘to subdivide’, etc.). Other Lati-
1706 nate deverbal prefixes are post- (postposer ‘to postpone’, etc.), pre´- (pre´voir ‘to foresee’, etc.), super- (superposer ‘to super(im)pose’, etc.), trans- (transporter ‘to transport’, ⫺ the popular outcome tres- is still present in lexicalized tre´passer ‘to die’), and some others. For prefixes in parasynthetic verbs see 3.2.2. Some of the Latinate prefixes just mentioned also occur before nouns and adjectives. Other Latinate denominal prefixes are bis- (bisaı¨eul ‘forefather’), co- (copilote ‘co-pilot’, etc.), ex- (ex-mari ‘ex-husband’, etc.), non- (non-intervention ‘non-intervention’, etc.), vice- (vice-chancelier ‘vice-chancellor’, etc.), etc. As far as the deadjectival prefixes are concerned, one must mention, among others, the negative prefixes in- (incroyable ‘incredible’, etc.) and anti- (anti-atomique ‘anti-atomic’, etc.), as well as the intensive prefixes archi-, extra-, super-, and ultra-. Many such “Latinate” formations, of course, were not borrowed directly from Latin but from other European languages, mainly English. Popular transmission, here too, is much rarer: minus- > me´(s)- (me´content ‘discontent’, etc.), and some others. Some French prepositions and adverbs have also developed prefixal uses: apre`s-midi ‘afternoon’, arrie`re-garde ‘rear guard’, avantgarde ‘vanguard’, contre-sens ‘nonsens’, entremets ‘entremets; lit. between-dishes’, sousofficier ‘non-commissioned officer’. As one can see, such formations may be exocentric (the apre`s-midi ‘afternoon’, for example, is not a midi ‘midday’). There is even one adjective that has been converted into a prefix: medius > mi- (la mi-fe´vrier ‘mid-February’, etc.). All in all, one can see that French prefixation closely resembles that of Latin and other European languages due to intensive borrowing from Latin and other European languages. 3.2.2. Parasynthesis Parasynthesis is a term used in Romance word-formation for what seems to be the simultaneous application of two word-formation rules to a base. Enrichir ‘to enrich’, e.g., where neither *enrich nor *richir are independently existing words of the language, is generally believed to be formed by the simultaneous adjunction of en- and -ir to the base riche ‘rich’. Since -ir marks the infinitive and hence is an inflectional ending, some prefer to define this kind of formation as formed through the simultaneous application of enprefixation and conversion, while others hold
XVIII. Morphologischer Wandel II: Fallstudien
that the whole derivational process is located in the ⫺ then category-changing ⫺ prefix. Under this last hypothesis, parasynthesis is reduced to simple prefixation. Parasynthesis is most prominent in verbs. The type is already present in Old Latin and became particularly frequent in Late Latin (Crocco Gale`as & Iacobini 1993): clarus ‘clear’ > acclarare ‘to clarify’, uncus ‘hook’ > inuncare ‘to hook’, etc. Originally the prefix of such verbs was a preposition, so that the whole could be viewed as a derivation on a phrasal base: per noctem ‘over night’ > pernoctare ‘to stay overnight’, etc. In the course of time, however, such formations were reanalysed as consisting of nominal or adjectival bases “circumfixed” by a combination of a generally desemanticized prefix and conversion. In French, e.g., the prepositional phrase corresponding to cadre ‘frame’ > encadrer ‘to frame’ is (mettre) dans un cadre, not en cadre, that of lune ‘moon’ > alunir ‘to land on the moon’ is (aborder) sur la Lune, not a` la Lune, and in courage ‘courage’ > de´courager ‘to discourage’ or in deadjectival formations such as long ‘long’ > allonger ‘to make longer’ no plausible prepositional phrase is in sight anyway. Parasynthetic verbs have been very productive throughout the history of French. 3.2.3. Suffixation French suffixation is so rich that it will only be possible to mention the most important categories and suffixes here. Nominal derivatives are by far the largest group. In a first subgroup, the base is also a noun and the derivative only implies a slight modification. Collective nouns did not form a well-established category in Latin. French (Baldinger 1950) inherited some Latin suffixes like -alia (Fr. -aille; pierraille ‘crushed stone’) or -eta (Fr. -aie; heˆtraie ‘beech forest’), but most formations in French arose through semantic extensions or reanalysis of other derivational patterns. Such was the case, among others, of -age (feuillage ‘leafwork’, etc.), which goes back to neuter nominalized adjectives in -aticum and of extensions from deadjectival and deverbal abstracts (argenterie ‘silverware’, gouvernement ‘government’, etc.). The characteristic Latin feminine ending -a (alternating with masculine -us) is reflected in French -e: Lat. ursus ‘bear’/ursa ‘she-bear’ > Fr. ours/ourse, etc. A suffix that gained great momentum in Old French was -esse (< Lat.
157. From Latin to French
-issa): comtesse ‘countess’, etc. In the 16th century, however, -eur took over the feminine ending -euse from -eux instead of traditional -eresse (Lindemann 1977). For loan-words, the Latin feminine ending -trix was adapted as -trice. Many feminine forms can be expressed only in a rather roundabout way in French, such as Madame le ministre; feminists, however, insist that the corresponding feminine forms should be imposed. One of the most salient traits of Romance word-formation, inherited from Latin, is its wealth of diminutive suffixes (cf. also Art. 99). Old French fits into this picture, while in Modern French only the diminutive suffix -ette is marginally productive in journalistic and commercial language, less so in spontaneous conversation. The abandonment of diminutive suffixes ⫺ augmentatives have never been very productive in French ⫺ and their substitution with the adjective petit ‘small’ was a process that began in Middle French and came to an end in the 17th century, after a short Renaissance in the poetry of the Ple´iade (cf. Eckert 1986: 284⫺318). The reasons for the agony of diminutives in Middle and early Modern French are difficult to pin down and have been the subject of much debate: homonymy of -et with several other suffixes, after final consonants having become mute, may have been one decisive factor. It should not be forgotten, however, that French has created a new reduplicative diminutive type (see 3.2.5), and that popular French and slang have several “evaluative” suffixes of dialectal origin like -uche (Baldinger 1997), which do not have any denotative meaning but only serve to give the word a particular, often slangy connotation. Another group of denominal nouns is that of status suffixes, i.e. suffixes that express the status of the person referred to by the base. Latin had quite a number of suffixes with that function (consulatus ‘consulate’, etc.), but only very few were directly passed on to French, among them -ı´a (mairie ‘office of mayor’, etc.). French however created new status suffixes like -age (baronnage ‘barony’, etc.) or later on reintroduced some of the Latin suffixes (professorat ‘professorship’, prefecture ‘charge of a pre´fet’, etc.). Many such formations have locative, temporal or collective extensions; so the pre´fecture, nowadays, is more commonly the place where the pre´fet has his office. Personal nouns also constitute an important class of denominal nouns. Among the
1707 names for traditional social roles we find several suffixes that were passed on directly from Latin to French: -ain (< Lat. -anus; chaˆtelain ‘lord of a castle’, etc.), -on (< Lat. -o, -onis; charron ‘cartwright’, etc.), -eron (from -ier ⫹ -on; forgeron ‘smith’, etc.), and, most important, -ier (cf. Roche´ 1998), which arose through an ellipsis in collocations such as faber ferrarius ‘smith’ (Lat. carbonarius > Fr. charbonnier ‘charcoal (charbon) burner’, etc.). In more modern professional spheres, -ier was ousted later on by learned -iste (< Lat. -ista): tractoriste ‘tractor driver’, etc. Equally learned are -ien (< Lat. -ianus; mathe´maticien ‘mathematician’, etc.) and -aire (< Lat. -arius; fonctionnaire ‘civil servant’, etc.). The suffix -iste also commonly refers to the follower of a person or doctrine: marxiste, etc. The same holds for the more colloquial -ard (from Germanic -hard; dreyfusard ‘follower of Dreyfus’, etc.), in one of its several uses. The last important group of denominal nouns is that of instrumental and locative nouns. The borderline between these two categories is not always very neat, especially with nouns referring to containers. We thus sometimes find the same suffix for both subgroups, as with -ier (masc.) and -ie`re (fem.), going back to Lat. -arius and -aria: cendrier ‘ashtray’, cafetie`re ‘coffee-pot’, etc. The gender of the derivative is in most cases the opposite one of that of the base. Purely locative are Old Fr. -oi (< Lat. -etum; sapinoi ‘fir wood’, etc.) and synonymous -aie, which has already been mentioned under the collective suffixes. The most important locative suffix is -ie, originally an extension of the homonymous status suffix: boulangerie ‘bakery’, etc. Through the reanalysis of words like tuilerie ‘tilery’, which could not only be referred plausibly to tuilier ‘tile-maker’ but also to tuile ‘tile’, a new suffix -erie was created. In the class of deverbal nouns we again find many of the suffixes and categories just treated. The category of agent nouns is dominated by -(at)eur (< Lat. -(at)or): vainqueur ‘winner’, exportateur ‘exporteur’, etc. The Latinate allomorph in -ateur is chosen if the corresponding action noun ends in -ation. From the 18th century onwards this suffix has also become by far the most important deverbal instrumental suffix (cf. Spence 1990): condensateur ‘condenser’, etc. The traditional instrumental suffixes like -ail (< Lat. -aculum; e´pouventail ‘scarecrow’, etc.), or -oir (< Lat. -orium; rasoir ‘razor’, etc.) have lost ground.
1708 Here too we find the place/instrument ambivalence noted with respect to the denominal formations: abrevoir ‘watering place’, baignoire ‘bathtub’, etc. The huge category of action nouns has undergone severe changes from Latin to French. The suffix -(at)io, -onis (applied to the past participle stem) survived as -aison, but was later on pushed aside by the learned -ation (cf. Schmitt 1988). Lat. -ura survives in Fr. -ure (ouverture ‘opening’, etc.). Latin action nouns in -us (attached to the past participle stem) survive in the feminine form (due to a reinterpretation as feminine singular nouns of neuter plurals; Alsdorf-Bolle´e 1970): mise ‘putting’, entre´e ‘going in’, sortie ‘going out’, etc. French formations like mise are interesting examples of form-function-anisomorphism: they are action nouns, but have the form of feminine perfect participles. The most productive action noun suffix during the history of French has been -ment, whose Latin antecedent -mentum had been relatively rare in this function until Late Latin. French -ance (naissance ‘birth’, etc.) goes back to another Late Latin creation, -ntia, which was due to a coalescence of the present participle in -nt- ⫹ -ia. Old Fr. -e¨iz (cf. Malkiel 1986), which survives in some Modern French derivatives like clapotis ‘ripple’, and Fr. -age (cf. Fleischman 1977) eventually go back to the Latin adjectival suffixes -at=¯ cius and -aticus. Action nouns in -erie like reˆverie ‘dreaming’ arose when speakers reanalysed denominal formations like reˆveur ⫹ -ie as deverbal (reˆv(er) ⫹ -erie). Another important means of forming action nouns is conversion (see 3.2.4). Latin had a lot of suffixes for forming deadjectival abstract nouns. Unstressed -ia was subject to phonetic erosion and only survives in few words like force (from V.Lat. fortia) as well as with bases going back to present participles (importance ‘importance’, etc.). It has been replaced by the stressed -ı´a, which underlies Fr. -ie: jalousie ‘jealousy’, etc. Deverbal -or, which had already been reanalysed as deadjectival in Latin, was passed on to French as -eur: hauteur ‘height’, etc. Lat. -tas is present both in popular derivatives like bonte´ and in learned ones like latinite´ ‘latinity’. The same is true for -itia, which we find as -esse (richesse ‘richness’, etc.), -ise (franchise ‘exemption’, etc.), and -ice (justice ‘justess’). Lat. -tudo was changed to -tume, as in amertume ‘bitterness’. The extremely polysemous suffix -erie is also found with deadjecti-
XVIII. Morphologischer Wandel II: Fallstudien
val abstracts like pruderie ‘prudery’. Last but not least, -isme (< Lat. -ismus) may also be used to form abstract nouns (conservatisme ‘conservatism’, etc.), besides its many other uses. Adjectival derivatives (cf. Gawełko 1977) may be classified in deadjectival, deverbal, and denominal both in Latin and French. The latter group may again be subdivided into qualitative and relational adjectives, though the suffixes are partly the same. Among the denominal adjectives the most important category is that of relational adjectives. Of the numerous Latin adjectives of this type only relatively few survived as such in Old French: Lat. carnalis > Fr. charnel, Lat. aquaticus > O.Fr. evage, Lat. marinus > Fr. marin, etc. Their decline was probably due, in part at least, to the fact that many adjectives had been converted into nouns. In the course of the centuries, and especially from the 18th century onwards, however, relational adjectives were borrowed massively from Latin, Neo-Latin, and other European languages, so that in present-day, particularly written French they again constitute a very prominent category. The same is true, essentially, for the subcategories of adjectives derived from place names (cf. Wolf 1964) and personal names (cf. Schweickard 1992). Nonrelational denominal adjectives express resemblance (-esque, etc.), possession (-u, etc.), and support (-iste, etc.), among others. Deverbal adjectives have witnessed great changes. Of Lat. -abilis/-ibilis only the former was passed on directly to Old French, while the latter was reintroduced later on through loan-words. The active sense of the suffix (durable ‘durable’, etc.) was more prominent in the old language, while later on it specialized in the passive meaning (faisable ‘feasible’, etc.), evincing completely the competing Old French suffix -(e)is. In Old French, denominal -able was also relatively prominent; this had arisen through the reanalysis of ambiguous formations like taillable ‘obliged to pay the taille (a tax)’, originally derived from the verb tailler ‘to subject to the taille’. Another prominent active deverbal suffix in Old French was -eux (Lat. -osus), later on reduced to a more marginal position by -(at)eur (Lat. -(at)or), which is the most important suffix with an active sense in Modern French, together with -ant (important ‘important’, etc.), which goes back to the Latin present participle. With respect to the originally deverbal -(at)if and -(at)oire it is noteworthy that
157. From Latin to French
many formations were reanalysed as denominal, like circulatoire, which does not mean ‘that circulates’ but ‘related to circulation’. Among the suffixes of Germanic origin, -ard was the one that gained the greatest importance. Deadjectival adjectives are rare in French. The Latin comparative suffix -ior has disappeared and has survived only in several residual forms like majeur (< Lat. maior), while the Latin superlative suffix -issimus has disappeared completely. Both have been replaced by analytic constructions: Lat. grandior ‘bigger’ ⬇ Fr. plus grand, Lat. grandissimus ‘the biggest’ ⬇ Fr. le plus grand, etc. The intensive use of Lat. -issimus, which was so firmly reestablished in Italian or Spanish, never became very productive in French (richissime ‘very rich, etc.). Equally low is the productivity of approximative suffixes like -ot (viellot ‘a bit old’, etc.) or -aˆtre (< Lat. -aster; blanchaˆtre ‘whitish’, etc.). Latin had four adverbial suffixes (cf. Karlsson 1981) and several adverbial phrases consisting of adjective ⫹ noun. None of the suffixes has remained transparent in French, but the adverbial phrase consisting of adjective (in the ablative) ⫹ mente ‘with an adj. mind’ gave rise to the very productive adverbial suffix -ment. In Old French, the possibility of omitting -ment after the first adjectival base of two coordinated adverbs (humble et doucement ‘lit. humb- and softly’) still betrayed the syntactic origin of the suffix. Of the Latin verbal suffixes, -icare and -iare had to succumb to the erosive force of the sound-laws: Lat. caballicare ‘to ride’ > Fr. chevaucher, Lat. acutiare ‘to sharpen’ > Fr. aiguiser, etc. The intensive type cantare (Fr. chanter) also became opaque when the base canere ‘to sing’ went out of use. A verbal suffix, on the contrary, destined to a great future was -idiare/-izare, borrowed from Greek. In the popular transmission it became -oyer, a suffix quite productive in Old French (guerre ‘war’ > guerroyer ‘to make war’, etc.). It has gained much more importance in the modern language since its reintroduction as Latinate -iser ‘-ize’. The suffix -ifier ‘-ify’ has also been reintroduced through latinisms. Besides these denominal and deadjectival suffixes, French also has a small series of deverbal suffixes that modify the Aktionsart of the base: sauter ‘to jump’ > sautiller ‘to make small and repeated jumps’, crier ‘to shout’ > criailler ‘to shout continuously’, tousser ‘to cough’ > toussoter ‘to cough slightly’, etc.
1709 3.2.4. Conversion Conversion, i.e. the transposition of a word from one part of speech to another without formal change, was not unknown to Latin, but became even more important in French. Most conversions yield a noun. Many Latin adjective-noun conversions ⫺ often of elliptical origin ⫺ survive in French as simple nouns, since the corresponding adjective has disappeared: Lat. alba ‘dawn’ (related to albus ‘white’) > Fr. aube, etc. In other cases, the conversion of relational adjectives into nouns has given rise to new nominal suffixes: Lat. -arius > Fr. -ier, Lat. -aticus > Fr. -age, etc. Adjective-noun conversions (cf. Malkiel 1938), of the type le rouge ‘lit. the red (colour)’, l’essentiel ‘lit. the essential (thing)’, le ridicule de sa conduite ‘lit. the ridiculous(ness) of his conduct’ or un curieux ‘a curious (person)’ have gained importance in Modern French, partly as a consequence of the imitation of Latin usage on the part of the Ple´iade poets. As far as verb-noun-conversions are concerned, in Old French the infinitive could freely be used as a noun, while in Modern French the process has become unproductive and only few lexicalized conversions of this type like un baiser ‘a kiss’ have survived. Action nouns formed by conversion were already present in Latin (pugnare ‘to fight’ > pugna ‘a fight’, etc.), but became much more important in French (cf. Lene´ 1899); they can be masculine (soutenir ‘to support’ > le soutien ‘support’, etc.) or feminine (chasser ‘to hunt’ > la chasse ‘hunting’, etc.). Agent and instrument nouns like le guide ‘guide’ (from guider ‘to guide’) or la beˆche ‘rake’ (from beˆcher ‘to rake’) are probably off-shoots of the action nouns. Noun-adjective conversion, which is also highly characteristic of French, is of Latin origin but has been greatly extended in French: une adulte`re ‘an adulteress’/une femme adulte`re ‘an adulterous woman’, une orange ‘an orange’/une robe orange ‘an orange dress’, etc. It is not always easy to distinguish such conversions from copulative noun-noun compounds. There is also a conspicuous number of noun-verb and adjective-verb conversions (-er and -ir are infinitive endings) in French: sel ‘salt’ > saler ‘to salt’, sale ‘dirty’ > salir ‘to make dirty’, etc. Both types are of Latin origin and have given rise to many new formations during the history of French, but they are subject to heavy restrictions. Only
1710
XVIII. Morphologischer Wandel II: Fallstudien
the first conjugation (-er) has remained productive. Adjective-adverb conversions of the type parler haut ‘to speak loudly’ may be traced back to Latin adverbs in -e which came to coincide with adjectives when the final [e] became mute. 3.2.5. Reduplication Latin used reduplication for the formation of the perfect tense of some verbs, but had no reduplicative word-formation rule. Reduplication is also unknown to Romance wordformation, with the exception of Portuguese and especially French, where a diminutive rule using partial reduplication has developed (cf. Rainer 1998). This rule, whose earliest traces may be found at the end of the Middle Ages or the beginning of Modern Times, developed through the imitation of child language in baby talk, but has since spread to colloquial language. It consists of the repetition of the onset and the nucleus of the base, which must always be a monosyllable or be shortened to a monosyllable: patte [pat] ‘paw’ > papatte [papat], Pompidou > Pompom, etc.
4.
References
Alsdorf-Bolle´e, Annegret (1970), Die lateinischen Verbalabstrakta der u-Deklination und ihre Umbildungen im Romanischen. Ph.D. Dissertation, Universität Bonn Baldinger, Kurt (1950), Kollektivsuffixe und Kollektivbegriff: Ein Beitrag zur Bedeutungslehre im Französischen mit Berücksichtigung der Mundarten. Berlin: Akademie-Verlag Baldinger, Kurt (1997), “Das Suffix -uche in der Volkssprache und im Argot”. In: Holtus et al. (eds.), 115⫺147 Bierbach, Mechthild (1982), Die Verbindung von Verbal- und Nominalelement im Französischen. Tübingen: Narr Bork, Hans Dieter (1968), Review of Meyer-Lübke (21966). Romanische Forschungen 80, 421⫺439 Bork, Hans Dieter (1990), Die lateinisch-romanischen Zusammensetzungen Nomen ⫹ Verb und der Ursprung der romanischen Verb-Ergänzungskomposita. Bonn: Romanistischer Verlag Crocco Gale`as, Grazia & Iacobini, Claudio (1993), “Parasintesi e doppio stadio derivativo nella formazione verbale del latino”. Archivio Glottologico Italiano 78, 167⫺199
Dees, Anthonij (1971), E´tude sur l’e´volution des de´monstratifs en ancien et en moyen franc¸ais. Groningen: Wolters-Noordhoff Eckert, Gabriele (1986), Sprachtypus und Geschichte: Untersuchungen zum typologischen Wandel des Französischen. Tübingen: Narr Encreve´, Pierre (1988), La liaison avec et sans enchaıˆnement. Paris: Seuil Fleischman, Susanne (1977), Cultural and Linguistic Factors in Word Formation: An Integrated Approach to the Development of the Suffix -age. Berkeley: University of California Press Gawełko, Marek (1977), E´volution des suffixes adjectivaux en franc¸ais. Breslau: Polskiej Akademii Nauk Geckeler, Horst (1996), “Gemeinromanische Tendenzen II: Flexionslehre”. In: Holtus et al. (eds.), 199⫺222 Hatcher, Anna Granville (1951), Modern English Word-Formation and Neo-Latin: A Study of the Origins of English (French, Italian, German) Copulative Compounds. Baltimore: The Johns Hopkins Press Heinz, Sieglinde (1982), Determination und Repräsentation im Altfranzösischen. München: Fink Höfler, Manfred (1967), Review of Meyer-Lübke (21966). Zeitschrift für romanische Philologie 83, 104⫺114 Höfler, Manfred (1972), Zur Integration der neulateinischen Kompositionsweise im Französischen. Tübingen: Niemeyer Holtus, Günter (1984), “L’emploi des formes surcompose´es dans les varie´te´s linguistiques du franc¸ais et l’attitude des grammairiens”. Französisch heute 15, 312⫺329 Holtus, Günter & Kramer, Johannes & Schweickard, Wolfgang (1997, eds.), Italica et Romanica. Festschrift für Max Pfister zum 65. Geburtstag, Vol. II. Tübingen: Niemeyer Holtus, Günter & Metzeltin, Michael & Schmitt, Christian (1996, eds.), Lexikon der Romanistischen Linguistik, Vol. II.1: Latein und Romanisch: Historisch-vergleichende Grammatik der romanischen Sprachen. Tübingen: Niemeyer Hunnius, Klaus (1977), “Frz. je: ein präfigiertes Konjugationsmorphem? Ein Forschungsbericht zur Frage der Prädetermination”. Archiv für das Studium der neueren Sprachen und Literaturen 224/ 129, 37⫺48 Hunnius, Klaus (1990), “Französisch: Flexionslehre”. In: Holtus, Günter & Metzeltin, Michael & Schmitt, Christian (eds.), Lexikon der Romanis-
157. From Latin to French
1711
tischen Linguistik, Vol. V.1: Französisch. Tübingen: Niemeyer, 59⫺71
ergänzte Auflage von J. M. Piel]. Heidelberg: Winter
Jänicke, Otto (1969), Review of Meyer-Lübke (21966). Vox Romanica 28, 329⫺339
Rainer, Franz (1998), “La re´duplication franc¸aise du type fifille d’un point de vue diachronique”. In: Ruffino, Giovanni (ed.), Atti del XXI Congresso di Linguistica e Filologia Romanza (Universita` di Palermo 18⫺24 settembre 1995). Tübingen: Niemeyer
Karlsson, Keith E. (1981), Syntax and Affixation: The Evolution of MENTE in Latin and Romance. Tübingen: Niemeyer Klingebiel, Kathryn (1989), Noun ⫹ verb Compounding in Western Romance. Berkeley etc.: University of California Press Koch, Peter (1993), “Le ‘chinook’ roman face a` l’empirie: Y a-t-il une conjugaison objective en franc¸ais, en italien et en espagnol et une conjugaison subjective pre´de´terminante en franc¸ais?”. In: Hilty, Gerold (ed.), Actes du XXe Congre`s International de Linguistique et Philologie Romanes, Vol. III. Tübingen, Basel: Francke, 169⫺190
Rainer, Franz (2002), “Convergencia y divergencia en la formacio´n de palabras de las lenguas roma´nicas”. In: Garcı´a-Medall, Joaquı´n (ed.), Aspectos de morfologı´a derivativa del espan˜ol. Lugo: Tris Tram, 103⫺133 Rainer, Franz (2003), “L’inte´gration des compose´s latins du type aurifer en franc¸ais”. In: Fradin, Bernard & Dal, Georgette & Hathout, Nabil & Kerleroux, Franc¸oise & Ple´nat, Marc & Roche´, Michel (eds.), Les unite´s morphologiques. Villeneuve d’Ascq: SILEX, 151⫺168
Kuen, Heinrich (1952), “Rückläufige Bewegungen in der Entwicklung der romanischen Sprachen zum analytischen Typus: dabis ⫺ dare habes ⫺ daras”. In: Festgabe Ernst Gamillscheg zu seinem 65. Geburtstag. Tübingen: Niemeyer, 140⫺163 [reprint: Kuen, Heinrich (1970), Romanistische Aufsätze. Nürnberg: Carl, 72⫺92]
Roche´, Michel (1998), “Deux e´tudes sur la de´rivation en -ier(e)”. Carnets de grammaire, Rapport nº 2. Toulouse: Universite´ de Toulouse-Le Mirail
Lausberg, Heinrich (1962), Romanische Sprachwissenschaft, Vol. III: Formenlehre. Berlin: de Gruyter
Schmitt, Christian (1988), “Funktionale Variation und Sprachwandel: Zum Verhältnis von ererbter und gelehrter Wortbildung im Spanischen und Französischen”. In: Albrecht, Jörn & Lüdtke, Jens & Thun, Harald (eds.), Energeia und Ergon: Sprachliche Variation, Sprachgeschichte, Sprachtypologie, Vol. II. Tübingen: Narr, 183⫺203
Lene´, Gustav (1899), Les substantifs postverbaux dans la langue franc¸aise. Uppsala: Almqvist & Wiksell Lindemann, Margarete (1977), Zum Suffixwechsel von -eresse zu -euse und -trice im Französischen. Tübingen: Narr Lindner, Thomas (2002), Lateinische Komposita. Innsbruck: Institut für Sprachen und Literaturen der Universität Innsbruck Loporcaro, Michele (1998), Sintassi comparata dell’ accordo participiale romanzo. Torino: Rosenberg & Sellier Malkiel, Jacques (1938), Das substantivierte Adjektiv im Französischen. Berlin: Speer & Schmidt Malkiel, Yakov (1986), “The Old French Verbal Abstracts in -e¨iz”. Zeitschrift für romanische Philologie 102, 1⫺39
Schlegel, August Wilhelm von (1818), Observations sur la langue et la litte´rature provenc¸ales. Paris: Librairie grecque-latine-allemande [reprint: (1971), Tübingen: Narr]
Schmitt, Christian (1996), “Euromorphologie: Perspektiven einer neuen romanistischen Teildisziplin”. In: Dahmen, Wolfgang & Holtus, Günter & Kramer, Johannes & Metzeltin, Michael & Schweickard, Wolfgang & Winkelmann, Otto (eds.), Die Bedeutung der romanischen Sprachen im Europa der Zukunft. Tübingen: Narr, 119⫺146 Schøsler, Lene (1984), La de´clinaison bicasuelle de l’ancien franc¸ais. Odense: University Press Schwegler, Armin (1990), Analyticity and Syntheticity: A Diachronic Perspective with Special Reference to Romance Languages. Berlin: Mouton de Gruyter
Martı´nez Moreno, Annette (1993), Sprachwandel und Irregularität: Morphosyntaktische Veränderungen im Bereich französischer Nominalkategorien. Berlin: Schmidt
Schweickard, Wolfgang (1992), “Deonomastik”: Ableitungen auf der Basis von Eigennamen im Französischen: Unter vergleichender Berücksichtigung des Italienischen, Rumänischen und Spanischen. Tübingen: Niemeyer
Meyer-Lübke, Wilhelm (21966), Historische Grammatik der französischen Sprache, Zweiter Teil: Wortbildungslehre [zweite, durchgesehene und
Spence, N. C. W. (1990), “Instrumental -eur/ -aleur/-euse/-abice”. The Modern language Review 85, 29⫺35
1712
XVIII. Morphologischer Wandel II: Fallstudien
Togeby, Knud (1980), “Romance Historical Morphology”. In: Posner, Rebecca & Green, John N. (eds.), Trends in Romance Linguistics and Philology, Vol. I: Romance Comparative and Historical Linguistics. The Hague, Paris, New York: Mouton, 105⫺155 Wanner, Dieter (1987), The Development of Romance Clitic Pronouns: From Latin to Old Romance. Berlin etc.: Mouton de Gruyter
Wolf, Heinz Jürgen (1964), Die Bildung der französischen Ethnica (Bewohnernamen). Geneva: Droz; Paris: Minard Wuest, Jakob (1997), “Zur Entstehung der analytischen Verbformen in den romanischen Sprachen”. In: Holtus et al. (eds.), 31⫺44
Franz Rainer, Vienna (Austria)
158. From Vedic to modern Indic languages 1. 2. 3. 4. 5. 6.
Introduction Old Indo-Aryan Middle Indo-Aryan New Indo-Aryan Transliteration and transcription References
1.
Introduction
Indo-Aryan, among the most richly documented of Indo-European language groups, is represented by Old, Middle and New IndoAryan languages (see Bloch 1965, Masica 1991, Cardona & Jain 2003 for historical surveys and Cardona & Jain 2003, eds. for sketches of individual languages). The oldest documents ⫺ composed centuries earlier than they are attested in writing ⫺ are the Vedas: the R ø gveda, the most ancient, certainly composed before the end of the second millennium B. C. and possibly in the third millennium; the Atharvaveda; the Sa¯maveda; and the Yajurveda, this last in two major groupings, called s´ukla (‘white’) and krø sønø a (‘black’), and various recensions. A variety of Old Indo-Aryan systematically close to a late Vedic stage is described by the grammarian Pa¯nø ini (see Art. 5), who distinguishes between the spoken language (bha¯søa¯) and usage proper to Vedic (chandas). These and other early Indo-Aryan dialects are included in a language termed sam ˙ skrø ta (Sanskrit, abbr. Skt.), the pure or polished language, opposed to other vernaculars, which were considered corruptions (apabhram ˙ s´a) of Sanskrit. These vernaculars make up the group of Prakrits (Skt. pra¯krø ta- ‘natural’), Middle Indo-Aryan languages and dialects (see Pischel 1965; Hinüber 2001). The earliest attested Prakrit (abbr. Pkt.) documents are the inscriptions of the emperor As´oka (3rd c. B. C.), which are
especially valuable because they appear in the different dialects spoken in areas of As´oka’s empire (see Oberlies 2003). The major texts of Therava¯da Buddhism are represented in the mixed language called Pa¯li; Jaina texts also are composed in Prakrit (see Geiger 1965; Hinüber 2001; Oberlies 2001). The exemplar of literary Prakrits, used in both Sanskrit dramas and independent works, is called Ma¯ha¯ra¯søtørı¯. The latest Middle IndoAryan stage is represented by various dialectal forms of Apabhram ˙ s´a (see Bubenik 2003). New Indo-Aryan documents date from approximately the twelfth century, and these languages cover the largest part of the Indian subcontinent, from Asamiya, Bangla, and Oriya in the east, to Gujarati, Marathi, and Konkani in the west. Hindi, spoken in the midlands and used as a lingua franca, is one of the official languages of the Republic of India. Indo-Aryan languages not only have an extremely rich store of literary sources from various eras, they have also been expertly described by grammarians of Sanskrit (see Art. 5) and of Middle Indo-Aryan dialects. The following sketch is meant to give an idea of the major developments in Indo-Aryan. Obviously, a great many details have been omitted, which are covered in grammars and studies of individual languages (for bibliography, see Masica 1991: 486⫺510 and the references given in individual chapters of Cardona & Jain 2003, eds.).
2.
Old Indo-Aryan
Old Indo-Aryan morphology is fairly typical of early Indo-European inflectional languages. It is possible also to trace developments from early Vedic to later stages (see Cardona 2003).
158. From Vedic to modern Indic languages
1713
2.1. Inflectional morphology 2.1.1. Nominal morphology The Old Indo-Aryan nominal system has three numbers (singular, dual, plural), three genders (masculine, feminine, neuter), and up to seven case forms plus a vocative. Pronouns and pronominally inflected adjectives have special characteristics. The dual shows the least number of contrasts, with three distinct forms: ‘nom/acc’, ‘instr/dat/abl’, ‘gen/ loc’. Nominative and accusative are identical for neuters; e.g. sg. phalam, du. phale, pl. phala¯ni ‘fruit’. Ablative and genitive singular are formally identical except for a-stems, and dative and ablative plural are identical for all stems. In addition to a variety of endings, nominals can have vowel alternation (ablaut); there are also nouns with heterogeneous stems. The greatest number of formal contrasts occurs in a-stems (e.g., deva- ‘god’, phala- ‘fruit’). The forms of deva- and the consonant-stem va¯c- are shown in Art. 5 (section 4.2.2, (9)⫺(10)). sena¯- ‘army’, agni- ‘fire’ (m. i-stem), va¯yu- ‘wind’ (m. u-stem), krø ti‘deed’, (f. i-stem), dhenu- ‘cow’, (f. u-stem), nadı¯- ‘river’ (f. ¯ı-stem), vadhu¯- (f. u¯-stem) ‘bride’, pitrø - ‘father’, ma¯trø - ‘mother’, kartrø ‘doer, agent’, ra¯jan- ‘king’, and asthi-/asthan‘bone’ (n. i-/n-stem) will illustrate further, as in Tab. 158.1.
of two’, itarat ‘other’, katarat ‘which of two?’, katamat ‘which of many?’. Demonstrative pronouns have degrees of deixis: idam ‘this’, etat ‘this, that’, tat ‘that’, adas ‘that (yonder)’ (all neuter nominative/ accusative singular). There are alternating stem forms: esøa-, esøa¯-, sa-, sa¯- in the nominative singular non-neuter (m. esøas, sas [esøa, sa before consonant], f. esøa¯, sa¯), eta-, ta- elsewhere, according to the model shown for sarva-. Forms for the proximate and distant demonstratives noted are as follows:
In accordance with Pa¯nø ini’s system, case forms are labeled as follows: 1. nominative, 1 a. vocative, 2. accusative, 3. instrumental, 4. dative, 5. ablative, 6. genitive, 7. locative (cf. Art. 5, section 4.2.2).
Indefinite pronouns are formed with elements postposed to formal interrogatives, as in kas´cit (nom. sg. m.) ‘some one’, kin˜cit ‘some thing’ (nom./acc. sg. n.), ko’pi ‘any one’. Moreover, some pronouns have orthotonic and enclitic alternant forms (see Art. 5: 4.2.2, (29)⫺(30)). Anaphoric pronominal forms such as anena (3rd sg. instr. m./n.), enam (acc. sg. m.) ‘this one’ are also atonic.
Stems such as agni-/agne- (prevocalic agnay-), va¯yu-/va¯yo- (va¯yav-), pitrø - (pitr-) /pitar-, kartrø -/kartar-/karta¯r-, ra¯jan-/ra¯ja¯n-/ra¯jn˜-/ra¯jashow vocalic alternation. The pronominal system includes personal pronouns (see Art. 5: 4.2.2, (29)⫺(30)), interrogative, relative, and demonstrative pronouns. Pronominals in -a, of the type sarva‘all’, ka- (interr.), ya- (rel.) have a particular set of forms: (1) (a) M. sg.: 4. sarvasmai, 5. sarvasma¯t, 7. sarvasmin; pl. 1. sarve, 6. sarvesøa¯m (the rest like deva-, phala-); (b) F. sg.: 4. sarvasyai, 5⫺6. sarvasya¯s, 7. sarvasya¯m; pl. 6. sarva¯sa¯m (the rest like sena¯-). A subset of pronominals in -a have a nominative-accusative singular neuter with -at instead of -am: anyat ‘other’, anyatarat ‘either
(2) (a) M. sg. 1. ayam ‘this’ asau ‘that’, 2. imam amum, 3. anena amuna¯, 4. asmai amusømai, 5. asma¯t amusøma¯t, 6. asya amusøya, 7. asmin amusømin; du. 1⫺2. imau amu¯, 3⫺5. a¯bhya¯m amu¯bhya¯m, 6⫺7. anayohø amuyos; pl. 1. ime amı¯, 2. ima¯n amu¯ n, 3. ebhis amı¯bhis, 4⫺5. ebhyas amı¯bhyas, 6. esøa¯m amı¯søa¯m, 7. esøu amı¯søu; (b) F. sg. 1. iyam asau, 2. ima¯m amu¯m, 3. anaya¯ amuya¯, 4. asyai amusøyai, 5⫺6. asya¯s amusøya¯hø , 7. asya¯m amusøya¯m; du. 1⫺2. ime amu¯, 3⫺5. a¯bhya¯m amu¯bhya¯m, 6⫺7. anayos amuyos; pl. 1⫺2. ima¯hø amu¯hø , 3. a¯bhihø amu¯bhihø , 4⫺5. a¯bhyahø amu¯bhyahø , 6. a¯sa¯m amu¯søa¯m, 7. a¯su amu¯søu; (c) N. 1⫺2 sg. idam adas; du. ime amu¯; pl. ima¯ni amu¯ ni.
2.1.2. Verbal morphology The Old Indo-Aryan verb system distinguishes the following tense and modal forms: present, future, aorist, imperfect, perfect; indicative, optative, subjunctive, conditional. There is a subtype of optative used in wishes (termed precative, benedictive) and a variety of future that originally was periphrastic. Forms called injunctive ⫺ equivalent to unaugmented imperfects and aorists ⫺ were also used. Singular, dual, and plural are distinguished in all persons. There is also a contrast between active and medio-passive. Some verbs are activa tantum or media tantum; e.g., as ‘be’ and a¯s ‘be seated’ have
1714
XVIII. Morphologischer Wandel II: Fallstudien
(a) Singular 1
1a
2
3
4
5⫺6
7
sena¯ agnis va¯yus krø tis dhenus nadı¯ vadhu¯s pita¯ ma¯ta¯ karta¯ ra¯ja¯ asthi
sene agne va¯yo krø te dheno nadi vadhu pitar ma¯tar kartar ra¯jan
sena¯m agnim va¯yum krø tim dhenum nadı¯m vadhu¯m pitaram ma¯taram karta¯ram ra¯ja¯nam asthi
senaya¯ agnina¯ va¯yuna¯ krø tya¯ dhenva¯ nadya¯ vadhva¯ pitra¯ ma¯tra¯ kartra¯ ra¯jn˜a¯ asthna¯
sena¯yai agnaye va¯yave krø tyai/krø taye dhenvai/dhenave nadyai vadhvai pitre ma¯tre kartre ra¯jn˜e asthne
sena¯ya¯s agnes va¯yos krø tya¯s/krø tes dhenva¯s/dhenos nadya¯s vadhva¯s pitur ma¯tur kartur ra¯jn˜as asthnas
sena¯ya¯m agnau va¯yau krø tya¯m/krø tau dhenva¯m/dhenau nadya¯m vadhva¯m pitari ma¯tari kartari ra¯jn˜i asthni
(b) Dual 1⫺2
3⫺5
6⫺7
sene agnı¯ va¯yu¯ krø tı¯ dhenu¯ nadyau vadhvau pitarau ma¯tarau karta¯rau ra¯ja¯nau asthinı¯
sena¯bhya¯m agnibhya¯m va¯yubhya¯m krø tibhya¯m dhenubhya¯m nadı¯bhya¯m vadhu¯bhya¯m pitrø bhya¯m ma¯trø bhya¯m kartrø bhya¯m ra¯jabhya¯m asthibhya¯m
senayos agnyos va¯yvos krø tyos dhenvos nadyos vadhvos pitros ma¯tros kartros ra¯jn˜os asthnos
(c) Plural 1⫺1a
2
3
4⫺5
6
7
sena¯s agnayas va¯yavas krø tayas dhenavas nadyas vadhvas pitaras ma¯taras karta¯ras ra¯ja¯nas asthı¯ni
sena¯s agnı¯n va¯yu¯ n krø tı¯s dhenu¯s nadı¯s vadhu¯s pitrør¯n ma¯trør¯s kartrør¯n ra¯jn˜as asthı¯ni
sena¯bhis agnibhis va¯yubhis krø tibhis dhenubhis nadı¯bhis vadhu¯bhis pitrø bhhis ma¯trø bhis kartrø bhis ra¯jabhis asthibhis
sena¯bhyas agnibhyas va¯yubhyas krø tibhyas dhenubhyas nadı¯bhyas vadhu¯bhyas pitrø bhyas ma¯trø bhyas kartrø bhyas ra¯jabhyas asthibhyas
sena¯na¯m agnı¯na¯m va¯yu¯ na¯m krø tı¯na¯m dhenu¯ na¯m nadı¯na¯m vadhu¯ na¯m pitrør¯nø a¯m ma¯trør¯nø a¯m kartrør¯nø a¯m ra¯jn˜a¯m asthna¯m
sena¯su agnisøu va¯yusøu krø tisøu dhenusøu nadı¯søu vadhu¯søu pitrø søu ma¯trø søu kartrø søu ra¯jasu asthisøu
Tab. 158.1: Nominal system in Old Indo-Aryan
active and middle inflexion, respectively: 3rd sg. pres. indic. as-ti, a¯s-te; 3rd pl. s-anti, a¯sate. Other verbs can have both inflexions, with a semantic difference; e.g., yaj-a-ti ‘ven-
erates, performs a sacrificial rite on another’s behalf’ : yaj-a-te ‘arranges for a sacrificial rite on his own behalf’, kar-o-ti ‘makes for someone else’, kur-u-te ‘makes for himself’ (see
158. From Vedic to modern Indic languages
Art. 5: 4.2.2, (11)). In addition, stems can have ablaut variants; e.g., as/s. Tense stems consist of bases alone or bases with affixes, either suffixed or infixed. In addition to the root type as-ti s-tas s-anti (impf. a¯sı¯t [archaic a¯s] a¯sta¯m a¯san) there are nine types of present-imperfect stems formed with different affixes, as follows (3rd sg., pl.): (3) (a) bhav-a-ti bhav-anti (abhavat abhavan) ‘be, become’: full-grade root with unaccented suffix -a-. (b) juho-ti juhv-ati (ajuhot ajuhavus) ‘offer oblations’, jihı¯t-e jih-ate ‘flee’: reduplicated stem with ablaut. (c) pusø-ya-ti pusø-yanti (apusøyat apusøyan) ‘thrive’: suffix ya. (d) su-no-ti su-nv-anti (asunot asunvan) su-nu-te su-nv-ate (asunuta asunvata) ‘press juice out of’: suffix -no-/-nufollows zero-grade root. (e) tud-a-ti tud-anti (atudat atudan) tuda-te tud-ante (atudata atudanta) ‘goad, wound’: zero-grade root with accented suffix -a-. (f) yu-na-k-ti yu-n˜-j-anti (ayunak ayun˜jan) yu-n˙-k-te yu-n˜-j-ate (ayun˙kta ayun˜jata) ‘join, connect, yoke’ : zerograde root with infixed -na-/-n-. (g) tan-o-ti tan-v-anti ‘stretch’, kar-o-ti: suffix -o-/-u- (see Art. 5: 4.2.2 (11)). (h) krı¯-nø a¯-ti krı¯-nø -anti (akrı¯nø a¯t akrı¯nø an) krı¯-nø ¯ı-te krı¯-nø -ate (akrı¯nø ¯ıta akrı¯nø ata) ‘buy’: zero-grade root with suffix -na¯-/nı¯-/-n-. (i) cor-ay-a-ti cor-ay-anti ‘steal’, kath-aya-ti kath-ay-anti ‘tell relate’: suffix -i-/ -e (-ay- before vowels). Similarly, there are several types of aorist stems, as in the following examples: (4) (a) ada¯-t ad-us (da¯/d ‘give’), abhu¯-t abhu¯v-an (bhu¯ ‘be, become, come into being’): root aorist. (b) apusø-a-t apusø-an (pusø ‘thrive’), agama-t agam-an (gam ‘go’): root followed by stem vowel -a-. (c) acu¯cur-a-t acu¯cur-an (cur ‘steal’), ajı¯jan-a-t ajı¯jan-an ‘brought into being’ (jan ‘be born’): reduplicated aorist with stem vowel -a-, the regular aorist corresponding to presents with -i of the type (3i) and causatives (jani: jan-ay-a-ti ‘brings into being’). (d) anai-sø-ı¯t anai-sø-us, mid. ane-sø-tøa anesø-ata (nı¯ ‘lead’); aka¯r-sø-ı¯t aka¯r-sø-us akrø -ta (1st sg. akrø -sø-i) akrø -sø-ata (krø
1715 ‘do, make’), abhait-s-ı¯t abhait-s-us abhit-ta abhit-s-ata (bhid ‘split’): s-aorist. (e) anam ˙ -s-ı¯t anam ˙ -sisø-us (nam ‘bow, bend’), aya¯-s-ı¯t aya¯-sisø-us (ya¯ ‘go’): stem with -sisø-, regular for bases in -a¯. (f) adhuk-søa-t adhuk-søa-n (duh ‘milk’): stem in -sa-, regular for anitø verbs with final consonants that in combination with a following -s- give a cluster -ksø-, as in -dhuksøa-; this verb originally had an s-aorist (3rd sg. mid. adugdha, 1st sg. mid. adhuksøi) which was the analogic source of this formation. The general future stem is formed with a suffix -sya-, as in da¯-sya-ti ‘will give’, kar-isøya-ti ‘will do, make’. There is also a complex future, with an additional suffix -ta¯s-, of the type: (5) 3rd sg. karta¯ karta¯rau karta¯ (act./mid.); 2nd sg. karta¯si karta¯sthas karta¯stha (act.), karta¯se karta¯sa¯the karta¯dvhe (mid.); 1st sg. karta¯smi karta¯svas karta¯smas (act.), karta¯he karta¯sve karta¯smahe (mid.). The source of this type is a periphrastic formation consisting of an agent noun in -trø and forms of the verb as, except for the third person forms. But this came to have the structure Pa¯nø ini ascribes to it: a suffix -ta¯s- is added. In the language Pa¯nø ini describes, the aorist (e.g., avadhı¯t ‘slew, has slain’) is used for a general past, including the day of reference; the imperfect (e.g., ahan ‘slew’) and perfect (e.g., jagha¯na ‘slew’) are used for a past excluding the day of reference, with the perfect further distinguished by being used if the act in question was not witnessed by the speaker. These contrasts are also found in certain Vedic texts. There is a similar contrast between the future types karisøya¯mi ‘I shall do, make’ and karta¯smi ‘I shall do, make (at some time after today)’. The conditional has an augmented stem with a suffix -sya- followed by secondary endings, of the type ada¯-sya- (3rd sg. ada¯syat ‘would give, would have given’), akar-isøya(akarisøyat ‘would do, would have done’). This is used in contrary-to-fact conditional sentences. Perfect stems are generally characterized by reduplication, ablaut, and a particular set
1716 of endings. For example, the perfect of krø ‘do, make’ is: (6) 3rd sg. caka¯ra cakratus cakrus (act.), cakre cakra¯te cakrire (mid.); 2nd sg. cakrø tha cakrathus cakra (act.), cakrø søe cakra¯the cakrø dø hve (mid.); 1st sg. caka¯ra/cakara cakrø va cakrø ma (act.), cakre cakrø vahe cakrø mahe (mid.). Not all perfects have reduplicated stems. The most notable exception is the perfect of vid ‘know’: veda vidatus vidus. A periphrastic perfect of the type gamaya¯m ˙ caka¯ra ‘caused to go’ is the norm for derived verbs such as causatives. There are also suppletive stems; e.g., abhu¯-t, avadh-ı¯t are the aorists corresponding to the presents as-ti ‘is’, han-ti ‘kills’. Imperative forms have particular endings in second and third person forms. To indicatives with -ti, -anti/-ati, -e (see (3)) correspond imperatives with -u instead of -i and -a¯m instead of -e in active and middle forms respectively. For example: as-tu, s-antu, juho-tu, juhv-atu; sunu-ta¯m, sunv-a¯ta¯m, sunv-ata¯m, sunv-a¯tha¯m, yaj-anta¯m. Second person singular forms have -hi, -dhi, -ø, and the corresponding middle ending is -sva. For example: i-hi ‘go’, juhu-dhi ‘offer oblation’, bhind-dhi ‘split’, paca ‘cook’, sunu ‘press juice’, kuru ‘do, make’; paca-sva, sunu-søva, kuru-søva. The second plural middle has -dhvam; e.g., sunudhvam. Imperatives of the type krı¯-nø ¯ı-hi ‘buy’ (see (3h)) are regular for vocalic bases; consonantal bases have corresponding imperative forms of the type grø h-a¯nø a (grø hnø a¯ti ‘grasps’). First person imperative forms contain -a¯- followed by -ni (1st sg. act.), -vahai, -mahai (1st st. du., pl. mid.), with -ai instead of -e; similarly, the first singular middle has -ai; e.g., bhav-a¯ni ‘I would be’, sunav-ai, sunav-a¯vahai, sunav-a¯mahai. Other imperative forms have secondary endings, as in bhavata¯m, (3rd du.), bhavatha¯m (2nd du.), bhavata (2nd pl.), bhava¯va (1st du.), bhava¯ma (1st pl.). Optative forms to stems in -a are characterized by -e-, and optatives from other stems have active forms with -ya¯-, middle forms with -ı¯(y)-. Third plural forms contain endings -us (act.), -ran (mid.), and -a is the first singular middle ending; otherwise, secondary endings follow the stems. E. g., pac-e-t pacey-us, pac-e-ta pac-e-ran pac-ey-a, sunu-ya¯t sunu-yus, sunv-ı¯-ta sunv-ı¯-ran sunv-ı¯y-a. Precative forms are radical and have endings like those of the optative, but the stem is characterized by -s- in the middle and active endings
XVIII. Morphologischer Wandel II: Fallstudien
have -s- before nonfinal sounds; e.g., su¯-ya¯-t su¯-ya¯s-ta¯m su¯-ya¯s-us, so-sø¯ısø-tøa so-sø¯ıy-a¯sta¯m so-sø¯ı-ran (su ‘press juice’). 2.2. Derivational morphology Early Indo-Aryan has a rich system of deriving bases from nominals and verbs. 2.2.1. Nominal derivation Derived nominals are denominative or deverbative. Nouns formed directly from verb roots include derivates signifying actions, agents, objects, instruments, and loci; e.g., krø -ti- ‘doing, deed’, kar-trø - ‘doer, agent’, karman- ‘action, object’, kar-anø a- ‘doing, means, instrument’ (base krø ), -dha¯nı¯- ‘receptacle’ (base dha¯- ‘put’). Action nouns in -tu supply what are traditionally called infinitives. What is etymologically an accusative singular in -tum (e.g., kar-tum) is the norm in the language Pa¯nø ini describes, but earlier Vedic also had other case forms of such derivates (dat. -tave, abl./gen. -tos). Moreover, forms in -tum alternate with datives of action nouns; e.g., bhok-tum, bhojana¯ya (dat. sg. of bhoj-ana-) ‘to eat’. There are also participles etymologically related to derivates with -tu-, of the type kar-tavya- ‘to be done, which can be done’, which alternate with derivates in -anı¯ya(e.g., kar-anø ¯ıya-) and other suffixes. In addition, there are participles with tense values; e.g., -ta- (krø -ta- ‘done, made’, ga-ta- ‘gone’), present and future participles in -at-/-ant(act.) and -a¯na-, -ma¯na- (med./pass.), such as kur-v-at- (m. nom. sg. kurvan, instr. sg. kurvata¯) ‘doing, making’, kurva¯nø a- ‘id.’, karisøyat- (karisøyan, karisøyata¯ ‘going to do’), and perfect participles in -vas-/-us- (act.), -a¯na- (mid.), such as cakrø vas- (cakrø va¯n [nom. sg. m.], cakrusøa¯ [instr. sg. m./n.]). There are, in addition, indeclinables with -tva¯ or -ya ⫺ the latter in nonnegative compounds ⫺ which are used with reference to an action that is performed prior to another; e.g., krø -tva¯ ‘after making, doing’, ni-søad-ya ‘after sitting’. Several kinds of nominals are derived with secondary derivative affixes (called taddhita). A series of affixes occurs in derivates signifying the property of being what a given nominal base denotes; for example, purusøa-tva-, purusøa-ta¯- signify the property of being a man (purusøa-), and prath-ima- designates the property of being broad (prø thu-). There is a large group of derivates that correspond to phrases of the type X-E Y, with which they alternate, where the values of X-E are case
158. From Vedic to modern Indic languages
forms of nominals and Y stands for a nominal whose meaning is attributed to the derivational affix. For example, there are patronymics like ga¯rgi- ‘son of Garga’, aupagava‘descendant of Upagu’: a case form of ga¯rgior aupagava- corresponds to and alternates with a phrase containing a genitive of gargaor upagu- and a form of putra- ‘son’, apatya‘offspring, descendant’, or a synonym of such a noun. Some taddhita derivates are formed from a more restricted group of nominals. Thus, ta¯vat- (nom. sg. m. ta¯va¯n) ‘this much’, ya¯vat‘as much’, kiyat- ‘how much?’ are formed from a subset of pronouns. Some such derivates correspond to case forms. For example, tatas ‘from that, thence’, tatra ‘in, at that, there’ correspond to and alternate with ablative and locative forms of tad- ‘that’. Other adverbial terms derived from pronouns are tatha¯ ‘thus’, yatha¯ ‘as’, katham ‘how?’. A particularly wide-spread taddhita affix is ka, which serves to form derivates with diminutive and other semantic shadings (e.g., putra-ka- ‘poor little child’, as´va-ka- ‘wretched horse’) and also occurs redundantly, as in avika-, synonymous with avi- ‘sheep’. Sanskrit is like other Indo-European languages in that it forms what are traditionally called comparatives and superlatives of adjectives by means of suffixes. For example, madhumat-tara- ‘quite full of honey’, priyatama- ‘most dear’, gar-ı¯yas- ‘weightier, more prolix (guru- ‘weighty, prolix’), gar-isøtøha‘most weighty, prolix.’ As can be seen, derivates with ¯ıyas, isøtøha are formed from roots, not to adjective stems, and in Vedic the type yaj-isøtøha- ‘who best performs a rite’ is wide spread. Moreover, while eliminating the use of derivates with these suffixes in agentive derivates formed from productive verb roots, Sanskrit developed the type pacati-tara¯m (⫽ pacati-tara-a¯m) ‘cooks very well’, pacatitama¯m ‘cooks extremely well’, with the suffixes -tara, -tama attached to finite verb forms. Compounds are of four general kinds: tatpurusøa (determinative), dvandva (copulative), bahuvrı¯hi (exocentric), and a type that is usually invariant (avyayı¯bha¯va). The first member of a tatpurusøa is regularly equivalent to a case form other than a nominative; e.g. tatpurusøas (nom. sg.), gra¯ma-gatas ‘gone to the village’ are equivalent to tasya purusøahø ‘his (gen. sg. tasya) man, servant’ and gra¯mam ˙ gatas. A subtype of tatpurusøa, called karmadha¯raya, has a first member coreferen-
1717 tial with the second, which it modifies, as in nı¯lotpalam ‘blue (nı¯la-) lotus’, equivalent to nı¯lam utpalam. Like tatpurusøa-, the compound bahuvrı¯hi- is an instance of the class: bahuvrı¯his (nom. sg.) is equivalent to bahur vrı¯hir asya, used of someone who has (asya ‘of this one’) much (bahu-) rice (vrı¯hi-). Dvandvas are equivalent to phrases with ca ‘and’; e.g., ma¯ta¯pitarau (nom./acc. du.) is the equivalent of ma¯ta¯ ca pita¯ ca ‘mother and father’. Avyayı¯bha¯va compounds are generally though not always invariant; e.g., upa¯gni ‘near the fire’. upa¯gni- is an obligatory compound in the sense that the constituents upa and agnes, are not used together in a syntagma that alternates with the compound. Similarly, there are tatpurusøa compounds such as kumbhaka¯ra- ‘pot maker’: the derivate -ka¯ra- does not occur outside of a compound. Certain nominals also have variants that occur only in compounds. For example, ra¯jan- is an nstem, but there is an a-stem -ra¯ja- found in compounds, as in maha¯ra¯ja- ‘great king’. Number words designating units also form compounds with terms for decades. (7) ‘11’ eka¯-das´a (eka- ‘one’) ‘12’ dva¯-das´a (dvi- ‘2’ [archaic nom./acc. du. dva¯]) ‘13’ trayo-das´a (< trayas-das´a) (tri- ‘3’ [nom. pl. m. trayas]) ‘14’ catur-das´a ‘15’ pan˜ca-das´a ‘16’ søo-dø as´a (søasø- ‘6’) ‘17’ sapta-das´a ‘18’ asøtøa¯-das´a (asøtøan ‘8’ [archaic nom./ acc. pl. asøtøa¯]) ‘19’ nava-das´a (also ekona-vim ø s´atis; u¯ navim ø s´atis [‘20 less one’]) ‘20’ vim ø s´atis ‘50’ pan˜ca¯s´at ‘51’ eka-pan˜ca¯s´at ‘52’ dvi-pan˜ca¯s´at ‘53’ tri-pan˜ca¯s´at ‘54’ catuhø -pan˜ca¯s´at ‘55’ pan˜ca-pan˜ca¯s´at ‘56’ søatø-pan˜ca¯s´at ‘57’ sapta-pan˜ca¯s´at ‘58’ asøtøa-pan˜ca¯s´at ‘59’ nava-pan˜ca¯s´at (ekona-søasøtøis) Sanskrit also has iterative sequences like dame` dame ‘in each house, in house after house’. The second element does not have a high-pitched vowel. In pada-pa¯tøhas (analyzed versions) of Vedic texts, such sequences
1718 are treated as compounds, but Pa¯nø ini and other grammarians treat them as sequences that involve repetition of a word. 2.2.2. Verb derivation Derived verbal bases are either deverbative or denominative. To the former group belong causatives, desideratives, and intensives. Causatives are formed with a suffix -i-, which generally conditions substitution by a vowel a¯ ai au in a preceding base. For example, the causatives of yaj ‘venerate, perform a rite’, krø ‘do, make’ are ya¯j-i- (3rd sg. pres. ind. ya¯j-aya-ti), ka¯r-i- (ka¯r-ay-a-ti). In addition, verbs in -a¯ and a small group of other verbs show a final -p preceding the causative suffix, as in stha¯p-i- ‘set up, found’ (stha¯ ‘be in place’), arp-i ‘cause to reach’ (rø ‘go to’). Desideratives have the suffix -sa-, usually with reduplication; e.g., vi-vak-søa- (vivaksøati) ‘wish to speak’. Intensives are generally formed to consonant-initial monosyllabic bases, with the suffix -ya- and reduplication; such bases are inflected medially; e.g., ya¯-yaj-ya- (ya¯yajyate) ‘repeatedly perform a rite’. There are also intensives of the type roravı¯ti ‘roars’, which involve reduplication without a suffix. Denominatives are regularly formed with the suffix -ya-, as in putrı¯-ya- (3rd sg. pres. indic. putrı¯yati) ‘wish for a son, treat as a son’, vrø søa¯-ya- (vrø søa¯yate) ‘act like a bull’, apsara¯-ya- (apsara¯yate) ‘act like an Apsaras’. In addition, verb endings could directly follow nominal bases, without a derivational suffix, a type found mainly in Vedic; e.g., bhisøaj- ‘healer’, bhisøak-ti ‘heals’. 2.3. Vedic features Early Vedic had variants that were eliminated in dialects that make up later Sanskrit. For example, Vedic dialects have a-stem forms of the types vı¯ryaa`¯ ‘heroic power’ (instr. sg.), uß bha¯ ‘both’ (nom./acc. du. m.), deß va¯sa`s ‘gods’ (nom. pl.), a¯yu`dha¯ ‘weapons’ (nom./acc. pl. n.), deß va¯m (gen. pl.). In early Vedic, a distinction is also made between derivative stems of the types vrø kı˘- ‘female wolf’ and devı˘- ‘goddess’. The first type has -s in the nominative singular (vrø kı¯s) and inflects like a consonant stem (e.g., acc. sg. vrø kya`m); the second lacks -s in the nominative singular (deß vı¯) and has alternative stems of the types devı¯-, devya¯-. Forms of the two types were later collapsed into a single type, illustrated by nadı¯- in Tab. 158.1. Sanskrit has particles such as a¯, which cooccur with particular case forms, as in a¯
XVIII. Morphologischer Wandel II: Fallstudien
pa¯tøaliputra¯t (abl. sg.) ‘up to Pa¯tøaliputra’. In Vedic, one also has a¯ redundantly postposed to a case form, as in dißva a¯ ‘from heaven’ (abl. sg. dißvas), aß bhra a¯m ˘˙ ‘in the cloud’ (loc. sg. aß bhre). Vedic shows traces of an aspectual contrast (perfective : imperfective), most clearly seen in negative imperatives of the types ma¯ va`dhı¯s ‘do not slay’ (perfective: 2nd sg. aor. avadh-ı¯s), meant to keep someone from initiating an act, versus ma¯ dı¯vyas ‘stop gambling’ (imperfective: 2nd sg. impfv. adı¯vya-s), meant to have someone cease what he or she is already doing. From earliest attested texts, however, the contrasts are primarily among temporal stems. Some perfect forms in early Vedic reflect the Indo-European use as stative (e.g., biß-bha¯y-a` ‘is afraid’ [3rd sg. pf.]), but from quite early on the perfect too is incorporated into a tense system. The subjunctive (e.g., as-a`-t, 3rd sg. subj. of as ‘be’) gradually goes out of use in Sanskrit, being supplanted by forms of the optative and present and future indicative. In Pa¯nø ini’s description, subjunctive forms are marked as particular to Vedic usage. A remnant of this type persists in first person imperative forms of the type karava¯nø i (1st sg.) ‘I ... do, make’ (e.g., kim ˙ karava¯nø i ‘what shall I do, what may I do?’). Similarly, the earlier injunctive remains in the negative imperative types ma¯ va`dhı¯s, ma¯ dı¯vyas (see above). In some instances, there are different stems which are historically related. For example, in Vedic krø ‘make, do’ has a root aorist of the type aka`r (3rd sg. act. < *a´kart [1st sg. aka`ram]), akra`n (3rd pl. < *a´kr-ant), in which the stem is simply the base (with possible augment), with either full grade (kar) or zero-grade (kr). There is also a type aka`r-at (3rd sg.), developed analogically from the athematic root aorist, in which the stem is invariant: akar-a-, with the stem vowel -a-. Both these stems are found in earliest Vedic texts. Later, however, the productive aorist is sigmatic: aka¯r-sø-ı¯t, with the stem suffix -sand lengthened grade of the base (ka¯r-). Forms of the earlier root aorist persist in the medio-passive; e.g., 3rd sg. akrø -ta ‘did, made (for himself)’ forms part of a paradigm that includes 1st sg. akrø -sø-i and 3rd pl. akrø -sø-ata. In early Vedic also, middle forms like stave ‘is praised’ could function as passives and coexisted with marked passives of the type stu¯-ya-te ‘is praised’. Later, the second type prevailed as the norm for passives.
158. From Vedic to modern Indic languages
A particularly noteworthy contrast between early Vedic and the later language which Pa¯nø ini describes concerns imperative and optative forms. The indicative (3rd sg.), imperative (2nd sg.) and optative (3rd sg.) forms in (8) represent the norm at Pa¯nø ini’s time: (8) (a) sad ‘sit down’: sı¯dati, sı¯da, sı¯det (b) s´ru ‘hear, listen’: s´rønø oti, s´rønø u, s´rønø uya¯t Such forms occur in earlier Vedic, which, however, also has nonindicative forms like 2nd sg. imper. sada, s´rßuß dhi, from the aorist stems. Early Vedic had comparable modal forms also to perfect stems (e.g., sasadya¯t [sad ‘sit’]). This abundance is also to be found in other categories. Thus, there is a large number of accusative, dative, ablative, and genitive forms used as infinitives that Pa¯nø ini explicitly recognizes as proper to Vedic usage only, and he also mentions Vedic absolutives in -tva¯ya, -tvı¯, -tvı¯nam in addition to those in -tva¯. Some formations that were eliminated in what came to be viewed as standard Sanskrit are reflected in Middle Indo-Aryan.
3.
Middle Indo-Aryan
3.1. Inflectional morphology 3.1.1. General Middle Indo-Aryan morphology differs from that of Old Indo-Aryan with respect to the number of contrasts observed. The dual is eliminated as a distinct category, and the use of ablaut variants is severely restricted both in the nominal and verbal systems. The number of nominal cases is diminished. In the verb system also the number of contrasts gradually lessens. The contrast between active and medio-passive endings is gradually eliminated, and the large number of stems within categories is reduced. 3.1.2. Nominal morphology The major differences between Old and Middle Indo-Aryan nominal morphology consists in the reduction of contrasts in the latter. The number of cases formally distinguished is reduced. In general, the dative is replaced by the genitive, except in expressions of purpose. The instrumental and ablative plural merge in the Pa¯li type purisehi ‘men’, which has the ending -hi found in other instrumentals (cf. Old Indo-Aryan -bhis). In addition, the con-
1719 trast between nominal and pronominal endings of the types -a¯t, -e, -a¯ya¯s; -sma¯t, -smin, -asya¯s in Old Indo-Aryan (see 2.1.1) tends to be levelled, so that Pa¯li has nominal forms of the type purisamha¯, purisamhi (purisa- ‘man’, cf. also As´okan vijitamhi ‘in the empire’), and pronominal ta¯ya, comparable to tamha¯, tamhi, and kan˜n˜a¯ya (kan˜n˜a¯- ‘girl’). Moreover, ablaut alternation of the type agni-/agne-, va¯yu-/va¯yo- is gradually eliminated. The loss of final consonants, along with analogic remodelings, results in the replacement of consonant-stems by vowel-stems, with alternants found in early stages. For example, Pa¯li gacchanto ‘going’, mahanto ‘great:nom.sg.m’, manassa ‘mind:gen.sg’, are inflected like Skt. devas, devasya (see Art. 5: 4.2.2, (10)) as opposed to Skt. maha¯n, manasas. These developments continued in other dialects and later times. As a consequence, there are paradigms of the type devo (see Art. 5: 5.2, (38)). According to the descriptions of Pra¯krit grammarians and literary sources, the following are among Pra¯krit forms equivalent to the forms of Sanskrit agni- ‘fire’, va¯yu‘wind’, pitrø - ‘father’, bhartrø - ‘husband’ (see Tab. 158.1; Pischel 1965: 266⫺76). Analogical extensions have been carried out, so that stems like piara-, bhatta¯ra- have been abstracted. The obliteration of formal contrasts continues to develop in later Middle Indo-Aryan, so that, for example, in Apabhram ˙ s´a nominative and accusative forms merged (e.g., dahamuhu ‘ten.faced:nom. sg.m’, caümuhu ‘four.faced:acc.sg.m’), as did the ablative and genitive (e.g., mahu ‘of, from me’), and etymologically instrumental forms could function as locatives, as in da¯hinø abha¯em ˙ ‘in the southern section’, nø iyanø iyatha¯˙˘ ‘in their proper places’. nø ehim m One additional point that is worth mentioning concerns the accumulation of affixes in some case forms. For example, -him ˙ to, -sum ˙ to (see Tab. 158.2(b)) obviously consist of -him ˙ , -su and an appended element -to. In Old Indo-Aryan, -tas was indeed used in forms that alternated with ablatives; e.g., gra¯matas ⫽ gra¯ma¯t ‘from the village’. Moreover, the R ø gveda already shows -tas appended to an actual case form: paß tsuß tas ‘at the feet’, with -tas following a locative plural. The extension of such a procedure has been carried further in Middle Indo-Aryan. Already in Old Indo-Aryan, the affix -ka was used not only in derivates with particular semantic shadings such as diminutive (e.g., putra-ka- ‘little child, poor child’) but also
1720
XVIII. Morphologischer Wandel II: Fallstudien
(a) Singular 1
2
3
5
6
7
aggı¯
aggim ˙
agginø a¯
agginø o/agissa
aggimmi
va¯u¯ pia¯ bhatta¯
va¯üm ˙ piaram ˙ bhatta¯ram ˙
va¯ünø a¯ piunø a¯/piarenø a bhatta¯renø a
aggı¯o/agginø o/aggı¯hi/ aggı¯him ˙ to va¯u¯o/va¯ünø o/va¯u¯him ˙ to
va¯ünø o/va¯üssa piunø o/piussa bhattunø o/bhattassa
va¯ümmi
(b) Plural 1
2
3
5
6
7
aggı¯/aggı¯o/ agginø o va¯u¯/va¯u¯o/ va¯ünø o piaro/piunø o
aggı¯/agginø o
aggı¯hi(m ˙)
agginø a(m ˙)
aggı¯su(m ˙)
va¯u¯/va¯ünø o
va¯u¯hi(m ˙)
va¯ünø a(m ˙)
va¯u¯su(m ˙)
piare/piunø o bhatta¯re/ bhatta¯ra¯/ bhatta¯ünø o
piu¯ nø am ˙/ piaranø a(m ˙) bhatta¯ra¯nø a(m ˙)
piaresu(m ˙)
bhatta¯ra¯/ bhatta¯ünø o
piarehi(m ˙ )/ piu¯hi(m ˙) bhatta¯rehi(m ˙ )/ bhatta¯u¯hi(m ˙)
aggı¯him ˙ to/ aggı¯sum ˙ to va¯u¯him ˙ to/ va¯u¯sum ˙ to piara¯him ˙ to/ piarehim ˙ to bhatta¯ra¯him ˙ to/ bhatta¯rehim ˙ to
bhatta¯resu(m ˙)
Tab. 158.2: Nominal system in Middle Indo-Aryan
with no difference from the base to which the suffix was added; e.g., avi-, avi-ka- ‘sheep’. This redundant affix and its variant -aka- became wide spread in Middle Indo-Aryan. For example, As´okan hakam ˙ ‘I’ is equivalent to aham; as a result of phonological developments, in Apabhram m ˘˙ . ˙ s´a, the form is haüm Similarly, Apabhram ø s´a nominative-accusative forms in -aü (m.) and -aüm m ˘˙ (n.) are developed from such enlarged bases, and feminines in -ı¯ developed from comparable enlarged stems (Old Indo-Aryan -ika¯). These are the sources of New Indo-Aryan -o, -a¯, -u ˙˘ , -ı¯ in variable nouns and adjectives (see m 4.1.1). The Middle Indo-Aryan pronominal system is systematically like that of Old Indo-Aryan (see 2.1.1). The major differences between the two have to do with dialectal variations, phonological changes, and analogic developments. Interrogative and relative pronouns are characterized by k- and j-, the latter due to a phonological change of y- ⫺ still found in early Middle Indo-Aryan (Pa¯li yo (nom. sg. m.) etc.) ⫺ to j-. Personal pronoun forms like Pa¯li amhehi (instr./abl. pl.), amhesu (loc. pl.) ‘us’ (Pkt. amhehim ˙ , amhesu) have amheas in the nom./acc. pl. amhe. Similarly, tumhehi, tumhesu ‘you’ (Pkt. tumhehim ˙ , tumhesu) have tumhe-, as in the nom./acc. tumhe; and
forms with yu- have been eliminated in favor of forms with tu-. Apabhram ˙ s´a again shows a considerable reduction in forms: ˙˘ , 2⫺3⫺7. maı¨m (9) (a) Sg. 1. haüm m ˘˙ , 5⫺6. mahu/majjhu; pl. 1⫺2. amhe/amhaı¨m ˘˙ , 3⫺7. amhehim m m ˘˙ 5⫺6. amhaham ˘˙ , 7. amha¯su; ˙˘ , 2⫺3⫺7. taı¨m ˙˘ /paı¨m (b) Sg. 1. tuhum m ˘˙ 5⫺ 6. tau/tujjha/tudhra; pl. 1⫺2. tumhe/ ˙˘ , 3/7. tumhehim tumhaı¯m m m ˘˙ , 6. tumhaham ˘˙ 7. tumha¯su. Middle Indo-Aryan also has demonstrative pronouns comparable to those of Old IndoAryan: sa-/ta- ‘that’ (e.g., Pa¯li so (nom. sg. m.), sa¯ (f.), tam ˙ (acc. sg. m./f.), te (nom./acc. m.), ta¯ (nom./acc. f.), and so on. Again various phonological and analogical changes account for differences between Old and Middle Indo-Aryan. For example, since final stops are lost, the Old Indo-Aryan contrast between the types Skt. sarvam ‘all’ and anyat ‘other’ is absent; thus Pa¯li an˜n˜a- ‘other’ has an inflection like that of sabba- ‘all’. Pa¯li imam ˙ ‘this:acc.sg.f’ as opposed to Skt. ima¯m shows a regular shortening of -a¯- in final syllable, but forms like imina¯, imamha¯, imassa, imamhi (m./n. sg. 3, 5, 6, 7), imehi, imesam ˙, imesu (m./n. pl. 3/5, 6, 7) and imaya¯, ima¯ya, imissa¯, imissam ˙ (f. sg. 3, 5, 6, 7), ima¯hi, ima¯-
158. From Vedic to modern Indic languages
sam ˙ , ima¯su (f. pl. 3/5, 6, 7) show a generalization of ima-. Similarly, Pa¯li forms like amu¯hi, amu¯sam ø , amu¯su, used for all genders, show a generalization of amu¯- at the expense of amı¯-. In late Middle Indo-Aryan, Apabhram ˙ s´a ˙˘ , jahim has taha¯m m m ˘˙ , jaha¯m ˘˙ , kaha¯m ˘˙ , tahim ˘˙ , ˙˘ equivalent to Old Indo-Aryan ablakahim m tives and locatives tasma¯t ‘from that ...’, yasma¯t ‘from which ...’, kasma¯t ‘from which ...?’, tasmin, yasmin, kasmin. 3.1.3. Verb morphology In its course of development, Middle IndoAryan also eliminated many contrasts found in the Old Indo-Aryan verb system. One now has a system with present, future, and preterit stems. The perfect and the periphrastic future are done away with. In addition, forms with active endings gradually replace forms with middle endings, as in presents like Pa¯li man˜n˜ati ‘thinks’, vuccati ‘is said’; contrast Skt. man-ya-te, uc-ya-te, with the middle ending -te. Ablaut variations are gradually eliminated in favor of single stems; e.g. Pa¯li e-ti (3rd sg.) ‘go-3.sg (goes)’, e-nti ‘go-3.pl (go)’ as opposed to Skt. e-ti, y-anti. In addition, the present stem tends to function as the basis for deriving stems of other verb forms; e.g., Pa¯li gacch-a-ti ‘goes’, fut. gacch-issa-ti, aor. agacch-i (agan˜ch-i), past participle gacch-itaas opposed to Skt. gacch-a-ti, gam-isøya-ti, agam-a-t, ga-ta-, which also have counterparts in Middle Indo-Aryan. In later Middle Indo-Aryan, moreover, the present stems themselves are generally reduced to two prevalent types, in -a- and -e-; e.g., kar-a-i ‘does’, kah-e-i ‘says, tells’ (Pa¯li katheti, Skt. kathayati). Further, these two types could alternate, as in sunø a-/sunø e (sunø a-i, sunø e-i) ‘hear, listen’. In early Middle Indo-Aryan, the prevalent preterital forms were sigmatic (e.g. Pa¯li agacch-is-am ˙ ‘I went’). However, in the course of history, Middle Indo-Aryan tended to eliminate verbally inflected preterital forms in favor of past participles of the type gaya-, gaa- ‘gone’, kaya-, kaa- ‘done, made’ (Skt. ga-ta-, krø -ta-). If such participles derive from transitive verbs, they agree in number and gender with nominals denoting objects, but derivates from intransitives agree with subjects. By the late Middle Indo-Aryan period, this was the norm. Future stems of the Pa¯li type karissati ‘do:fut:3.sg (will do)’ represent phonological outcomes of Old Indo-Aryan -sya- (Skt. kari-
1721 søyati). Due to a phonological change, moreover, later Prakrit has not only a future with -ssa- but also one with -ha-, as in pucchiham ˙ ‘I will ask’. Suppletion of course occurs also in Middle Indo-Aryan. For example, the past participle corresponding to the Prakrit present ja¯- (3rd sg. ja¯-i) ‘go’ is gaya-/gaa-. Moreover, as a result of phonologic and analogic developments, there are other instances where suppletion can justifiably be invoked although etymologically a single root is involved. For example: ne-i ‘leads’ (phonologically < naya-ti) : nı¯-ya- (< nı¯-ta-), sunø a-i ‘hears, listens’ (generalized stem sunø a- < s´rø-nø v-a- [cf. Skr. s´rønø vanti]) : su-ya-. Etymologically, the present and participle forms contain the same base. Middle Indo-Aryan maintains not only indicative but also imperative and optative forms, with varying developments at particular stages in different dialects. For example, Pa¯li gaccha, gaccha¯hi ‘go:imp.2.sg’, gacchatha ‘go:imp.2.pl’, karotu ‘go:imp.3.sg (should go)’, karontu ‘go:imp.3.pl’, gaccheya ‘go:opt. 3.sg’. Middle Indo-Aryan also has passives, again with various dialect forms. Passives like Pa¯li vuccati represent developments of passives with the suffix -ya-. In addition, there are passives with a suffix -iya-/-ı¯ya-, as in Pa¯li har-iya-ti/har-ı¯ya-ti ‘is taken away’, Pkt. pucch-ijja-i ‘is asked’ (pucch-a-ti, puccha-i ‘asks’). Moreover, new passive aorists based on present stems were developed, as in Pa¯li chijjim ˙ su ‘were cut down’ (pres. chijjanti ‘are cut down’). 3.2. Derivational morphology 3.2.1. Nominal derivation Nominal forms derived from verbs are typologically like those of Old Indo-Aryan, with variation: action nouns such as Pa¯li/Pkt. karanø a- ‘doing’; agent nouns such as Pkt. katta¯r(nom. sg. katta¯, acc. sg. katta¯ram ˙ , etc.; see Tab. 158.2); present and past participles (e.g., gacchanta-; cf. 3.1.1), gacchita-, gaya-, gaa-); gerundives (e.g., Pa¯li ka¯tabba- ‘to be done’, dassanı¯ya- ‘worthy of being seen’; Skt. kartavya-, dars´anı¯ya-); absolutives (e.g., Pa¯li nikkham-itva¯ ‘after leaving’); infinitives (e.g., Pa¯li pappotum ˙ ‘obtain’; Pkt. soum ˙ ‘hear, listen’). As can be seen, the rules of distribution differ from those of Old Indo-Aryan: nikkhamitva¯ has -tva¯ after a complex with a preverb and pappo-tum has tum added to the present stem, but Sanskrit has nisøkramya,
1722 with -ya, and pra¯ptum, in which the derivational suffix follows the verb root. In addition, Middle Indo-Aryan has some derivates not found in standard Sanskrit; for example, Prakrit absolutives with -u¯ nø a/-iu¯ nø a, as in da¯u¯ nø a ‘after giving’, genø h-iu¯ nø a ‘after taking’. Apabhram ˙ s´a has a series of different gerundives, absolutives, and infinitives. Hemacandra (4.438⫺441; cf. Vaidya 1958) enumerates the following: gerundives in -ievvaüm m ˘˙ (e.g., karievvaüm m m ˘˙ ⫽ kartvayam), -evvaüm ˘˙ (sa˙˘ ⫽ sodø havyam [sah ‘bear’]), -eva¯ hevvaüm m (soeva¯ ⫽ svapitavyam [svap ‘sleep’]); absolutives in -i (ma¯ri ⫽ ma¯rayitva¯ ‘after first killing’), -iu (adø ohiu ‘without plunging into’), -ivi (cumbivi ⫽ cumbitva¯ ‘kissing’), -avi (vichodø avi ‘letting go of’), -eppi (jeppi ⫽ jitva¯ ‘conquering’), -eppinø u (deppinø u ⫽ dattva¯ ‘giving’), -evi (levi ‘taking’), -evinø u (jha¯evinø u ⫽ dhya¯tva¯ ‘meditating on’); infinitives in -evam ˙ (devam ˙ ⫽ da¯tum ‘give’), -anø a (karanø a ⫽ kartum ‘do’), -anø aham ˙ , -anø ahim ˙ (bhun˜janø aham ˙ , bhun˜janø ahim ˙ ⫽ bhoktum ‘enjoy’), -eppi (jeppi), -eppinø u (caeppinø u ‘abandon’), -evi (pa¯levi ⫽ pa¯layitum ‘maintain’), -evinø u (levinø u). As can be seen, some absolutive and infinitive forms are identical. Moreover, the infinitives in question are etymologically absolutives in origin. For the use of absolutives as infinitives, compare also pievae lagga¯ ‘they began to drink’, with the locative singular of a gerundive of pi ‘drink’ (Bhayani 1953, ed.: 70). The formation of derivates signifying degrees of qualities by means of suffixes continues in Middle Indo-Aryan. For example: Pa¯li piya-tara- ‘dearer’, sat-tama- ‘best’; Pkt. dadø ha-yara- ‘firmer’, pia-ama- ‘dearest’. As in Old Indo-Aryan, so too in Middle Indo-Aryan there are derivates formed from pronouns. For example: Pa¯li ya¯vat- ‘as much’, ta¯vat- ‘that much’, yatha¯ ‘as’, tatha¯ ‘thus’, katham ˙ ‘how?’ and their Prakrit equivalents ja¯vaı¨a-, ta¯vaı¨a-, jaha¯/jaha, taha¯/taha, kaham ˙. Middle Indo-Aryan also has compounds comparable to those of Old Indo-Aryan. E.g. tatpurusøa: Pa¯li aggi-sikha¯- ‘flame of fire’, suvanø nø a-ka¯ra- ‘goldsmith’, Pkt. ahinda- (ahi ⫹ inda) ‘king of snakes’, kumbha-a¯ra- ‘pot maker’; bahuvrı¯hi: Pa¯li rattakkha- (ratta ⫹ akkha) ‘red-eyed’, Pkt. canø dø am ˙ su- (canø dø a ⫹ am ˙ su) ‘(hot-ray, i.e.) the sun’; dvandva: Pa¯li ajeløaka¯- (aja ⫹ eløaka) ‘goats and sheep’, Pkt. amma¯-piyara- ‘parents’; avyayı¯bha¯va: Pa¯li abhi-dosam ‘at evening time’, Pkt. upa-gan˙gam ˙ ‘near the Ganga¯’. One difference between Old and Middle Indo-Aryan compounds is due to phonologi-
XVIII. Morphologischer Wandel II: Fallstudien
cal behavior (see Davane 1956: 144⫺151). Such phonological differences are evident also in other instances. For example, the Apabhram ˙ s´a equivalents of Skt. ya¯drø s´a- ‘of what sort’, ta¯drø s´a- ‘of that sort’ (Pa¯li ya¯disa-, ta¯disa-, Pkt. ja¯risa-, ta¯risa-) are jaı¨sa-, taı¨sa-, and corresponding to the Old Indo-Aryan number words for ‘11’ to ‘19’ (see 2.2.1, (7)), Apabhram ˙ s´a has: (10) ‘11’ ‘12’ ‘13’ ‘14’ ‘15’ ‘16’ ‘17’ ‘18’ ‘19’
ea¯raha ba¯raha teraha caüddaha panø nø araha soløaha dahasatta atøtøha¯raha egunø avim ˙ sa¯
The first vowels of the taddhita derivates Pa¯li nemittika-, Pkt. nø emittia- ‘one who knows the omens’, Pa¯li/Pkt. koma¯ra- ‘juvenile’ are e, o, as opposed to ai, au of the corresponding Sanskrit derivates naimittika-, kauma¯ra-. This is due to the change of ai au to e o. 3.2.2. Verb derivation Causatives generally contain -e- (< -aya-) as in Pa¯li ka¯r-e-ti, Pkt. ka¯r-e-i ‘has ... do’. As there are alternant noncausative present stems with -a/-e (Pkt. sunø a-i/sunø e-i ‘hears, listens’), so are there causatives with -a- and -e-; e.g., pa¯dø e-i/pa¯dø a-i ‘causes to fall’. Further, due to sound changes, consonant alternation came to be linked with causative derivation, as in todø a-i ‘breaks (trans.)’ versus tutøtøa-i ‘breaks (intr.)’; contrast Skt. trutøyati : trotøayati. A type with -a¯p-ay-a- was gaining in productivity from early Middle Indo-Aryan on. Thus, As´oka uses lekh-a¯pi-ta- ‘caused to be inscribed’, and causatives with this extended affix are formed to present stems, as in Pkt. pucch-a¯ve- ‘have ... ask’. In addition, there are alternative causatives in -a¯ve-/-a¯va-, e.g., kar-a¯ve-/ kar-a¯va- ‘have ... do, make’. Apabhram ˙ s´a also has causatives such as bham-a¯dø -a‘cause to turn’. Desideratives and intensives of the Pa¯li types bubhukkha- (3rd sg. pres. bubhukkhati) ‘wish to eat’, Pkt. susu¯ssa(susussaı¨) ‘attend to, serve’ and Pa¯li can˙kama(can˙kamati), Pkt. cakkamma- (cakkammaı¨ ) ‘stride’ reflect the Old Indo-Aryan types bubhuksøa- (bubhuksøate), s´us´ru¯søa- (s´us´ru¯søate), can˙kramya- (can˙kramyate). Denominatives are formed, as in Old Indo-Aryan, with a suffix -ya-, e.g., Pa¯li sukha¯yati, Pkt. suha¯adi ‘is happy’. In addi-
158. From Vedic to modern Indic languages
tion, there are denominatives such that endings follow a nominal directly; e.g., missaı¨ ‘mixes’.
4.
New Indo-Aryan
New Indo-Aryan grammatical systems represent results of developments whose beginnings appear in earlier stages, but the outcome makes this stage considerably different from Old Indo-Aryan. 4.1. Nominal morphology 4.1.1. Nominal system The old system of inflection, whereby a single inflected form signalled both case and number, has been greatly reduced, with an increasing use of postpositions. In a large part of the Indo-Aryan area, there is an opposition between direct forms, used independently, and oblique forms. The latter occur with postpositions and connectives ⫺ postposed elements that function in the manner of genitive endings in the earlier stages ⫺ followed by postpositions. The following examples from Gujarati and Hindi will illustrate: (11) Gujarati (a) chokro ‘boy:dir.sg’ chokrı¯ ‘girl:dir.sg’ chokra¯ (o) ‘boy:dir.pl’ chokrı¯ (o) ‘girl:dir.pl’ (b) chokra¯ ‘boy:obl.sg’ chokrı¯ ‘girl:obl.sg’ chokra¯ (o) ‘boy:obl.pl’ chokrı¯ (o) ‘girl:obl.pl’ (c) chokra¯-ne (obj.) ‘the boy, to the boy’ chokra¯-thı¯ ‘from, by the boy’ chokra¯-ne-ma¯tøe ‘for the boy’ chokra¯-nı¯-pa¯chalø ‘behind the boy’ chokra¯-nı¯-sa¯the ‘with the boy’ chokrı¯-ne ‘the girl, to the girl’ chokrı¯-thı¯ ‘from, by the girl’ chokri-ne-ma¯tøe ‘for the girl’ chokra¯-ni-pa¯chalø ‘behind the girl’ chokrı¯-nı¯-sa¯the ‘with the girl’ (d) chokra¯ (o)-ne chokra¯ (o)-thı¯ chokra¯ (o)-ne-ma¯tøe chokra¯ (o)-nı¯-pa¯chalø chokra¯ (o)-nı¯-sa¯the chokrı¯ (o)-ne chokrı¯ (o)-thi chokrı¯ (o)-ne-ma¯tøe chokri(o)-nı¯-pa¯chalø chokrı¯ (o)-nı¯-sa¯the
1723 (12) Hindi (a) ladø ka¯ ‘boy:dir.sg’ ladø kı¯ ‘girl:dir.sg’ ladø ke ‘boy:dir.pl’ ladø kiya¯m ˘˙ ‘girl:dir.pl’ (b) ladø ke ‘boy:obl.sg’ ladø kı¯ ‘girl:obl.sg’ ˙˘ ‘boy:obl.pl’ ladø kom m ladø kiyom m ˘˙ ‘girl:obl.sg’ (c) ladø ke-ko (obj.) ‘the boy, to the boy’ ladø ke-se ‘from, by the boy’ ladø ke-ke-liye ‘for the boy’ ladø ke-ke-pı¯che ‘behind the boy’ ladø ke-ke-sa¯th ‘with the boy’ ladø kı¯-ko ‘the girl, to the girl’ ladø kı¯-se ‘from, by the girl’ ladø kı¯-ke-liye ‘for the girl’ ladø kı¯-ke-pı¯che ‘behind the girl’ ladø kı¯-ke-sa¯th ‘with the girl’ ˙˘ -ko (d) ladø kom m ˙˘ -se ladø kom m ˙˘ -ke-liye ladø kom m ˙˘ -ke-pı¯che ladø kom m ˙˘ -ke-sa¯th ladø kom m ladø kiyom m ˘˙ -ko ladø kiyom m ˘˙ -se ladø kiyom m ˘˙ -ke-liye ladø kiyom m ˘˙ -ke-pı¯che ladø kiyom m ˘˙ -ke-sa¯th There are also nominals that do not have different direct and oblique forms throughout. For example, Hindi a¯dmı¯ ‘man’ has an ˙˘ , but a¯dmı¯ occurs oblique plural a¯dmiyom m otherwise; a feminine noun such as mez ‘ta˙˘ and an oblique ble’ has a direct plural mezem m plural mezom m ˘˙ . Gujarati kotø (m.) ‘coat’, va¯t (f.) ‘matter, story’, ka¯m (n.) ‘work’ show no variation between direct and oblique forms. The contrast between direct and oblique forms is not always observed elsewhere also. For example, eastern languages like Bangla lack the contrast in the singular. On the other hand, even in such instances, the contrast appears for certain forms. For example, Bangla chele-ra ‘boys’ (nom. pl.), chele-d-er ‘of the boys’, chele-d-er-ke ‘(to) the boys’ (colloquial cheleder). As can be seen, New Indo-Aryan languages agree in the general opposition of direct versus oblique forms and the use of postpositions but differ with respect to the exact forms of the nominals and postpositions. Moreover even where postpositions are the norm, inflectional forms occur. For example, Hindi has a vocative plural of the type ladø ko ‘boys’, ladø kiyo ‘girls’, and Panjabi has abla-
1724 tive singular (e.g., khetom m ˘˙ ‘from the field’) ˙˘ ‘in the and locative plural (e.g., khetı¯m fields’) forms for subsets of nouns. In addition, Gujarati has a plural marker -o-, which can be omitted if plurality is otherwise shown (e.g., by use of a number word) in a context. Further, the connective -na¯-/-ne-/nı¯ can be omitted in Gujarati in certain instances. Comparable differences are seen in other languages. The status of postpositions is like that of the English genitive s in that they regularly follow phrases, as in Gujarati pela¯ ma¯nø as-na dı¯kra¯-ne, Hindi us a¯dmı¯-ke betøe-ko ‘that man’s son’, where the postpositions -ne and -ko follow phrases with oblique forms. In addition, more than one postposition can be attached to a nominal. For example, in Gujarati cphis-e-thı¯ ‘from the office’, Hindi gharmem m ˘˙ -se ‘from the house’, the postpositions -thı¯, -se follow complexes that include the locatival postpositions -e, -mem m ˘˙ . Such usage has parallels in earlier Indo-Aryan (see 3.1.2). Indeed, the use of postpositions goes back to earliest Indo-Aryan (see 2.3), and R ø gvedic samı`ddhasya ... puß rasta¯ t ‘... in front of the lighted (fire)’ (see Bloch 1965: 159) is equivalent to the Hindi type a¯g-ke sa¯mne ‘before the fire’, in which sa¯mne ‘before’ with the connective -ke follows a¯g ‘fire’. What sets the New Indo-Aryan system apart is that it involves direct forms, which occur independently, in opposition to oblique forms that occur with postpositions. In the western area of Indo-Aryan, languages like Gujarati, Marathi, and Konkani have three formal genders; e.g., Gujarati chokro ‘child:dir.sg.m (boy)’, chokrı¯ ‘child:dir.sg.f (girl)’, chokrum m ˘˙ ‘child:dir.sg.n’ (chokra¯m ˘˙ ‘child:dir.pl.n’); adjectives na¯no, na¯nı¯, na¯num m ˘˙ ‘small’. In other areas, two-genders contrast, as in Hindi ladø ka¯ versus ladø kı¯ ‘girl’, chotøa¯, chotø¯ı ‘small’. In the east, on the other hand, languages like Bangla lack such formal gender distinctions. 4.1.2. Pronominal system The New Indo-Aryan pronominal system has reflexes of the earlier types: personal, demonstrative, interrogative, indefinite, relative. E.g., Hindi maim m ˘˙ ‘I’ (mujhe ‘me’, mujh-se ˙˘ ‘us’, ‘from me, by me’, etc.), ham ‘we’ (hamem m ham-se etc.); kaun ‘who?’, kya¯ ‘what?’ (obl. ˙˘ ), koi ‘some sg. kis-, pl. kin-, pl. obj. kinhem m one’ (obl. kisı¯-), jo ‘who, which’ (obl. sg. jis-, ˙˘ ). In some areas, reflexpl. jin-, pl. obj. jinhem m ive pronouns occur obligatorily in certain
XVIII. Morphologischer Wandel II: Fallstudien
syntactic configurations; e.g., Hindi apna¯ (obl. pl. apne), apnı¯. In Indo-Aryan, as in many other areas of Indo-European, personal pronouns retain morphological complexities that are eliminated elsewhere. Thus, Hindi has maim m ˘˙ , also ˙˘ -ne, opposed to used in the agentive maim m mujh-, and the derivate equivalent to a noun with -ka¯ is mera¯ (f. merı¯): mera¯ betøa¯ ‘my son’, merı¯ betø¯ı ‘my daughter’, mere liye ‘for me’. Correspondingly, tum ‘you’ pairs with tumha¯ra¯ (tumha¯rı¯ ) ‘your’. Gujarati hum m ˘˙ ‘I’, mε˜ (agentive), ma¯rum m ˘˙ and tame ‘you’ (also ˙˘ has comparable variants. agentive), tama¯rum m New Indo-Aryan languages also show the effects of regularization. For example, Old Gujarati had forms such as the accusative singulars mujha ‘me’, tujha ‘you’, which have been replaced by mane, tane in modern Gujarati. New Indo-Aryan languages differ with respect to degrees of deixis; e.g., Hindi yah (obl. sg. is-, pl. in-) ‘this’, vah (obl. sg. us-, pl. un-) ‘that’; Bangla e ‘this one’ (obj. sg. e-ke, gen. sg. e-r), nom. pl. e-ra¯, obj./gen. pl. e-der), se [se] ‘he, she it’ (obj. sg. ta¯-ke, gen. sg. ta¯r, nom. pl. ta¯-ra¯, obj./gen. pl. ta¯-der), o ‘that one’ (o-ke, o-r, o-ra¯, obj./gen. pl. o-der); Gujarati a¯ ‘this’, e (agentive sg. inø e, pl. imnø e), te ˙˘ (m. (agentive sg. tinø e, pl. timnø e), ‘that’, pelum m pelo, pela¯, etc.) ‘yonder one’. Languages of the west also distinguish inclusive and exclusive first person plural pronouns, as in Gujarati ame ‘1.pe’ versus a¯pnø e ‘1.pi’ ‘we’; Marathi a¯mhi, a¯panø . In general, modern Indo-Aryan observes a contrast of degrees of respect or distancing conveyed by the use of different second person pronouns: honorific, ordinary, intimate. Thus, Hindi tum is a non-distancing pronoun used towards close friends, one’s wife and such; a¯p is the more generally used, respectful pronoun; and tu¯ is intimate. Similarly, Bangla has three degrees of formality represented in the use of tui, tumi, a¯pni, but these are singular pronouns with plural counterparts (tora¯ tomra¯ a¯pna¯ra). In Gujarati, on the other hand, tame is the pronoun generally used, tum m ˘˙ is used as the intimate pronoun, and a¯p is not generally used. If someone is addressed with this, it implies either extreme respect, even fawning, or that one is making fun of the person. The contrast between ordinary and honorific can be shown by formal contrasts for the third person also, as in Bangla: neutral se [se] versus formal tini (pl. ta¯ra, ta¯m ˘˙ ra¯). Even in those languages that do not have comparable formal distinctions among
1725
158. From Vedic to modern Indic languages
third person pronouns, there are strategies to signal the difference: plural forms are used ⫺ including the neuter plural for respectful reference to a woman (as in Gujarati) ⫺ along with plural agreement in verb forms that take gender-number affixes, and an honorific second person pronoun can be used, as in Hindi (a¯p). 4.2. Verb morphology The sources of the verbal systems of New Indo-Aryan languages are of course found in late Middle Indo-Aryan (Apabhram ˙ s´a) and early stages of New Indo-Aryan. By the late Middle Indo-Aryan period, as noted earlier (see 3.1.3), verbally inflected preterit forms had been eliminated in favor of nominal forms. Nevertheless, present indicative forms of the types karaı¨ ‘does’, kaheı¨ ‘says’ continued in use, right down into the early New Indo-Aryan period, and such forms are still used in some languages as presents (e.g., see (18)). In the course of development, however, these too came not to be used as present indicative forms. For example, Gujarati and Hindi have the type kar-i, kar-e ‘might do, may do’ (Old Gujarati, Old Hindi kar-a-i), but this does not function as a present indicative. In effect, the system represented in the greater part of New Indo-Aryan languages is based on a contrast between imperfective and perfective, along with temporal auxiliary forms. The following Gujarati examples will illustrate: ˙˘ kar-i (13) (a) mohan s´um m Mohan what do-3 ‘what might Mohan do?’ ˙˘ kar-i ch-i (b) mohan s´um m Mohan what do-3 aux-3 ‘what is Mohan doing?/what does Mohan do?’ ˙˘ kar-t-o (14) (a) mohan s´um m Mohan what do-ipfv-m.sg ha-t-o aux-ipfv-m.sg ‘what was Mohan doing?’ (b) mohan-e s´um m m ˘˙ kar-y-um ˘˙ Mohan-agt what do-pfv-n.sg ‘what did Mohan do?’ The imperfective stem kar-t- is followed by a gender-number element, as is the temporal auxiliary stem ha-t-. The imperfective stem kar-t- appears also in negative sentences of the type:
˙˘ (15) mohan kas´um m nathı¯ kar-t-o Mohan something neg do-ipfv-m.sg ‘Mohan is doing nothing at all’ but in positive sentences of the type (13 b), one has what I call the neutral form kar-i with the temporal auxiliary form ch-i. The ergative type (14 b) has the perfective stem kary- along with a gender-number element. The imperfective and perfective markers appear also in jam-t- and kar-y- of examples such as in (16). (16) (a) jam-t-ı¯ wakhat-e eat-ipfv-f time-loc ‘at the time to eat, at dining time’ (b) ka¯m kar-y-a¯ pachı¯ work do-pfv-obl after ‘after doing the work’ Other languages have comparable system, with variations. Thus, Hindi distinguishes a continuous from a non-continuous, as in (17). (17) (a) mohan kya¯ ka¯m kar-t-a¯ Mohan what work do-ipfv-m.sg h-ai aux-3.sg ‘what work does Mohan do?’ (b) mohan kya¯ kar rah-a¯ Mohan what do remain-pfv.m.sg h-ai aux-3 ‘what is Mohan doing?’ Although modern Hindi has the type kar-e, it does not use this in conjunction with a temporal auxiliary to form present forms, as in Gujarati. Habitual and continuative presents contrast elsewhere also, in different ways. For example, in Bangla personal endings follow the base or the affix -ch-: (18) (a) tumi ki kar-o [kcro] you what do-2.neutral ‘what do you do?’ (b) tumi ki kar-ch-o [kcrcho] you what do-ipfv-2.neutral ‘What are you doing?’ The ergative type (14 b) also shows variations in different languages with respect to agreement phenomena. For example, contrast the following examples from Gujarati (19) and Hindi (20): (19) mi˜ tama¯r-ı¯ dı¯kr-ı¯-ne jo-ø-ı¯ 1.sg.agt your-f child-f-obj see-pfv-f ‘I saw your daughter’
1726 (20) maim m betø-ı¯-ko ˘˙ -ne a¯p-kı¯ 1.sg-agt you-poss.f child-f-obj dekh-a¯ see-pfv.m.sg ‘I saw your daughter’ In (19), the perfective jo-ı¯ (I jo-y-ı¯ ) is feminine, agreeing with dı¯kr-ı¯-ne, but in (20) the verb agreement is suspended, since betø-ı¯-ko has the postposition -ko in the object phrase. The agreement appears in an example like (21), where the object phrase kita¯b ‘a book’ does not have -ko. (21) maim m ˘˙ -ne kita¯b padø h-ı¯ 1.sg-agt book read-f ‘I read a book’ In languages of the east, on the other hand, perfective forms show person number agreement of non-ergative systems. The following illustrate the difference (Gujarati, Hindi, Bangla): (22) (a) hum m ˘˙ gayo/gaı¯ ‘I went’ ame gaya¯/gaı¯ ‘we went’ tame gaya¯/gaı¯ ‘you went’ e gayo/gaı¯ ‘he/she went’ e gaya¯/gaı¯ ‘they went’ (b) maim m ˘˙ gaya¯/gaı¯ ‘I went’ ˙˘ a¯p gae/gaı¯m ˙˘ ‘we went’ ham gae/gaı¯m vah gaya¯/gaı¯ ‘he/she went’ ve gae/gaı¯m ˘˙ ‘they went’ (c) a¯mi/a¯mra¯ gelum (gila¯m) ‘I/we went’ tumi/tomra¯ gele ‘you went’ se [se]/ta¯ra¯ gela [gilo] ‘he/she/they went’ The perfectives gayo, gai, gaya, gaı¯, and so on in (22 a-b) agree in gender and number with the subjects, but in (22 c) the verb forms differ according to person; similarly, Bangla a¯mi karlum [korlum] and so on from the base kar ‘do, make’. In addition, New Indo-Aryan languages have future stems with various formants. For example, Gujarati has a future with -s´- that shows person-number differences: kar-is´ ‘I shall do’, kar-s´-i ‘he/she/they will do’, kar-is´um m ˘˙ ‘we will do’, kar-s´-o ‘you will do’. Bangla kar-b-o [korbo] (1st person) kar-b-e [korbe] (ordinary 2/3 person) ‘will do’ has person markers following the future suffix -b-. Hindi, on the other hand, has a future with -g- that shows not only person-number but also gender differences: sg. kar-um m ˘˙ -ga¯-/ı¯ ‘I shall do’, kar-e-ga¯/ı¯ ‘he/she shall do’, pl. karem m ˘˙ -ge/ı¯. Forms of the type Hindi kar-e, Gujarati kar-i have modal functions. They are used,
XVIII. Morphologischer Wandel II: Fallstudien
for example, in the protasis of a conditional sentence as well as in sentences like (23) from Hindi (23 a) and Gujarati (23 b). (23) (a) maim m kya¯ kar-um m ˘˙ ˘˙ 1.sg.agt what do-1.sg ‘what am I to do?’ (b) hum m s´um m m ˘˙ ˘˙ kar-um ˘˙ 1.sg.agt what do-1.sg ‘what am I to do?’ There are also forms used in contrary-to-fact conditional sentences, as in the following examples from Gujarati (24 a) and Hindi (24 b): (24) (a) ma¯rı¯-pa¯se paisa¯ hc-t to me-by money be-cond then tama¯rı¯-sa¯the ja¯-t you-with go-cond (b) agar mere-pa¯s paise ho-t-e if by me money be-cond-m.pl to a¯p-ke sa¯th ja¯-t-a¯ then you-obl with go-cond-m.sg ‘If I had money I would go with you.’ In Gujarati, forms of the type hct, ja¯t have -t following a verbal base; in Hindi, -t- is followed by a gender-number affix: m. sg. hota¯, pl. hote. The forms in question, however, can be neutral with respect to time reference, as is the Sanskrit conditional of the type abhavisøyat. Historically, the forms with -t- reflect the use of present participles in conditional sentences, a usage that goes back to late Middle Indo-Aryan. In addition, New Indo-Aryan languages have formally distinct imperatives. For example, the imperatives of the bases baitøh, bas [bcs/bos], bes [bis] ‘sit’ linked to the use of the pronouns Hindi tu¯ tum a¯p, Bangla tui tumi a¯pni, Gujarati tum m ˘˙ tame are: (25) (a) Hindi: baitøh, baitøh-o, baitøh-iye (b) Bangla: bas, bas-o, bas-u¯ n (c) Gujarati: bis, bis-o The first consists simply of the base; others have particular affixes, as shown. In addition, Hindi allows the use of imperatives in -o with the pronoun a¯p: a¯p a¯o. Various other forms function as imperatives, with semantic shadings. For example, Hindi has imperatives with -iega¯, as in a¯-iega¯, which are used to attenuate the feeling of command, and Gujarati has forms with -je, -jo (the latter more deferential), as in a¯v-je, a¯v-jo; the last cited form is also the common expression for ‘goodbye’.
158. From Vedic to modern Indic languages
In addition, there are third person imperative forms, e.g., Bangla kar-uk, Gujarati kar-o ‘... should do’. Moreover, some languages distinguish tenses for imperatives; thus, the Bangla forms in (25 b) are present imperatives, and the corresponding future imperatives are: kar-is, kar-o, kar-ben. The very polite imperative Gujarati type kar-s´-o-ji ‘(you will) please do’ contains a formal future. As in earlier stages, there is suppletion in New Indo-Aryan, and the situation is comparable to that of Middle Indo-Aryan (see 3.1.3). Thus, the Hindi preterit forms tha¯ the ˙˘ (m. sg., pl., f. sg., pl.) of the temporal thı¯ thı¯m auxiliary correspond to present forms with h-: hu¯m ˘˙ (1st sg.), hai (2nd sg. [ti¨], 3rd sg.), ho ˙˘ (2nd [a¯p], pl.); in Gujarati, (2nd [tum]), haim m the preterit forms corresponding to the present ch- (hum m m m ˘˙ chum ˘˙ , tum ˘˙ chi, e chi, ame/a¯pnø e chie, tame chc, e chi) are hato, hata¯, hatı¯, ˙˘ . The following examples of infinitives hatum m and perfectives will serve further to illustrate: (26) (a) Hindi de-na¯ : di-ya¯ ‘give’ ja¯-na¯ : ga-ya¯ ‘go’ (b) Gujarati ˙˘ : dı¯-dhum a¯p-vum m m ˘˙ ‘give’ ja-vum m m ˘˙ : ga-yum ˘˙ ‘go’ As expected, levelling has also taken place; for example, in standard Gujarati the perfec˙˘ ‘see, look’ is jo-yum ˙˘ , though tive of jo-vum m m ditøtøhum m ˘˙ occurs in dialects. New Indo-Aryan languages have passive formations, but these occupy a special status, for two reasons. Passive sentences that correspond to actives usually involve obligatory omission of an agent expression. Passive sentences with agent expressions have particular semantics: that one is or is not capable of doing something, that one has done something accidentally, and are most commonly negative. Moreover, passive formations are not limited to transitive verbs. Passives are formed both with particular affixes and by periphrasis. For example, Gujarati has a passive suffix -a¯-, as in kar-a¯- ‘be done’, av-a¯(passive to a¯v- ‘come’). In Hindi, on the other hand, a passive consists of a perfective form of the verb followed by a form of ja¯ ‘go’, as in kiya¯ ja¯-, cala¯ ja¯- (cal-na¯ ‘move’). 4.3. Derivation 4.3.1. Nominal derivation New Indo-Aryan languages have a large variety of nominals derived from verbs, such as Hindi ladø -a¯¯ı ‘fighting’ (ladø -na¯ ‘fight-inf’), bah-a¯v ‘flow’ (bah-na¯ ‘flow-inf’), ruk-a¯vatø
1727 ‘obstruction’ (rok-na¯ ‘stop-inf’). Of particular interest are derivates that are generally classed as infinitives and absolutives. The examples from Hindi (27 a) and Gujarati (27 b) both of which mean ‘Allow him to go’, have the oblique forms ja¯ne, java¯ in construction with an imperative of de- ‘give’. (27) (a) us-ko ja¯-ne do this-obj go-inf give:imp ‘allow him to go’ (b) i-ne ja-va¯ dc this-obj go-inf give:imp ‘allow him to go’ ˙˘ (e.g., ja¯-na¯, Action noun forms in -na¯, -vum m ja-vum m ˘˙ ) serve as lexical entries for verbs in dictionaries of Hindi and Gujarati. Oblique forms of such derivates appear also with postpositions, as in Hindi ja¯ne-ke liye, Gujarati java¯ ma¯tøe ‘in order to go’. Hindi (28) and the equivalent Gujarati sentences (29) contain masculine kar-na¯, feminine padø h-nı¯, and neuter kar-vum m ˘˙ , feminine wa¯m ˘˙ c-vı¯ in concord with the nouns ka¯m, kita¯b and ka¯m, ccpdø ¯ı. (28) (a) mujhe ka¯m kar-na¯ to.me work do-acnnr.m padø -a¯ fall-pfv.m.sg ‘I had to work’ (b) a¯p-ko yah kita¯b padø h-n-ı¯ you-to this book read-adjr-f padø -eg-ı¯ fall-fut-f ‘you will have to read this book’ (29) (a) ma¯re ka¯m kar-vum m ˘˙ to.me work do-acnnr.n padø -y-um m ˘˙ fall-pfv-m.sg ‘I had to work’ (b) tama¯re a¯ ccpdø -ı¯ wa¯m ˘˙ c-vı¯ you.to this book read-acnnr.f padø -s´-i fall-fut-f ‘you will have to read this book’ In these examples, the verbal derivates function as adjectival modifiers, in the manner that gerundives function in earlier Indo˙˘ pa¯nø ¯ı Aryan. Similarly, Gujarati pı¯-va¯-num m ‘drinking water’ has an adjectival derivate pı¯va¯num m ˘˙ (n. sg.) construed with pa¯nø ¯ı ‘water’. In Hindi, an oblique stem of the type ja¯-ne is followed by the suffix va¯la¯ to form agentive nouns such as ja¯neva¯la¯ ‘one who is going, is about to go’, rahneva¯la¯ ‘one who dwells (in an area, resident of a place)’.
1728
XVIII. Morphologischer Wandel II: Fallstudien
Another type of infinitive appears in kar-ı¯ s´ak ‘be able to do’ of Gujarati:
from Hindi (32 a) and Gujarati (32 b), with equivalent meanings:
(30) a¯ ka¯m a¯m kar-ı¯ s´ak-a¯-y this work thus do-inf can-pass-3 ‘this work can be done this way’
(32) (a) aisa¯ vaisa¯ jaisa¯ (rel.) kaisa¯ itna¯ utna¯ ˙˘ (b) a¯vum m evum m ˘˙ ˙˘ (rel.) jevum m kevum m ˘˙ ˙˘ a¯tølum m etølum m ˘˙
The equivalent modern Hindi infinitive has the simple base (kar): kar sak- ‘be able to do’. A derivate in -ı¯ occurs also as an absolutive in Gujarati: (31) ka¯m kar-ı¯-ne ghir ja¯-c work do-abs-to home go-2 ‘do your work before you go home’ Other New Indo-Aryan languages have comparable constructions, with differences in how the absolutives are formed. For example, in Hindi, kar follows a verb base (e.g., ja¯-kar) other than kar; this takes -ke (kar-ke), which can be used with other verbs also. Attributive participles corresponding to perfectives are formed in different manners. For example, Hindi has the construction type padø hı¯ huı¯ kita¯b ‘the book that has been read’, in which a perfective is followed by a form of the perfective of ho- (hua¯, hue, huı¯, huı¯m ˘˙ ) agreeing in number and gender with the modified noun. In Gujarati, on the other hand, the equivalent construction has a suffix -el- after the perfective stem, as in wa¯m ˘˙ celı¯ ccpdø ¯ı ‘a book that has been read’, pidhelo ma¯nø as ‘drunken man’. The formation with -loccurs elsewhere and is doubtless also incorporated in verb forms of the type given in (22 c). Denominative formations involve affixation and composition. There is a large number of suffixes used in forming substantives and adjectives. Some examples from Hindi will suffice to illustrate: lamb-a¯¯ı ‘length’ (lamba¯ ‘long’); bac-pan ‘childhood’ (bacca¯ ‘child’); duka¯n-da¯r ‘shopkeeper’ (duka¯n ‘shop); tøaiksi-va¯la¯ ‘taxi driver’; dha¯rmik ‘religious’ (dharma ‘religion’). There are also prefixes, though less numerous. For example: Hindi anu-kram ‘sequence’; gair-ha¯zir ‘absent’ (ha¯zir ‘present’); be-ka¯r ‘unemployed, useless’. Many such derivates involve elements and processes borrowed from Sanskrit and Perso-Arabic sources. Thus, dharma is a Sanskrit borrowing and dha¯rmik involves a Sanskrit vowel replacement along with the affix -ik (Skt. -ika). As in earlier stages of Indo-Aryan, there are particular derivates from pronouns, illustrated here by the following examples
‘of this sort’ ‘of that sort’ ‘of which sort’ ‘of what sort?’ ‘this much’ ‘that much’ ‘of this sort’ ‘of that sort’ ‘of which sort’ ‘of what sort?’ ‘this much’ ‘that much’
There are also adverbial counterparts: Hindi aise ‘in this way’, vaise ‘in that way’, kaise ˙˘ ‘here’, vaha¯m ˙˘ ‘there’, jaha¯m ‘how?’, yaha¯m ˘˙ ˙˘ ‘where?’, jab ‘when’, kab ‘where’, kaha¯m ‘when?’. The equivalent Gujarati derivates are a¯m, im, jim, kim, ahim m ˘˙ , tya¯m ˘˙ , jya¯m ˘˙ , kya¯m ˘˙ , jya¯re, kya¯re. Synchronically, these involve affixation, but historically a form like Hindi jaisa¯ reflects an earlier compound, and ˙˘ and similar as well as Gujarati Hindi yaha¯m im, jim reflect earlier case forms (cf. 3.1.2). Compounds such as Hindi, Gujarati ta¯rghar ‘telegraph office’ and Hindi, Gujarati ma¯-ba¯p ‘parents’ conform to the tatpurusøa and dvandva compound types. Compounds of the bahuvrı¯hi type are commonly Sanskrit (e.g., Hindi pragati-s´¯ıl ‘one with progressive character’). The productivity of composition and of different compound types varies across languages. There are also compound-like sequences of closely connected units. Particularly noteworthy because it is not limited to Indo-Aryan is the “echo formation” of the type Hindi pustak vustak, Gujarati ccpdø ¯ı bcpdø ¯ı, Bangla boi øtoi ‘book and such’: A nominal is repeated with a change of its first sound to form a sequence meaning ‘x and related matter’. 4.3.2. Verb derivation New Indo-Aryan languages also continue the use of morphologically derived causatives, reflecting both the earlier type Old Indo-Aryan trotøayati, Middle Indo-Aryan todø aı¨ ‘breaks’ and the more prevalent type with a morpheme that reflects the Middle Indo-Aryan generalization of -a¯p-ay- (Middle Indo-Aryan -a¯p-e-, -a¯v-e-). For example: (33) (a) khulkhul‘open (intr.)’
: :
khol- (Hindi) khol- (Gurajati) ‘open (trans.)’
1729
158. From Vedic to modern Indic languages
tøu¯tøtu¯tø‘break’ (b) karkar‘do, make’ calca¯l‘move, go’ padø h‘read, study’ ˙˘ cva¯m ‘read’
: : : : : : : :
todø - (Hindi) todø - (Gurajati) ‘break’ kar-a¯- (Hindi) kar-a¯v- (Gurajati) ‘have ... do, make’ cal-a¯- (Hindi) cal-a¯v- (Gurajati) ‘cause to move, to go’ padø h-a¯ (Hindi) ‘teach’ van˜c-a¯v- (Gurajati) ‘have ... read’
A noteworthy characteristic of modern languages of India is the use of complexes usually referred to as compound verbs, involving particular verbs, which have been labelled vector verbs. The complex commonly involves particular semantic features. For example, Hindi ma¯r- means ‘beat’ but ma¯r dø a¯lmeans ‘kill’; cf. also: bol- ‘speak’ : bol utøh‘blurt out’. There are also oppositions of the type dekh de- ‘let someone have a look’ : dekh le- ‘have a look’. In addition, aspectual features and restrictions concerning possible negation are linked to the use of compound verb complexes.
5.
Transliteration and transcription
I generally follow the standard system of transliteration (see Art. 5), but where ai, a¯i, and so on are sequences of vowels ⫺ as in some Middle Indo-Aryan forms ⫺ I translit˙˘ following erate aı¨ and so on. In addition, m a vowel indicates nasalization. For modern Indo-Aryan, I use a combination of transliteration and transcription. dø represents the flap commonly represented by ør. Since vocalic [l] does not occur in Middle and New Indo-Aryan languages, øl represents retroflex [l]. Although in some languages, such as Bangla, Gujarati, and Marathi, high vowels do not contrast by length, I still write long vowels in accordance with spelling conventions. If the script is ambiguous and an unambiguous representation is required, a transription is given. For example, in Hindi the graphemes ai au represent [i] [c], so that a transliteration is used, but in Gujarati the same symbols (e, o) are used to represent [e] [o] and [i] [c], so that I use IPA symbols for the lower vowels.
6.
References
Bhayani, Harivallabh Chunnilal (1953, ed.), Paumacariu (A Pura¯nø ic Epic in Apabhram ø s´a), Composed by Kavira¯ja Svayambhu¯deva. Bombay: Bharatiya Vidya Bhavan (Singhi Jain Series 34) Bloch, Jules (1965), Indo-Aryan from the Vedas to Modern Times [English edition, largely revised by the author and translated by Alfred Master; Original French: Paris, 1934]. Paris: Adrien Maisonneuve Bubenik, Vit (2003), “Pra¯krits and Apabhram ø s´a”. In: Cardona & Jain (eds.), 204⫺249 Cardona, George (2003), “Sanskrit”. In: Cardona & Jain (eds.), 104⫺160 Cardona, George & Jain, Dhanesh (2003), “General Introduction”. In: Cardona & Jain (eds.), 1⫺45 Cardona, George & Jain, Dhanesh K. (2003, eds.), The Indo-Aryan Languages. London: Routledge Davane, G. V. (1956), Nominal Composition in Middle Indo-Aryan. Poona: Deccan College (Deccan College Dissertation Series 11) Geiger, Wilhelm (1965), Pa¯li Language and Literature, authorised English translation by Batakrishna Ghosh. Calcutta: University of Calcutta [Original German: Strassburg, 1916] Hinüber, Oscar von (2001), Das ältere Mittelindisch im Überblick. Wien: Verlag der Österreichischen Akademie der Wissenschaften (Österreichische Akademie der Wissenschaften, Philosophisch-historische Klasse, Sitzungsberichte 467, Veröffentlichungen zu den Sprachen und Kulturen Südasiens, Heft 20) Masica, Colin (1991), The Indo-Aryan Languages. Cambridge: Cambridge University Press Oberlies, Thomas (2001), Pa¯li, A Grammar of the Language of the Theravaada Tipitøaka with a Concordance to Pischel’s Grammatik der PrakritSprachen. Berlin, New York: Walter de Gruyter (Indian Philology and South Asian Studies 3) Oberlies, Thomas (2003), “As´okan Prakrit and Paali”. In: Cardona & Jain (eds.), 161⫺203 Pischel, Richard (1965), Comparative Grammar of the Pra¯krø it Languages, translated from the German by Subhadra Jha¯. Delhi: Motilal Banarsidass. [Original: Strassburg, 1900] Vaidya, P. L. (1958), Pra¯krit Grammar of Hemacandra, Being the Eighth adhya¯ya of his SiddhaHema-S´abda¯nus´a¯sana. Poona: Bhandarkar Oriental Research Institute. (Bombay Sanskrit and Prakrti Series No. LX)
George Cardona, Philadelphia (U.S.A.)
1730
XVIII. Morphologischer Wandel II: Fallstudien
159. From Archaic Chinese to Mandarin 1. 2. 3. 4. 5. 6. 7. 8. 9.
Introduction Affixation in Old Chinese Prosodic alternation ⫺ Type A and B syllables Syllabic prefixes and suffixes in Old Chinese Reduplication Compounding New flectional suffixes New derivational affixes References
1.
Introduction
The non-phonetic nature of the Chinese script places peculiar difficulties in the way of the historian of the language. The native tradition of rhyme dictionaries and rhyme tables, supplemented by comparison of modern dialects and other indirect evidence, makes it possible to reconstruct the phonological system of Late Middle Chinese (LMC), the koine of the Tang dynasty (A. D. 618⫺906), and Early Middle Chinese (EMC), the standard common to the educated classes of North and South during the period of division from the fall of Western Jin in A. D. 316 to the reunification by Sui in 589, with a fair degree of confidence. The first reconstruction of Early Middle Chinese on a modern scientific basis (under the name Ancient Chinese) was achieved by Bernhard Karlgren (1915⫺ 1926) and is still widely used. The newer reconstructions of Late and Early Middle Chinese cited in this article are those of Pulleyblank (1984; 1991a). For Old Chinese, the language of the Chinese classics and other early texts, the evidence is much less satisfactory. The reconstruction (under the name Archaic Chinese) by Karlgren (1940; 1957) based on the rhyming categories of the earliest poetic anthology, the Shijing (terminus ante quem 600 B. C.) and the evidence of phonetic loans in the script, is commonly cited but has been much criticized (e.g. Li 1971; 1974⫺1975; Pulleyblank 1962; 1977⫺1978; 1991c). Tentative reconstructions cited here are preceded by an asterisk. References to Modern Chinese in this article are to the current standard called variously Mandarin, National Language (Guoyu) or Common Language (Putonghua) based on the vernacular of the capital, Beijing. The traditional characterization of Chinese as monosyllabic is not far off the mark.
As one distinguished linguist put it, “The socalled ‘monosyllabic myth’ is in fact one of the truest myths in Chinese mythology” (Chao 1968: 139). Though much of the vocabulary of Modern Chinese consists of disyllabic or polysyllabic lexemes these are mostly compounds of monosyllables that have the status of morphemes. Exceptions are some foreign loanwords, words like me´iguı` ‘rose’, and cho´uchu´ ‘hesitate’ that are reduplicative in form though there are no simpler morphemes to which they can be referred, and a small residue of unanalyzable disyllabic or polysyllabic forms like wu´gong ‘caterpillar’ and ha`shimaˇ ‘Chinese forest frog’ (cf. ha´ma ‘frog’). The traditional term “isolating” is also not inappropriate. Grammatical relationships are mostly conveyed by rules of word order and particles. In the modern language there are a few noun-forming suffixes, the plural or collective suffix -men for personal pronouns and animate nouns, and a handful of aspect suffixes for verbs, but even these affixes often do not have the obligatory character that such elements commonly have in inflected languages. It was not always so. Isolated traces of ancient morphological processes survive to the present. Thus cha´ng ‘long’ and zhaˇng ‘grow; elder; chief’ are written with the same graph and must be from the same root. Zha¯ng ‘stretch’, zha`ng ‘swell’ and zha`ng ‘curtain (i.e. something stretched)’, which are written with the graph for cha´ng ‘long’ with additional semantic determinatives, must also be related. In other cases obsolete variant readings of the same graph associated with differences in meaning are recorded in old dictionaries and commentaries, or sound similarities suggest etymological relationships between words even when there is no graphic connection. Early attempts to analyze morphological processes in Old Chinese (Conrady 1896) could make little progress before phonological reconstruction had been put on a sound basis. Serious analysis of the morphology of grammatical particles also began with Karlgren (1920), who observed that the first person pronouns, wu´ and woˇ, had different distribution in texts of the classical period and suggested that wu´ had originally been used for the nominative and genitive and woˇ for the accusative. While the interpretation of the distinction as one of case on the Indo-Euro-
159. From Archaic Chinese to Mandarin
pean model is unsatisfactory (Kennedy 1956), the general conclusion that these and other pronouns and particles were subject to morphological alternations associated with changes in grammatical function is well established. Karlgren also assembled sets of words that appeared to be related in both sound and meaning which he called “word families” and pointed out a number of recurrent patterns (Karlgren 1933; 1949: 70⫺101; 1956). He concluded that Proto-Chinese must have had formal word classes and regular systems of inflection and word derivation but was unable to find clearcut associations between patterns of phonetic alternation in Old Chinese as he reconstructed it and specific semantic or grammatical functions and concluded that by the time of the earliest texts the derivational processes had been too much obscured by phonetic change to yield to analysis. The study of morphological processes can, however, suggest alternative solutions to problems of Old Chinese phonology. Comparisons with other Sino-Tibetan languages, especially Tibetan, which in its classical form preserves a rich (though imperfectly understood) system of affixation in verbal morphology as well as word formation, can also suggest hypotheses to explain patterns of alternation in Chinese. The proposals that are put forward below are based on a thoroughly revised reconstruction of Middle Chinese (Pulleyblank 1984; 1991a) and on new ideas about the reconstruction of Old Chinese that are still being tested and revised.
2.
Affixation in Old Chinese
2.1. Departing Tone as derivational ⫺ the suffix *-s In Middle Chinese syllables were divided into four categories, labelled Level, Rising, Departing, and Entering, that correspond to the tones of modern spoken forms of Chinese. The Entering Tone in particular, which consisted of syllables ending in a stop consonant, was not “tonal” in a phonetic sense, and it is now thought that the Rising and Departing Tones had also developed between Old and Middle Chinese out of segmental features at the end of the syllable, a final glottal stop in the case of the Rising Tone and a suffix *-s in the case of the Departing Tone (Haudricourt 1954 a; 1954 b; Pulleyblank 1962; 1973 b; 1978; Mei 1970). During the Late Middle Chinese period a further split occurred into
1731 Upper and Lower registers conditioned by the devoicing of initial obstruents, giving a schematic system of eight tone classes which have been simplified or further elaborated in various ways in different dialects. The four tones of Modern Mandarin correspond roughly to the Middle Chinese categories as follows: (1) ma¯ ‘mother’ Upper Level, (2) ma´ ‘hemp’ Lower Level, (3) maˇ ‘horse’ Rising, (4) ma` ‘curse’ Departing. Neutral tone on particles and the second morphemes of some compounds is unmarked. Even in the modern language there are many examples of words originally in Level, Rising or Entering Tone with cognate forms in the Departing Tone. Mo´ ‘rub, grind’ and mo` ‘mill’, written with the same character, show such an alternation between Level and Departing. Similar alternations between Rising and Departing and between Entering and Departing are haˇo ‘good’ and ha`o ‘love’, e` ‘bad’ and wu` ‘hate’. In the last example both words now have the Departing Tone but e` ‘bad’, Cantonese c6k, Sino-Japanese aku, ended in -k in Middle Chinese. Though the semantic relationships associated with this pattern are quite various, it seems clear that, in general, forms in Departing Tone must be derivatives of forms in the other categories. It has sometimes been thought that the conditioning factor was always a matter of tone (e.g. Wang 1958). In the case of the Entering Tone, however, this would imply that variation in pitch could lead to the loss of a final stop consonant, a kind of phonetic change for which it is difficult to find a parallel. Comparison with cognate Austroasiatic languages led to the discovery that the Vietnamese tone corresponding to the Middle Chinese Departing Tone originated in the loss of final *-h, from an earlier *-s, and to the suggestion that Chinese might have had a parallel history (Haudricourt 1954 a; 1954 b). A parallel was also drawn with the -s suffix that figures prominently in Tibetan verbal morphology (Forrest 1960) and there is much evidence that a final sibilant survived in certain classes of Departing Tone syllables throughout the Han period (206 B. C. ⫺ A. D. 220) and as late as ca. A. D. 500 in the southern form of Early Middle Chinese (Pulleyblank 1962; 1973 b; 1979). The semantic relationships corresponding to Departing Tone derivation are quite varied and difficult to classify (Downer 1959; compare also Chou 1962: 50⫺92). It is possible that more than one original suffix is involved. Apart from its
1732 use as a marker of the Perfect in certain classes of verbs, the suffix -s in Tibetan appears in other forms where its meaning is hard to determine and it is nearly in complementary distribution with another suffix -d. 2.2. Voicing of initial obstruents ⫺ *a˘ as a prefix There are many alternations involving a contrast in Early Middle Chinese between voiceless and voiced initial obstruents (Karlgren 1949: 90⫺92; Downer 1959: 263; Chou 1962: 50⫺92). Typical examples are jia`n EMC kin h ‘see’, xia`n EMC ¥ inh ‘appear’, zhuˇ EMC tCuawk ‘attach’, shuˇ EMC dZuawk ‘be attached, belong’. It has been proposed (Pulleyblank 1973 a: 114⫺116; 1989) that the voiced alternants in such pairs reflect a prefix a˘-, a non-syllabic, pharyngeal glide that caused voicing, or in some dialects, voiced aspiration of a following obstruent, cognate to the Tibetan prefix, hø a-cˇhun˙, and to the prefix ? a˘ - of Written Burmese. It typically had the effect of changing transitive verbs into intransitives or to adjectives describing a state. In Classical Tibetan the prefix hø - is associated especially with the so-called “present” stem in verbs, which has an aspectual rather than temporal meaning. It expresses what is universally the case, is going on at the present time, or is treated as going on in a narrative referring to the past (Hahn 1974: 61). The ? a˘of Burmese is used especially to derive nouns from verbs, e.g. wak ‘to halve’, ? a˘wak ‘half’. It seems to be the same morpheme *a, with “introvert” meaning, that was involved in the e/a ablaut (2.3) and it could, perhaps, also occur as a suffix (2.4). 2.3. Ablaut ⫺ *a as an infix As a corollary to the hypothesis that, like that of modern Mandarin (Hartman 1944; Hockett 1947; 1950; 1955: 88; Martin 1957; Chao 1968; Cheng 1973; Chen 1976; Pulleyblank 1981; 1984), the vowel system of Old Chinese was based on a two-way ‘vertical’ contrast in rhyme vowels, /e/ and /a/, with features of frontness and roundedness encoded underlyingly as glides, /j/, /w/ and /}/, it has been proposed that there was a morphological system of ablaut between the vowels /e/ and /a/, associated with an extrovert/introvert contrast in meaning. This will account for alternations such as ta´n EMC dem ‘talk about (trans.)’, ta´n EMC dam ‘talk (intrans.); conversation’; sı` EMC zI’, < *-e` ¥? ‘resemble’, xia`ng EMC zIan’, < *-a`n? ‘imi-
XVIII. Morphologischer Wandel II: Fallstudien
tate, represent; image’; qiu¯ EMC khuw < *-xwe` ¥ ‘hill’, qu¯ EMC khIa˘ < *-xwa` ¥ ‘large hill, mound, abandoned city’ (Pulleyblank 1963; 1965; 1973 a; 1989). In spite of the example of Mandarin, vertical vowel systems of this kind have been regarded as unnatural (Ting 1975: 32, citing Jakobson & Halle 1971) but the development of nonlinear phonology in the seventies has made such an analysis more acceptable (Anderson 1978). A further point that has become a commonplace of recent phonological theory is that shwa vowels are often best analyzed as epenthetic, inserted by rules of syllabification, rather than underlyingly present in the lexicon (cf. Anderson 1978; 1982). If Old Chinese is analyzed in this way, the *e/a ablaut can be interpreted as a process of affixation, i.e. as infixation of *a, rather than alternation between two ends of a polar opposition and is easier to understand as a normal linguistic process. Much work remains to be done on identifying and analyzing the role of the *e/a ablaut. It plays a role in function words as well as families of content words. For example, the prohibitory negative wu´ EMC mua˘ ‘do not’ and wu´ EMC mua˘ ‘not have, there is no ...’ were already homonyms by the fourth or third century B. C. but were originally written with different graphs, implying that the former had the vowel *e and the latter had the vowel *a. The particles zhu¯ EMC tCIa˘ < *ta` ¥ ‘the collectivity of, those of a certain class’, and zheˇ EMC tCia’, < *ta` 6¥? (?), pronoun substitute for the head of a noun phrase, must be related by way of the *e/a ablaut to zhı¯ EMC tCI < *te`(¥), the particle that functions both as the mark of subordination before the head of a noun phrase and as a general third person object pronoun. 2.4. Rising Tone ⫺ *a as suffix? Oppositions within word families between the Middle Chinese Level and Rising Tones are less frequent than those involving the Departing Tone but are by no means rare. Zhaˇng EMC trIan’ ‘grow; elder’, contrasting with zha¯ng EMC trIan ‘stretch’ is one such case. As noted above, there is good evidence that the Rising Tone reflects an earlier final glottal stop. Though glottal stop ? - was also an initial consonant in Middle Chinese, there is reason to think that it was not a lexically distinctive phoneme in Old Chinese but rather an automatic accompaniment of vowel onset (Pulleyblank 1991c). In this case, the source of final glottalization also becomes
159. From Archaic Chinese to Mandarin
problematical. Parallelism suggests that it may have been an automatic feature after an open vowel, e.g. *-an (> Level Tone) vs. *-ana? > *-an? (> Rising Tone). In this case the affix that changed ‘stretch’ into ‘grow’ would be a suffix *-a, presumably the same morpheme that, as a prefix, was responsible for voicing the initial in cha´ng EMC drIan ‘long’. In both cases the derived form had a less active, more introvert, meaning, though in the case of the prefix the result was an adjective, while in the case of the suffix the result was an intransitive verb. Another example of Rising Tone corresponding to a clearly defined difference in meaning is zheˇng EMC tCiajn’ ‘in good order; put in order; whole’, which is related both graphically and phonetically to zhe¯ng EMC tCiajn ‘to set right, regulate’, now found only in the sense of zhe¯ng yue` ‘regulating moon, the month that starts the year’ and, with an additional semantic indicator in the graph, in zhe¯ng ‘mount a (punitive) expedition against, attack; tax’, as well as in the more common derivatives in Departing Tone, zhe`ng EMC tCiajn h ‘correct, right’, and zhe`ng EMC tCiajn h ‘government’. 2.5. Prefix *s In Tibetan s- clearly has a causative function and this is also found in other Tibeto-Burman languages, for example, Burmese, where *s- has given rise to aspiration of the following consonant (Benedict 1972: 105 ff.). Traces of the same prefix can, perhaps, be seen in Old Chinese in a case such as sı` EMC zIh < *sle`ks ‘feed; food, fodder’, from shı´ EMC Zik < *le`k ‘eat’ (Pulleyblank 1973 a:117), but parallel examples are few and the phonology of Old Chinese *s- clusters is still quite uncertain. 2.6. Initial aspiration Though Middle Chinese had a three-way manner contrast for obstruents, the voiced stops were of secondary origin, partly derived from the effect of the prefix *a˘- and partly from the ‘hardening’ of voiced continuants in Type A syllables (*l- > d-, *w- > ¥w-), leaving only a two-way contrast between plain voiceless and voiceless aspirates in Old Chinese. Classical Tibetan also has a three-way contrast in obstruents between plain voiceless, voiceless aspirates and voiced, but the two voiceless series are in complementary distribution after prefixes and only the aspirates occur freely without a prefix as syllable onsets. Plain voiceless stops in this position are
1733 mostly confined to special categories of words, including reduplicative forms, foreign loanwords and, in some cases, doublets of forms with prefixes such as kog-pa ~ skogpa ‘shell, rind, bark’ or pags-pa ~ lpags (in compounds) ‘skin, hide’ (Li 1933; Benedict 1972: 20). Benedict concluded that only the distinction of voice was pertinent and aspiration of the voiceless initials was allophonic in Tibeto-Burman but the comparative evidence will equally well support the hypothesis that the original contrast was between plain voiceless and voiceless aspirates and that voicing of the non-aspirates in Classical Tibetan was secondary. In that case, alternations that we find in Tibetan verbal morphology between voiced stops in Present and Future and the corresponding plain voiceless or voiceless aspirate in Perfect and Imperative, as in hø gebspa, bkab, dgab, khob ‘cover’ imply that an original plain voiceless stop became aspirated as a result of some morphological process. The source of this aspiration cannot be the prefix s-, which survives in Tibetan, but there could have been another prefix, such as s´, x, or h, that did not survive in Classical Tibetan (Shafer 1950: 721 f.). An alternative suggestion, also by Shafer, is that aspiration could have resulted from the coming together of two identical plain unaspirated stops: p-p- > ph-, k-k- > kh-, etc., a process that is alleged to occur in the morphophonemics of Tsangla, that is, from a kind of initial reduplication. In Chinese aspiration also plays a role in word families. In a few cases, such as pa`n EMC phanh ‘divide’ vs. ba`n EMC panh ‘half’ or kuo` EMC khwak ‘expand, stretch’, vs. guaˇng EMC kwan’ ‘wide’ (the variation in final consonants is also to be noted) it appears to be associated with causative formations and, in the absence of clear evidence as to the phonology of the prefix *s in Chinese, one might conjecture that this was the source of aspiration. In other cases the semantic relationship between the unaspirated and the aspirated form is much less clear, as in qı¯ EMC khia˘ ‘one-footed’, cf. jı¯ EMC kia˘ ‘odd number’, qı´ EMC gia˘ ‘strange’; cıˇ EMC ts hia˘’ ‘this’, cf. zı¯ EMC tsI ‘this’; che¯ EMC tC hia ‘carriage’, also ju¯ EMC kIa˘ ‘carriage’ (variant pronunciations of the same character). Comparison with Tibetan suggests that aspiration probably arose independently of prefixed *sin Chinese also. 2.7. Prefix (or infix?) *r The Tibetan prefix r- with parallels in Kachin, Bodo-Garo and Mikir (Benedict 1972: 109 f.) may have an Old Chinese paral-
1734 lel in cases like zhı` EMC tri h ‘cause to arrive, bring’, derived from zhı` EMC tCi h ‘arrive at, reach’, and chu` EMC tr hwit ‘expel’, derived from chu¯ EMC tC hwit ‘go out’ (Pulleyblank 1973 a:118). Though the Middle Chinese form might seem to imply a *Cr- cluster, in which the derivation would have to be interpreted as infixation, a prefixed *r- could have given the same result. 2.8. Presyllables *ke-, *te-, *pe-, *meAssuming that Tibetan orthography was based on the conventions of the Indian alphabet on which it was modelled, the Tibetan prefixes g, d, b, m, which are written on the line before the root initial rather than on top of it, were probably originally followed by a shwa vowel and formed enclitic presyllables. Isolated parallels to this are found in modern Chinese dialects, e.g. colloquial Pekingese ge¯bei or ge¯bo ‘arm’, which must be related to Classical bı` EMC pjia˘h ‘arm’ (Yang 1977⫺ 1978: 292⫺294). Another trace of this may be the so-called “softened” initials in Northwestern Min dialects. Thus, Kienyang sometimes has zero or v- instead of p-, corresponding to Middle Chinese p- and b-, l- corresponding to Middle Chinese t-, d-, ts-, dz-, and zero or h- corresponding to Middle Chinese k- and g- (or ¥-). The common Min word for ‘cockroach’ has a presyllable in Fuzhou ka 6 sak 8, Amoy ka 1 tsua? 8, but a “softened” initial in Kienyang lue 8, Chungan luai 8, suggesting that in at least some cases the “softening” of the initials may have resulted from weakening in intervocalic position, with subsequent loss of the presyllable (Norman 1986: 383 f.). A similar “softening” of initial obstruents in Vietnamese corresponds to presyllables in other Mon-Khmer languages, e.g. ga` ‘chicken’ Ruc reka; va´i ‘cotton’ Ruc kupal (Haudricourt 1965; Thompson 1976: 1131⫺ 1133) and such “softened” initials also occur in early Vietnamese loans from Chinese (Pulleyblank 1981: 279). In particular, Vietnamese has vai ‘shoulder’ corresponding to Chinese bı` EMC pjia˘ h, colloquial ge¯bo, ge¯bei ‘arm’. In a number of cases the Min softened stops correspond to voiced stops in Yao loanwords which, in turn, correspond to prenasalized stops in Miao, which has led to the suggestion that prenasalization was also the general source of the phenomenon in Chinese (Norman 1986: 381⫺383). Prenasalization in Miao-Yao could, however, simply be the effect of a preceding unaccented shwa vowel derived from a lost presyllable. The hø a-cˇhun˙
XVIII. Morphologischer Wandel II: Fallstudien
prefix in Tibetan, written in Tibetan script as a presyllable, had the effect of prenasalizing obstruents in that language. 2.9. Dental suffixes *-n, *-t, *-l Besides *-s, Tibeto-Burman had the dental suffixes *-n and *-t whose precise functions are difficult to analyze (Benedict 1972: 98⫺ 103). There is also internal Chinese evidence for *-n and *-t as suffixes. In a number of sets of particles in the preclassical and early classical language *-n seems to be associated with durative or continuative aspect, while *-t is associated with punctual or perfective aspect (Graham 1983) and derivatives in *-n or *-t can be recognized among ordinary content words as well (Pulleyblank 1991b: 29⫺33). It seems likely that *-l could also function as a suffix in word formation. 2.10. Miscellaneous affixation A distributive suffix *-k appears in a number of pronominal forms: huo` EMC ¥wek < *we´k ‘(there is) some one, something’ from yoˇu EMC wuw’ < *we´? ‘have; there is/are’; mo` EMC mak < *ma´k ‘(there is) no one, nothing’ from wu´ EMC mua˘ < *ma` ¥ ‘not have, there is/are not’; shu´ EMC dZuwk ‘which’ from shuı´ EMC dZwi ‘who’; ge` EMC kak ‘each’, cf. juˇ EMC kIa˘’ < *ka`y? ’ ‘all’, jie¯ EMC keIj ‘all’, ju¯ EMC kua˘ ‘both’, etc. These particles are placed between the subject and the verb, or in front of the verb without an expressed subject, and have the effect of selecting among possible subjects: ‘some of X’, ‘none of X’, ‘which of X?’, ‘each of X’. The same suffix can be recognized in the early classical demonstrative pronoun shı´ EMC dZik < *-e`k ‘this’, with a dialectal variant showing palatalization of the final consonant shı´ EMC Zit (for dZit), which is related to the common demonstrative shı` EMC dZia˘’ ‘this’ and has the function of recapitulating a preceding subject and giving it contrastive emphasis (Pulleyblank 1960). A type of alternation of which there are several examples is between Old Chinese *-n and either the corresponding voiced continuants *-¥ or zero: de˘ng EMC ten’ < *ten? ’step, degree’, also read EMC tej’ < *te¥? , wa´ng EMC muan < *ma`n ‘there is no; disappear, die; lose’, wu´ EMC mua˘ < *ma`(¥) ‘there is no’; waˇng EMC wuan’ < *wa`n? ‘go’, yu´ EMC wua˘ < *wa`(¥) ‘go to; to’; fa¯ng EMC puan < *pa`n ‘just now’, fuˇ EMC pua˘’ < pa`(¥)? ‘begin, for the first time’ (Pulleyblank 1962: 232 f.).
1735
159. From Archaic Chinese to Mandarin
3.
Prosodic alternation ⫺ Type A and B syllables
A phonological alternation that was probably not the result of affixation is the contrast between what have been called Type A and Type B syllables. In Karlgren’s reconstructions this phonological contrast is represented by the absence (Type A) or presence (Type B) of a medial yod, -i4-. Though the contrast between syllable types that it marks is real and important, the interpretation in terms of a medial segmental element is a basic error in his system (Pulleyblank 1984; 1992 a). In EMC Type A syllables had a mid or low nuclear vowel, while Type B syllables had a high vowel, i, I, or u, either as the sole element in the nucleus or as the first mora of a diphthong, ia, Ia, or ua. Since one can in general derive corresponding A and B pairs from every Old Chinese rhyme category, it seems likely that the Middle Chinese contrast reflects an earlier prosodic contrast in Old Chinese which put prominence either on the second mora of the syllable (Type A) or the first (Type B). There is a parallel in Sizang (Siyin) Chin, a Tibeto-Burman language of Burma, which distinguishes within sequences ending in -VV, -VC, and -VVC between a syllabic peaking that falls upon either the first or the second mora, e.g.: pa5i ‘to lose’ vs. pai5 ‘to go’; be5l ‘to take shelter under’ vs. bel5 ‘(superlative particle)’; ta5m ‘jungle’ vs. tam5 ‘to number, many’ (Stern 1963; Pulleyblank 1992 b). The process of diphthongization affecting Type B syllables that gave rise to the Early Middle Chinese contrast seems to have occurred around the beginning of the present era and there is comparative evidence that it spread as an areal feature to Vietnamese and other languages that were within the Chinese sphere of influence at that time (Pulleyblank 1992 b). Type A and B syllables are marked by an acute or a grave accent respectively in Old Chinese reconstructions, e.g. *-a´n > EMC -an (Type A), *-a`n > EMC -Ian or -ian, depending on the initial (Type B). There are alternations between the two syllable types both in grammatical particles: a¯n EMC ? an < *? a´n ‘where? how?’ (Type A), ya¯n EMC ? ian < *? a`n ‘where? how?’ (Type B) (with no apparent difference in meaning); nuo` EMC nak < *na´k ‘agree, say yes’, ruo` EMC MIak < *na`k ‘thus; conform to, accord with’ ⫺ and content words: bia¯n EMC pin < *pja´n and EMC pjian < *pja`n ‘plait; compile’. Compare also ma¯o EMC maIw <
*mra´w ‘cat’, also read EMC miaw < *mra`w with corresponding forms in some modern dialects, e.g. Changsha miau, as well as mau. There is a more marked semantic difference between Type A and Type B in naˇi EMC nej’ < *ne´ ¥? ‘then, thereupon’ and e´r EMC MI < *ne´ ¥ ‘then, and’. The stronger form, naˇi, is a connective between independent clauses or sentences and can also serve to emphasize a following noun predication, while e´r is a colourless connective between serial verbs.
4.
Syllabic prefixes and suffixes in Old Chinese
Besides processes that changed the forms of monosyllables, there are certain constructions in the preclassical language that have been described as ‘prefixation’ that involve syllables represented by separate graphs. Yoˇu, which ordinarily means ‘have; there is …’ sometimes appears in front of nouns, especially proper names, where it seems to be part of the noun phrase though what, if anything, it adds to the meaning is not clear (Wang 1958: 219 f.). There are also certain prefixed syllables that mark distinctions of aspect in a following verb (Chou 1962: 248 ff.; Graham 1983: 56; Pulleyblank 1986: 3). They may be compared to the aspect suffixes of Mandarin (see below). A formative element that was used in Classical Chinese and is also found in the modern language is -ra´n ‘-like, -ly’, a derivative in -n of ru´ ‘is like’, used as a suffix forming adverbs of manner, e.g. huı`ra´n ‘kindly’ from huı` ‘to favour; kindness’. Unlike purely colloquial suffixes, it retains its full tone in the modern language, a mark of its literary origin.
5.
Reduplication
A process that was already present in the ancient language and has continued to be active until the present is complete or partial reduplication (cf. Art. 57). Examples of reduplication used in onomatopoeia, expressive adjectives and adverbs, and names of plants and small animals are gua¯n’gua¯n, ‘cry of a bird’, xia¯oya´o ‘free and unfettered, carefree’, zhı¯zhu¯ ‘spider’. Of more interest from the point of view of word formation are expressions such as: xu¯yu´ EMC sua˘ jua˘ ‘a moment’, from xu¯ EMC sua˘ ‘wait’, and co¯ngro´ng EMC ts huawn juawn (note the aspiration) ‘at leisure, unhurriedly’, from the root found in zo`ng EMC
1736
XVIII. Morphologischer Wandel II: Fallstudien
tsuawn h ‘release; loose’ and co´ng EMC dzuawn ‘follow’; pu´fu´ EMC bcbek < *ba´be`k ‘crawl’. In a few cases the analysis of such disyllabic forms as reduplicative breaks down, e.g. hu´die´ EMC ¥c dip ‘butterfly’ (Kennedy 1955), but the second syllable does occur separately and is probably cognate to Tibetan leb in phye-ma-leb ‘butterfly’ (where phye-ma means ‘flour’ and refers to the flourlike scales on a butterfly’s wing), and written Burmese lip-pra ‘butterfly’. Certain productive categories of full reduplication that are normal in the modern language ⫺ familiar terms of address like ge¯ge ‘elder brother’, distributives like re´nre´n ‘person:rdp (everybody)’, tia¯ntian ‘every day’, and repetition of a verb to express tentative aspect as in ka`nkan ‘have a look’ ⫺ are not found in the classical language.
6.
Compounding
Compounding, by which word sequences arising out of regular syntactic processes of the language are lexicalized with special meanings (Lyons 1977: 535), which is so prominent in modern Chinese, was already in progress in the classical language. There are many disyllabic expressions whose meaning cannot be inferred transparently from that of the separate elements, e.g.: baˇixı`ng ‘common people’ (literally: ‘hundred surnames’), guaˇre´n ‘I (term of self-reference for the ruler of a state)’ (literally: ‘solitary man’), ju¯ nzıˇ ‘gentleman’ (literally: ‘son of a lord’), shu`jıˇ ‘almost, nearly, it is to be hoped that’ (literally: ‘many-few’). Unfortunately there is no way to tell whether there were any phonological correlates. In modern Mandarin the second member of a compound is frequently toneless, which binds the two syllables together into a single tone-bearing unit. Since there is no clue to it in the writing system, it is difficult to know when this appeared historically. It is not found in more conservative dialects such as Cantonese.
7.
New flectional suffixes
7.1. The plural suffix -men In modern spoken Chinese -men is used to mark the plural of personal pronouns: woˇmen ‘we’, za´nmen ‘we (inclusive)’, nıˇmen ‘you’, ta¯men ‘they’ (only when referring to persons, otherwise simply ta¯). In this sense it clearly fits the definition of an inflectional suffix. Its use with nouns is not strictly comparable. It
has been called a mark of the indefinite plural (Kupfer 1980: 56), but according to another interpretation, it forms collective nouns (cf. Art. 101), which would make it a derivational suffix (Chao 1968: 244 f.). With nouns it cannot be used when a definite number is specified: liaˇngge qia´ngdao ‘two bandits’, not *liaˇngge qia´ngdaomen, while with pronouns it can, though it is also possible to omit it: woˇmen liaˇngge or woˇ liaˇngge, woˇ liaˇ, ‘we two’. The etymology of -men is obscure. In the literary language certain words like cha´i, shuˇ, ca´o, deˇng, be`i meaning ‘class, category, group’ or the like are used from time to time with both nouns and pronouns to indicate a collectivity but there is nothing approaching a regular grammatical process of plural marking. The suffix -men, written in a variety of ways that seem to indicate that it was an innovation in the vernacular without a recognized antecedent, begins to appear in colloquial texts from about the tenth and eleventh centuries (Lü 1955; Wang 1958: 272 f.). From an early period it appears to have had the allomorphs -me and -m, still found in present-day usage. Out of the latter have developed polite forms of the second and third person pronouns used without reference to number, nıˇn < nıˇm and ta¯n < ta¯m, by the regular shift of syllable final -m to -n that occurred around the end of the fifteenth century (Lü 1955: 213). In epistolary style, though not in speech, one can even create new plural forms, nıˇnmen and ta¯nmen (Chao 1968: 244 f., 640 f.). 7.2. The verbal suffixes -le, -zhe, -guo The perfective suffix -le and its homonym, the sentence final particle le, are derived from liaˇo ‘end, finish’, which is still current as a monosyllabic verb in certain expressions and is also found in such combinations as liaˇojie´ ‘finish, settle’, liaˇoju´ ‘end, solution, settlement’, zho¯ngliaˇo ‘end (of a period)’, etc. The full form liaˇo is also found as what Chao calls a ‘phase complement’ to other verbs, that is, a complement that expresses the phase of an action of the preceding verb, rather than some result in the action or goal like an ordinary resultative complement (Chao 1968: 449). This may have been the earliest development in its progress towards grammaticalization. Its use as a perfective suffix seems to have become established in the vernacular by the tenth and eleventh centuries, by which time it had also come into use as a sentence final particle, taking over the functions of the classical final particle yıˇ (Wang 1958: 304⫺307, 447 f.). Unfortunately
159. From Archaic Chinese to Mandarin
textual evidence does not enable us to trace the steps by which the full form liaˇo became reduced phonetically to le in the suffix and sentence final particle. The suffix -zhe for progressive aspect is similarly related both graphically and etymologically to zha´o EMC drIak ‘come in contact with, touch’, which also occurs both in the form zha´ole, with the perfective suffix -le, as an ordinary resultative complement ‘hit the mark, touched the essential point’, and, with optional neutral tone, as a phase complement, translated by Chao as ‘touched, got at, successful after an attempt’ (Chao 1968: 447). As an aspect suffix it immediately follows the verb and cannot be separated from it by an object (Chao 1968: 248⫺250). The suffix zhe was a little later in appearing than le but there is evidence for it already in the eleventh century (Wang 1958: 309⫺311). The verb guo` ‘pass’ can also be used, with optional neutral tone, as a complement, in which case it can be followed by the perfective suffix -le, as in chı¯guo`le fa`n ‘have had one’s meal’, but it can also be used as a suffix in the neutral tone to mean ‘completed action at an indefinite time in the past’, as in me´i chu¯guo ya´ng ‘has never been abroad’ (Chao 1968: 252 f.). This usage can also be found by the eleventh century (Wang 1958: 312). Two compound suffixes, -qilai, from qıˇlai ‘rise, get up’, and -xiaqu, from xia`qu ‘go down, descend’, which one scholar would prefer to relegate to the domain of word formation or syntax (Kupfer 1980: 57), are used as suffixes for inchoative and successive aspect respectively (Chao 1968: 251 f.).
8.
New derivational affixes
8.1. Noun prefixes A-, as in a¯ge¯ ‘elder brother’, is found widely in modern Chinese dialects (though not in Peking Mandarin) as a prefix with proper names and words of relationship, conveying a feeling of endearment or familiarity. Though it is not found in texts of the classical period, it occurs quite early in postclassical times (Wang 1958: 222) and is probably an inheritance outside the literary tradition from Sino-Tibetan. Benedict (1972) argued that, as a prefix of familiarity in Tibeto-Burman, ? awas a variant of the hø a-cˇhun˙ prefix of Tibetan that has been discussed in 2.2 and this seems very likely. The difference would be that, in the one case, *a- acquired a glottal stop onset and remained a distinct syllable, while in the other, it was a non-syllabic prefix, that in
1737 Chinese voiced a following obstruent initial and then disappeared but in Tibetan was still pronounced as an onglide at the time when the script was invented and afterwards realized as prenasalization (Pulleyblank 1986). Although it does not appear as a separate syllable in the classical language, the prefix of familiarity has left its trace in the Middle Chinese voiced initial of standard Chinese fu` “father” EMC bua’
1738 tury, however, there are unambiguous examples in which e´r is attached to nouns referring to inanimate objects, apparently with a diminutive sense. Unlike zıˇ, which lost any diminutive force that it may have once had at an early date, e´r has retained this connotation, combined with an implication of endearment or familiarity (Wang 1958: 227⫺ 229; Chao 1968: 230). It is still a productive suffix in the modern language. Except for some speakers of Manchu background who still pronounce it as a separate, toneless syllable -er (Chao 1968: 140), the phonological reduction has gone farther for this suffix than for other flexional and derivational suffixes of modern Mandarin, so that it has become simply the consonant -r which is attached to the preceding syllable with various morphophonemic changes. Etymologically, suffixal -r also has two other sources in Mandarin, a reduced form of rı` ‘day’ in jı¯nr ‘today’, zuo´r ‘yesterday’, etc., and a reduced form of lıˇ ‘inside’ in the place words zhe`r ‘there’, na`r ‘there’ and naˇr ‘where’ (Chao 1968: 229). A noun-forming suffix of less frequent occurrence, for which there are examples from the fifth century onwards, is -tou, derived from to´u ‘head’, as in shı´tou ‘stone’, guˇtou ‘bone’ (Wang 1958: 230 f.). The suffix -ba, apparently derived from the verb ba¯ ‘cling’, is found in a few nouns such as weˇiba ‘tail’, nı´ba ‘mud, clay’ (Chao 1968: 244). The classical particle zheˇ which functions as a pronominal substitute for the head of a noun phrase has been borrowed into the modern spoken language as suffix for agent nouns, as in (xı¯nwe´n) jı`zheˇ ‘(newspaper) reporter’, la´odo`ngzheˇ ‘worker’. It shows its literary origin, however, in that it retains its tone and cannot be combined with purely non-literary verbs, such as chı¯ ‘eat’ or pa´o ‘run’. The true modern equivalent of zheˇ is de, which, like zheˇ in the classical language, is not only a suffix for agent nouns, as in zhaˇngsha´orde ‘chef’, but, more generally, a nominalizer of a noun-modifier phrase with the head omitted or without a specified head, e.g. Ta¯ shi co´ng Zho¯ngguo´ la´i de ‘He is one who comes from China’ instead of Ta¯ shi co´ng Zho¯ngguo´ la´i de re´n ‘He is a man who comes from China’. In the modern language the particle of subordination in de re´n is homonymous with the nominalizing particle de that is used when re´n is omitted. As a particle of subordination de is a variant of the classical particle zhı¯ that has developed separately in the colloquial language. Nominalizing de can similarly be regarded as the colloquial
XVIII. Morphologischer Wandel II: Fallstudien
development of zheˇ. For the original morphological connection between zhı¯ and zheˇ see 2.3. 8.3. Other derivational suffixes Shi, a reduced form of the copula shı` ‘is’, is a suffix in a number of particles: yao`shi ‘if’, ruo`shi ‘if’, do¯ushi ‘all, in all cases’, keˇshi ‘but’, etc. (Chao 1968: 722 f.). Chao notes that in the case of ruo`shi the two components can still be separated in ruo`bushi ‘if (is) not’, showing that the meaning of the suffix as a copula is not completely attenuated. Prepositions, or “coverbs”, in Chinese are a special class of transitive verbs that form phrases that modify a main verb rather than standing on their own. They normally do not take aspect particles since they take their aspect from the main verb. In some cases, however, we do find them with the aspect suffixes -zhe or -le with no apparent change in meaning or with some specialized meaning that cannot be directly related to the usual force of the particle. Thus, we`i ‘for; because of; for the sake of’ also has the forms we`izhe and we`ile. Authorities differ as to the precise shades of meaning that these forms convey.
9.
References
Anderson, Stephen R. (1978), “Syllables, Segments and the Northwest Caucasian Languages”. In: Bell, Alan & Hooper, Joan Bybee (eds.), Syllables and Segments. Amsterdam, New York: North Holland Publishing Company, 47⫺58 Anderson, Stephen R. (1982), “The Analysis of French Shwa”. Language 58, 534⫺573 Benedict, Paul K. (1972), Sino-Tibetan: A Conspectus. Cambridge: Cambridge University Press Chao, Yuen Ren (1968), A Grammar of Spoken Chinese. Berkeley, Los Angeles: University of California Press Chen, Matthew Y. (1976), “From Middle Chinese to Modern Peking”. Journal of Chinese Linguistics 4, 113⫺277 Cheng, Chin-chuan (1973), A Synchronic Phonology of Mandarin Chinese. The Hague: Mouton Chou, Fa-kao (1962), Zhongguo gudai yufa, Gouci bian (A Historical Grammar of Ancient Chinese, Part II: Morphology). Taipei: Academia Sinica Conrady, August (1896), Eine indochinesische Causativ-Denominativ-Bildung und ihr Zusammenhang mit den Tonaccenten. Leipzig: Harrassowitz Downer, Gordon, B. (1959), “Derivation by Tone Change in Classical Chinese”. Bulletin of the School of Oriental and African Studies 22, 258⫺290
159. From Archaic Chinese to Mandarin Forrest, R. A. D. (1960), “Les occlusives finales en Chinois archaı¨que”. Bulletin de la Socie´te´ de Linguistique de Paris 55, 228⫺239 Graham, A. C. (1983), “Yün and Yüeh as Verbs and as Particles”. Acta Orientalia Hafniensia 44, 33⫺71 Hahn, Michael (1974), Lehrbuch der Klassischen Tibetischen Schriftsprache. Bonn: Indologisches Seminar Hartman, Lawton M. (1944), “The Segmental Phonemes of the Peiping Dialect”. Language 20, 28⫺42 Haudricourt, Andre´ G. (1954 a), “Comment reconstruire le chinois archaı¨que”. Word 10, 351⫺364 Haudricourt, Andre´ G. (1954 b), “De l’origine des tons en Vieˆtnamien”. Journal Asiatique 242, 69⫺82 Haudricourt, Andre´ G. (1965), “Les mutations consonantiques des occlusives initiales”. Bulletin de la Socie´te´ de Linguistique de Paris 60, 160⫺172 Hockett, Charles F. (1947), “Peiping Phonology”. Journal of the American Oriental Society 67, 253⫺ 267 Hockett, Charles F. (1950), “Peiping Morphophonemics”. Language 26, 63⫺85 Hockett, Charles F. (1955), A Manual of Phonology. Baltimore: Waverly Press (Memoir 11 of the International Journal of American Linguistics) Jakobson, Roman & Halle, Morris (1971), Fundamentals of Language. The Hague: Mouton Karlgren, Bernhard (1920), “Le Proto-Chinois langue flexionelle”. Journal Asiatique 15, 205⫺233 Karlgren, Bernhard (1933), “Word Families in Chinese”. Bulletin of the Museum of Far Eastern Antiquities 5, 1⫺120 Karlgren, Bernhard (1940), Grammata Serica. Taipei: Ch’eng-Wen [reprinted from: Bulletin of the Museum of Far Eastern Antiquities 12, 1⫺471]
1739 the Institute of History and Philology, Academia Sinica 28, 273⫺281 [reprinted in: Li, Tien-yi (1964, ed.), Selected Works of George A. Kennedy. New Haven/CT: Far Eastern Publications Yale University, 434⫺442] Kupfer, Peter (1980), “Morphemklassen und Wortstrukturen im modernen Chinesischen”. Sprachwissenschaft 5, 53⫺72 Li, Fang-kuei (1933), “Certain Phonetic Influences of the Tibetan Prefixes upon the Root Initials”. Bulletin of the Institute of History and Philology, Academia Sinica 4, 135⫺157 Li, Fang-kuei (1971), “Shang-ku yin yen-chiu”. Tsinghua Hsuehpao n.s. 9, 1⫺61 Li, Fang-kuei (1974⫺1975), “Studies on Archaic Chinese”. Monumenta Serica 31, 219⫺287 Lü, Shuxiang (1955), “Shuo men” (A Study of the Plurality Suffix meˆn, with English Summary). In: Lü, Shuxiang, Hanyu yufa lunwenji. Beijing: Kexue chubanshe, 145⫺168, 212⫺213 Lyons, John (1977), Semantics, Vol. 1⫺2. Cambridge: Cambridge University Press Martin, Samuel E. (1957), “Problems of Hierarchy and Indeterminacy in Mandarin Phonology”. Bulletin of the Institute of History and Philology, Academia Sinica 29, 219⫺229 Mei, Tsu-lin (1970), “Tones and Prosody in Middle Chinese and the Origin of the Rising Tone”. Harvard Journal of Asiatic Studies 30, 86⫺110 Norman, Jerry (1986), “The Origin of the ProtoMin Softened Stops”. In: McCoy, John & Light, Timothy (eds.), Contributions to Sino-Tibetan Studies. Leiden: E. J. Brill, 375⫺384 Pulleyblank, Edwin G. (1960), “Studies in Early Chinese Grammar, Part I”. Asia Major 8, 36⫺67
Karlgren, Bernhard (1949), The Chinese Language. New York: Roland Press
Pulleyblank, Edwin G. (1962), “The Consonantal System of Old Chinese”. Asia Major 9, 58⫺144, 206⫺265
Karlgren, Bernhard (1956), “Cognate Words in the Chinese Phonetic Series”. Bulletin of the Museum of Far Eastern Antiquities 28, 1⫺18
Pulleyblank, Edwin G. (1963), “An Interpretation of the Vowel Systems of Old Chinese and Written Burmese”. Asia Major 10, 200⫺221
Karlgren, Bernhard (1957), Grammata Serica Recensa. Stockholm: Museum of Far Eastern Antiquities [reprinted from: Bulletin of Far Eastern Antiquities 29, 1⫺332]
Pulleyblank, Edwin G. (1965), “Close/Open Ablaut in Sino-Tibetan”. In: Milner, G. B. & Henderson, E. J. A. (eds.), Indo-Pacific Linguistic Studies [⫽ Lingua 14], 230⫺240
Kennedy, George A. (1955), “The Butterfly Case”. Wennti 8 [reprinted in: Li, Tien-yi (1964, ed.), Selected Works of George A. Kennedy. New Haven/ CT: Far Eastern Publications Yale University, 274⫺322]
Pulleyblank, Edwin G. (1973 a), “Some New Hypotheses Concerning Word Families in Chinese”. Journal of Chinese Linguistics 1, 111⫺125
Kennedy, George A. (1956), “Zai lun wu wo” (The Classical Pronoun Forms ngo and nga). Bulletin of
Pulleyblank, Edwin G. (1973 b), “Some Further Evidence Regarding Old Chinese -s and Its Time of Disappearance”. Bulletin of the School of Oriental and African Studies 36, 368⫺373
1740
XVIII. Morphologischer Wandel II: Fallstudien
Pulleyblank, Edwin G. (1977⫺1978), “The Final Consonants of Old Chinese”. Monumenta Serica 33, 180⫺206
Pulleyblank, Edwin G. (1991c), “The ganzhi as Phonograms and Their Application to the Calendar”. Early China 16, 39⫺80
Pulleyblank, Edwin G. (1978), “The Nature of the Middle Chinese Tones and Their Development to Early Mandarin”. Journal of Chinese Linguistics 6, 173⫺203
Pulleyblank, Edwin G. (1992 a), “European Studies on Chinese Phonology: the First Phase”. Paper for the International Symposium on the History of European Sinology, Taipei, April 1992
Pulleyblank, Edwin G. (1979), “Some Examples of Colloquial Pronunciation from the Southern Liang Dynasty”. In: Bauer, Wolfgang (ed.), Studia SinoMongolica: Festschrift für Herbert Franke. Wiesbaden: Franz Steiner Verlag, 315⫺328
Pulleyblank, Edwin G. (1992 b), “Type A and B Syllables in Middle Chinese and Their Origin in Old Chinese”. Paper for the 25th International Conference on Sino-Tibetan Languages and Linguistics, Berkeley, 14⫺18 October 1992
Pulleyblank, Edwin G. (1981), “Some Notes on Chinese Historical Phonology”. Bulletin de l’Ecole Franc¸aise d’Extreˆme-Orient 69, 277⫺288
Pulleyblank, Edwin G. (2000), “Morphology in Old Chinese”. Journal of Chinese Linguistics 28, 26⫺51
Pulleyblank, Edwin G. (1984), Middle Chinese: A Study in Historical Phonology. Vancouver: UBC Press
Shafer, Robert (1950⫺1951), “Studies in the Morphology of Bodic Verbs”. Bulletin of the School of Oriental and African Studies 13, 702⫺724, 1017⫺ 1031
Pulleyblank, Edwin G. (1986), “The Locative Particles yü, yü, and hu”. Journal of the American Oriental Society 106, 1⫺12
Stern, Theodore (1963), “A Provisional Sketch of Sizang (Siyin) Chin”. Asia Major 10, 222⫺278
Pulleyblank, Edwin G. (1989), “Ablaut and Initial Voicing in Old Chinese Morphology: *a as an Infix and Prefix”. Proceedings of the Second International Conference on Sinology: Section on Linguistics and Paleography. Taipei: Academia Sinica, 1⫺ 21 Pulleyblank, Edwin G. (1991a), Lexicon of Reconstructed Pronunciation in Early Middle Chinese, Late Middle Chinese, and Early Mandarin. Vancouver: UBC Press Pulleyblank, Edwin G. (1991b), “Some Notes on Morphology and Syntax in Classical Chinese”. In: Rosemont, Henry Jr. (ed.), Chinese Texts and Philosophical Contexts: Essays Dedicated to Angus C. Graham. La Salle, Illinois: Open Court, 21⫺45
Thompson, Laurence C. (1976), “Proto-Viet-Muong Phonology”. In: Jenner, Philip N. & Thompson, Laurence C. & Starosta, Stanley (eds.), Austroasiatic Studies. Honolulu: University Press of Hawaii, 1131⫺1203 Ting, Pang-hsin (1975), Chinese Phonology of the Wei-Chin Period. Taipei: Academia Sinica [Institute of History and Philology, Academia Sinica, Special Publications, no. 65] Wang, Li (1958), Hanyu shigao, Vol. 1⫺3. Beijing: Kexue chubanshe Yang, Paul Fu-mien S. J. (1977⫺1978), “Prefix Kein Modern Chinese Dialects and Proto-Chinese”. Monumenta Serica 33, 286⫺299
Edwin G. Pulleyblank, Vancouver (Canada)
160. From Classical Arabic to the modern Arabic vernaculars 1. 2. 3. 4. 5. 6.
History of the Arabic language Morphology of the verb Morphology of the noun Demonstratives and interrogatives Pidginization and creolization in Arabic References
1.
History of the Arabic language
For the purposes of this article the term “Classical Arabic” will refer to the type of
Arabic that was used in the Arabian peninsula during the pre-Islamic period in poetry and the QurÅa¯n and that was to remain the standard language throughout the entire Arabo-Islamic history. Its modern form is usually called “Modern Standard Arabic”; this variety is distinguished from Classical Arabic mainly by its lexicon and phraseology, as well as minor syntactic adaptations. According to most Western Arabists, in the pre-Islamic
160. From Classical Arabic to the modern Arabic vernaculars
period Classical Arabic was an intertribal and supratribal language, based on the dialects of the Bedouin in the central and eastern parts of the peninsula (cf. Rabin 1951). In this view, the “poetico-QurÅa¯nic koine”, as it is sometimes called (cf. Zwettler 1978: 97⫺188), was never used in everyday communication. The vernacular of the Arab tribes is assumed to have differed from the Classical type of Arabic in lexicon and phonology, but chiefly in the absence of declensional and modal markers (Åi¤ra¯b). The proponents of this view point to the evidence found in the inscriptions of early forms of Arabic, such as Lihyanite (cf. Müller 1982; Robin 1992). According to Corriente (1976) the Åi¤ra¯b-less dialects were mainly found in the periphery of the peninsula, namely in those areas where contact with other languages took place. The Arabs themselves view the history of their language in a different way. According to them, there was only one Arabic language in the pre-Islamic period, the ¤Arabiyya. It is true that the Arab grammarians mention a considerable amount of differences between tribal dialects (lug˙a¯t), but these concern almost without exception lexical and phonetic features. In this account, the Arabic language was “corrupted” by the contact with the inhabitants of the conquered territories after the conquests. This resulted in a drive for the preservation of the language by the grammarians, thanks to whom the Classical language remained the standard language for cultural, political, and above all, religious purposes. The Arabs emphasize the vulgarity and decadence of the language of the sedentary population as against the purity of the language of the Bedouin. Some Western Arabists have followed this view of the development of the language (cf. Nöldeke 1904; Blau 1977), and it has been re-interpreted in terms of the acquisition process of the language, which may be compared with processes of pidginization and creolization (cf. Versteegh 1984; 1997: 102⫺112; but see Holes 1995: 19⫺24). In this article, the type of Arabic that became the colloquial language after the period of the conquests will be called “New Arabic” or “Arabic dialects”. In the contemporary Arab world this colloquial language is usually called the “common, vulgar language” (¤a¯mmiyya) as against the “correct, Classical language” (fusøhø a¯). Both varieties serve as the extremes in a continuum of speech variation, which was described as one of the classic cases of diglossia by Charles Ferguson
1741
(1959 a). Even according to those Arabists who maintain that some of the features of New Arabic were already present in the posited pre-Islamic colloquial language, the bulk of the linguistic changes were introduced during the period of the conquests. Explanations for the development of the colloquial language are complicated by the occurrence in the dialects of common differences that set them apart from the Classical language, as well as differences among them. The common features of New Arabic are explained either by a theory of monogenetic origin (cf. Ferguson 1959 b), or by a general drift, which is compared to similar trends in other Semitic languages, such as Neo-Aramaic and Amharic. Others draw the attention to subsequent processes of convergence, which led to greater similarity between the dialects (cf. Cohen 1970; Diem 1978). The differences between the various dialect areas are sometimes explained by substratal influence (cf. Diem 1979; Versteegh 1997: 104⫺107). In some instances, the dialects exhibit analogous processes of change with varying lexical realization, such as the development of an analytical possessive construction and a system of aspectual markers in the verb: each dialect has its own characteristic genitive exponent and aspectual particles. In view of this, the most reasonable account for the emergence of the dialects seems to be an explanation in terms of a polygenetic origin. Accordingly, the similarity of the development they went through must be the result of the similarity in acquisition process of the language by the inhabitants of the conquered territories. This may explain some of the common features. Other features, especially in the lexicon, may have spread by subsequent convergence. An important common factor in the development of New Arabic is the influence of the Classical language, which was pervasive in all Arabophone areas, except in the so-called language islands (Malta, Cyprus, Anatolia, Uzbekistan, Afghanistan). The development in the latter dialects is characterized by the fact that from a certain point in time onwards they no longer underwent the influence of the Classical standard, so that elements of the original structure could be maintained. Changes in the structure of the dialects of the language islands did occur under the influence of the adstratal languages (e.g., Italian in the case of Maltese, Uzbek in the case of Uzbekistan Arabic, Turkish and Kurdish in the case of Anatolian Arabic).
1742 Within the Arabic dialects the dialects of the Bedouin occupy a special position (cf. Rosenhouse 1984). The Arab grammarians asserted that only the Bedouin spoke a pure Arabic, with all the inflectional features preserved. According to them, the Bedouin only gradually succumbed to the corruption of sedentary speech. This constitutes an important argument for the thesis according to which the ¤Arabiyya was, indeed, of the Classical type, with only minor regional differences. The main criteria that distinguish Bedouin from sedentary dialects are: the realization of Classical /q/ (Bedouin voiced [g], [gy] etc.; sedentary voiceless [q], [Å] etc.); the realization of the Classical interdentals (Bedouin [t], [dß ], [d døß ]; sedentary [t], [d], [dø ]); the realization of some morphological features in Bedouin dialects (e.g., the ending -in, -an as a marker of indefiniteness; gender distinctions in the plural of the verb and the pronoun; internal passive; causative). The characterization of a dialect as Bedouin or sedentary is a linguistic one, since there have been many secondary processes of bedouinization and sedentarization, for instance in the dialects of Baghdad, where the Muslim dialect was bedouinized, whereas the Christians and the Jews continued to speak their sedentary dialects (cf. Blanc 1964). Geographically, the contemporary dialects are divided into the Western and the Eastern dialects. The main isogloss between these two areas is that of the 1st person singular and plural of the imperfect verb (cf. 2.1). The Western dialects include the dialects of North ø assa¯niyya dialect of Mauretania, Africa: the H Moroccan (for the dialect of Rabat, cf. Harrell 1962; Caubet 1993), Algerian, Tunisian, and Libyan. The Eastern dialects are usually divided into four major groups: Egyptian Arabic (cf. Behnstedt & Woidich 1982⫺1999; Woidich 1990), including Sudanese and Chad Arabic, with outliers stretching into Nigeria (cf. Owens 1993); Syro-Lebanese dialects (for the dialect of Damascus, cf. Grotzfeld 1965); Mesopotamian dialects (cf. Jastrow 1978; for the dialect of Baghdad, cf. Erwin 1963), including the dialects of language islands in Anatolia and Uzbekistan; and the dialects of the Arabian peninsula: Yemeni (cf. Behnstedt 1985/1992; Watson 1993), dialects of SaudiArabia (cf. Prochazka 1988; Ingham 1982; 1994), and Eastern Arabic Gulf dialects (cf. Holes 1990). Within each area, there is a division into Bedouin and sedentary dialects.
XVIII. Morphologischer Wandel II: Fallstudien
The description of the Arabic dialects is complicated by the diglottic situation. On the one hand, the existing continuum between dialect and standard and the constant interference between the extremes make it very difficult to elicit ‘pure’ dialect forms. Especially in the presence of a researcher speakers tend to use the prestigious variants more often than they normally do. As a result, the description of dialects often reflects Classical structures, and researchers usually interpret New Arabic structures in terms of Classical Arabic. On the other hand, speakers actually use the prestigious variants much more frequently nowadays because of increased education and exposure to the language of the media. A case in point is the analytical possessive construction, which for most speakers competes with the Classical construction even in informal speech. In this connection, the dialects of the language islands may serve as a point of reference: to some degree, they have escaped the influence of the Classical language and are, therefore closer to the original dialect structure. In general, scholars maintain that the development of the Arabic dialects (apart from those in the language islands and from the attested cases of pidginization/creolization in the Arab world, cf. 5) is not very spectacular compared to the development in, for instance, the Neo-Aramaic dialects (cf. Diem 1978). To some extent, this is correct because of the influence of the Classical language, but on the whole, such a view tends to obscure the most salient traits of New Arabic, especially in the field of syntax. Most scholars would agree that the development of the dialects is characterized by a process of reduction and a tendency towards greater analyticity vis-a`-vis the Classical language. Partly because of changes in stress and accent (cf. Janssens 1972) New Arabic has lost the final short vowels, some of which served as the core of the inflectional endings. This has led some scholars to assume that the disappearance of the inflectional system may be explained phonologically. On the whole, however, this is not acceptable (cf. Diem 1991) since the inflectional system has broken down in its entirety, and the changes involved are clearly structural ones.
2.
Morphology of the verb
Classical Arabic has a verbal system with distinctions for number (singular, dual, plural), gender (masculine, feminine in the 2nd and
1743
160. From Classical Arabic to the modern Arabic vernaculars
3rd person), person (1st, 2nd, 3rd). Most verbs consist of three radicals that are usually indicated by the three consonants f-¤-l. Traditionally, a perfect (fa¤ala) and an imperfect (yaf¤a/i/ulu) verb are distinguished, along with an imperative (if¤al). Morphologically the perfect, which denotes past actions with a punctual, perfective or stative aspect, is marked by suffixes, and the imperfect, which denotes actions with a durative or habitual aspect, is marked by prefixes and suffixes. The imperfect verb has a three-way distinction between indicative, subjunctive (mainly after conjunctions) and apocopate or jussive (mainly for adhortative and prohibitive functions), which are marked by a change in endings, e.g. yaf¤al-u (indicative), yaf¤al-a (subjunctive), yaf¤al (jussive).
⫺
⫺ ⫺
⫺
2.1. The sound verb The paradigm for sound verbs (i.e., those not containing w or y as one of their radicals) is as follows (cf. Tab. 160.1). In all or most dialects the following developments have taken place, usually as a result of reduction or analogous change: ⫺ the category of the dual has disappeared without leaving any traces; ⫺ in most sedentary dialects the gender distinction has disappeared in the 2nd and 3rd person plural, in Western dialects even in the singular; ⫺ the modal distinctions in the imperfect verb have disappeared through the loss of the short vowel endings and the ending -na; ⫺ in most dialects the personal endings -tu/ -ta have merged in -t; for the feminine ending of the 2nd person singular a new
‘to write’ perfect
imperfect
singular 1st 2nd masc. 2nd fem. 3rd masc. 3rd fem.
katabtu katabta katabti kataba katabat
1st 2nd masc. 2nd fem. 3rd masc. 3rd fem.
Åaktubu taktubu taktubı¯na yaktubu taktubu
Tab. 160.1: The sound verb in Classical Arabic
⫺
⫺
ending -tı¯ (> -ti) has emerged; in the Western dialects -ta/-ti have merged in -tı¯ (> -ti), whereas the 1st person has become -t; the personal prefixes in the imperfect verb have the vowel -i- instead of Classical -a- in almost all dialects; this phenomenon (called taltala) was already known in some of the pre-Islamic dialects (cf. Rabin 1951: 61); the imperfect verb has developed a system of aspectual markers (cf. 2.4); some dialects have achieved greater regularity for the perfect plural endings -tum/ -u¯, e.g., in some Egyptian dialects katabtum, katabum; other dialects have -tu/-u; Western dialects have -tum/-tı¯w; all Western dialects are characterized by a 1st person singular prefix of the imperfect verb ni-, plural ni-..-u; the existence of a mixed paradigm in the Egyptian delta (aktib/niktibu) shows that the earlier explanation of this development as a contraction from Åana Åaktubu ‘I write’ > naktub is probably wrong (cf. Behnstedt 1978: 69); in Classical Arabic verbal bases of the form fa¤ala alternate with verbal bases of the pattern fa¤ila (for non-permanent states, e.g., g˙adø iba ‘to be angry’) and fa¤ula (for permanent states, e.g., hø asuna ‘to be good’); in the dialects only two bases have been preserved without semantic distinction, namely fa¤al and fi¤il; in Moroccan and Algerian Arabic there is only one base, f¤el; in Classical Arabic the perfect bases correlate with imperfect patterns: fa¤ala verbs have imperfect yaf ¤ulu or yaf ¤ilu (in
dual katabtuma¯ kataba¯ katabata¯ taktuba¯ni yaktuba¯ni taktuba¯ni
plural katabna¯ katabtum katabtunna katabu¯ katabna naktubu taktubu¯ na taktubna yaktubu¯ na yaktubna
1744
XVIII. Morphologischer Wandel II: Fallstudien
Classical
Egyptian (Cairo)
‘Iraqi (Baghdad)
kataba/yaktubu dø araba/yadø ribu haraba/yahrabu ¤amila/ya¤malu labisa/yalbasu sˇariba/yasˇrabu
katab/yiktib dø arab/yidø rib harab/yihrab ¤amal/yi¤mil libis/yilbis sˇirib/yisˇrab
kitab/yiktib ødß urab/yud ødß rub hirab/yuhrub ¤imal/yi¤mal libas/yilbas sˇirab/yisˇrab
‘to ‘to ‘to ‘to ‘to ‘to
write’ hit’ flee’ work’ dress’ drink’
Tab. 160.2: Perfect/imperfect patterns in Classical Arabic and two dialects
some phonological contexts yaf¤alu); fa¤ila verbs regularly have yaf¤alu; and fa¤ula verbs regularly have yaf¤ulu; in the dialects reflexes of these three patterns occur, but often with a different distribution (cf. Tab. 160.2). As an example of one of the most conservative dialects, in which only the modal endings have disappeared, we quote here the verbal paradigm of a Saudi-Arabian Bedouin dia˙ a¯mid (cf. Prochazka 1988: 27) (cf. lect, G Tab. 160.3): ‘to write’
singular plural 1st 2nd masc. 2nd fem. 3rd masc. 3rd fem.
katabt katabt katabti katab katabat
katabna katabtum katabtunnah katabaw katabnah
imperfect 1st 2nd masc. 2nd fem. 3rd masc. 3rd fem.
aktub tiktub tiktubı¯n yiktub tiktub
niktub tiktubu¯ n tiktubnah yiktubu¯ n yiktubnah
perfect
˙ a¯mid (SaudiTab. 160.3: Verbal paradigm of G Arabia)
As an example of an innovative dialect with complete disappearance of gender distinction, except in the 3rd person singular we quote the verbal paradigm of sedentary Moroccan (Rabat) (cf. Tab. 160.4). 2.2. The weak verb In Classical Arabic the conjugation of the weak verbs (i.e., those containing a w or y as one of their radicals) differs from that of the sound verbs as the result of various morphophonological rules. The following types may be distinguished (I, II, III indicating the three radicals of the verbal stem) (cf. Tab. 160.5):
singular plural perfect
1st 2nd masc. 2nd fem. 3rd masc. 3rd fem.
imperfect 1st 2nd masc. 2nd fem. 3rd masc. 3rd fem.
ktebt
ktebna
ktebti
ktktebtiw
kteb ketbet
ketbu
nekteb tekteb tketbi yekteb tekteb
nketbu tketbu yketbu
Tab. 160.4: Verbal paradigm of sedentary Moroccan (Rabat)
In New Arabic the following developments in the conjugation of the weak verbs may be noted; they all tend towards greater regularity: ⫺ in most dialects the type Iw has become sound, e.g., in Egyptian the verbs wazan/ yiwzin ‘to weigh’, wa¤ad/yiw¤id ‘to promise’, wiÅi¤/yiwÅa¤ ‘to fall down’; ⫺ the two types IIIw and IIIy have merged in all dialects, e.g., in Egyptian misˇi/ misˇe¯t/yimsˇi ‘to walk’ (IIIy, e¯ < ay) and da¤a/da¤e¯t/yid¤i ‘to invite’ (IIIw); the same development took place in the perfect of the laqiya group, e.g., Egyptian laÅa/laÅe¯t/ yilÅa ‘to meet’; ⫺ the type II ⫽ III has merged with the II measure of the IIIy verbs, e.g., in Egyptian radda/radde¯t ‘to return’ like the II measure of masˇa: masˇsˇa/masˇsˇe¯t. A tendency may be noted in the dialects towards greater similarity between the endings of the weak and those of the sound verbs, especially in the case of the type IIIy. In many cases this similarity was achieved by
1745
160. From Classical Arabic to the modern Arabic vernaculars
underlying Iw IIw IIy
|wa¤ada| |qawama| |sayara| |hawifa| IIIw |da¤awa| IIIy |masˇaya| |laqiya| II ⫽ III |radada|
3rd person perfect
3rd person imperfect
1st person imperfect
wa¤ada qa¯ma sa¯ra ha¯fa da¤ a¯ masˇa¯ laqiya radda
ya¤ idu yaqu¯mu yası¯ru yaha¯fu yad¤ u¯ yamsˇ¯ı yalqa¯ yaruddu
wa¤adtu qumtu sirtu hiftu da¤ awtu masˇaytu laqı¯tu radadtu
Tab. 160.5: Types of weak verbs in Classical Arabic
assigning the sound endings to the weak verb, but alternatively, there are dialects in which the weak endings are transferred to the sound verb, either in the perfect (e.g., Muslim dialect of Baghdad) or in the imperfect (e.g., Jewish dialect of Baghdad), as shown in Tab. 160.6 (where e¯ < ay, o¯ < aw).
Similar developments have taken place in French loans in Moroccan Arabic (cf. Heath 1989: 108⫺112).
Classical
Egyptian Muslim Jewish (Cairo) Baghdad Baghdad
katabu¯ 3rd plural masˇaw 3rd plural yaktubu¯ na 3rd plural masc. yansayna 3rd plural masc. taktubı¯na 2nd singular fem. tansayna 2nd singular fem.
katabu
kitbaw
katbu
2.3. The verbal measures In Classical Arabic the verb system may be enlarged to form semanto-syntactic categories, the so-called verbal measures. Usually the following measures are recognized as productive stem formations (the main function indicated as far as possible, although in many cases the meaning is fixed lexically):
misˇu
masˇaw
masˇaw
II
kattaba
yiktibu
yikitbu¯ n yketbo¯n
yinsu
yinsu¯ n
yenso¯n
tiktibi
tikitbı¯n
tketbe¯n
III IV V VI VII VIII IX X
ka¯taba Åaktaba takattaba taka¯taba inkataba iktataba (ihø marra) istaktaba
tinsi
tinsı¯n
tense¯n
Tab. 160.6: Weak and sound endings in Arabic dialects
The development of the weak endings was important in another respect as well. In Maltese Arabic the pattern of the weak verbs has proliferated to such an extent that foreign loans, especially from Italian, which often ended in a vowel, could be integrated easily (cf. Mifsud 1995), e.g., iddependa/jiddependi ‘to depend’ (< Italian dipendere), importa/ jimporta ‘to matter’ (< Italian importare).
intensive, causative, denominative transitive to direct object causative reflexive-intransitive to II reciprocal to III reflexive-passive reflexive-intransitive to I colours, bodily defects request, opinion of I
Tab. 160.7: The verbal measures in Classical Arabic
In most dialects the II and III measures have been preserved, although only the II measure is productive (as intensive, denominative or causative); the III measure is lexically fixed and functions just as an alternative vowel pattern, e.g., Moroccan qtel ‘to kill’/qatel ‘to fight’. The IV measure has disappeared from almost all dialects and is replaced by the II measure, e.g., Moroccan kteb ‘to write’/ketteb ‘to make so. write’ (Classical Arabic Åaktaba), or by a periphrastic expression. Only in some Bedouin dialects in the East a IV measure is still productive, e.g., De¯r iz-Zo¯r (Syria) hø addøß ar/yihø døß ur ‘to be present’, ihddøß ar/yihddøß ir ‘to
1746 produce’ (Fischer & Jastrow 1980, eds.: 46; cf. Ingham 1994: 77 for Nagˇdı¯ Arabic). In Classical Arabic, the IX measure was strictly limited to verbs denoting colours and bodily defects; in the dialects this form has often been replaced by the II measure, e.g., in the Gulf dialects bayyad ‘to make white’ or ‘to be white’ (Fischer & Jastrow 1980, eds.: 71); other dialects have preserved a restricted IX measure, e.g., Egyptian ihø marra ‘to be red’. In Maghrebine dialects a new form has developed, e.g., Moroccan hø mar ‘to be red’. The X measure exists only in isolated lexical items, often loans from Classical Arabic, e.g., Moroccan stehø sen ‘to approve’, sted¤a ‘to summon’. Each verbal measure has a passive in Classical Arabic, marked by a characteristic vowel pattern (so-called ‘internal passive’), e.g., I measure kataba/yaktubu, passive kutiba/yuktabu; VIII measure iktataba/yaktatibu, passive uktutiba/yuktatabu. In most dialects the internal passive has disappeared (Retsö 1983), but some Bedouin dialects have preserved a passive formation, e.g., in Saudi-Arabia (Rista¯q) from gatal ‘to kill’ gtil/yugtal (Fischer & Jastrow 1980, eds.: 117, 262; cf. Ingham 1994: 26⫺28). In the H ø assa¯niyya dialect of Mauretania a new internal passive has developed for derived stems, e.g., from søahø hø ahø (II) ‘to correct’ usøahø hø ahø /yusøahø hø ahø (Zavadovskij 1981: 35 f.). In the other dialects some of the verbal measures have become specialized for the passive (cf. Retsö 1983). In some dialects the VII measure is used for this purpose, e.g., in H ø assa¯niyya from sˇrab ‘to drink’ nsˇrab ‘to be drunk’ (Zavadovskij 1981: 34). In other dialects the Classical V and VI measures with their intransitive meaning are used for the passive, building on the Classical pattern of verbs like g˙ayyara II ‘to change st.’, tag˙ayyara V ‘to be changed, to change [intrans.]’. In these dialects the marker ta- has become a passive marker which may even be used with the base stem (measure I). Examples are Moroccan kteb (I) ‘to write’/ttekteb (tI, with secondary initial gemination) ‘to be written’; semma (II) ‘to call’/tsemma (tII) ‘to be called’; Egyptian misik ‘to catch’/itmasak ‘to be caught’. In the tIII stems the Classical reciprocal meaning has usually been preserved, e.g., Moroccan tqatel ‘to fight each other’. The vowel patterns of the derived measures show a tendency towards a uniform vocalization. In Egyptian, for instance, the per-
XVIII. Morphologischer Wandel II: Fallstudien
fect and the imperfect of the II measure have acquired identical vowels: Classical Arabic kallama/yukallimu ‘to talk’, Egyptian kallim/ yikallim. In the tII and tIII measures the vocalization is identical: Classical Arabic ta¤allama/yata¤allamu ‘to learn’, taqa¯bala/yataqa¯balu ‘to be met, to meet each other’, Egyptian it¤allim/yit¤allim; itÅa¯bil/yitÅa¯bil (Jastrow 1982: 136). 2.4. The aspectual markers The most radical departure from the structure of Classical Arabic is the marking of verbal aspect. The main function of the two verbal forms, perfect and imperfect, in Classical Arabic seems to have been aspectual (Denz 1982: 71 f.; but cf. Eisele 1999). They may be combined with the auxiliary verb ka¯na to indicate past perfectivity and past durativity or habituality, respectively, e.g., ka¯na qad kataba ‘he had written’, ka¯na yaktubu ‘he was writing, he used to write’. There are two particles that are connected with aspect/tense: qad indicates perfectivity and sawfa (sa-) indicates future tense. Neither particle has any reflex in New Arabic. In the dialects a new form of aspectual marking is used. The old imperfect, stripped of its modal endings, has become a modal form itself, e.g., Egyptian tisˇrab Åahwa ‘would you like to have coffee?’. Durative and future/volitional aspect are indicated by clitic particles, whose etymology is not always clear, but seems to be connected with the notion of ‘being at, sitting’ for the durative and ‘going’ or ‘desiring’ for the future. The particles are highly characteristic for each dialect (Czapkiewicz 1975) (cf. Tab. 160.8).
Egyptian Arabic Moroccan Arabic Syrian Arabic Baghdad (Muslim)
durative
future
bika¤amda-
hø ag˙abirahø -
Tab. 160.8: Aspectual markers in Arabic dialects
Apart from the geographical variation the markers vary with regard to their semantic domain: the Syrian marker bi-, for instance, is used for a volitional aspect (probably < yibg˙i ‘to desire’), whereas certain future actions are indicated in this dialect with a marker rahø - (< ra¯yihø ‘going’). Another example is the indication of habitual or general statements. In some dialects, for instance
1747
160. From Classical Arabic to the modern Arabic vernaculars
Egyptian and Moroccan, the durative marker is used for habitual actions, whereas in Syrian the bi- marker is used for that purpose. In ¤Iraqi Arabic (cf. Erwin 1963: 335⫺344) the durative particle is restricted to a certain period of time, whereas the modal imperfect is used for general and habitual statements. The dialects also differ with regard to the combinability of the aspectual markers. In most dialects the durative marker may be combined with the past marker ka¯n to indicate past durativity, e.g., Egyptian ka¯n byiktib ‘he was writing’; the combination ka¯n ⫹ future marker to indicate irreality is not available in all dialects: in Egyptian ka¯n hø ayiktib means ‘he was going to write (but he didn’t)’, but in Moroccan the combination ka¯n g˙a- is impossible. In Classical Arabic the active participle fa¯¤ilun may be used to indicate durativity: huwa ka¯tibun ‘he is writing, he is going to write’ (Denz 1982: 72 f.). In many dialects the participle is used to indicate a resultative or perfective aspect, e.g., Egyptian ana na¯si ‘I have forgotten’, ana wa¯kil ‘I have eaten’. In Uzbekistan Arabic the participle has become an integral part of the verbal system: it is used instead of the perfect for punctual actions in the past, and it takes both subject and object suffixes, e.g., zo¯rib ‘he hit’, zorbinı¯ ‘I hit’, zorbinı¯k ‘I hit you’ (Fischer 1961: 255). 2.5. The verbal complex In Classical Arabic pronominal suffixes referring to the object may be added to the verb, e.g., katabtu-hu ‘I wrote it’. There are no restrictions with regard to the person; although fairly unusual, combinations of suffixes may occur with ditransitive verbs, e.g., Åa¤lamtuka-hu ‘I let you know it’ (the preferred form is with the particle Åiyya¯: Åa¤lamtu-ka Åiyya¯hu). The indirect pronominal object is indicated with the preposition li-, e.g., katabtuhu la-ha¯ ‘I wrote it to her’. The indirect object is not incorporated in the verbal form, and it can be separated from it, e.g., qultu da¯lika laka ‘I told you that’. In general, the modern dialects have preserved the form of the pronominal suffixes. Noteworthy is the disappearance of the -hof the 3rd person masc. sing. in virtually all dialects, e.g., Egyptian katabt-o ‘I wrote it’. In some dialects this leads to a secondary stress contrast with suffixless forms, e.g., Syrian ka´tabu ‘they wrote’ vs. katabu´ ‘they wrote it’. The normal morphophonemic rules lead to epenthesis and vocalic switch, e.g., Syrian
yaktubu-hu > yiktbo > yeketbo, or Egyptian (with stressed epenthetic vowel) yiktı´bo. An innovation is the development of the indirect object as a clitic to the verbal form, e.g., Egyptian katab-lı´na ‘he wrote to us’. In most dialects a direct pronominal object of the 3rd person may be combined with an indirect object suffix, usually without constraints as to the person of the indirect object, e.g., Egyptian iddet-ha¯-lo ‘I gave her to him’, bi-ywarri-ho-lhom ‘he shows him to them’. In Anatolian Arabic (Daragözü) the indirect object precedes the direct object, which may only be 3rd person, e.g., weddı¯-lı¯a ‘bring her to me!’, weddı¯-la¯-en ‘bring them to her!’ (Jastrow 1973: 68). For the development of subject suffixes with the participle in Uzbekistan Arabic see 2.4. As we have seen in 2.4 the aspectual markers are added clitically as prefixes to the verbal form. The negative circumfix ma¯ ... sˇ may be added to the verbal form as well, so that morphologically rather complex forms may result. As an example we may quote Moroccan ma-g˙a-nekteb-o-lek-sˇ ‘neg-fut-I.writeit-to.you-neg (I won’t write it to you)’ or Egyptian ma-bi-tgib-ha¯-lhom-sˇ ‘neg-duryou.bring-her-to.them-neg (you are not bringing her to him)’.
3.
Morphology of the noun
In Classical Arabic substantives and adjectives are inflected for case (nominative, genitive, accusative), number (singular, dual, plural) and (in)definiteness. There is a two-way distinction between masculine and feminine nouns, the feminine nouns being characterized by the ending -at (pausal form -ah), and for some categories of nouns, by -a¯, -a¯Å. Indefiniteness is marked by the ending -n (the so-called nunation), definiteness by the definite article al- and the absence of nunation. 3.1. Case The singular declension is as follows (cf. Tab. 160.9):
nominative genitive accusative
masculine ‘man’
feminine ‘city’
ragˇul-u-n ragˇul-i-n ragˇul-a-n
madı¯n-at-u-n madı¯n-at-i-n madı¯n-at-an
Tab. 160.9: Singular declension of the noun in Classical Arabic
1748
XVIII. Morphologischer Wandel II: Fallstudien
In New Arabic the feminine ending has become -a in sedentary dialects; the form with -at (> -it) appears before pronominal suffixes and in possessive constructions (Egyptian Åo¯da ‘room’, Åo¯dtak ‘your room’, Åo¯dit ilbe¯t ‘the room of the house’). Bedouin dialects retain the pausal form -ah, -ih. In some dialects a redistribution of the feminine ending has taken place, depending on the preceding consonant, e.g., Syrian mnı¯hø a ‘good:f’ vs. kbı¯re ‘big:f’. The alternative endings -a¯, -a¯’ have disappeared, e.g., Egyptian hø amra ‘red:f’ vs. Classical hø amra¯Åu. Since the short vowel endings of Classical Arabic have all disappeared in New Arabic, there is no inflection for case in the singular. The fact that in the sound plural, too, only one case ending has been preserved (cf. 3.2) and the fact that the nunation as marker of indefiniteness has disappeared completely (apart from a few traces in Bedouin dialects), clearly show that the development is a morphological one (cf. Diem 1991). The syntactic function of the cases has been transferred to syntactic devices: the word order in the dialects has become more or less fixed to SVO; the adverbial and adnominal functions of the accusative have been transferred to other constructions, and the possessive function of the genitive has been taken over by the genitive exponent. The emergence of the genitive exponent constitutes one of the most obvious changes in the structure of the language towards greater analyticity, e.g., Egyptian il-be¯t bita¯¤ il-wizı¯r ‘the house of the minister’ instead of Classical Arabic bayt-u l-wazı¯r-i. The etymological origin of the genitive exponent is usually a noun meaning ‘possession’, ‘belonging to’, ‘thing’. Just as with the aspectual markers of the verb (cf. 2.4) the choice of the exponent is highly characteristic for each dialect area (cf. Harning 1980) (cf. Tab. 160.10): Syrian Arabic Egyptian Arabic Moroccan Arabic Baghdad Arabic (Muslims) Sudanese Arabic Cypriot Arabic Sa¤ı¯dı¯ Egyptian Chad Arabic Anatolian Arabic (Daragözü) Maltese
taba¤ bita¯¤ d-, dyal ma¯l hø aqq sˇa´yt sˇug˙l hana lı¯l ta
Tab. 160.10: Genitive exponents in the Arabic dialects
Although not strictly relevant for the morphological development, two remarks should be added here. In most dialects, the analytical possessive construction coexists with the Classical construction. The choice for the latter is in some cases obligatory, due to semantic constraints. In Moroccan Arabic, for instance, the analytical construction is not used with terms of kinship, e.g., m øˆ m øˆ -i ‘my mother’ (not *l-umm dyal-i). In most dialects there is a distinction between alienable and inalienable possession (cf. Art. 103), e.g., Egyptian il-lahø m ibta¯¤i ‘my piece of meat’ vs. lahø m-i ‘my flesh’. Apart from these semantic distinctions there is a high degree of sociolinguistic variation in the use of the possessive construction. The current diglossia has (re-)introduced the Classical construction as an alternative in the dialect speech community. Not surprisingly, the language islands do not exhibit such variation. In Maltese Arabic, for instance, the analytic construction with the genitive exponent ta is obligatory, except for a few isolated expressions and terms of kinship. Conversely, in some of the more conservative Bedouin dialects the analytic genitive is hardly ever used. In some dialects the disappearance of the formal marker for indefiniteness has led to the development of an indefinite article. Examples are ¤Iraqi fat (< Classical Arabic fard ‘individual’), also in Uzbekistan Arabic (which has lost the definite article), Moroccan Arabic wahø d (< Classical Arabic wa¯hø id ‘one’), Cypriot Arabic ehen/ehte (< Classical Arabic Åahø ad ‘one’; cf. Borg 1985: 145). In a few Bedouin dialects the nunation has not disappeared completely since there is a generalized nunated ending, acting as a connector between noun and attribute and indicating indefiniteness without marking the syntactic function, e.g., in Nagˇdı¯ Arabic (Saudi-Arabia) the indefinite marker -in (cf. Ingham 1982: 53⫺56), kalmit-in ga¯lo-ha¯-li ‘a word which they said to me’, gizÅ-in min-h ‘a part of it’; in Uzbekistan Arabic the same marker -in is used, e.g., fat hø agˇart-in kabı¯ra ‘a big stone’ (cf. Fischer 1961: 244, who derives this marker from Classical Arabic Åayna ‘where?’). Andalusian Arabic seems to have known the ending -an as a connective element (cf. Corriente 1977: 121). In Yemeni Arabic a contrast exists between bayt-u ‘a house’ and al-bayt ‘the house’ (cf. Fischer & Jastrow 1980, eds.: 120).
160. From Classical Arabic to the modern Arabic vernaculars
3.2. Number Classical Arabic has three numbers, singular, dual, and plural. The dual is marked by the endings -a¯ni (nom.)/-ayni (gen./acc.) for masculine, and -ata¯ni/-atayni for feminine nouns. One of the characteristics of the modern dialects is that they have lost the dual in the verb and the pronoun. In most dialects the formation of dual nouns is not productive either, except for a limited category of nouns. In Egyptian and Syrian, for instance, the dual is current with terms of kinship, e.g., binte¯n ‘two daughters’, but it may be used occasionally with other nouns as well. In Moroccan the dual is only used with expressions indicating a period of time, e.g., yumayn ‘two days’, sa¤tayn ‘two hours’. In some dialects the ending -ayn is also used for paired body parts, the so-called pseudo-dual (cf. Blanc 1970), e.g. Syrian ¤e¯ne¯n ‘eyes’, ¯ıde¯n ‘hands’. The difference between this pseudo-dual and the original dual is obvious in the contrast between ¤e¯ne¯n, which functions as a plural, and ¤e¯nte¯n ‘two eyes’, which functions semantically as a dual. In combination with pronominal suffixes the pseudo-dual loses the ending -n, as in Classical Arabic ⫺ compare ¯ıde¯k ‘your hands’ with binte¯nak ‘your two daughters’. In Classical Arabic there are two kinds of plural. The so-called sound plural is formed with the help of suffixes: -u¯ na (nom.)/-ı¯na (gen./acc.) for masculine, and -a¯tun/-a¯tin for feminine nouns. The sound plural is used almost exclusively with some adjectives and all active participles, e.g., gˇamı¯lu¯ na/gˇamı¯la¯tun ‘beautiful’; muslimu¯ na/muslima¯tun ‘Muslims [act. part. IV measure]’. The second type of plural (the so-called broken plural) consists in a rearrangement of the vowel pattern of the singular. More than thirty different patterns are used, some of them connected with specific singular patterns, but by no means exclusively. A few examples: ragˇulun/rigˇa¯lun ‘man’; kita¯bun/kutubun ‘book’; wazı¯run/wuzara¯’u ‘minister’; øsadı¯qun/Åasø diqa¯Åu ‘friend’; tøa¯libun/tøulla¯bun ‘student’. The broken plurals are declined as singular nouns, except for the ones ending in -u (diptotic plurals) which have their gen./acc. in -a when they are indefinite. As we have seen in 3.1, New Arabic has lost its declensional system: the sound plural is marked by the oblique ending -ı¯n for masculine and -a¯t for feminine nouns. In the iso-
1749
lated dialects there is a tendency to generalize these endings. In Uzbekistan Arabic (Fischer 1961: 243), for instance, the ending -ı¯n is reserved for male persons, e.g., uhwı¯n ‘brothers’ (Classical Arabic Åihwatun), zugˇ¯ın ‘husbands’ (Classical Arabic Åazwa¯gˇun), whereas the ending -a¯t is used for female persons and all other plurals, e.g., ra¯sa¯t ‘heads’ (Classical Arabic ruÅu¯sun), uha¯t ‘sisters’ (Classical Arabic Åahawa¯t). In Cypriot Arabic (cf. Borg 1985: 180) -a¯t is the general plural ending for a large number of inanimate nouns, even for those which etymologically are broken plurals, e.g., huma¯t ‘kinds of meat’ (Classical Arabic lahø m, plural luhø u¯m). The usual plural ending for adjectives in the dialects is -ı¯n/-a¯t. For adjectives of the type fa¤¯ıl, e.g., kabı¯r ‘big’, all dialects have developed a new form, fi¤¯ıl; the plural for this type has become fu¤a¯l (Classical Arabic fi¤a¯l), e.g., Egyptian kuba¯r, whereas substantives with the same plural ending in Classical Arabic retain fi¤ a¯l, e.g., gamal/gima¯l ‘camel’. Most dialects have retained a number of broken plural types, although with less variation than in Classical Arabic. In some cases the dialects have expanded the domain of certain types of plurals. The two patterns fa¤ a¯lı¯l and fa¤ a¯lil for quadriliteral nouns have merged in most dialects into fa¤ alı¯l, e.g., Egyptian sˇababı¯k (Classical Arabic sˇaba¯bı¯ku) from sˇubba¯k ‘window’. In Western dialects the resulting pattern is f¤a¯lil, e.g., Moroccan sˇebbak/sˇbabek. The plural pattern Åaf¤a¯l has become identical in many dialects with the pattern fi¤a¯l, due to the disappearance of the pretonic vowel, e.g., Moroccan qlem/qlam ‘pen’ (Classical Arabic qalamun/Åaqla¯mun). The plural pattern fa¤ a¯lı¯ > fa¤ a¯li, which in Classical Arabic was reserved for the singular types fa¤ la¯, fa¤ la¯Åu, has become much more general, e.g., Egyptian hø a¯ra/hø awa¯ri ‘alley’ (Classical Arabic hø a¯ra/hø a¯ra¯t). In one category of nouns, expressing professional activities, the ending -a¯t has been replaced with the feminine singular (and collective) ending -a, e.g., Syrian lahø hø a¯m/lahø hø a¯me. Weak nouns of this type have the ending -iyya, e.g., Egyptian hø ara¯mi/hø ara¯miyya ‘thief’. The latter ending became generalized in foreign loans, e.g., Syrian sˇo¯fe¯r/sˇo¯fe¯rı¯ye ‘driver’. In most dialects new patterns for the plural have developed. In Yemeni Arabic, for instance, the plural af¤u¯l is found, e.g. ¤amm/ a¤mu¯m ‘uncle’ (cf. Fischer & Jastrow 1980, eds.: 91). Another new plural pattern in Ye-
1750 meni Arabic is fi¤wal, fu¤wal for singulars of the type fa¤¯ıl, e.g., tøarı¯q/tøurwaq ‘street’. This pattern has sometimes been associated with the plural qetwo¯l, qetyo¯l in Modern South Arabian, but according to others, it is a case of common Semitic stock rather than direct substratal influence (cf. Diem 1979: 64⫺75). 3.3. Nominal patterns As we have seen in the case of the broken plurals in the modern dialects, New Arabic has preserved the triradical system of Classical Arabic. Alongside the inflectional patterns Classical Arabic used a large number of derivational patterns, which were partly productive, partly fixed. In the case of singular nouns, some types appear to constitute semantically constrained categories, e.g., the type fa¤¯ıl, which is prevalent in adjectives. Clear examples of derivational patterns are the following: Åaf ¤alu (colours), fa¤¤a¯ l (occupations), mif ¤a¯l (instruments), maf ¤ala, maf ¤ala (places of action). In New Arabic most of these patterns have been preserved, and in some cases they have regained productivity through the efforts of the Arab Academies to create a pure Arabic lexicon with the help of derivational patterns (cf. Ali 1987). This development stimulated the creation of new word patterns in the dialects, or the expansion of old patterns. A case in point is the pattern fu¤a¯l, which has become the normal pattern for illnesses, e.g., søufa¯r ‘jaundice’ (from asø far ‘yellow’). In a few cases the interference from other languages led to the formation of completely new patterns; an example is the pattern tafe¤¤alet indicating professions in Moroccan Arabic, which was borrowed from Berber, e.g., tanezˇzˇaret ‘carpentry’. The most serious threat to the triradical structure of Arabic was the enormous influx of foreign loans. In Classical and Modern Standard Arabic most loans were integrated more or less fully into the system, e.g., faylasu¯f/fala¯sifa ‘philosopher’ (< Greek philo´sophos); film/Åafla¯m. In the modern dialects of North Africa the borrowing of French loans as a rule did not cause any major deviation from the triradical structure since most of these could be integrated. As examples we may quote from Algerian Arabic (cf. HadjSadok 1955) forø sˇe¯tøa/frø a¯sˇa¯tø (< French fourchette), borø dı¯l/brø a¯del (< French bordel). In cases where an internal plural is impossible
XVIII. Morphologischer Wandel II: Fallstudien
most loans receive the plural ending -a¯t or -iyye. The existence of a large number of quadriliteral consonants in Classical Arabic facilitated this incorporation of foreign loans. In Moroccan Arabic a similar situation obtains. A detailed analysis (Heath 1989) has shown the amount of segmental conversion and adaptation that was needed in order to incorporate the nouns borrowed from French into the Arabic system. Noteworthy is the reanalysis of final vowels as the feminine ending -a: antøirø nø a (< French l’internat) is feminine in Moroccan Arabic, but dusˇ (< French douche) is masculine. The former example also shows the re-analysis of the French article as part of the noun; other examples of this are libø a (< French les bas), plural libø a¯t, and lutøusøtøup (< French l’autostop). In Maltese, the incorporation of Italian loans has led to a completely new system of nominal patterns and to a proliferation of patterns based on Arabic weak stems (cf. Mifsud 1995). Due to the structure of Italian words, the morphology of the nominal patterns had to be restructured, and in many words the pattern has come to control only part of the word, e.g., umbrella/umbrelel (< Italian umbrella), gverra/gverer ‘war’ (< Italian guerra). The reduction in variation also applies to the patterns of the verbal nouns. In Classical Arabic the verbal noun (the so-called masdar) of the base stem occurs in a large variety of forms, e.g., dø arb ‘hitting’, huru¯gˇ ‘going out’, daha¯b ‘going away’; some verbs may have more than one masdar. In the derived measures, the verbal nouns have a fixed form. In the dialects there is a general tendency to reduce the number of masdar types (Jastrow 1982: 139) and some of the more isolated dialects have preserved only one type. In Uzbekistan Arabic all verbal nouns end in -a¯n, e.g., gatala¯n ‘killing’ (Classical Arabic qatl), dahala¯n ‘entering’ (Classical Arabic duhu¯l). The verbal nouns of the derived measures in this dialect use the same ending attached to the participle, e.g., inkasar (VII), verbal noun minkisriya¯n (Fischer 1961: 257). Likewise, in Daragözü Arabic f ¤¯ıl is the pattern for all verbal nouns, e.g., ftı¯hø ‘opening’, even in the derived measures, e.g., tammam (II), verbal noun temmı¯m ‘completing’ (Jastrow 1973: 53). In the Maghrebine dialects, which have preserved a much larger number of patterns for the verbal noun, the pattern f ¤¯ıl < fa¤¯ıl is also much more frequent than in Classical Arabic.
1751
160. From Classical Arabic to the modern Arabic vernaculars
4.
Demonstratives and interrogatives
4.1. Demonstratives A good example of the enormous variation between the Arabic dialects, within an identical frame of development, is provided by the deictic system. Classical Arabic has two series of demonstrative pronouns, one for near, and one for far deixis. The canonical forms for near deixis are ha¯da¯, fem. ha¯dihi, pl. ha¯Åula¯Åi; for far deixis the forms are da¯lika, tilka, u¯la¯Åika. These forms are indeclinable, except for the dual, which has two cases (masc. nom. ha¯da¯ni; gen./acc. ha¯dayni; fem. ha¯ta¯ni/ha¯tayni). The deictic element -t- in the feminine tilka is also found in an alternative form of the demonstrative pronoun for the near deixis, ta¯, tı¯, which is, however, seldom used. All modern dialects have at least two series of demonstrative pronouns (cf. Fischer 1959). The near demonstrative pronouns either continue the Classical form or they lose the element ha¯-. The forms for the far deixis are all new developments and the Classical forms are not found in any dialect. The demonstratives for the near deixis are sometimes found in two forms, a short one for attributive use, and a long one for substantivized use, e.g., Algerian Arabic (Djidjelli) had vs. hada, hadi, hadu. With regard to the form of the near deixis demonstratives the dialects may be divided into three groups: those which use only forms without ha¯- (e.g., Egyptian, Eastern Sudanese); those which use only forms with ha¯- (e.g., Tunisian, Palestinian, Lebanese); and those which use both (e.g., Moroccan, Algerian, Saudi-Arabian, Andalusian Arabic). In the third group the forms with ha¯- are usually less emphatic than the ones without. In general the number of different forms in the dialects is staggering. As an example we quote here the demonstrative series from a number of dialects that may be regarded as typical. They all exhibit in varying combinations the deictic elements -h-, -d-, -k-, -l- (cf. Tab. 160.11). The pronouns for the far deixis in Egyptian Arabic are fairly uncommon and used almost exclusively in situations of contrasting deixis. In Cairene there is an alternative set of forms for the near deixis consisting of a combination of demonstrative and personal pronoun (dawwat, diyyat, do¯lat); these forms have become slightly stigmatized. In some dialects there is a third series for middle deixis, consisting of a combination between the de-
near deixis Saudi Arabian (Mecca)
Lebanese Arabic Egyptian Arabic Moroccan Arabic Daragözü Arabic Cypriot Arabic
far deixis
ha¯da¯, da¯
(ha)da¯k(a) ha¯di, dı¯ (ha)dı¯k(a) (ha)do¯l(a) (ha)dola¯k(a) ¯ hä, häda hada¯k(i) hay, haydi haydı¯k(i) haw, hawdi hawdı¯k(i) da dukha di dikha do¯l(a) dukham(a) hada hadak hadi hadik hadu haduk a¯za a¯k ayzi ayk(i) o¯zi o¯k adß a adß ak adi adik alli allik
Tab. 160.11: Demonstratives in the Arabic dialects
monstratives of the near and the far deixis, ˇ iblah (Yemen) for instance in the dialect of G ha¯dka¯h, ha¯dkı´h, ha¯dku´h; in the dialect of Behø za¯nı¯ (Iraq) ha¯ka¯, hayka¯, hawka¯ (cf. Fischer & Jastrow 1980, eds.: 116, 151); and in the dialect of Daragözü ukka¯, ukkı¯, ukko¯ (cf. Jastrow 1973: 41). 4.2. Interrogatives The interrogatives of Classical Arabic man ‘who?’ and ma¯ ‘what?’ have not been preserved in the majority of Arabic dialects. For ‘who?’ many dialects have a form that goes back to a non-attested form mı¯n (except for Moroccan which has sˇkun, probably going back to a combination Åayyu sˇayÅ in yaku¯ n ‘what thing is it?’); for ‘what?’ almost all dialects have opted for a reflex of Åayyu sˇayÅ in (huwa) ‘which thing?’, e.g., Moroccan asˇ, sˇnu, Egyptian Åe¯, Chad Arabic sˇinu, Syrian sˇu¯, e¯sˇ. The development of a new interrogative ‘what?’ illustrates a process of periphrasis that is particularly obvious in the interrogative adverbs. For ‘how much?’, for instance, many dialects have a form that goes back to a substantive meaning ‘form’ or ‘measure’, e.g., Syrian adde¯sˇ (< *qadr Åayyi sˇayÅ in ‘which measure?’), Cypriot Arabic a´sˇka (< *Åayyu sˇayÅ in qadr). Likewise, in the interrogative adverb ‘how?’ we find Syrian sˇlo¯n (< *Åayyu sˇayÅ in lawn ‘which colour?’), Egyptian
1752
XVIII. Morphologischer Wandel II: Fallstudien
izzayy (< *Åayyu sˇayÅ in zayy ‘which appearance?’) and Cypriot Arabic a´sˇsˇik (< *Åayyu sˇayÅ in sˇakl ‘which form?’). These different realizations of the interrogative adverbs seem to point at a polygenetic origin in different contact situations. They all represent the product of an attempt at communication in which salient and semantically transparent forms are preferred over opaque ones such as Classical kam ‘how much?’ and kayfa ‘how?’.
⫺
⫺ ⫺ ⫺
5.
Pidginization and creolization in Arabic
There is one documented case of pidginization and subsequent creolization in Arabic (cf. also Art. 153). The structure of this dialect, (Ki-)Nubi, is an interesting example of a nonIndo-European based creole. Moreover, this dialect plays an important role in the discussions about the developmental processes in the central Arabic dialects (cf. Heine 1982; Versteegh 1984: 113⫺128; Prokosch 1986; Owens 1990). The origin of Nubi lies in the 19th century when a pidginized variety arose in the Anglo-Egyptian army camps in Upper Egypt as the principal means of communication between the Arabic-speaking subalterns and the multilingual recruits from the Southern Sudan, most of whom were speakers of Nubian. It is possible that in the formation of this pidgin elements from an earlier Arabic trade language in the region were used. Some of the soldiers followed the English army to Uganda and Kenya, where they settled and married indigenous women. The children who were born in these mixed marriages creolized the pidginized variety. Nowadays there are (80,000 speakers of Nubi, all of them living in Uganda and Kenya. In the Southern Sudan the pidginized variety has remained in use as a lingua franca under the name of ‘Juba Arabic’. In the cities it tends to become creolized by children from bilingual marriages (cf. Versteegh 1993). Nubi exhibits the classic features of other creole languages (Wellens 2003): ⫺ as in all creole languages there is a complete breakdown of the inflectional system of the original language; in the pidginized variety nouns had only one form for singular and plural, in Nubi there seems to be a new plural marker -a´, which is, however, frequently omitted (cf. Heine
⫺
⫺
⫺
1982: 29); there are a few fossilized broken plurals, e.g., ka´a´fir/kufara´ ‘infidel’; adjectives have no inflection for gender or number, e.g., ma´ria u´e-de´ kebı´r ‘this woman is big’; there seems to be a plural marker -ı´n, but this marker, too, is often omitted (Heine 1982: 33); there is an analytical collective form, e.g., nas-yala´ ‘(group of) children’ (< na¯s ‘people’); there are many compound nouns, e.g., ra´su-laka´ta ‘tree-top’ (lit. ‘head ⫹ tree’); there is no reflex of the Arabic article, but the deictic form de´ (< Egyptian da) has taken over some of the functions of the article, e.g., laka´ta de´ kebı´r ‘the [above mentioned] tree is big’; a new series of demonstratives has developed: u´e-de´, plural do´l-de´ for the near deixis; na´a-de´ for the far deixis; the verb has only one form, diachronically connected with the 3rd plural of the perfect or with the imperative, e.g., a´alim ‘to teach’ (< imperative ¤allim?), a´surubu ‘to drink’ (< imperative isˇrab?); the ending -u seems to function as a transitivity marker, possibly derived from the 3rd person singular object suffix -hu (cf. Versteegh 1984: 124); from verbs a verbal noun is derived by a change in stress (or possibly tone) pattern, e.g., ka´rabu´ ‘to spoil’, verbal noun kara´b; this is probably an innovation that arose at the creolization stage; there is a system of aspectual markers that is remarkably similar to that of other creole languages; at the same time it is not completely different from the system in central Arabic dialects (cf. Tab. 160.12).
‘to come’ ru´a´ bi ru´a´ kala`(s) ru´a´ ka´a´n ru´a´ gı´ ru´a´ bi gı´ ru´a´ ka´a´n gı´ ru´a´ ka´an bi ru´a´
neutral form future perfective pluperfect durative future durative past durative hypothetical, irrealis
Tab. 160.12: Aspectual markers in Nubi (Heine 1982: 38)
In Juba Arabic exposure to the prestige variety of Khartoum Arabic has led to a change
160. From Classical Arabic to the modern Arabic vernaculars
in the verbal system. The personal prefixes of the imperfect verb in Khartoum Arabic are first re-interpreted as alternative aspectual markers; with continued contact they are assigned the function of indicating person (cf. Mahmud 1979). In this way, the dichotomy perfect/imperfect of the central Arabic dialects is re-instated in Juba Arabic.
6.
References
Ali, Abdul Sahib Mehdi (1987), A Linguistic Study of the Development of Scientific Vocabulary in Standard Arabic. London & New York: Kegan Paul International Behnstedt, Peter (1978), “Zur Dialektgeographie des Nildeltas”. Zeitschrift für arabische Linguistik 1, 64⫺92 Behnstedt, Peter (1985/1992), Die nordjemenitischen Dialekte, Vol. I: Atlas; Vol. II: Glossar. Wiesbaden: L. Reichert
1753
Denz, Adolf (1982), “Die Struktur des klassischen Arabisch”. In: Fischer (ed.), 58⫺82 Diem, Werner (1978), “Divergenz und Konvergenz im Arabischen”. Arabica 25, 128⫺147 Diem, Werner (1979), “Studien zur Frage des Substrats im Arabischen”. Der Islam 56, 12⫺80 Diem, Werner (1991), “Vom Altarabischen zum Neuarabischen: Ein neuer Ansatz”. In: Kaye, Alan S. (ed.), Semitic Studies in Honor of Wolf Leslau, Vol. I. Wiesbaden: O. Harrassowitz, 297⫺308 Eisele, John C. (1999), Arabic Verbs in Time: Tense and Aspect in Cairene Arabic. Wiesbaden: O. Harrassowitz Erwin, Wallace M. (1963), A Short Reference Grammar of Iraqi Arabic. Washington, D.C.: Georgetown University Press Ferguson, Charles A. (1959 a), “Diglossia”. Word 15, 325⫺340 Ferguson, Charles A. (1959 b), “The Arabic Koine”. Language 25, 616⫺630
Behnstedt, Peter & Woidich, Manfred (1982⫺ 1999), Die ägyptisch-arabischen Dialekte, Vol. 1⫺5. Wiesbaden: L. Reichert
Fischer, Wolfdietrich (1959), Die demonstrativen Bildungen der neuarabischen Dialekte: Ein Beitrag zur historischen Grammatik des Arabischen. The Hague: Mouton
Blanc, Haim (1964), Communal Dialects in Baghdad. Cambridge/MA: Harvard University Press
Fischer, Wolfdietrich (1961), “Die Sprache der arabischen Sprachinsel in Uzbekistan”. Der Islam 36, 232⫺263
Blanc, Haim (1970), “Dual and Pseudo-dual in the Arabic Dialects”. Language 46, 42⫺57
Fischer, Wolfdietrich (1982, ed.), Grundriss der arabischen Philologie, Vol. I: Sprachwissenschaft. Wiesbaden: L. Reichert
Blau, Joshua (1977), The Beginnings of Diglossia: A Study of the Origin of Neo-Arabic. Malibu: Undena Borg, Alexander (1985), Cypriot Arabic. Wiesbaden: F. Steiner Caubet, Dominique (1993), L’arabe marocain. Vol. 1⫺2. Paris, Louvain: Peters Cohen, David (1970), “Koine`, langues communes et dialectes arabes”. In: Cohen, David (ed.), Etudes de linguistique se´mitique et arabe. The Hague, Paris: Mouton, 105⫺125 Corriente, Federico C. (1976), “From Old-Arabic to Classical Arabic through the Pre-Islamic Koine: Some Notes on the Native Grammarians’ Sources, Attitudes and Goals”. Journal of Semitic Studies 1, 62⫺98 Corriente, Federico C. (1977), A Grammatical Scketch of the Spanish Arabic Dialect Bundle. Madrid: IHAC Czapkiewicz, Andrzej (1975), The Verb in Modern Arabic Dialects as an Exponent of the Development Processes Occurring in Them. Wrocław: Wydawnictwo Polskiej Akademii Nauk
Fischer, Wolfdietrich & Otto Jastrow (1980, eds.), Handbuch der arabischen Dialekte. Wiesbaden: O. Harrassowitz Grotzfeld, Heinz (1965), Syrisch-arabische Grammatik (Dialekt von Damaskus). Wiesbaden: O. Harrassowitz Hadj-Sadok, Mahammed (1955), “Dialectes arabes et francisation linguistique de l’Alge´rie”. Annales de l’Institut des Etudes Orientales d’Alger 13, 61⫺ 97 Harning, Kerstin E. (1980), The Analytical Genitive in the Modern Arabic Dialects. Göteborg: Acta Universitatis Gothoburgensis Harrell, Richard S. (1962), A Short Reference Grammar of Moroccan Arabic. Washington: Georgetown University Press Heath, Jeffrey (1989), From Code-Switching to Borrowing: A Case Study of Moroccan Arabic. London, New York: Kegan Paul International Heine, Bernd (1982), The Nubi-Language of Kibera: An Arabic Creole. Berlin: D. Reimer
1754 Holes, Clive (1990), Gulf Arabic. London, New York: Routledge Holes, Clive (1995), Modern Arabic: Structures, Functions and Varieties. London, New York: Longman Ingham, Bruce (1982), North East Arabian Dialects. London, New York: Kegan Paul International Ingham, Bruce (1994), Najdi Arabic: Central Arabian. Amsterdam, Philadelphia: J. Benjamins Janssens, Gerard (1972), Stress in Arabic and Word Structure in the Modern Dialects. Leuven: Peeters Jastrow, Otto (1973), Daragözü: Eine arabische Mundart der Kozluk-Sason-Gruppe (Südostanatolien). Nürnberg: H. Carl Jastrow, Otto (1978), Die Mesopotamisch-Arabischen Qeltu-Dialekte, Vol. I: Phonologie und Morphologie. Wiesbaden: F. Steiner Jastrow, Otto (1982), “Die Struktur des Neu-Arabischen”. In: Fischer (ed.), 128⫺141 Mahmud, Ushari Ahmad (1979), Variation and Change in the Aspectual System of Juba Arabic. Ph.D. Dissertation, Georgetown University Mifsud, Manwel (1995), Loan Verbs in Maltese: A Descriptive and Comparative Study. Leiden: E. J. Brill Müller, Walter M. (1982), “Das Altarabische und das klassische Arabisch”, “Das Altarabische der Inschriften aus vorislamischer Zeit”. In: Fischer (ed.), 17⫺36
XVIII. Morphologischer Wandel II: Fallstudien Prokosch, Erich (1986), Arabische Kontaktsprachen (Pidgin- und Kreolsprachen) in Afrika. Graz: Institut der Sprachwissenschaft der Universität Graz Rabin, Chaim (1951), Ancient West-Arabian. London: Taylor’s Foreign Press Retsö, Jan (1983), The Finite Passive Voice in Modern Arabic Dialects. Göteborg: Acta Universitatis Gothoburgensis Robin, Christian (1992), L’Arabie antique de KaribÅ ¯ıl a` Mahomet: Nouvelles donne´es sur l’histoire des Arabes graˆce aux inscriptions. Aix-en-Provence: Editions Edisud Rosenhouse, Judith (1984), The Bedouin Arabic Dialects. Wiesbaden: O. Harrassowitz Versteegh, Kees (1984), Pidginization and Creolization: The Case of Arabic. Amsterdam: J. Benjamins Versteegh, Kees (1993), “Leveling in the Sudan: From Arabic Creole to Arabic Dialect”. International Journal of the Sociology of Language 99, 65⫺97 Versteegh, Kees (1997), The Arabic Language. Edinburgh: Edinburgh University Press [2nd, revised edition, 2001] Watson, Janet C. E. (1993), A Syntax of S ø an¤a¯nı¯ Arabic. Wiesbaden: O. Harrassowitz Wellens, Ineke (2003), An Arabic Creole: The Nubi Language of Uganda. Ph. D. diss., University of Nijmegen
Nöldeke, Theodor (1904), “Das klassische Arabisch und die arabischen Dialekte”. In: Nöldeke, Theodor (ed.), Beiträge zur semitischen Sprachwissenschaft. Strassburg: K. J. Trübner, 1⫺14
Woidich, Manfred (1990), Ahlan wa sahlan: Eine Einführung in die Kairoer Umgangssprache. Wiesbaden: L. Reichert
Owens, Jonathan (1990), “East African Nubi: Bioprogram vs. Inheritance”. Diachronica 7, 217⫺250
Zavadovskij, Jurij N. (1981), Mavritanskij dialekt arabskogo jazyka (Chassaniia). Moscow: Nauka
Owens, Jonathan (1993), A Grammar of Nigerian Arabic. Wiesbaden: O. Harrassowitz
Zwettler, Michael (1978), The Oral Tradition of Classical Arabic Poetry: Its Character and Implications. Columbus: Ohio State University Press
Prochazka, Theodore (1988), Saudi Arabian Dialects. London, New York: Kegan Paul International
Kees Versteegh, Nijmegen (The Netherlands)
161. Tok Pisin
1755
161. Tok Pisin 1. 2. 3. 4. 5. 6. 7.
Pidgins and development Historical background History of research Documentation Internal history of Tok Pisin word formation Theoretical consequences References
1.
Pidgins and development
Pidgin languages, because of their relatively short history and rapid development, are ideal test cases for general linguistics. Tok Pisin (from English talk pidgin also referred to as Neo-Melanesian or New Guinea Pidgin) is a particularly important example because (a) it is one of the very few languages whose entire linguistic development can be documented in some detail; (b) it illustrates two types of developments: that from simple jargon to complex creole and that of a conservative rural language to an anglicised urban one; (c) it illustrates both the spontaneous development of new morphology and the influence of deliberate acts of standardisation and planning. While the development of inflectional morphology in Tok Pisin has been widely used (Sankoff 1977; 1980) in the debate about substratum influence and universals of linguistic development as well as the testing of other theoretical proposals, data from Tok Pisin derivational morphology have rarely been employed.
2.
Historical background
Like the other two big Melanesian pidgins, Solomon Islands Pijin and Vanuatu Bislama, the history of Tok Pisin began outside its present area of currency. Between 1879 and 1914, thousands of New Guineans were recruited to work on the plantations of Samoa, as well as a much smaller number for a much shorter time (1880⫺1885) on the Queensland sugar plantations. In Samoa, Papua New Guineans picked up the plantation jargon that had developed among the multilingual workforce from Kiribati, Vanuatu and elsewhere and, from about 1890 onward, they also influenced this plantation language lexi-
cally and grammatically. Whilst it was a reasonably stable pidgin, Samoan Plantation Pidgin English remained a rather rudimentary language with no inflectional morphology, circumlocution and borrowing being the main sources of lexical expansion. As the Melanesian workers returned from Samoa to New Guinea and as a local plantation economy developed, the number of Tok Pisin speakers in former German New Guinea (which comprised of Northeastern New Guinea and the Bismarck Archipelago but not Papua) rapidly increased. From a few hundred in 1880, the numbers had grown to about 15,000 in 1914. Today there are about 2 million speakers of Tok Pisin, i. e. more than 50 percent of the population of independent Papua New Guinea. With the increase in speakers, Tok Pisin became more useful and was employed in a growing number of new situational contexts. It also began to acquire first language speakers. The main new contexts were (a) the gold industry from 1930 onwards, (b) Christianisation and missionisation, also from about 1930 onwards, (c) extension of Australian administrative control from 1920 onwards and, most important of all, (d) increasing use of Tok Pisin as the language of inter-tribal communication, replacing previously established local pidgins, bilingualism or silent barter. During the Second World War Tok Pisin was used extensively in the information campaigns of the allied forces and the Japanese. After 1945 a Tok Pisin-medium press was set up and it also became a medium of education (until its condemnation by the United Nations in 1953). In spite of negative official policies Tok Pisin developed into the de facto language of village councils, new political parties in pre-independence days and the language of political self determination after independence. In the first seventy years of its existence intertribal marriages were rare, but, after 1945, new non-traditional settlements developed (mission stations, inter-tribal cargo cult centres, boarding schools and towns), significantly increasing the number of intertribal marriages. Creolisation of Tok Pisin began around 1950 and this process has accelerated greatly since. In most large urban areas, children have grown up with Tok Pisin as their first language.
1756
3.
XVIII. Morphologischer Wandel II: Fallstudien
History of research
The study of the morphology of Tok Pisin is a very recent phenomenon, mainly because until very recently there was little morphology to study. Early missionary grammars (surveyed by Mühlhäusler 1985 a) contain occasional references to apparent patterns of morphology, and in Hall (1943) a more detailed statement of zero derivation is given. The absence of productive reduplication is noted, an observation that was confirmed subsequently in Laycock (1970) and Wurm (1971), both of whom identified some other productive word formation devices. It was only when investigators turned to data from speakers who had acquired Tok Pisin after the Second World War that larger areas of morphological regularities became evident. Thus, the beginnings of verbal inflections are noted in Sankoff (1977), and in Mühlhäusler (1979) a dramatic increase in derivational productivity in the early 1970s is documented. Coinciding with political independence, there was a rash of activity in the area of lexical planning (surveyed by Wurm et al. 1977) and some detailed studies of recent developments in Tok Pisin morphology have been prepared by Romaine (1990). Smith (2002) contains valuable information on recent trends in creolised Tok Pisin and the continuing uncertainty about its social position and grammatical standards.
4.
Documentation
The emergence of inflectional categories is relatively easy to document because of their high token frequency. There are, however, some problems for those who rely on the written text, including: (a) the failure of recorders to pick up phonologically reduced forms; (b) the lack of clear indications of what is an affix and what is an independent form (e. g. should wanfela botol be analyzed as ‘one fellow bottle’ or not); (c) the emergence of norms for writing, (e. g. mipela ‘we’) can conceal new developments, such as the morphologically reduced mitla. The development of derivational morphology is much more difficult to document, as many lexical items are of a very low token frequency. However, research on Tok Pisin is in the fortunate position of being able to draw
on extensive lexicographical work by German missionaries. In my own studies I have used these to compile a very large corpus, supplemented by observational data and questionnaires. An important consideration in all this is the locus of lexical regularities, as with a rapidly developing language such as Tok Pisin one cannot assume that a complex morphological form is the result of the application of any rule at all. For instance, for some speakers the compound adjective bolhet ‘bald’ forms part of the pattern: (1) bikhet guthet
‘conceited’ ‘intelligent’
i. e. a pattern someone i gat adjective noun > adjective noun. For others, it is interpreted as het i olsem bol ‘the head is like a ball’, just like adjectives of the type sakman ‘the man is like a shark (sharp customer)’. For yet others bolhet may simply be an unanalysed simple lexical entry. It has been found that speakers who learnt Tok Pisin before World War II do not have the same number of productive patterns as younger speakers. It has also been observed that they will produce fewer morphologically complex forms and, in some instances, Melanesian-type forms produced by older speakers are being replaced by more English derived ones among younger speakers, as in kot ren ‘rain coat’ being replaced by renkot. To what extent such clear differences in grammar affect the communication is not understood. However, one suspects that the postulation of a common code is not a satisfactory model for a dynamic system such as Tok Pisin.
5.
Internal history of Tok Pisin word formation
The internal history of Tok Pisin is most conveniently portrayed by referring to a number of idealised stages. It should be kept in mind that the processes are at least partly continuous and that at most times speakers of different developmental stages can interact. The stages are: first found (a) Jargon 1870 (b) Stabilisation 1890 (c) Expansion 1925 (d) Creolisation 1950 (e) Post pidgin/post creole 1960
1757
161. Tok Pisin
5.1. The Jargon stage Jargons are distinguished from proper pidgin languages by their excessive instability, extremely impoverished expressive power and high context dependence. As they are individual rather than social solutions to the problem of crosslinguistic communication, it would seem to make little sense to postulate ideal speaker/hearers. Rather, one is to assume that interaction is characterised by different communication strategies and resulting misunderstandings and failures to communicate. At this stage, speakers of the lexifier language English (i. e. the language which provided most of the lexical roots) continue to use their own inflectional morphology, though they also make unsystematic simplifications characteristic of foreigner talk. Local speakers tend not to have any productive inflectional patterns, though fossilised inflections are in evidence, such as English plural -s in: (2) masis anis
‘match’ ‘ant’
or the genitive affix -s in (3) boskru ‘boat’s crew’ the German plural -en in (4) binen
‘bee’
and the German infinitive -en in (5) beten singen
‘pray’ ‘sing’
Some of these endings have since found their way into the grammar of Tok Pisin. The derivational morphology of English is similarly eroded in the jargon stage, and morphologically complex forms typically end up with simple roots, as in the case of (6) bilinat trausel katuana
‘betelnut’ ‘tortoise shell’ ‘guard of honour’
Differences between indigenes’ and expatriates’ perceptions of morphological relatedness of words persist for long periods of time. As late as 1970 one could note differences such as: expatriate
indigenous
gloss
go ⫺ go ap daun-daun bilo
go ⫺ kwap daun-tambilo ⫺
‘go-climb’ ‘down-hold’ ‘of a ship’
Tab. 161.1: Expatriates’ and indignes’ words
The jargon phase is characterised both by a very reduced and mostly nominal lexicon, and considerable pressure on speakers to say more than the language can express adequately. In fully developed languages there tends to be constructional iconicity (cf. Art. 30) of the type: (a) all central concepts to be expressed by lexical stems; (b) all complex concepts to be expressed by complex or derived forms; (c) all percepts and other singularities to be expressed by phrases, sentences or other syntactic means. In early jargon texts, one encounters many examples of speakers/hearers resorting to either non-verbal means of expression (e. g. by pointing to the chest to indicate ‘chest’, an item for which there is no word, but one that was used frequently as chest infections were common in the plantation context), or else, the formation of ad hoc circumlocutions, though this device was limited by the absence of syntactic regularities and lexical items useful in defining other items. Some early examples include: (7) kas bilong pulimap wara ‘cask for filling with water (barrel)’ bokis bilong man i dai ‘box of person who died (coffin)’ wara bilong maus ‘water of mouth (saliva)’ lata bilong haus ‘ladder of house (stairs)’ Pointing and circumlocution were individual ad hoc solutions and most of these were probably unrecorded. 5.2. The stabilisation stage Stabilisation of a pidgin language refers to its becoming a linguistic system separate from its lexifier languages, and the first language of its users. Stabilisation in the lexical component is manifested (a) in the development of socially sanctioned conventions about lexical information, (b) the reappearence of lexical structure, and (c) the use of stable syntax as a means of deriving new lexical items. Although additions to the lexical inventory do occur, they filter slowly into the language without causing severe structural disruption or major changes in lexical organisation. The main function of lexical additions is to con-
1758
XVIII. Morphologischer Wandel II: Fallstudien
solidate the pidgin as a means of communication in a relatively narrow field of interlingual contacts. The stabilisation of Tok Pisin was not a sudden event but a continuing process, starting with the stabilisation of Samoan Plantation Pidgin on the Samoan plantations and culminating in the crystallisation of a flexible and stable lingua franca spoken throughout German New Guinea. The lexicon which developed during this period derived its separate identity from the impact of Tolai and German and, to a lesser degree, Malay and Samoan, which replaced English as Tok Pisin’s principal lexifier languages between 1880 and 1920. The effect of stabilisation on inflectional morphology was relatively small. It manifested itself mainly in word class marking and number marking in pronouns. The form -pela which occurred in a number of jargons variably as a kind of adjective marker or noun classifier became a marker of attributive, but not predicative monosyllabic adjectives in stable Tok Pisin as in: (8) gutpela man smolpela pik bikpela tais
‘a good fellow, a good man’ ‘a small pig’ ‘a big swamp’
Similarly, -im (from English him) became a marker of transitive verbs as in: (9)
lukim tanim ritim
‘to see’ ‘to turn’ ‘to read’
An incipient marking of intransitives by means of reduplication as in Tolai, the language spoken around Rabaul, the former capital, remained unsuccessful and is now found only in a handful of items including: (10) waswas wasim singsing singim lukluk lukim
‘bathe’ ‘wash someone’ ‘dance’ ‘sing a song’ ‘watch’ ‘see someone’
Pronominal number marking in Tok Pisin is the result of an encounter of two solutions, the Samoan Plantation Pidgin one using ol after pronouns as in: (11) me ol you ol (em)ol
‘we (excl. & incl.)’ ‘you (pl.)’ ‘they’
and a different tradition employing -pela for this purpose.
Most present-day variants of Tok Pisin employ (12) mi pela ‘we (excl.)’ you pela ‘you’ (em)ol ‘they’ Ol rather than -pela was subsequently adopted as a nominal plural marker (see Mühlhäusler 1981). Tense, aspect and modality in stable Tok Pisin continued to be marked by fully stressed adverbials and auxiliaries. Developments of word formation devices were more spectacular, suggesting that the expansion of the lexicon was more important than stylistic flexibility or syntactic refinements. By the time German control came to an end in 1919, the lexical inventory of Tok Pisin comprised perhaps 750⫺1,000 items, most of them lexical bases. Throughout the stabilisation stage, borrowing from outside sources remained the most important means of increasing Tok Pisin’s referential adequacy. However, in contrast with the lexicon of Jargon English, that of stabilised Samoan Plantation Pidgin and Tok Pisin was no longer an unstructured list. Instead, one can observe the development of lexical structures manifested in (a) the emergence of tightly structured semantic fields; (b) the use of stable syntax in circumlocutions; and (c) the development of a small number of regularities of lexical derivation. With regard to the last category, most complex forms were still phrase level items as: (a) names for different kinds of people: (13) man bilong pekato man bilong kamda man bilong stia man bilong limlimbur
‘a sinner’ ‘a carpenter’ ‘a helmsman’ ‘an idler’
(b) names for certain kinds of buildings: (14) haus haus haus haus haus
kuk boi kot pekpek marit
‘kitchen’ ‘hut for labourers’ ‘court house’ ‘toilet’ ‘married quarters’
(c) distinction of sex in humans and animals: (15) hos man hos meri pikinini man pikinini meri dok man
‘stallion’ ‘mare’ ‘young boy’ ‘young girl’ ‘male dog’
1759
161. Tok Pisin
dok meri pik man pik meri kakaruk man kararuk meri
‘bitch’ ‘boar’ ‘sow’ ‘cock’ ‘hen’
(d) antonyms of adjectives and intransitive verbs: (16) gut nogut strong nostrong stret nostret nap nonap
‘good’ ‘bad’ ‘strong’ ‘feeble’ ‘correct’ ‘incorrect, unacceptable’ ‘sufficient’ ‘insufficient’
Another important innovation at this stage are conventions for changing the subcategorial status of words. Thus nouns referring to trees can also refer to the fruit from these trees, thus bata ‘avocado tree’ can also mean ‘avocado pear’, saksak ‘sago palm’ and ‘sago’, or muli ‘orange tree’ and ‘orange’. The name of a sound is derived from the name of the instrument with which it is produced. Examples are tandok ‘horn trumpet’ and ‘signal for beginning or ending work’; kundu ‘drum’ and ‘sound of a drum’; visil, kaviv ‘flute’ and ‘sound of a flute’, and so forth. Next we can observe a third convention, which was less generally maintained than the first two: a noun referring to a material can at the same time refer to something typically made out of this material, e. g. purpur ‘flower, purpur bush’ and ‘grass skirt’, ain ‘metal’ and ‘flat-iron, anvil’, kapa ‘shiny metal’ and ‘washing copper’, let ‘leather’ and ‘belt’. Finally a very small number of causative verbs were formed by adding the transitivity marker im to intransitive verbs. As late as 1940 observers remark on the absence of word formation in Tok Pisin. There are certainly few signs of the sometimes allegedly vigorous role of reduplication in early pidgin development. 5.3. The expansion stage The realisation that pidgins can be classified not only in terms of their social functions, but also in terms of their structural complexity, is a relatively recent one, stemming from the work of Samarin (1971), Mühlhäusler (1974), and Todd (1974). It was found that, as pidgins are used in new functions, their linguistic complexity also increases. Tok
Pisin’s structural expansion appears to have taken place at a time when the number of first language speakers was diminutive. Whilst these few first language speakers may have contributed to its structural development, it appears that the bulk of innovations originated among second language speakers. Thus, the view that significant structural expansion of a pidgin only occurs with its creolisation is not borne out. Tok Pisin’s structural expansion after 1920 is defined by two main developments, namely the emergence of sentence embedding and discourse structure and the development of a derivational lexicon. Developments in the area of inflectional morphology remained modest, as phonological reduction and consequent clitisation, or affixation, is not favoured by second language speakers of Tok Pisin. As the age at which Tok Pisin was learnt decreased progressively from 1920 onwards, some of the processes usually associated with creolisation are found among fluent younger speakers of Tok Pisin, particularly those that have grown up in the towns. Derivational morphology on the other hand experienced a dramatic expansion during this phase. This asymmetry of development in the inflectional and derivational areas is an index of the relative unimportance of stylistic flexibility and importance of referential adequacy in a pidgin language. The following general tendencies in the growth of the derivational lexicon have been observed: (a) a continued increase in the number of productive lexical regularities; (b) an increase in the distance between lexical structures and syntactic structures, in particular a tendency for lexical items to shift down from higher to lower lexical size levels; (c) an increase in the numbers of lexical items accounted for by individual lexical regularities, i. e. increase in productivity. Thus, instead of the handful of word formation patterns found in 1920, 57 such patterns were identified by Mühlhäusler (1979), involving conversion, compounding and reduplication, enabling speakers to create hundreds of new morphologically complex words. This increase in productivity of the lexicon appears to have been governed by a number of principles, particularly those of gradual downshifting and an increase in derivational complexity.
1760
XVIII. Morphologischer Wandel II: Fallstudien
Downshifting, i. e. the shift from phrase to word level lexical items can be illustrated with examples such as: 1945
present day Tok Pisin
lam wokabaut ‘hurricane lantern’ manki bilong masta ‘servant (male)’ mekim hariap ‘to speed up someone’ hatpela wara ‘soup, hot water’ mani pepa ‘paper money’ wara bilong skin ‘sweat’ man bilong stil ‘thief’ man save ‘knowledgeable person expert’
wokabautlam mankimasta hariapim hatwara pepamani skinwara stilman saveman
Tab. 161.2: Downshifting
This trend is continuing today. One consequence is the decreasing reliance of speakers on circumlocutions. An interesting though insufficiently understood phenomenon is the creation of morphologically complex nonce words (cf. Downing 1977), e. g. hetman ‘a ghost that was a head only’ in a ghost story, rather than the lexicalised item hetman ‘village chief’. Increasing derivational complexity is illustrated by a loss of morphological transparency. For example, one can contrast the early emergence of cumulative compounds of the type: (17) papamama manmeri
‘parents’ ‘people’
with the late appearance of: (18) wantok wankaikai wansmok wanmak wanlain
‘someone who speaks the same language’ ‘mess mate’ ‘someone who shares a smoke’ ‘someone of the same size’ ‘someone working in the same labour line’
In addition, many derivational regularities have become considerably more productive over the years. Thus, no compounds ending
in the agent suffix -man ‘person, doer’ are found before 1920 though lexical phrases of the form [man bilong Vint] expressing ‘someone who usually does what is referred to by the verb’ are documented in fair numbers for the mid 1920s, including: (19) man bilong singaut ‘noisy person, beggar’ man bilong slip ‘sleepy, lazy person’ man bilong stil ‘thief’ The only word-level items at this point are sutman ‘policeman’, and fulman ‘jokester’. Most commonly nouns with an agentive meaning are fully lexicalised. Typical examples from an unpublished 1924 vocabulary by Leo Brenninkmeyer are kuskus ‘clerk, writer’, ridima ‘redeemer’, kamda ‘carpenter’ and kundar ‘acolyte’. For the mid-1930s the authors of a dictionary (Wörterbuch n. d.: 53) remark that “-man as the suffix of verbs forms agent nouns” (author’s translation). However, only a few word-level items are listed: (20) wasman sikman daiman stilman
‘watchman’ ‘patient’ ‘dead, dying man’ ‘thief’
Phrase-level items listed in the Wörterbuch (n. d.) include: (21) man bilong toktok ‘talkative person’ man bilong save ‘wise, knowledgeable person’ man bilong pait ‘warrior, fighter’ man bilong pret ‘fearful person’ The only additional items culled from Kutscher (1940) are: (22) rabisman lesman
‘destitute person’ ‘loafer’
It is interesting to note that Kutscher does not list ‘thief’ as stilman, as had been done in the Wörterbuch (1935) but as man bilong stil. The three trends identified in the derivational morphology of expanded Tok Pisin continue to dominate the development of the next stage, creolisation. 5.4. The creolisation stage Creolisation (i. e. the language becoming adopted as a first language) in Tok Pisin began in the early 1950s but has only become a large scale phenomenon since urbanisation, from the 1970s onward. Today an estimated 50,000 children speak Tok Pisin as their first or main language, mainly in towns such as
1761
161. Tok Pisin
Port Moresby, Lae or Madang. The main changes once Tok Pisin becomes a first language appear to lie in the area of inflectional morphology. Many of these changes are the result of the growing tempo at which Tok Pisin is now spoken and the consequent phonological reductions. As yet the underlying morphological structure of complex forms does not appear to have been affected, though quite significant deeper restructuring may occur within a few generations. Generally speaking one can observe two trends: (a) that from lesser to greater morphological synthesis; (b) that from lesser to greater fusion. These combined trends make for a dramatic increase in morphological complexity and result in communication problems between speakers of creolised and non-creolised varieties as well as problems with the existing spelling system. Typical examples of such problems include possessive pronouns, indirect object pronouns and future/irrealis modality in verbs (cf. Romaine & Wright 1982; Mühlhäusler 1983; 1985 b). In traditional second language varieties of Tok Pisin the pattern for expressing the possessive pronoun is: (23) bilong mi bilong yu bilong em
‘my’ ‘your’ ‘his/her’
and so on. In progressive creolised varieties these forms are reduced to blomi, bloyu, blem, and blusat (from bilong husat) ‘whose’. Most speakers would produce such eroded forms as well as intermediate forms such as belomi and so on. As regards the dative case of pronouns, in traditional varieties one finds: (24) long mi long yu long em
‘to me’ ‘to you’ ‘to him/her’
In creolised varieties these appear as lemi, leyu and lem respectively. Note that in at least two examples there is a loss in regularity (instead of bilong as the possessive marker, we now have blo or blu, depending on what noun or pronoun follows) as well as a loss of transparency: lem is not as transparent and easy to recognise as long em. A final example of the loss of morphological transparency is that of future/irrealis mo-
dality in verbs. Labov’s (1985) demonstration that pidgins and creoles develop verbal affixes out of free-standing adverbials or auxiliaries is borne out by evidence from Tok Pisin. Most commentators, e. g. Sankoff (1980), have postulated a developmental hierarchy such as: (a) baimbai mi kam ‘I shall come’, with a fully stressed sentence initial adverbial; (b) bambai mi kam, exhibiting phonological reduction; (c) bai mi kam, with a fully stressed monosyllabic adverbial; (d) mi bai kam, with a shift of the adverbial to a position in front of the main verb; (e) mi bai kam, with further phonological reduction and finally: (f) mi bakam, with clitisation of the future marker. Whilst examples of all of the above constructions can still be found, most speakers use these forms variably, but not necessarily in an implicationally ordered manner. Further research is needed to determine what the role of creolisation is in the development of inflectional morphology. The effects of creolisation on derivational morphology as characterised by Mühlhäusler (1979) are: (a) an increase in the productivity of existing patterns of word formation; (b) a greater relaxation of earlier restrictions on word formation rules, i. e. a move from context-sensitive to context-free; (c) a relaxation of the constraint on the recursive application of word formation rules. These three combined trends make for a much more powerful derivational lexicon. Note that all of the above trends have already begun with fluent second language speakers of expanded pidgin. We are thus dealing with a quantitative rather than a qualitative difference. I shall now turn to more specific examples which demonstrate these trends. Increased productivity of derivational morphology in Malabang creolised Tok Pisin can be illustrated with the occurrence of new abstract nominals derived from verb or adjective bases. For example: (25) no ken gat sot bilong wara ‘there won’t be a shortage of water’ I gat planti harim bilong Tok Pisin ‘there are many dialects of Tok Pisin’
1762
XVIII. Morphologischer Wandel II: Fallstudien
maski mi gat sting nogut ‘although I have a bad infection’ i gat planti kain kolim bilongen ‘there are many different names for it’ rait bilong mi i no klia tumas ‘my writing is not very clear’ karim bilong dispela meri i hat tru ‘the act of childbirth is very painful for this woman’ dilim bilong yu i no stret ‘your way of dealing (cards) is not correct’ em i lukim ron bilong olgeta pis ‘he looked at the movement of the fish’ A second instance of functional change is the direct derivation of causative verbals from verb or adjective bases by means of the affix -im, which in Malabang Tok Pisin appears to apply to all adjective or intransitive verb bases and to be fully recognised as a stylistic variant of the periphrastic construction [mekim N i Adj/V]: (26) wara i stingim ol plang ‘the water makes the planks rot’ dispela kaikai i switim maus bilong mi ‘this food gives my mouth a pleasant taste’ meri i bonim pikinini ‘the woman gave birth to a child’ em i wok long raunim diwai ‘he is busy making a piece of wood round’ meri i smatim em yet ‘the girl dolled herself up’ A further example are verbs derived from nouns referring to instruments. In second language Tok Pisin these can be formed as long as they are not used as instruments for cutting, a contextual constraint that may have been introduced from Tolai. Whilst forms such as akisim ‘to cut with an axe’ and naipim ‘to cut with a knife, to stick a knife into’ are not documented for non-creole varieties, they occur freely in the Malabang creole of Manus Island. Next, one finds the relaxation of derivational constraints on the formation of new word-level items. The disappearance of the constraint specifying that words should not consist of more than two morphemes can be seen in the following examples observed in spontaneous conversations:
(27) nek-tai-im ‘tie a necktie around someone’ han-kap-im ‘put handcuffs on’ kol-sisel-im ‘cut with a cold chisel’ pinis-taim-man ‘returned labourer’ bik-het-pasin ‘stubbornness’ More important, and hardly ever observed in non-creolised varieties of Tok Pisin, is the high incidence of multiple derivation, i. e. the operation of derivational processes on derived lexical items, as in: (28) huk ‘hook’ kuk ‘boil, cook’
> huk ‘to go fishing’ > kukim ‘cook something’
> huk ‘fishing’ > kukim ‘cooking method’
5.5. The post-pidgin and post-creole stages Post-pidgin and post-creole development, i. e. the language coming under heavy influence from English, can be dealt with together, since in the towns where these developments are encountered, the difference between first and second language varieties of Tok Pisin is not great, particularly among young speakers. Generally speaking, what one expects to find at this stage are the beginnings of a postcreole continuum. However, in the case of Tok Pisin, one cannot easily identify the operation of certain, generally accepted principles such as “new forms are adopted before new meanings”. Since there is a growing number of speakers who are bilingual in English and Tok Pisin and as access to the acrolect continues to increase with the introduction of new media such as T.V. and video, in at least some cases, contact with the acrolect has led not to a gradual approximation, but to new morphology distinct from that of the systems in contact. As English is not a highly inflected language, the number of inflectional morphemes taken on by speakers of urban Tok Pisin remains limited. Most notable is the very considerable increase in the use of plural -s, replacing or supplementing the conservative ol ⫹ noun construction. My own data (Mühlhäusler 1981) suggest that the adaptation of the -s does not follow any known developmental path, such as the animacy hierarchy. However, other researchers have found evidence of such patterns emerging in the speech
1763
161. Tok Pisin
of young urban speakers. Possessive -s and verbal agreement -s are absent in Tok Pisin as is the past tense -ed. A number of examples of -ing were found, particularly in written texts. They appear to be the result of a reinterpretation of the transitive/causative suffix -im, rather than a straight borrowing of the English continuous marker -ing: Consider the following examples from private letters: (29) mi no man bilong spiking English ‘I do not speak English’ ol meri i ken draiving kar ‘the women can drive a car’ dispela toktok i helibing yu ‘this talk supports you’ mi laik askim yu long wanpela samtim ‘I want to ask you for something’ Most writers would prefer spikim, draivim, helipim, and samting instead of the highlighted forms. The effects of renewed massive contact with its original lexifier language has brought about some important changes in the derivational morphology of Tok Pisin. Generally speaking, the current preferred strategy of many urban speakers is to borrow a word from English rather than make use of existing derivational regularities. In some cases, existing complex items are being replaced by borrowed ones, with the resulting decline in the productivity of derivational morphology and an increase in suppletion. The following table illustrates how borrowing has replaced previously existing regularities with suppletion as effect: rural urban Tok Pisin Tok Pisin
meaning
bekim bekim bungim bung hariap hariap subim subim tingtink tingting hevi hevi peim pe
‘answer’ ‘the answer’ ‘gather’ ‘gathering place’ ‘hurry up’ ‘speed’ ‘shove’ ‘force’ ‘think’ ‘idea’ ‘heavy’ ‘weight’ ‘pay’ ‘pay, wages’
bekim ansa kolektim bung hariap spit subim pos tingting aidia hevi wait peim wes
Tab. 161.3: Suppletion as effect of borrowing
A second trend is the reordering of the elements of nominal compounds to make them conform more closely to the English model: rural urban Tok Pisin Tok Pisin
meaning
tok ples sit bet masta rot haus morata gris pik
‘local language’ ‘bed sheet’ ‘road constructor’ ‘house constructed of bush materials’ ‘pig grease, lard’
plestok betsit rotmasta moratahaus pikgris
Tab. 161.4: Reordering in nominal compounds
As yet, no productive pattern directly resulting from English has been observed. However, if other creoles are anything to go by, then these will appear in Tok Pisin before too long. The ultimate fate of Tok Pisin would not seem to be a linguistic matter. It will depend on the degree to which Tok Pisin is needed by its speakers. With the remergence of regional lingue franche in the provinces and English in the towns, the role of Tok Pisin remains unclear.
6.
Theoretical consequences
The evidence presented here raises a number of questions including: (a) the question of substratum influence; (b) the question of linguistic universals; (c) the question of structure and function. These questions can only be addressed if proper attention is paid to the temporal dimension. Large scale changes in morphological power and productivity can be observed over a very brief period of time. Comparing arbitrarily selected points of this developmental continuum with the grammar of a particular substratum or superstratum language could be quite misleading because of the numerous changes the language has undergone. There is also the analytical problem of how one can identify morphological material across two or more languages. As regards inflectional morphology, the sole successful example of substratum influence are a few (subsequently lexicalized) examples of nominal plurals being indicated by means of reduplication, as in Tolai: (30) sipsip meme
‘sheep’ ‘goat’
1764 The bulk of the innovations in this area are the result of internal developments such as increasing tempo, reassignment of stress, reinterpretation of word boundaries and similar well known phenomena. These developments parallel those found in other creoles such as those discussed by Labov (1985). For derivational morphology, Mosel (1980) has prepared a painstaking comparison between the fifty-seven word formation patterns of Tok Pisin, identified by Mühlhäusler (1979), and those of Tolai. Mosel’s findings include: (a) in the vast majority of cases (40 out of 57 patterns) there are no close parallels between Tok Pisin and its alleged substratum languages; (b) in many cases where there are parallels, other forces such as universals or reanalysis seem to have been at work simultaneously. Not mentioned by Mosel but highly significant, is that many of the similarities between Tok Pisin and Tolai developed at a time when Tolai’s importance as a lexifier language of Tok Pisin had already declined. In the foundation years of Tok Pisin (1880⫺1910) virtually no word formation was in evidence. There is another temporal or developmental factor that weakens substratum arguments, i. e. the observable fact that the productivity of derivational regularities is very low until well into the expansion phase of the language. This would seem to indicate that substratum languages contribute a few bases for analogical extension rather than productive lexical rules. There are numerous other ways in which derivational morphology can arise. These are discussed by Mühlhäusler (1979: 257⫺273). Among these, grammatical reanalysis is a particularly important process. The question of universals arises mainly in connection with the sequence in which morphological regularities are added to the grammar of a language such as Tok Pisin. The presence of substratum or superstratum sources in itself is neither a necessary nor a sufficient condition for the development of new morphology. Rather, it appears that its growth is governed by language independent principles (such as causatives will be derived first from static verbs and all verbs before adjectives or nominals or the emergence of plural affixation will proceed along the accessibility and animacy hierarchies). Insufficient developmental data for other pidgins and
XVIII. Morphologischer Wandel II: Fallstudien
other types of second language acquisition are available to warrant firmer conclusions. The presence of universal trends in the development of pidgin morphology needs to be regarded as independent of the question of the ontological status and sources of such universals (cf. Art. 118, 153). It certainly does not follow that the postulated hierarchies of morphological development are bioprogrammatic (Bickerton 1981). Functional explanations for the development of pidgin morphology remain underexplored, though it is widely acknowledged that the structural expansion of these languages is a response to increased functional requirements. The often heated debate about “substratum or universals” has failed to ask to what end such processes need to operate. As has been pointed out above, pidgins, like other languages, serve to express a number of concepts and generalities as well as percepts or singularities. The latter are expressed by syntactic means, for instance: dispela naispela meri i sindaun long ret pela sia ‘this nice girl who is sitting in the red chair’. For concepts, however, one needs lexical items such as: (31) meri
‘woman, girl’
Typically, basic concepts are expressed by word bases. Complex concepts are expressed by morphologically complex lexical units such as in: (32) singelmeri develmeri pamukmeri
‘single girl’ ‘spirit woman’ ‘prostitute’
It is the growing need for more complex concepts that has brought into being greater lexical power. At the same time, because pidgins are second languages they need to remain learnable, and enriching the lexicon by borrowing more and more lexical bases from substratum or superstratum languages is therefore costly in terms of learnability. The emergence of lexical regularities of the type described in this article is an answer to this problem. This view is in harmony with the observation that pidgins begin not as simple languages, but as highly complex, heavily lexical and impoverished ones, and that simplicity in the sense of regularity or grammaticality emerges only in the process of their becoming referentially and expressively more adequate. Many morphological questions still remain unanswered. A corpus of linguistic and func-
161. Tok Pisin
tional data at present being compiled in connection with the international project on “Languages of Intercultural Communication in the Pacific” (see Mühlhäusler 1986) is likely to make available information which can illuminate further the issues raised here.
7.
References
Bickerton, Derek (1981), Roots of Language. Ann Arbor: Karoma Downing, Pamela (1977), “On the Creation and Use of English Compound Nouns”. Language 53: 810⫺842 Hall, Robert A. Jr. (1943), Melanesian Pidgin English: Grammar, Texts, Vocabulary. Baltimore: Linguistic Society of America Kutscher, P. (1940), Wörterbuch Deutsch⫺PidginEnglish. Vunapope: Rabaul Labov, William (1985), On the Adequacy of Natural Languages. Trier: Linguistic Agency, University of Trier Laycock, Donald C. (1970), Materials in New Guinea Pidgin (Coastal and Lowlands). Canberra: Pacific Linguistics Mosel, Ulrike (1980), Tolai and Tok Pisin: The Influence of the Substratum on the Development of New Guinea Pidgin. Canberra: Pacific Linguistics Mühlhäusler, Peter (1974), Pidginization and Simplification of Language. Canberra: Pacific Linguistics Mühlhäusler, Peter (1979), Growth and Structure of the Lexicon of New Guinea Pidgin. Canberra: Pacific Linguistics Mühlhäusler, Peter (1981), “The Development of the Category of Number in Tok Pisin.” In: Muysken, Pieter (ed.) Generative Studies on Creole Languages. Dordrecht: Foris, 35⫺84 Mühlhäusler, Peter (1983), “The Development of Word Formation in Tok Pisin”. Folia Linguistica 17, 463⫺487 Mühlhäusler, Peter (1985 a), History of the Study of Tok Pisin. In: Wurm & Mühlhäusler (eds.), 15⫺33 Mühlhäusler, Peter (1985 b), “Internal development of Tok Pisin”. In: Wurm & Mühlhäusler (eds.), 75⫺166
1765 Mühlhäusler, Peter (1986), “Pidgins and Creoles of Australia and the Pacific”. Australian Journal of Linguistics 6, 181⫺199 Romaine, Suzanne (1990), “The Use of bai in Young Children’s Creolized Tok Pisin”. In: Verhaar, John (ed.), Melanesian Pidgin and Tok Pisin. Amsterdam: Benjamins, 187⫺203 Romaine, Suzanne & Wright, Fiona (1982), “A Short Note on Short Forms in Tok Pisin”. Journal of Pidgin & Creole Languages 2, 64⫺67 Samarin, William J. (1971), “Salient and Substantive Pidginization”. In: Hymes, Dell (ed.) Pidginization and Creolization of Language. Cambridge: Cambridge University Press, 117⫺140 Sankoff, Gillian (1977), “Variability and Explanation in Language and Culture: Cliticization in New Guinea Tok Pisin”. In: Saville-Troike, Muriel (ed.), Linguistics and Anthropology. Washington, D. C: Georgetown University Press, 59⫺73 Sankoff, Gillian (1980), “Variations, Pidgins & Creoles”. In: Valdman, Albert & Highfield, Arnold (eds.) Theoretical Orientations in Creole Studies, New York: Academic Press, 139⫺164 Smith, Geoff P. (2002), Growing up with Tok Pisin. London: Battlebridge Publications Todd, Loreto (1974), Pidgins and Creoles. London: Routledge & Kegan Paul (Language and Society Series) Wörterbuch mit Redewendungen Pidgin-Englisch⫺ Deutsch. (Pidgin-English⫺German dictionary, with idioms). (n. d.) [1935] Mimeo Wurm, Stephen A. (1971), New Guinea Highlands Pidgin: Course Materials. Canberra: Pacific Linguistics Wurm, Stephen A. & Mühlhäusler, Peter (eds. 1985), Handbook of Tok Pisin. Canberra: Pacific Linguistics Wurm, Stephen A. & Mühlhäusler, Peter & Laycock Donald C. (1977), “Language Planning and Engineering in Papua New Guinea”. New Guinea Area Languages and Language Study (Canberra: Pacific Linguistics) C-40, 1151⫺1177
Peter Mühlhäusler, Adelaide (Australia)
XIX. Psycholinguistische Perspektiven Psycholinguistic perspectives 162. Mentale Repräsentation morphologischer Strukturen 1. 2.
4. 5.
Mentale Repräsentationen Das mentale Lexikon und das Lexikon der Grammatik Beobachtungsbereiche und Forschungsmethoden Einige Grundprobleme der Forschung Zitierte Literatur
1.
Mentale Repräsentationen
3.
Der Ausdruck “mentale Repräsentationen” bezeichnet ein Kernkonzept für die Frage nach der Struktur, dem Format und der Funktionsweise kognitiver Prozesse. In den Kognitionswissenschaften hat der Begriff unterschiedliche Ausformungen erhalten, die schon früh in der Geistesgeschichte nachweisbar sind (Scheerer 1992). Die beiden Grundpositionen des symbolverarbeitenden Ansatzes und des Konnektionismus sind dabei beide durch ihren Bezug auf Computerorganisation und Programmiersprachen gekennzeichnet. Im symbolverarbeitenden Ansatz wird die Manipulation von Symbolen als zentrale Aufgabe kognitiver Prozesse angesehen; mentale Repräsentationen sind durch Regelanwendung zustande gekommene syntaktisch wohlartikulierte und semantisch interpretierbare Verkettungen von Symbolen (vgl. u. a. Pylyshyn 1984). Die Grundeinheiten in den konnektionistischen Ansätzen sind dagegen subsymbolischer Natur: Es gibt nur Knoten (nodes) und ihre mannigfachen Verknüpfungen (connections); mentale Repräsentationen können sich nur als emergente Eigenschaften von Netzwerken insgesamt ergeben, und zwar als multivariate Aktivierungsmuster. Die Diskussion zwischen den beiden Grundpositionen ist derzeit nicht entschieden, vgl. die Reaktionen auf den expositorischen Artikel “On the proper treatment of connectionism” in der Zeitschrift Behavioural and Brain Sciences (Smolensky 1988). Dabei spielt sich die Debatte neuerdings vornehmlich auf dem Gebiet der Modellierung sprachverarbeitender
Prozesse ab: Der symbolverarbeitende Ansatz findet ja eine vorbildliche Ausformung in den Regeln und Repräsentationsebenen generativer Grammatiken. Morphologie ist erst relativ spät Gegenstand psycholinguistischer Analysen geworden. Zunächst modellierte man auf der Basis linguistischer Beschreibungen (Halle 1973) den Worterkennungsvorgang als einen Zerlegungsprozeß, in dem zuerst die Bausteine des Wortes bestimmt werden, weil nur so auf das mentale Lexikon zugegriffen werden kann (s. 2); z. B. engl. re-tell besteht aus dem Präfix re- und dem Stamm tell, beides lexikalische Einträge, die durch eine Regel “Verb J Präfix plus Stamm” zu der komplexen Einheit retell führen (Taft & Forster 1975). Die Diskussion dieses in der Folgezeit vielfach modifizierten Ansatzes wird mit wenigen Ausnahmen auf der Grundlage der symbolverarbeitenden Position geführt (vgl. zusammenfassend Taft 1988 und Art. 163). Erst in den letzten Jahren werden konnektionistische Modelle mit linguistischen Erklärungsansätzen für spezielle Erscheinungen wie Kasussynkretismus in Zusammenhang gebracht (Carstairs & Stemberger 1988) oder die Ähnlichkeit des alten Terminus Analogie (z. B. Paul 51920; Art. 148) mit dem konnektionistischen Grundschema hervorgehoben (Derwing & Skousen 1989). Neuerdings ist der Erwerb der Flexionsmorphologie zu einem zentralen Gegenstand in der Konnektionismusdiskussion geworden. In Rumelhart & McClelland (1986) war gezeigt worden, daß Übergeneralisierungen vom Typ engl. goed (statt des suppletiven went), die seit langem (z. B. Berko 1958) als prototypische Evidenz für regelbasiertes Verhalten galten, auch in einem konnektionistischen Netzwerk ohne Regeln durch back propagation erzeugt werden können. Die Debatte darüber ist nicht abgeschlossen, u. a. weil nicht klar ist, in welcher Weise konnektionistische Systeme den un-
162. Mentale Repräsentation morphologischer Strukturen
markierten Fall erkennen und als solchen lernen (vgl. u. a. Clahsen & Rothweiler 1992; Hare et al. 1995).
2.
Das mentale Lexikon und das Lexikon der Grammatik
2.1. Das Lexikon in modernen Grammatiktheorien In der Auffassung der neueren Grammatiktheorien erzeugt die Syntax reguläre abstrakte Strukturen, die durch die Einsetzung lexikalischer Elemente individuiert werden. Es ist das Geschäft der lexikalischen Komponente einer Grammatik, lexikalische Repräsentationen zu erzeugen; diese (und nicht die lexikalischen Minimaleinheiten) sind es, die in syntaktische Strukturen eingesetzt werden. Für diesen Zweck spezifiziert die morphologische Theorie den Begriff “mögliche Wortform” einer Sprache. Dementsprechend wird von den meisten Linguisten angenommen, daß zumindest reguläre Flexionsformen wie etwa im Paradigma der schwachen Verben des Deutschen nicht jede einzeln im Lexikon der Grammatik aufzuführen sind, sondern durch eine Regel “Verbform ⫽ Verbstamm plus Endung” generiert werden. Lexikalische Einheiten sind also nur der Stamm, z. B. sag-, sowie die einzelnen Endungen, so daß bei der lexikalischen Einsetzung die grammatische Form “Verb ⫹ 2. Person Singular Präteritum Indikativ Aktiv” identifiziert werden kann mit, z. B., sag- ⫹ -test. In neueren Wortstrukturtheorien wird ferner angenommen, daß auch die Wortbildung weitgehend analog zur Syntax organisiert ist (vgl. unter vielen anderen Olsen 1986). Deutsche Adjektivableitungen mit dem Suffix -bar oder dreiwertige deutsche Verben mit dem Präfix be- lassen sich informal so beschreiben: Sei X ein deutsches transitives Verb. Dann ist X-bar ein Adjektiv mit der Bedeutung “kann (ge)-x-t werden”, z. B. hörbar, bewohnbar, ableitbar, etc. Sei X ein deutsches transitives dreiwertiges Verb mit Subjekt, direktem Objekt und direktionaler Präpositionalphrase. Dann ist be-X-en ein transitives dreiwertiges Verb gleicher Bedeutung mit Subjekt, Objekt und einer Präpositionalphrase mit der Präposition mit, wobei Objekt und Präpositionalphrase bei Basis
1767 und Ableitung vertauscht sind, vgl. belegen, bemalen, bepflanzen etc.
Diese Strukturbeschreibungen treffen sowohl auf eine Anzahl existierender deutscher Wörter zu als auch auf Wörter, die wir gegenwärtig kaum in einem auch umfangreichen Lexikon finden wie z. B. drückbar, zerrbar, ionisierbar oder bekugeln, betüpfeln, besprayen, wobei die Regeln auch kombiniert werden können wie in betüpfelbar oder (mit weiteren Regeln) Unbetüpfelbarkeitssymptomatik. Solche lexikalischen Regeln für die Generierung von Ableitungen und Zusammensetzungen spezifizieren lexikalische Repräsentationen, d. h. das, was beim Vorgang der lexikalischen Einsetzung in die durch die Syntax generierten Strukturen eingesetzt wird. Unabhängig vom konkreten Format, in dem derlei Regeln zu formulieren sind, erlaubt diese Konzeption die Aufrechterhaltung der klassischen Lexikonkonzeption, wonach konkrete Lexikoneinträge nur das enthalten, was nicht durch Regeln abgeleitet werden kann. Dazu gehören aufgrund des Arbitraritätsprinzips zunächst einmal alle Morpheme einer Sprache. Daneben muß es freilich aufgrund des Phänomens der Lexikalisierung auch komplexe Lexikoneinträge geben. Dazu gehören z. B. Wörter, die bestimmte Affixe aufweisen, aber dennoch offenbar nicht unter die einschlägigen Regeln fallen, im Beispielbereich etwa Wörter mit unikalen Morphemen wie begehren, urbar sowie lexikalisierte Bildungen wie befolgen, kostbar etc. Ein besonders instruktives Beispiel ist das Wort befehlen. Obgleich es weder wortsyntaktisch noch semantisch irgendwie zu der o. a. Regel paßt, muß es als präfigiertes starkes Verb gekennzeichnet werden, denn das Partizip heißt nicht *gebefehlt oder *gebefohlen, sondern befohlen, eine für Präfixverben typische Bildungsweise (ohne das Partizipialaffix ge-). Die lexikalische Repräsentation von befehlen muß also einerseits die Struktur “be-fehl-en” spezifizieren, andererseits klar machen, daß fehl- hier nicht der gleiche Stamm wie z. B. in Fehl-er ist. Es gibt also im Lexikon der Grammatik Einträge für befehl-, befolg-, begehr- etc., da diese Wörter zwar als präfigiert anzusehen sind, aber nicht der o. a. be-Regel folgen. Dagegen gibt es keinen Lexikoneintrag für z. B. bepflanz-; für dieses gibt es aber eine lexikalische Repräsentation, da die o. a. Regel auf pflanz- anwendbar ist, was in bepflanzen resultiert.
1768 2.2. Das mentale Lexikon Die Tatsache, daß die Wortform sagtest im Deutschen analysiert werden kann als bestehend aus dem Stamm sag und dem Affix -test, wobei letzteres die 2. Person Singular des Präteritums bezeichnet, und daß alle regulären deutschen Verben diese Form so bilden, besagt zunächst einmal nichts darüber, ob jemand, der diese Form in einem Satz äußern möchte, in seinem Gedächtnis erst den Stamm sag sucht, dann das Suffix für “2. Person Singular des Präteritums regulärer Verben” rauskramt, beide zusammenleimt und schließlich diese Gebilde produziert, oder ob ein Leser sie beim Lesen in ihre Teile zerlegen muß, bevor er sie verstehen kann (vgl. Günther 1988 b), denn “the existence of a logical system in the language does not necessarily imply that the mental structures that develop for dealing with this system will reflect this logic” (Sandra 1994: 245). Ein zentrales Merkmal linguistischer Theoriebildung ist es ja, daß von der realen Sprachtätigkeit systematisch abgesehen wird, u. a. von Bedingungen der Produktion und Perzeption von Äußerungen sowie von ihrer Modalität. Diese aber spielen eine zentrale Rolle bei psycholinguistischen Fragestellungen. Das Konzept des mentalen Lexikons bezeichnet den bei der menschlichen Sprachverarbeitung beim Sprechen und Hören, Lesen und Schreiben benutzten Speicher an sprachlichen Elementen. Der Ausdruck ist eine Metapher, wonach der Speicher sprachlicher Elemente im Gedächtnis nach Art eines Lexikons organisiert ist; eine gute Zusammenfassung des Forschungsstandes findet sich in Aitchison (21994). Lexikalischer Zugriff (lexical access) besteht in der Identifikation eines Reizes mit einer gespeicherten lexikalischen Einheit und ist definiert als derjenige Moment, in dem z. B. beim Lesen die auf dem Papier stehenden Buchstabenfolge WORT als das Wort Wort im Gedächtnis identifiziert wird. Nach den meisten Konzeptionen stehen in just diesem Moment schlagartig sämtliche Informationen zu dieser Lexikoneinheit zur Verfügung, also seine Bedeutung(en), verwandte Wörter, seine Aussprache, seine grammatischen Eigenschaften (Geschlecht, Flexion, Valenz), etc. Wie sind morphologisch komplexe Wörter im mentalen Lexikon gespeichert, und wie erfolgt lexikalischer Zugriff: auf morphologisch komplexe Wörter direkt, oder muß das Wort zuerst morphologisch zerlegt werden? Die Konsequenz aus beiden Ansätzen ist deutlich:
XIX. Psycholinguistische Perspektiven
Nach der oben schon erwähnten sog. Dekompositionshypothese enthält das Lexikon nur Stämme, Affixe und Wortbildungsregeln, und es muß immer erst zerlegt werden (Taft & Forster 1975); nach der Theorie vollständiger Einträge (sog. full listing hypothesis, Butterworth 1983) kann direkt auf die vollständig gespeicherten komplexen Wörter zugegriffen werden. Beide Ansätze sind dem symbolverarbeitenden Ansatz verpflichtet. Alternative Vorstellungen, die seit dem Beginn der 80er Jahre im konnektionistischen Umfeld publiziert worden sind, gehen dagegen von einem allgemeinen Erregungsgeschehen aus, in dem der magische Moment des lexikalischen Zugriffs als solcher entfällt; vielmehr ergibt er sich in einem ständig sich wandelnden Aktivationsgeschehen.
3.
Beobachtungsbereiche und Forschungsmethoden
Der Frage nach der mentalen Repräsentation und kognitiven Verarbeitung morphologisch komplexer sprachlicher Einheiten wird auf verschiedene Weise nachgegangen. Zusammenhängende Theorien lassen sich nur durch die Integration einer Vielzahl von Befunden aus verschiedenen Beobachtungsbereichen formulieren; dies sind u. a. linguistische morphologische Theorien, die Beobachtung und Analyse von Sprechfehlern und Sprachstörungen (Aphasien), die Beobachtung und Analyse des Spracherwerbs und seiner Störungen sowie die experimentelle Untersuchung der Produktion und Perzeption von Wörtern. Eine Darstellung von Theorien und Befunden in den genannten psycholinguistischen Bereichen geben die fünf folgenden Artikel (Art. 163⫺167). Es soll deshalb hier nur die jeweilige Forschungslogik gekennzeichnet und dazu eine Beschreibung der wichtigsten experimentellen Methoden gegeben werden. 3.1. Beobachtungsbereiche Die Analyse von Sprechfehlern gehört seit dem Ende des 19. Jahrhunderts zu den wichtigsten Quellen psycholinguistischer Theoriebildung aufgrund der schon in einer klassischen Studie vorgetragenen Erkenntnis, “daß man sich nicht regellos verspricht ... (die Sprechfehler) müssen durch konstante psychische Kräfte bedingt sein” (Meringer & Mayer 1895: 9). Weil Sprechfehler strukturiert sind und klassifiziert werden können, lassen sie Rückschlüsse auf den normalen
162. Mentale Repräsentation morphologischer Strukturen
Sprachproduktionsprozeß zu. Sprechfehler, in denen morphologische Elemente falsch arrangiert sind, vgl. wenn ein begründeter Bedacht versteht, im Wolken nordig oder er hat mich zum Skileiden eingelaufen (Beispiele aus Levelt 1989 und Leuninger 1993) werden als Beleg dafür angesehen, daß bei der Sprachproduktion komplexe Wörter aus den Morphemen zusammengesetzt werden. Für eine systematische Aufarbeitung der Sprechfehlerliteratur vgl. Levelt (1989); eine sehr knappe Zusammenfassung theoretischer Grundüberlegungen findet sich in Leuninger (1993) (s. auch Art. 164). Auch die Analyse von Sprachstörungen folgt der Logik, durch die Erklärung der Störungsursache den nicht direkt beobachtbaren Sprachmechanismus zu modellieren (s. Art. 167). Dabei werden spezifische sprachliche Ausfälle zum Beleg bestimmter theoretischer Konzepte der psycholinguistischen Morphologieforschung herangezogen. Das Phänomen des Agrammatismus hat besondere Aufmerksamkeit gefunden, weil ein Aspekt der Symptomatik, das Weglassen von Flexionsaffixen, diese aphasischen Störung prima facie als eine flexionsmorphologische Störung zu kennzeichnen scheint (Goodglass & Berko 1960). Einer sehr vielbeachteten Theorie zufolge handelt es sich aber um eine phonologisch bedingte Störung (Kean 1979). Neuere Arbeiten sehen eher eine syntaktische Störung, d. h. die morphologischen Ausfälle sind Folge und nicht Ursache des Agrammatismus (z. B. Grodzinsky 1990; Art. 167). Ein anderes Syndrom ist die sog. phonologische Dyslexie (die Patienten können u. a. Wörter lesen, aber keine Pseudowörter). Caramazza et al. (1988) interpretieren die Ergebnisse einer Untersuchung solcher Patienten als Beleg für ihren Ansatz eines Vollformenlexikons, in dem Wörter morphologisch zerlegt sind (s. 4.1). In der Spracherwerbsforschung gehört die Entwicklung der Morphologie zu den zentralen Themen. In allen Sprachen der Welt beginnen Kinder mit sog. Einwortsätzen zu sprechen, d. h. unflektierten Einzelwörtern. Flexionsformen werden in einer gewissen Reihenfolge erworben, wobei die Komplexität der signalisierten Bedeutung und die des sprachspezifischen Formensystems, die Eindeutigkeit der Laut-Bedeutungs-Zuordnung und andere Faktoren eine Rolle spielen (Slobin 1985, Hrsg.; Art. 165). Die Fähigkeit von Kindern, erworbene Flexionsregularitäten zu (über)generalisieren und auch auf unpas-
1769
sende Stämme sowie auf Nonsenswörter anzuwenden (z. B. singte, bringte, beste statt sang, brachte, fegte ⫺ Berko 1958, ein Klassiker der psycholinguistischen Morphologieliteratur) wird in der Regel als Beleg für eine morphologische Dekomposition im o. g. Sinne interpretiert. Die Fähigkeit zur Wortbildung durch Derivation bildet sich später heraus; auch hier kommt es zu Übergeneralisierungen in Fällen, in denen in der Erwachsenensprache ein bestimmter anderer Ausdruck konventionalisiert ist (wie das o. g. er beste statt fegte). Unklar ist, welche Relevanz Erwerbsdaten für die Modellierung mentaler Repräsentationen und den Zugriff auf morphologisch komplexe Wörter beim Erwachsenen haben (vgl. Sandra 1994: 260⫺264). Die experimentelle Untersuchung der Produktion und Perzeption morphologisch komplexer Wörter hat seit der initialen Studie vor 20 Jahren (Taft & Forster 1975) einen stürmischen Aufschwung genommen. Denn eine psychologische oder psycholinguistische Theorie kann nur in dem Maße als validiert angesehen werden, als sie empirisch ⫺ und das heißt experimentell ⫺ bestätigt ist; vgl. zu den methodischen Grundlagen Günther (1989). Eine Zusammenfassung der Befunde und Argumentationen in der älteren experimentellen Morphologieforschung bietet Henderson (1985); eine problemorientierte neuere Darstellung findet sich bei Sandra (1994) (s. auch Art. 163). Im folgenden sollen die wichtigsten Untersuchungsmethoden aufgezählt werden. 3.2. Methoden In der psycholinguistischen Morphologieforschung stehen bislang Experimente zur visuellen Worterkennung im Vordergrund. In solchen Experimenten wird der Versuchsperson ein einzelnes Wort visuell dargeboten. Bei den in der älteren Literatur beschriebenen Experimenten bedient man sich eines Tachistoskops. Mithilfe dieses Geräts ist es möglich, visuelle Reize beliebig kurz darzubieten. Die Aufgabe der Versuchsperson besteht darin, den nur als kurzen Lichtblitz gezeigten Reiz (schriftlich oder mündlich) zu benennen. Ausgewertet wird meist entweder die Fehlerzahl oder die Anzahl wiederholter Darbietungen, die nötig ist, um den Reiz korrekt zu identifizieren. Diese Methode wurde z. B. bei der Untersuchung der Verarbeitung englischer flektierter Wortformen angewandt (Gibson & Guinet 1971). Eine andere Aufgabenstellung besteht darin, gleichzeitig zwei
1770 Reize darzubieten und danach zu fragen, ob diese gleich oder verschieden sind. Sie fand z. B. bei der Untersuchung starker und schwacher Verbflexion Verwendung (Jarvella & Snodgrass 1974). Ebenfalls häufig ist die Messung der Zeit, die eine Versuchsperson braucht, um ein unter verschiedenen Bedingungen gezeigtes Wort auszusprechen (z. B. Schreuder et al. 1989). Die überwiegende Mehrzahl der in der neueren wissenschaftlichen Literatur publizierten Experimente zur Verarbeitung morphologisch komplexer Wörter aber bedient sich der Technik der lexikalischen Entscheidung. Hier bekommt die Versuchsperson auf einem Bildschirm eine Buchstabenfolge dargeboten und hat die Aufgabe, so schnell wie möglich zu entscheiden, ob diese Folge ein Wort ist oder nicht; gemessen wird die Reaktionszeit. Die Popularität dieser Aufgabenstellung in den letzten 20 Jahren hat mit dem Konstrukt des lexikalischen Zugriffs zu tun. Zur Untersuchung der Vorgänge beim lexikalischen Zugriff können im Grunde nur solche Aufgaben herangezogen werden, bei denen man sicher sein kann, daß auch tatsächlich lexikalischer Zugriff erfolgt (Coltheart 1978). Diese Sicherheit gibt es z. B. nicht, wenn die Versuchsperson die Aufgabe hat, von einem Paar von Wörtern zu sagen, ob diese gleich oder verschieden sind: Das kann sie tun, ohne ihr mentales Lexikon zu konsultieren, indem sie einfach Buchstaben vergleicht. Anders bei der lexikalischen Entscheidung, jedenfalls beim Einhalten bestimmter Randbedingungen: Daß die Buchstabenfolge FEST ein deutsches Wort ist, DEST dagegen nicht, kann man nur entscheiden, wenn man FEST im mentalen Lexikon gefunden hat, da die Buchstabenfolge DEST durchaus ein deutsches Wort sein könnte. Es ist typisch für die lexikalische Entscheidungsaufgabe, daß die Reaktionszeiten auf Wörter kürzer sind als die auf sog. Pseudowörter wie DEST, die orthographisch und phonologisch möglich, aber in der Sprache nicht realisiert sind. Die Logik solcher Experimente besteht darin, die Reaktionszeit auf zwei verschiedene Typen von Wörtern miteinander zu vergleichen, von denen die Verarbeitung des einen Typs durch einen morphologischen Zerlegungsprozeß nach der Theorie behindert /gefördert würde, die des anderen aber nicht. Ein Beispiel bieten sog. pseudopräfigierte Wörter: Das bekannteste englische Beispiel ist sublime, ein deutsches Beispiel wäre gering. Wenn man annimmt, daß automatisch
XIX. Psycholinguistische Perspektiven
zerlegt wird, so erhält man mit sub-lime (bzw. ge-ring) neben dem Pseudopräfix einen Bestandteil, der zwar im mentalen Lexikon aufzufinden ist (lime, ring), aber nicht Stamm des gezeigten Wortes ist; es muß reanalysiert werden, wodurch es zu Reaktionszeit-Verlangsamungen im Vergleich zu “echt” präfigierten Wörtern kommt (Taft 1981). Eine quasi umgekehrte Logik findet sich in Experimenten, die sich zusätzlich der sog. Priming-Technik bedienen. Werden einzelne Reize während des Experiments zweimal gezeigt, so zeigt ein Wort bei der Wiederholung eine signifikant kürzere Reaktionszeit verglichen mit der ersten Darbietung (Scarborough et al. 1977). Zwischen der ersten Darbietung (prime) und der zweiten (target) können unterschiedlich viele andere Kontrollreize eingefügt werden. Häufig werden die Reize auch ohne Pause hintereinander dargeboten, wobei nur auf den zweiten Reiz reagiert wird. Dieser Ansatz des Bahnens durch Wiederholung (repetition priming) hat in der Morphologieforschung der letzten Jahre eine besondere Rolle gespielt. Es konnte nämlich gezeigt werden, daß ein Bahnungseffekt auch dann auftritt, wenn nicht die gleiche Wortform als Bahnungsreiz dargeboten wird, sondern eine morphologisch verwandte Form, z. B. walked vor walk; dies wurde als Beleg dafür angesehen, daß in der Tat der Stamm wiederholt verarbeitet wird und nicht zwei einzelne Formen (Stanners et al. 1979). Neuerdings spielt auch die Modalität eine Rolle; so wird die Reaktionszeit auf einen visuellen Reiz gemessen, dem ein auditives Prime voranging (cross modal priming, vgl. z. B. Marslen-Wilson et al. 1994; für eine Diskussion der wichtigsten früheren Ergebnisse vgl. Monsell 1985, für neuere Daten Sandra 1994 und Art. 163). Eine andere Untersuchungstechnik ist die Beobachtung der Augenbewegungen beim Lesen (Inhoff & Rayner 1996). Durch verschiedene Meßmethoden läßt sich zeigen, ob morphologische Eigenschaften der gelesenen Wörter das Augenbewegungsmuster tangieren, z. B. ob bei pseudopräfigierten Wörtern abweichende Bewegungsmuster zu beobachten sind (Lima 1987). Ein Vorteil dieser Methodik besteht darin, daß Sätze oder auch Texte gelesen werden können, nicht nur einzelne Wörter (s. 4.3). ⫺ Da die meisten psycholinguistischen Experimente zur Verarbeitung morphologisch komplexer Wörter nach wie vor im Bereich der visuellen Worterken-
162. Mentale Repräsentation morphologischer Strukturen
nung angesiedelt sind, kann für weitere Methoden einschließlich kritischer Diskussion auf Smith (1996) verwiesen werden.
4.
Einige Grundprobleme der Forschung
Die psycholinguistische Morphologieforschung ist am Beginn des 21. Jahrhunderts nur 30 Jahre alt. Anders als in anderen Bereichen, in denen während der 80er Jahre ein systematisches Abschütteln der Dominanz linguistischer Theorien zugunsten einer Synthese psychologischer und linguistischer Theoriebildung notiert werden konnte (Tanenhaus 1988), ist in der Morphologieforschung deren Verhältnis nach wie vor unklar (Henderson 1988). Dies hat eine besonders ausgeprägte Divergenz der Ansätze zur Folge. Im folgenden sollen die wichtigsten Fragestellungen und Hauptprobleme der gegenwärtigen Diskussion knapp kritisch gekennzeichnet werden. 4.1. Dekomposition vs. Vollformenlexikon Dies ist die klassische Fragestellung: Werden morphologisch komplexe Wörter bei der Sprachwahrnehmung erst zerlegt, bei der Produktion erst zusammengesetzt, ehe sie verstanden/produziert werden können, oder sind sie als “Ganze” für die Wahrnehmung und Produktion verfügbar; enthält das mentale Lexikon ganze Wörter, deren morphologische Struktur die Verarbeitung nicht beeinflußt (Butterworth 1983), oder nur Stämme und Affixe, deren struktureller Zusammenhang durch Dekompositionsregeln bestimmt ist (Taft & Forster 1975)? Praktisch alles, was zwischen 1975 und 1985 in der psycholinguistischen Literatur zu morphologischen Aspekten der Sprachverarbeitung publiziert wurde, ist dieser Frage gewidmet (vgl. Henderson 1985). Die aus der linguistischen Diskussion (Halle 1973) in die Psychologie importierte Dichotomie von Dekomposition vs. Vollformenlexikon war ursprünglich durch Ökonomieerwägungen motiviert. Es schien unrealistisch, daß alle Formen im menschlichen Gedächtnis aufgelistet sein sollten; die linguistisch basierten Regeln boten eine Möglichkeit, den Speicherbedarf zu reduzieren. Betrachtet man die Metaphorik wissenschaftlichen Sprachgebrauchs, so zeigt sich eine enge Bindung an den jeweils erreichten Stand der Technik. Das Streben nach Speicherplatzmi-
1771
nimierung der frühen Dekompositionsmodelle entsprach den Kleincomputern der ersten Generation: So verblüffend (damals) ihre Geschwindigkeit schien, so klein war der Speicher ⫺ es stand für maximal 6000 Wörter Speicherplatz zur Verfügung, und nur etwa das Fünffache ließ sich auf Diskette speichern. Trotzdem gab es Disketten mit Lexika von knapp 200000 Wörtern, die geschickt organisiert waren und nur durch spezielle Software mit “Auspackregeln” lesbar waren. Demgegenüber ist großer Speicher in Form von Festplatten oder CD-ROM heute überhaupt kein Problem mehr. Statt der Reduzierung von Speicherplatz ist heute Geschwindigkeit des Zugriffs auf große Datenmengen gefragt. In weniger als 300 Millisekunden ruft der durchschnittliche Erwachsene praktisch fehlerfrei aus einem Wortschatz von 50000 bis 400000 Wörtern das richtige Wort ab (Überblick zu solchen Zahlen bei Aitchison 21994: 5⫺8) und fügt es in syntaktische und semantische Strukturen ein. Auch bislang für unvorstellbar gehaltene Datenmengen scheint das menschliche Gedächtnis aufnehmen zu können, wenn sie strukturiert abgelegt werden (s. 4.2). Sofern also Prozeduren wie die morphologische Zerlegung allein durch Ökonomieüberlegungen motiviert sind, sind sie abzulehnen, da sie Zeit kosten. Befürworter morphologischer Dekomposition müssen deshalb zeigen, daß ihr Modell schnellere Verarbeitung ermöglicht als ein Vollformenlexikon. Es ist instruktiv, diese Entweder-Oder Diskussion in Parallele mit der Frage des phonologischen Rekodierens beim Lesen zu stellen: Involviert der Lesevorgang notwendig eine Phase des phonologischen Rekodierens, oder liest zumindest der erfahrene Leser “direkt”, d. h. ohne die Notwendigkeit phonologischer “Übersetzung”? Weil es empirische Argumente für beide Auffassungen gab, entwikkelte sich die Idee des Pferderennenmodells (horse race model), womit eine über 10 Jahre dauernde Diskussion (nicht unbedingt voll befriedigend) abgeschlossen wurde: Danach wird immer beides versucht, nämlich phonologisches Rekodieren und direkter Zugriff; was perzipiert wird, ist (primär) von der Geschwindigkeit der Routen abhängig (Humphreys & Evett 1985; die Schwächen einer solchen Konzeption werden in Henderson (1982) ausführlich diskutiert). Dem entspricht das Ergebnis der Diskussion über Decomposition vs. Full Listing: Extremen Vertretern der einen Richtung (Taft &
1772 Forster 1975) und der anderen (Butterworth 1983) entgegnen etwa Laudanna & Burani (1985) oder Frauenfelder & Schreuder (1992), daß auch hier ein Wettbewerb stattfinde: Danach sind im Lexikon sowohl Stämme und Affixe als auch volle Wörter enthalten; zwei parallele Prozesse suchen entweder nach dem einen oder dem anderen, und der schnellere obsiegt. 4.2. Lexikalischer Zugriff und mentale Repräsentation Betrachtet man die Literatur, so ist bemerkenswert, daß die Frage nach der Rolle morphologischer Strukturen fast ausschließlich als Problem des lexikalischen Zugriffs diskutiert worden ist (Sandra 1994). Dabei wurden paradoxerweise entgegengesetzte Positionen für den gleichen Sachverhalt bemüht. Das Modell von Forster (1976) sieht einen Hauptspeicher aller Wörter vor; nur für den lexikalischen Zugriff muß zerlegt werden. In der Tat ist bei der Weiterentwicklung zum sog. BOSS-Modell (vgl. Taft 1991) nicht mehr die rein morphologische Zerlegung wichtig, sondern nurmehr eine linguistisch völlig unmotivierte Zwitterstruktur aus anlautender Silbe und Morphem. Caramazza et al. (1988) dagegen sehen alle Wörter in der Zugriffsdatei; der Hauptspeicher bietet die Wörter in zerlegter Form. Frauenfelder & Schreuder (1992) schließlich nehmen eben zur Erklärung der Geschwindigkeit doppelte Repräsentation (und doppelten Zugriff) an. Vom linguistischen Standpunkt aus betrachtet ist diese Vorstellung besonders problematisch: Sie verlegt systematisches morphologisches Wissen in unklare psychologische Prozesse. Eine alternative Konzeption könnte so aussehen, daß zwar das Lexikon volle Einträge enthält, diese aber morphologisch zerlegt und angeordnet sind (Caramazza et al. 1988); gemäß den Anordnungsprinzipien des Lexikons sind damit implizit auch die Morpheme gelistet. Bei der Verarbeitung liegt dann weder prä- noch postlexikalische Dekomposition vor: Es ist Bestandteil des sprachlichen Wissens, daß Haus-tür, freund-lich, sag-t etc. Einheiten sind, die aus zwei Teilen bestehen. Die zerlegte Struktur der Lexikoneinträge impliziert dabei nicht, daß morphologische Strukturen immer in die Syntax “mitgenommen” werden (z. B. bei Argumentvererbung, s. Art. 83), schließt das aber auch nicht aus (vgl. zu dieser Frage Anderson 1992 und die Rezension durch Carstairs 1993).
XIX. Psycholinguistische Perspektiven
Es ist offensichtlich, daß konnektionistische Modelle eine solche Konzeption systematisch inhärieren; aber auch symbolverarbeitende Ansätze könnten sie aufnehmen. In der psychologischen Diskussion war die Möglichkeit der internen Strukturiertheit von vollständig gespeicherten Wörtern lange Zeit nicht gesehen worden. Alle Befunde, die einen Einfluß morphologischer Struktur auf die Verarbeitung belegen (z. B. der Pseudopräfixeffekt, s. 3.2), lassen sich in einem Full Listing-Modell mit interner morphologischer Strukturierung gut erklären: Der Verarbeitungsvorteil entsteht hier dadurch, daß die morphologische Information bei der Verarbeitung genutzt wird; dieser automatische Vorgang führt bei pseudopräfigierten Wörtern (gering) zu Verlangsamungen. Dies ist weniger trivial als es zunächst scheint. Beispielsweise ist faktisch unbegrenzter Speicher wenig nützlich, wenn die gespeicherten Daten nicht auch schnell abrufbar sind. Wenn das Suchsystem, wie bei einer Tonkassette, nur linear abfragen kann, weil die Daten eindimensional linear und nicht weiter strukturiert angeordnet sind, ist kaum zu verstehen, wie solche Geschwindigkeiten erreicht werden sollen. Abgesehen von konnektionistischen Modellen sind im Allgemeinen lineare Abfragemechanismen vorgesehen worden (prototypisch: Forster 1976; vgl. kritisch Sandra 1994). Neuerdings ist die Forschung zum mentalen Lexikon stärker mit der Ermittlung seiner Organisationsprinzipien befaßt (Aitchison 21994). Offenbar spielt sowohl die links/rechts Orientierung als auch die von rechts nach links (Reimfähigkeit) eine Rolle. Primärassoziationen, Clusterbildung, semantisches Priming etc. zeigen, daß auch die semantischen Beziehungen zwischen den Wörtern Anordnungsprinzipien bilden ⫺ und es gibt noch viele andere (vgl. Aitchison 21994). Auf die alte Metapher des mentalen Lexikons als Bibliothek angewendet bedeutet das, daß z. B. alle Einheiten, die mit pe anfangen, in einem Regal stehen; die mit pel stehen links von denen mit, z. B., pet. Die reimenden Einheiten stehen vor bzw. hintereinander (Sprung, Leitung, jung vor Geiz, Reiz, Schweiz, aber hinter prima, firma, Rheuma) und die semantische Dimension ist in der Vertikalen ausgedrückt: Die Pflanzenwörter (z. B. Pelargonie, Petunie) stehen unter den Tieren (Pelikan, Perlhuhn), etc. Es ist offensichtlich, daß die Metapher zusammenbricht, sobald mehr als drei Dimensionen auftreten; dies aber ist der Fall: mor-
162. Mentale Repräsentation morphologischer Strukturen
phologische, syntaktische, konnotative Merkmale gehören ebenfalls zur Systematik des mentalen Lexikons. Im Sinne herkömmlicher Datenbanksysteme läßt sich diese vieldimensionale Anordnung der Einheiten als Indizierung begreifen: Jede Einheit ist u. a. sowohl bezüglich ihrer linksläufigen als auch ihrer rechtsläufigen Organisation, ihrer semantischen Kategorie, ihrer syntaktischen Klasse und ihrer alphabetischen Position indiziert. Alle Einheiten könnten intern strukturiert sein. Eine solche Zerlegtheit läßt die Indizierung der einzelnen Lexikoneinheiten nach Morphemen zu. Im konnektionistischen Ansatz würde man davon reden, daß ein Eintrag mit allen gleich beginnenden, allen reimenden, allen semantisch ähnlichen Einträgen etc. Verknüpfungen aufweist, und zwar unterschiedlich starke nach Maßgabe der Ähnlichkeit. Es ist ungeklärt, inwieweit dies lediglich Notationsvarianten ein und derselben Struktur sind oder aber inhaltlich verschiedene Lösungen. 4.3. Flexion vs. Wortbildung, Produktivität, Lexikalisierung Die Unterscheidung zwischen Flexion und Wortbildung ist in den frühen psycholinguistischen Arbeiten praktisch nicht berücksichtigt worden (Henderson 1985). Obgleich die exakte Grenzziehung im Einzelfall problematisch ist, so daß bisweilen von einem Kontinuum statt einer Dichotomie gesprochen wird (Bybee 1985), ist in der Mehrzahl der Fälle die linguistische Zuordnung problemlos: Flexion ist automatisch, verändert Wortart und lexikalische Bedeutung nicht, Wortbildung dagegen ist idiosynkratisch und verändert häufig Wortart und Bedeutung. In Analogie zur segmentalen Ebene visueller Worterkennung, für die eine Ebene abstrakter Buchstabenrepräsentationen angenommen wird (Coltheart 1981), wird von vielen Forschern im Falle der Flexion eine abstrakte Repräsentation angenommen (z. B. Marslen-Wilson et al. 1994), aus der dann bei der Äußerung die konkrete Form gebildet wird ⫺ eine Form des Paradigmas steht im Lexikon, die anderen werden daraus für die konkrete Verwendung “abgeleitet”. Experimentelle Daten legen nahe, als solche Basisform die lexikalische Zitierform anzusehen (Günther 1988a). Dabei ist nicht Speicherplatzökonomie der Grund für eine solche Annahme (s. 4.1). Es ist vielmehr der Umstand, daß auch Wörter, die der Sprecher nicht kennt oder selbst erfindet, sofort flektiert
1773
werden können (Mugdan 1977), also bereits bevor sie Teil des mentalen Lexikons sind. Teilweise wird auch davon ausgegangen, daß häufige Formen als Ganze gespeichert und abgerufen, seltene jedoch abgeleitet werden (z. B. Frauenfelder & Schreuder 1992). Für die Unterscheidung von Flexion und Derivation auch bei der mentalen Repräsentation spricht der Umstand, daß Bahnungseffekte zwischen Wortformen desselben Lexems immer wieder gefunden wurden (also z. B. zwischen engl. walking und walked, zuerst von Stanners et al. 1979); solche Effekte gibt es nicht zwischen derivierten Bildungen (wie friendly und friendship, Marslen-Wilson et al. 1994). Es ist bekanntlich möglich, aus den vorhandenen Morphemen gänzlich neue Kombinationen zu bilden, die jedoch von den Hörern sofort verstanden werden wie das betüpfelbar (s. 2.1), das man sofort zu unbetüpfelbar oder gar Unbetüpfelbarkeitssymptomatik erweitern kann, die ebenfalls direkt verstehbar sind. Man spricht davon, daß die Affixe wie be-, -bar, -keit, un- oder die nominale Komposition im Deutschen produktiv sind. Dieser alte Begriff linguistischer Wortbildungsforschung ist erst in jüngster Zeit Gegenstand psycholinguistischer Studien geworden (Baayen 1992; Frauenfelder & Schreuder 1992). Dabei ist es bemerkenswert, daß sich in der Linguistik Stimmen mehrten, dieses Konzept als grundsätzlich außerhalb des Bereichs linguistischer Morphologieforschung, als reine Angelegenheit der Sprachverwendung anzusehen (Motsch 1988). Zwar ist das Vorkommen gänzlich “neuer” Bildungen, d. h. solcher, die man nie gehört oder gesehen hat, erheblich seltener als die Literatur suggeriert, doch scheinen Hörer und vor allem Leser solche Wörter, wenn sie ordnungsgemäß gebildet sind, ohne Probleme zu verstehen. Die Produktivität von Wortbildungsmustern stellt ein zentrales Problem für Listenmodelle dar, die die Prozeßkomponente auf den reinen Zugriff beschränken wollen. Typisch dafür ist der Ansatz von Aitchison (21994), Neubildungen durch einen völlig unzureichend spezifizierten lexical tool kit zu erfassen, einer Art Zusatzeinrichtung, die keinen Bezug auf die im Lexikon enthaltene Morphologie nimmt (vgl. kritisch Berg 1989). Der Ansatz von Baayen (1992) zur Berücksichtigung von Produktivität in einem psycholinguistischen Modell besteht darin, Häufigkeiten von Token in Relation zur Zahl der
1774 vorkommenden Typs zu stellen; die so ermittelten Werte determinieren in seinem ZweiWege-Modell, ob ein Wort dekomponiert wird oder lexikalischer Zugriff auf das ganze Wort erfolgt (d. h. also eine Zwei-Wege-Theorie der morphologischen Dekomposition); eine besondere Rolle spielen dabei ad-hocBildungen, die definitionsgemäß nicht in einer Liste aufgeführt sein können. Auch in der kritischen Weiterentwicklung des Modells durch Frauenfelder & Schreuder (1992) spielt Vorkommenshäufigkeit eine zentrale Rolle; allerdings berücksichtigen sie auch die Durchsichtigkeit von Wortbildungen. Hinzugefügt muß allerdings werden, daß bisher experimentelle oder andere empirische Daten fehlen, die belegen, daß eine Neubildung wie betüpfelbar tatsächlich auf die gleiche Weise erkannt bzw. produziert wird wie etwa bemalbar, d. h. daß bei der Produktion oder Erkennung von Neubildungen derselbe Mechanismus aktiv wird wie bei “alten” Bildungen (vgl. auch Sandra 1994). Auch die Lexikalisierung von Wörtern, quasi die umgekehrte Dimension, hat in der psycholinguistischen Morphologieforschung bislang kaum Beachtung gefunden. So werden etwa in Taft (1981) als präfigierte Stimuli ausschließlich englische Wörter herangezogen, die als völlig lexikalisiert gelten müssen (z. B. approach, reproach). Es scheint jedoch aus linguistischer Sicht äußerst unwahrscheinlich, daß, wie auch immer sie im mentalen Lexikon repräsentiert sein und daraus abgerufen werden mögen, lexikalisierte Bildungen wie Zeitung, Handschuh oder anfangen auf die gleiche Weise verarbeitet werden wie durchsichtige Wörter wie Bergung, Hausschuh oder abfangen. In der Tat weisen Ergebnisse von Marslen-Wilson et al. (1994) in diese Richtung: Zwischen lexikalisierten Bildungen und dem Stamm treten keine Bahnungseffekte auf, wohl aber bei durchsichtigen Bildungen: Engl. government bahnt govern, aber department hat keine bahnende Auswirkung auf depart. 4.4. Untersuchungstechniken und Modalitäten Die Mehrzahl der psycholinguistischen Untersuchungen zur Rolle der Morphologie in der Sprachverarbeitung bedient sich der visuellen lexikalischen Entscheidungsaufgabe. Diese Methode ist, wie verschiedene Befunde der letzten 10 Jahre nahelegen, in besonderer
XIX. Psycholinguistische Perspektiven
Weise strategieabhängig: Die Versuchsperson versucht, aus den gebotenen Reizen bestimmte Prinzipien für ihre Antwortstrategie abzuleiten. So läßt sich der Pseudopräfixeffekt zum Verschwinden bringen, wenn alle Wörter des Experiments präfigiert sind (Rubin et al. 1979; Günther 1987); Bahnungseffekte werden um so stärker, je mehr “gute” Primes im Material auftauchen, etc. Es ist deshalb denkbar, daß eine Versuchsperson beim lexikalischen Entscheidungsexperiment Dinge tut, die sie bei der “normalen” Produktion und Perzeption von Sprache gerade vermeidet. Dafür spricht u. a. die Reaktionszeit: Geht man davon aus, daß die durchschnittliche Zugriffszeit zu einem Wort 220 msec beträgt (vgl. Aitchison 21994: 8), so klafft zwischen diesem Wert und den ca. 500 msec lexikalischer Entscheidungsaufgaben eine Lücke, die nicht primär der Auslösung und Exekution der motorischen Reaktion zugeschrieben werden kann. Es liegt nahe, an teilautomatisierte metasprachliche Determinanten der Reaktionszeiten zu denken. Dies wird verstärkt durch die Tatsache, daß es sich hier (wie in vielen anderen Aufgaben) um die Erkennung einzelner Wörter handelt. Dieser Umstand mag Strategien und Verhaltensweisen fördern, die in der “normalen” Sprachproduktion und -perzeption fehlen. In der Tat ist die Erkennung isolierter Wörter eine Aufgabe, die selbst in auditiven Experimenten Schriftlichkeit involviert oder zumindest voraussetzt; es ist eine in genuiner Mündlichkeit nicht vorkommende Tätigkeit, mit einzelnen Wörtern zu hantieren. Es mag also der Fall sein, daß ein Großteil der experimentellen Befunde zur Morphologie eine Ebene schriftbezogener Metasprachlichkeit betreffen, die in der “normalen” Sprachproduktion und -rezeption eine untergeordnete Rolle spielt. Das dürfte insbesondere für Befunde gelten, in denen lexikalisierte Bildungen dennoch morphologische Effekte zeigen (z. B. Taft 1981). Sandra (1994: 264 f.) zitiert Überlegungen, wonach morphologische Effekte simple Konsequenzen orthographischer Redundanz sind. Wichtiger scheint die Beobachtung, daß entwickelte Alphabetschriften grundsätzlich zur Morphologisierung tendieren. Ein wesentliches Ergebnis der neueren Orthographieforschung ist die Beobachtung des Prinzips der Morphemkonstanz in der geschriebe-
1775
162. Mentale Repräsentation morphologischer Strukturen
nen Sprache, die es im Gesprochenen so nicht gibt (s. Art. 175). In der Tat ist es ja ein wesentliches Geschäft von Phonologie und Morphophonemik, verschiedene Lautformen auf ein und dieselbe lexikalische Einheit zurückzuführen. Im Geschriebenen (Deutsch oder Englisch) ist dagegen eine solche Verschiedenheit die (eigens zu erklärende) Ausnahme. Das Morphemkonstanzprinzip ist im Englischen, der (bislang) hauptsächlich untersuchten Sprache in der psycholinguistischen Morphologieforschung, besonders ausgeprägt; es gibt gute Gründe anzunehmen, daß der Zusammenhang von sign und signal, von concern (Verb) und concern (Sustantiv) etc. dem Kind erst nach Erwerb der geschriebenen Sprache zur Verfügung steht. Unter diesen Umständen ist es bemerkenswert, daß morphologische Fragestellungen in erster Linie in schriftlichen Aufgabenstellungen untersucht worden sind. Bis ca. 1985 war die psycholinguistische Morphologieforschung vollständig auf die visuelle Domäne beschränkt; auch der jüngste Überblicksartikel beschränkt sich ausdrücklich auf diesen Bereich (Sandra 1994); Untersuchungen gesprochener Sprache wie Tyler et al. (1988) bleiben die Ausnahme. 4.5. Die Rolle der Diachronie Die Frage nach der mentalen Repräsentation morphologisch komplexer sprachlicher Einheiten ist im derzeitigen Stadium kaum auch nur ansatzweise zu beantworten. Dies liegt in einer besonderen Diskrepanz der verschiedenen Forschungsansätze. Im Gegensatz etwa zur Erforschung der Satzverarbeitung ist das Verhältnis zwischen linguistischen und psychologischen Ansätzen nach wie vor ungeklärt (Henderson 1988). Auf der einen Seite sind vor allem die frühen Modelle stark von linguistischen Überlegungen zur Struktur des Lexikons der Grammatik beeinflußt, deren Zielsetzung die systematische Ausblendung von Aspekten der Sprachtätigkeit bedingte, dem wesentlichen Forschungsgegenstand psychologischer Forschung. Paradoxerweise hat die Konzentration (um nicht zu sagen Reduktion) auf das (psychologische) Problem des lexikalischen Zugriffs Modelle hervorgebracht, die (zumindest für dieses Problem) sprachliche Strukturen als möglicherweise irrelevant erscheinen lassen (Sandra 1994). Auf der anderen Seite sind bis in jüngste Zeit klassische Fragestellungen sprachwissen-
schaftlicher Forschung (Produktivität, Lexikalisierung) dort praktisch nicht zur Kenntnis genommen worden. Dabei könnte gerade dieser Bereich zum Modellfall für das sinnvolle Miteinander linguistischer und psychologischer Forschung werden. Das folgende Zitat kennzeichnet den Sachverhalt, wenn Henderson (1982: 65) zu bedenken gibt, “how much clarifications of a psychological problem can be achieved before we have done a single experiment. Indeed, it seems to me folly to take up a psychological position before examining in some detail what the language allows and encourages. The linguistic evidence does not provide the psychological answers, but it goes a long way toward clarifying the questions”. Die ältere linguistische Literatur zu Flexion und Wortbildung und ihren historischen Veränderungen wimmelt geradezu von psychologisierenden Erklärungen. Insofern scheint die systematische Auswertung diachroner Morphologie- und Wortbildungsforschung auch und gerade für die psycholinguistische Modellierung der mentalen Repräsentation morphologisch komplexer Wörter von Interesse: Welche Typen von Veränderungen gibt es, und wie lassen sich diese auf Aspekte rezenter Modelle abbilden?
5.
Zitierte Literatur
Aitchison, Jean (21994), Words in the Mind: An Introduction to the Mental Lexicon. London: Blackwell Anderson, Stephen R. (1992), A-Morphous Morphology. Cambridge: Cambridge University Press Baayen, Harald (1992), “Quantitative Aspects of Morphological Productivity”. In: Booij & van Marle (Hrsg.), 109⫺149 Berg, Thomas (1989), “Rezension zu J. Aitchison, Words in the Mind”. Lingua 79, 342⫺346 Berko, Joan (1958), “The Child’s Learning of English Morphology”. Word 14, 150⫺177 Booij, Geert & Marle, Jaap van (1992, Hrsg.), Yearbook of Morphology 1991. Dordrecht: Kluwer Butterworth, Brian (1983), “Lexical Representation”. In: Butterworth, Brian (Hrsg.), Language Production, Bd. II: Development, Writing and Other Language Processes. London: Academic Press, 257⫺294 Bybee, Joan L. (1985), Morphology ⫺ A Study of the Relation between Meaning and Form. Amsterdam: Benjamins
1776
XIX. Psycholinguistische Perspektiven
Caramazza, Alfonso & Laudanna, Alexander & Burani, Cristina (1988), “Lexical Access and Inflectional Morphology”. Cognition 28, 297⫺332
Günther, Hartmut (1988 a), “Oblique Word Forms in Visual Word Recognition”. Linguistics 28, 583⫺600
Carstairs, Andrew (1993), “Morphology without Word-internal Constituents: A Review of Stephen R. Anderson’s A-Morphous Morphology”. In: Booij, Geert & van Marle, Jaap (Hrsg.), Yearbook of Morphology 1992. Dordrecht: Kluwer. 209⫺233
Günther, Hartmut (1988 b), “Kreativität, Produktivität, Lexikon ⫺ Zur Kompatibilität und Inkompatibilität linguistischer und psychologischer Konzepte in der Morphologieforschung”. In: Knobloch, Clemens (Hrsg.), Kommunikation und Kognition. Münster: Nodus, 157⫺172
Carstairs, Andrew & Stemberger, Joseph P. (1988), “A Processing Constraint on Inflectional Homonymy”. Linguistics 26, 601⫺618 Clahsen, Harald & Rothweiler, Monika (1992), “Inflectional Rules in Children’s Grammars: Evidence from German Past Participles”. In: Booij & van Marle (Hrsg.), 1⫺34 Coltheart, Max (1978), Lexical Access in Simple Reading Tasks. In: Underwood, Geoffrey (Hrsg.), Strategies of Human Information Processing. London: Academic Press, 151⫺217 Coltheart, Max (1981), “Disorders of Reading and their Implications for Models of Normal Reading.” Visible Language 15, 245⫺286 Derwing, Bruce L. & Skousen, Royal (1989), “Morphology in the Mental Lexicon: A New Look at Analogy”. In: Booij, Geert & Marle, Jaap van (Hrsg.), Morphology Yearbook 2, 55⫺71 Ellis, Andrew W. (1985, Hrsg.), Progress in the Psychology of Language, Bd. 1⫺2. London: LEA Forster, Kenneth I. (1976), “Accessing the Mental Lexicon”. In: Wales, Roger J. & Walker, Edwin (Hrsg.), New Approaches to Language Mechanisms. Amsterdam: North-Holland, 139⫺174 Frauenfelder, Uli & Schreuder, Rob (1992), “Constraining Psycholinguistic Models of Morphological Processing and Representation: The Role of Productivity”. In: Booij & van Marle (Hrsg.), 165⫺193 Gibson, Eleanor J. & Guinet, L. (1971), “Perception of Inflections in Brief Visual Presentations of Words”. Journal of Verbal Learning and Verbal Behaviour 10, 182⫺189 Goodglass, H. & Berko, Joan (1960), “Agrammatism and Inflectional Morphology”. Journal of Speech and Hearing Research 3, 257⫺267 Grodzinsky, Yosef (1990), Theoretical Perspectives on Language Deficits. Cambridge/MA: MIT Press Günther, Hartmut (1987), “Zur Repräsentation morphologisch komplexer Wörter im internen Lexikon”. In: Asbach-Schnitker, Brigitte & Roggenhofer, Johannes (Hrsg.), Studien zur Wortbildung und Historiographie der Linguistik. Festschrift für H. E. Brekle. Tübingen: Narr, 195⫺209
Günther, Hartmut (1989), “Experimentelle Morphologieforschung”. In: Günther (Hrsg.), 9⫺26 Günther, Hartmut (1989, Hrsg.), Experimentelle Untersuchungen zur deutschen Flexionsmorphologie. Hamburg: Buske Günther, Hartmut & Ludwig, Otto (1996, Hrsg.), Schrift und Schriftlichkeit: Ein interdisziplinäres Handbuch internationaler Forschung, Bd. II, Berlin et al.: de Gruyter Halle, Morris (1973), “Prolegomena to a Theory of Word Formation”. Linguistic Inquiry 4, 3⫺16 Hare, Mary & Elman, Jeff. L. & Daugherty, K. G. (1995), “Default Generalisation in Connectionist Networks”. Language and Cognitive Processes 10, 601⫺630 Henderson, Leslie (1982), Orthography and Word Recognition in Reading. London usw.: Academic Press Henderson, Leslie (1985), “Towards a Psychology of Morphemes. In: Ellis (Hrsg.), Bd. I, 15⫺72 Henderson, Leslie (1988), “Bericht über den Workshop ‘Linguistic and Psychological Approaches to Morphology’, Cambridge (1987)”. Linguistische Berichte 116, 357⫺363 Humphreys, Glyn W. & Evett, Lindsay J. (1985), “Are there Independent Lexical and Nonlexical Routes in Word Processing? An Evaluation of the Dual Route Theory of Reading.” Behavioural and Brain Sciences 8, 689⫺740 Inhoff, Albrecht W. & Rayner, Keith (1996), “Das Blickverhalten beim Lesen”. In: Günther & Ludwig (Hrsg.), 942⫺957 Jarvella, Robert J. & Snodgrass, John G. (1974), “Seeing Ring in Rang and Retain in Retention: On Recognizing Stem Morphemes in Printed ”. Journal of Verbal Learning and Verbal Behaviour 14, 590⫺598 Kean, Marie-Louise (1979), “Agrammatism: A Phonological Deficit?”. Cognition 7, 69⫺84 Laudanna, Alessandro & Burani, Cristina (1985), “Address Mechanisms to Decomposed Lexical Entries”. Linguistics 25, 575⫺592
162. Mentale Repräsentation morphologischer Strukturen Leuninger, Helen (1993), Reden ist Schweigen, Silber ist Gold ⫺ Gesammelte Versprecher. Zürich: Ammann
1777
Levelt, Willem J. M. (1989), Speaking: From Intention to Articulation. Cambridge/MA: MIT Press
Scarborough, D. L. & Cortese, C. & Scarborough, H. S. (1977), “Frequency and Repetition Effects in Lexical Memory”. Journal of Experimental Psychology: Human Perception and Performance 3, 1⫺17
Lima, Susan D. (1987), “Morphological Analysis in Sentence Reading”. Journal of Memory and Language 26, 84⫺99
Scheerer, Eckart (1992), Mentale Repräsentationen: Tutorial Paper. Bielefeld: ZiF, Arbeitsgruppe Brain & Mind
Marslen-Wilson, William & Komisarjevsky Tyler, Lorraine & Waksler, Tachelle & Older, Lianne (1994), “Meaning and the English Mental Lexicon”. Psychological Review 101, 1⫺32 Meringer, Rudolf & Mayer, Karl (1895), Versprechen und Verlesen: eine psychologisch-linguistische Studie. Stuttgart: Göschen (Nachdruck Amsterdam 1982) Monsell, Stephen (1985), “Repetition and the Lexicon”. In: Ellis (Hrsg.), 147⫺195 Motsch, Wolfgang (1988), “On Inactivity, Productivity and Analogy in Derivative Processes”. In: Motsch, Wolfgang (Hrsg.), The Contribution of Word-Structure-Theories to the Study of Word Formation (Linguistische Studien Reihe A 179), 1⫺30 Mugdan, Joachim (1977), Flexionsmorphologie und Psycholinguistik. Tübingen: Narr Olsen, Susan (1986), Wortbildung im Deutschen: Eine Einführung in die Theorie der Wortstruktur. Stuttgart: Kröner Paul, Hermann (51920), Prinzipien der Sprachgeschichte. Halle: Niemeyer [11880] Pylyshyn, Zeon Walter (1984), Computation and Cognition: Toward a Foundation for Cognitive Science. Cambridge/MA: MIT Rubin, Gary S. & Becker, C. A. & Freeman, R. H. (1979), “Morphological Structure and its Effect on Visual Word Recognition”. Journal of Verbal Learning and Verbal Behaviour 18, 757⫺767 Rumelhart, David E. & McClelland, James L. (1986), “On Learning the Past Tense of English Verbs”. In: Rumelhart, David E. & McClelland, James L. (Hrsg.), Parallel Distributed Processing. Cambridge/MA: MIT Press Sandra, Dominiek (1994), “The Morphology of the Mental Lexicon: Word Structure Viewed from a Psycholinguistic Perspective”. Language and Cognitive Processes 9, 227⫺269
Schreuder, Rob & Jarvella, Robert & Job, Raimo & Sandström, Göran (1989), “Morphologische Strukturen im mentalen Lexikon”. In: Günther (Hrsg.), 27⫺40 Slobin, Dan (1985, Hrsg.), The Crosslinguistic Study of Language Acquisition, Bd. 1⫺2. Hillsdale: LEA Smith, Philipp T. (1996), “Methods in Reading Research”. In: Günther & Ludwig (Hrsg.), 932⫺942 Smolensky, P. (1988), “On the Proper Treatment of Connectionism”. Behavioural and Brain Sciences 11, 1⫺74), Stanners, R. F. & Neiser, J. J. & Hernon, W. P. & Hall, R. (1979), “Memory Representation for Morphologically Related Words”. Journal of Verbal Learning and Verbal Behaviour 18, 399⫺412 Taft, Marcus (1981), “Prefix Stripping Revisited”. Journal of Verbal Learning and Verbal Behaviour 20, 289⫺297 Taft, Marcus (1988), “A Morphological Decomposition Model of Lexical Representation”. Linguistics 26, 657⫺668 Taft, Marcus (1991), Reading and the Mental Lexicon. London: LEA Taft, Marcus & Forster, Kenneth I. (1975), “Lexical Storage and Retrieval of Prefixed Words”. Journal of Verbal Learning and Verbal Behaviour 14, 638⫺647 Tanenhaus, Michael K. (1988), “Psycholinguistics: An Overview”. In: Newmeyer, Frederic (Hrsg.), Linguistics. The Cambridge Survey, Bd. III. Cambridge: Cambridge University Press, 1⫺37 Tyler, Lorraine K. & Marslen-Wilson, William & Rentoul, James & Hanney, Peter (1988), “Continuous and Discontinuous Access in Spoken Word Recognition: The Role of Derivational Prefixes”, Journal of Memory and Language 27, 368⫺381
Hartmut Günther, Köln (Deutschland)
1778
XIX. Psycholinguistische Perspektiven
163. Speech production and perception 1. 2. 3. 4. 5. 6. 7. 8.
1.
Introduction Processing models History of the question Important variables Sources of evidence Results of empirical studies Conclusion References
Introduction
Controversy surrounds the role of morphology in speech perception and production. The crucial issue is the following: do morphemes affect speech processing? That is, are words such as girls, boys, girlish, boyish, treated as if they were simple sequences of sound, or does their internal make-up affect the way in which they are produced or perceived? A common assumption is that items which are derived by a regular rule are (dis)assembled in the course of speech, in line with Chomsky and Halle’s claim (1968: 12) that “regular variations are not matters for the lexicon which should contain only idiosyncratic items”. The dissimilarity between regular and irregular forms has been confirmed by neural imaging (Jaeger et al. 1996; Marslen-Wilson & Tyler 1998). But a straightforward division between the two types is problematical, for two reasons. First, it is unclear which forms should be regarded as ‘regular’ and which ‘irregular’. There seems to be a cline, with regular items at one end, and irregular at the other. In the middle, a plethora of semi-regular forms are found (Bybee & Slobin 1982; Bybee & Moder 1983; Bybee 1995 a; Pinker 1999). They are governed by rules, but rules with a number of conditions attached. For example, English has a productive suffix -ness which forms abstract nouns from adjectives e.g. happiness, slowness. Yet if the adjective ends in [l], the preferred ending is -ity, not -ness e.g. mobility, futility. Such examples do not easily fit into either the category of ‘regular’ or ‘irregular’ behaviour. Second, common sense cannot decide the issue: there is a trading relationship between processing and storage which has to be discovered. Words ready-assembled in the mental lexicon would presumably allow faster processing, but would take up considerable
memory capacity. (Dis-)assembling words would slow down processing, but require less memory. Our current knowledge of the mind is insufficient to allow a priori judgments.
2.
Processing models
Three main processing models have been discussed intensively over the last quarter century ⫺ though this oversimplifies the numerous proposals found: (a) Whole-word look-up, often known as the full-listing hypothesis (e.g. Butterworth 1983). According to this model, speakers/ hearers look up whole words, which they routinely treat as indivisible sequences. Morphological affixes are firmly attached to their stems in the mental lexicon: girls, girlish, boys, boyish are four different lexical entries. Morphological divisions are potentially detectable by speakers/ hearers, but this analytic ability is not brought into use during normal speech processing. (b) On-the-spot (dis)-assemblage, sometimes referred to as the decompositional hypothesis (e.g. Taft et al. 1986). According to this model, speakers/hearers routinely (dis)-assemble complex words. Morphological affixes are detached from their stems, and morphological (de)composition is part of the normal processing procedure. The stems girl and boy constitute the main lexical entries, and morphological affixes are listed as sub-entries or additional notes. (c) Mixed processing (e.g. Laudanna & Burani 1985; Burani & Caramazza 1987). According to this model, processing can be carried out either via the full form, or via the component morphemes. Lexical entries have a double format, in which each word occurs both as a whole and in morphemes. A variant of this, which overlaps with (a), is whole-word look-up with automatic ‘word-parsing’: morphological affixes are attached to their stems, but morphological make-up is automatically accessed whenever the word is considered, as if, metaphorically, the lexical entry was stored in hyphenated form: girl-s, boy-s, girl-ish, boy-ish.
1779
163. Speech production and perception
These viewpoints all assume that morphological make-up is potentially available during processing. The critical question is the extent to which (de)composition is bound to occur. Schreuder & Baayen (1995) summarise similarities and differences between these and other newer variants, such as AAM (Augmented Addressed Morphology) model (Burani & Laudanna 1992), MRM (Morphological Race Model) (Baayen et al. 1997).
3.
of this dispute can be found in Butterworth 1983; Cutler 1983; Henderson 1985.) Early work tended to conflate evidence from different processes, different stages of processing, different media, different languages, different types of morphology, and different degrees of productivity. Later work on the topic showed increasing sophistication, as researchers refined their methodological tools, and showed that it may be unrealistic to rely on one global model for morphological processing.
History of the question
The role of morphology in speech processing has been discussed primarily by psychologists, since theoretical linguists have traditionally been more interested in a speaker’s ‘competence’ (internal system, or ‘representation’) than in ‘performance’ (how that system is used, or ‘access’) (Sandra 1994). The morphology question surfaced in psychology in the 1970s, largely as a result of a prefix-stripping claim made by Taft & Forster (1975). They discovered that readers rejected non-words such as *juvenate relatively slowly. This slowness, they suggested, might be because -juvenate is listed as an entry in the mental lexicon, though with an additional note specifying the obligatory attachment of re-. *Juvenate was therefore initially assumed by readers to be a word, and was eventually rejected only after the extra notes had been checked. Words, the researchers concluded, are normally stored as stems, with prefixes which have to be stripped off in comprehension. This claim was backed up with further experiments: ‘pseudo-prefixed’ words ⫺ words which look as if they have a prefix, but in fact do not, such as precipice ⫺ take extra processing time because readers mistakenly strip off the prefix, and look for a non-existent *cipice (Taft 1981). Later, the notion of stripping was extended to include suffixes also. The assumption that the mental lexicon contains primarily stems which can be accessed only after affixes have been stripped off was strongly maintained by some (e.g. Taft et al. 1986) ⫺ though Taft has now modified his views (Taft 1994). Affix-stripping claims were however disputed by others even from the start. According to its detractors, the experimental task may have induced special strategies, wordstress was ignored, and affix-stripping effects, if they exist, may involve at most only written words. (Useful summaries of the early years
4.
Important variables
Over the years, psycholinguists have become increasingly aware of the large number of variables which need to be considered. The following dozen are perhaps the most basic. They can be divided broadly into linguistic factors and processing factors: Linguistic factors (a) inflection vs. derivation vs. compounding (b) prefixes vs. suffixes vs. infixes vs. circumfixes (c) morphemes vs. syllables (d) isolation vs. polysynthesis (morphemes per word) (e) agglutination vs. fusion (segmentability) (f) bound vs. free forms Processing factors (g) (h) (i) (j) (k) (l)
perception vs. production speech vs. reading/writing representation vs. process access vs. selection productive vs. unproductive processes frequent vs. infrequent words
Tab. 163.1: Important variables
4.1. Linguistic factors (a) Inflection vs. derivation vs. compounding. In the literature, the main distinction discussed is that between inflection (e.g. drive-s) and derivation (e.g. driv-er) (cf. Art. 38). Compounding has received less attention. In general, inflectional morphemes provide extra information about existing words (e.g. The drivers loaded the vans with bananas) while derivational
1780
(b)
(c)
(d)
(e)
morphemes form new words from existing ones (e.g. The driver unloaded the perishable goods). This distinction is not necessarily always clearcut (e.g. English -ly as in quickly is a well-known borderline case), and in some languages the boundary is more blurred than others. But in languages such as English, differences of treatment are mostly clear: derivational morphemes precede inflectional ones, as in driv-er-s, not *drive-s-er; nois-i-est, not *nois-est-y (Aronoff 1976; Spencer 1991). In consequence, inflection and derivation must be analysed separately wherever possible. Prefixes vs. suffixes vs. infixes vs. circumfixes (cf. Art. 54⫺55). Affixes are traditionally divided into prefixes, suffixes, infixes, and (sometimes) circumfixes. In practice, almost all psycholinguistic work has related to prefixes (e.g. un-, dis- as in uncover, discover) and suffixes (e.g. -ful, -ous, as in harmful, perilous). Suffixes are considerably more common than prefixes in the languages of the world (Cutler et al. 1985; Hawkins & Cutler 1988), and beginnings of words are treated differently from endings (e.g. Marslen-Wilson 1989; Segui & Zubizaretta 1985). Morphemes vs. syllables. The syntactic unit ‘morpheme’ interacts with the phonological unit ‘syllable’. It sometimes coincides (e.g. care-less) and sometimes clashes (e.g. lead-er vs. lea-der). A further possibility is that morphemes may not have any status as ‘morphemes’, but may simply represent familiar phonic or orthographic chunks (Seidenberg & McClelland 1989). Isolation vs. polysynthesis. The number of morphemes per word varies from one (isolation, e.g. sing, seven) to several (polysynthesis, e.g. polyunsaturated: polyun-satur-ate-d). This variation occurs partially within a single language, but more importantly, across languages, where languages vary from low to high on an index of synthesis (Comrie 21989). Agglutination vs. fusion. The segmentability of morphemes varies from straightforward divisibility (agglutination, e.g. unwanted: un-want-ed) to complete interweaving (fusion, e.g. bit ‘bite.past’). This variation occurs both within and across languages, where languages score differently on an index of fusion (Comrie 21989).
XIX. Psycholinguistische Perspektiven
(f) Bound vs. free forms. Morphemes may be free (able to occur on their own as words, e.g. sing) or bound (obligatorily attached to another, e.g. -ed in wanted). This leads to degrees of affixation, e.g. good#ness has a suffix less tightly attached than erot⫹ic, since good is a word but *erot is not. Further complications are noted in Giegerich (1999), Plag (2003). 4.2. Processing factors (g) Perception vs. production. Speech recognition, both auditory and visual, is known to be partly a matter of informed guesswork: hearers/readers reach conclusions based on outline clues. Some morphological information may therefore be redundant, and less carefully attended to. But production requires attention to the full form. It is therefore unreasonable to extrapolate from one process to the other. (h) Speech vs. reading/writing. In speech, stress pattern and syllable structure are available to guide processing, whereas orthographic forms involve purely visual shapes. For example, the orthographic forms decide, decapitate, decontaminate, decorate all have a visual similarity for the first three letters, yet the spoken forms provide different, and possibly stronger, clues to their morphological make-up. The reverse happens with some inflections: for example, phonologically conditioned allomorphs of the English plural morpheme are spelled with an s, but pronounced either [s] (cats), [z] (dogs) or [iz] (horses). Discussions of morphology therefore need to be modality-specific (Chialant & Caramazza 1995). (i) Mental representation vs. process. Considerable discussion as to the general architecture of the ‘mental lexicon’ has taken place. Some researchers assume the existence of a single ‘master-file’ with word representations based on some abstract form sometimes referred to as a lemma, which must be altered in some way when used in speech (Marslen-Wilson 1989; Bock & Levelt 1994; MarslenWilson 1993; Marslen-Wilson & Zhou 1999). This lemma level (if it exists) is intermediate between lexical-semantic representations and modality-specific lexical representations. The morphemic representations within this hypothesized master-file must be distinguished from the ac-
163. Speech production and perception
tual forms used in speech processing. Others have argued that it is unnecessary to postulate a modality-neutral lemmalevel of representation (Caramazza 1997). (j) Access vs. selection. Recent discussions mostly assume that perception and production are two-stage processes. Lexical access involves multiple activation of lexical possibilities, followed by a narrowing down and final selection (e.g. MarslenWilson 1989; Aitchison 32003). Treatment of morphology may therefore alter, depending on the stage in progress. (k) Productive vs. unproductive processes. Productive processes which currently can be used to form new words (e.g. non- as in non-modular, non-slip) need to be distinguished from historical unproductive ones (e.g. in- as in impermeable, indecent). Productive morphemes are more easily attached/detached, and productivity also (usually) implies transparency and regularity (Bauer 2001). (l) Frequent vs. infrequent words. Morphemes within frequently used words may become welded together (e.g. would, should where the final -d was in origin a past tense marker). This may mean that similarly structured words are not necessarily processed in the same way. Bybee (1985; 1995 a; 1995 b) pointed out that words of high frequency are more likely to be stored whole, while regular words of lower frequency are more likely to be formed by combination. The issue of frequent vs. infrequent words also overlaps with that of familiar vs. unfamiliar words (Chialant & Caramazza 1995). The above variables indicate that morphological processing is a complex issue: numerous interacting factors cause complications even within a single language, and there is potentially considerable variation across languages.
5.
Sources of evidence
Broadly, evidence which might decide the issues can be divided into two types: experimental (preferred by psychologists) and naturalistic (favoured by linguists). Ideally, the results of the two types would coincide. 5.1. Experimental evidence The majority of experiments have probed speech perception, with either auditory or (more usually) visual stimuli. Experiments
1781 have typically involved the measurement of response times (‘response latencies’), on the assumption that if a morphologically complex word (e.g. bluish) takes longer to recognize than a simple word (e.g. yellow), then the complex word has probably involved some type of decomposition. A lexical decision task is a widely-used technique (“Tell me as quickly as possible if this is a word or not”). The subject responds either verbally, or by pressing a button. This task can arguably be made subtler by the use of priming (prior presentation of another word). In its simplest form, the prime-word is the same as the ‘target’ (word on which the decision is being made). Such repetition priming speeds up response to the target, suggesting that previous access to a word in the mental lexicon facilitates its later retrieval. Any strong facilitatory effect is usually regarded as evidence that the prime-word and target are closely connected, and may even share a lexical entry. It is therefore one proposed way of checking closeness between morphologically related pairs (e.g. creep and crept, deduce and induce). A common technique is to compare the time taken to recognize a repeated word (e.g. keep and keep) with that taken to recognize a word with an extra affix (e.g. keep and kept). Cross-modal priming is a further (controversial) refinement. It involves presenting the prime-word in a different modality from the target (e.g. an auditorily presented word followed by a visually presented one), on the (controversial) assumption that different modalities obligatorily involve accessing a neutral-between-modalities version of the word in the mental lexicon (Marslen-Wilson 1989). Masked priming is another variant, in which the prime-word is preceded by hash marks (#) equal to the number of letters in it, which supposedly makes memorization of the prime-word less likely (Forster & Davis 1984). But various doubts have been raised about priming, since facilitatory effects may simply be due to one-off strategies devised for the experimental task: priming is unable to distinguish between online access and post-access procedures, and similar or related words can in some cases have inhibitory effects, rather than facilitatory ones (Monsell 1985). Furthermore, it is often impossible to tell whether a delay in recognition is due to independent representation (e.g. keep and kept might have separate lexical entries) or de-
1782 composition (kept might need to be decomposed). Gating is a further variant of a basic lexical decision task. This involves chopping off the word at successive points, until a ‘recognition point’ is reached. Its value in morphological processing is that it taps online procedures, and can (supposedly) reveal which parts of the word are essential for identification. Overall, therefore, psycholinguistic experimentation has increased in sophistication over the years, though it is still something of a blunt instrument in that the results can often be interpreted in a variety of ways. 5.2. Naturalistic evidence These data involve deviations or variations mostly in production, which fall into a variety of categories. (a) Slips of the tongue. These are involuntary deviations from a speaker’s intended utterance. They are rule-governed, in that such errors show recurring patterns. Morphologically, their value rests on the assumption that morphological errors might represent assemblage which occurs in the course of speech production. e.g. concision for ‘precision’ might be taken as evidence that con- was mistakenly added to -cision instead of pre- ⫺ though care needs to be taken to eliminate alternative explanations (e.g. concision might be a blend of conciseness and precision, or might be a late example of the archaic word concision). A further problem is that it is not always possible to distinguish normal production routines from ‘back-up’ procedures, which come into play when the normal mechanisms break down. An error such as deduceful for ‘deductive’ could suggest that it is normal to assemble a derivative of deduce when required, or it might indicate that the speaker could not think of the ‘standard’ word, and so performed an emergency wordformation operation. (b) The errors of aphasics. These are mistakes made by patients suffering from serious speech disorders. They are often similar in nature to ordinary ‘slips of the tongue’, though tend to be more numerous and more bizarre. The link between aphasics and normal speakers was pointed out by Freud (1953: 13): “The paraphasia in aphasic patients does not dif-
XIX. Psycholinguistische Perspektiven
fer from the incorrect use and the distortion of words which the healthy person can observe in himself in states of fatigue and divided attention”. Researchers hope to find patients with selective impairments, with some types of morphological affixes omitted, but others retained. However, care must be taken to consider alternative explanations, especially as people with damaged brains may adopt unpredictable compensatory strategies. (c) Normal speech variation. Individuals are sometimes inconsistent in their treatment of inflection of the same lexical items. This is partly due to sociolinguistic factors, such as a shift from casual to formal styles, and partly uncertainty as to how to inflect particular words, as when obscurer interchanges with more obscure and more obscurer. Such inconsistency may reflect assemblage procedures. In conclusion, evidence on morphological processing is rarely clearcut, whether it is experimental or naturalistic. The data can often be interpreted in more than one way. It is therefore necessary to look for convergence: areas in which different types of evidence apparently collude to provide a plausible account. These two sources can sometimes be supplemented by the observations of theoretical linguists. However, care must be taken when using the latter, as the aim of many theoretical linguists is to describe phenomena in a maximally economical way: this may not be the way the mind works.
6.
Results of empirical studies
The majority of work on morphological processing has been carried out on English, so English will feature prominently in the account below, even though conclusions drawn from English cannot automatically be carried over to other languages. 6.1. Inflectional morphology English inflectional morphology is relatively straightforward, in that a fairly small number of inflectional suffixes are found, which are all easily detachable. Various clues suggest that these inflections are added on in the course of speech production: (a) Slips of the tongue involving phrasal verbs (verbs composed of a base form
163. Speech production and perception
(b)
(c)
(d)
(e)
(f)
(g)
and a separable particle). Examples such as point outed, wash upped instead of pointed out, washed up suggest that the basic lexical item is point out, wash up, and that the ending has been added on in the wrong place (Aitchison 1987; 32003). Correct morphophonological form on slips of the tongue such as: Take the freezes out of the steaker (steaks out of the freezer). Here the plural morpheme which was originally intended for steaks has been adapted to its new stem freeze. Ability of normal speakers to immediately use any new lexical item with appropriate inflections, e.g. they blanded out from the recently popularized verb bland out ‘to become conformist’. Ability of fluent aphasics to inflect nonsense words appropriately, e.g. She wikses a zen from me, with a pair of loyses or whatchemecallem (Butterworth & Howard 1987). Frequent failure of agrammatic aphasics to inflect words, except in the case of words which are more usually found in their inflected form (e.g. eyes). Inconsistent inflection of phrasal nouns such as mother-in-law, as in mother-inlaws (more usually mothers-in-law) suggests a lexical item mother-in-law, with inflection added on in the course of speech. Sociolinguistic variation between inflected and uninflected forms e.g. we see, we sees.
Overall, the evidence suggests that regular inflections are added as utterances are produced, though words which are more usually found in their inflected form may be fully listed with the inflection already added. A similar conclusion was reached by Chialant & Caramazza (1995). This is in line with arguments by Cutler et al. (1985) that the greater the syntactic role played by the suffix, the more likely it is that stem and affix will be separately processed. Relatively little information is available about the comprehension of spoken inflected forms, though decomposition is supported by an aphasic patient who was found to have a selective deficit for inflectional morphology (Tyler & Cobb 1987). Most of the information about written language relates to comprehension, where the evidence is consistent with decomposition. An early study showed that when subjects
1783 were briefly shown a visual representation of inflected forms, they tended to perceive them as two units (Gibson & Guinet 1971). Priming studies have shown that prior exposure to a regularly inflected form of a word (sings) speeds up recognition time of the base sing as much as prior presentation of the word itself, though this is not true of irregular inflections (sang) (Stanners et al. 1979). This suggests that the subject is indeed accessing the same word, and ignoring the suffix. However, questions have been raised about whether these effects are genuine morphemic ones, rather than a by-product of visual or semantic similarity. Some studies have therefore tried to tease out the effects of formal similarity (e.g. ribbon ⫺ rib), semantic similarity (e.g. pain ⫺ ache), and morphemic relatedness (e.g. sings ⫺ sing, sang ⫺ sing) (Henderson et al. 1984; Napps & Fowler 1987; Napps 1989). The results suggest that “Priming among morphemic relatives cannot be explained as the combined effects of semantic and formal priming ... morphemic priming is a separate dimension along which two words can be related” (Napps 1989: 736). Taken as a whole, evidence for (de)composition is fairly strong in the case of English inflectional suffixes, and further evidence is summarized in Cutler (1983). However, these English results are not necessarily generalizable to inflection in other languages ⫺ though Dutch and German appear to behave similarly (Jarvella & Meijers 1983). But in fusional languages, such as Serbo-Croat, inflectional forms appear to be fully listed: one inflected form seems to be at the centre of a constellation of related forms. A basic entry, the ‘nucleus’ has a cluster of ‘satellites’ around it (Lukatela et al. 1980; Lukatela et al. 1987; Feldman & Fowler 1987). This suggests that the English results may not be specific to inflectional morphology, but to the transparency and segmentability of inflectional morphemes in English, in particular to the fact that they are added to existing free forms. Inflectional prefixes (sometimes regarded as part of a circumfix) exist in Dutch and German, and the evidence here is for (de)composition (Jarvella & Meijers 1983). The prefix ge-, for example, is transparent and easily detachable. 6.2. Derivational morphology: prefixes Evidence for prefix-stripping in English derivational morphology is slim and unreliable. One early study of production found a
1784 number of apparently switched prefixes (e.g. advice for device) suggesting that prefixes are added on in the course of speech (Fay 1977). However, this study did not look at prefix maintenance, nor did it control for stress, or tease out the different types of prefixes. Another study which analysed speech errors involving words containing common initial sequences (e.g. con-, de-, in-, dis-) found that initial sequences were mostly maintained, and that almost half of the examples of apparent prefix switching were possibly blends e.g. dislayed for delayed in: The trains were disrupted and dislayed (Aitchison 1983⫺ 1984). A small residue was more plausibly interpreted as unstressed syllables which were ‘faint’ in memory, rather than detached prefixes (e.g. deferred for referred). In short, for all examples of prefix change “it is possible to suggest alternative, and arguably more plausible explanations than one which involves the notion of a mental lexicon with detached prefixes” (Aitchison 1983⫺1984:70). However, this study did not cover productive prefixes before word boundaries (e.g. un- as in unhappy) ⫺ though the rarity of reported errors such as *dishappy, *untasteful perhaps indicates that prefixes are firmly attached to words, even when they are combined with free lexemes, since the productive possibilities are wide-ranging. Attested examples of wrongly attached negatives such as: I disregard this as precise (I regard this as imprecise) do not result in non-words, so cannot be taken as firm evidence for on-the-spot attachment. In short, there is no evidence of prefix attachment for existing words ⫺ but plenty of evidence that it is possible to coin new words with the use of productive prefixes (e.g. recently coined unbundle ‘to sell subsidiary companies’). In spoken speech perception, Taft et al. (1986) claim to have found evidence of prefix stripping in a task which asked subjects to make lexical decisions in the case of various types of auditorily presented non-words: *dejoice (real prefix, real stem) vs. *tejoice (nonprefix, real stem) vs. *dejouse (real prefix, non-stem) vs. *tejouse (non-prefix, nonstem). They found that response latencies were slow in the case of real prefixes, and interpreted this as evidence that the prefix had been stripped, and the stem examined. Responses were slowest of all if the non-word contained a real prefix and a real stem, suggesting that the stem had been found in the mental lexicon, before realization dawned
XIX. Psycholinguistische Perspektiven
that the prefix-stem combination was non-existent. But this ingenious task indicates only that speakers try to make sense of a word in any way they can before discarding it: it possibly probes strategies for coping with oddities, rather than normal processing. A gating procedure (chopping off the word at various points until a ‘recognition point’ is reached) was used by Tyler et al. (1988) in a lexical decision task. They concluded that the presence of a prefix, whether stressed or unstressed, did not slow down processing: there was no evidence that lexical access was delayed until the stem could be identified, noting that words are simply recognized too early for the system to be waiting that long. The general conclusion therefore is that prefix (de)composition occurs in speech production and recognition primarily as a backup strategy, when the ‘normal’ processes have failed. Extensive literature exists on the topic of visual recognition of prefixes, spawned by the original Taft & Forster (1975) prefix-stripping paper. As already noted, this has been heavily criticized by various researchers (e.g. Henderson et al. 1984). In general, Taft & Forster’s claim that certain pseudo-prefixes (e.g. pre- from precipice) are stripped off is thought by many to be a task-induced strategy, or a post-access phenomenon. At the most, prefix stripping is a preliminary attempt to divide a visual word into chunks, rather than a standard decoding procedure. Relatively little information exists on the written production of prefixed words, though one puzzling study reported that a dysgraphic patient made more errors in writing non-affixed words than in productively-affixed ones (Badecker et al. 1990). Overall, prefix stripping in English appears to be unlikely, though may be used as an occasional reading strategy. Of course, English does not necessarily behave in the same way as other languages. Reports of reading experiments suggest that Italian, like English, has its prefixes attached, but that Dutch derivational prefixes may be detachable (Jarvella & Meijers 1983; Jarvella et al. 1987). 6.3. Derivational morphology: suffixes The unpredictability of English suffixes indicates that the majority are likely to be on fully listed words. Even productive and supposedly regular suffixes may show a high degree of unpredictability, as Henderson (1985)
1785
163. Speech production and perception
points out: he draws attention to agentive -er, and notes irregularities even in a small and circumscribed area such as manufacturers/ vendors (Henderson 1985: 39): (1) bake (V) butcher (V) fruit (N) hat (N) boot (N)
baker butcher fruiterer hatter ?
bakery (butchery?) ? ? ?
In addition, there are various ‘baseless derivatives’. For example, conflagration is apparently ‘baseless’, in that there is no stem conflagrate or -flagrate. Such examples count against the proposal that composition is a normal procedure. In speech production, slips of the tongue show relatively few ‘pure’ suffix errors. In one study, overall maintenance of the suffix was 81 Percent in a set of malapropisms (similar sound errors) which contained one of six common suffixes as target, error or both (e.g. cyclonic for ‘syphonic’, actual for ‘active’) (Aitchison 1987). Where suffix errors occurred, the malapropism usually involved other alterations as well (e.g. malicious for ‘malignant’, relative for ‘relevant’). The few examples of ‘pure’ suffix change could better be explained by procedures such as blending (e.g. audial for ‘auditory’, probably from ‘audio-visual’ or ‘auditory ⫹ visual’, contential for ‘contentious’, when the subject reported hovering between contentious and controversial) (vgl. Art. 91). An alternative explanation for some errors is a back-up procedure when the target is not wholly available: this is also a proposed explanation for a number of suffix errors such as orangine for ‘orangier’ made by a French aphasic (Pillon et al. 1991). According to the authors, the patient could only partially retrieve the desired phonological form, and produced errors which sometimes looked like derivational suffix errors, but were in fact provisional attempts to fill in the gaps in his partially retrieved form. Experimental evidence is somewhat inconclusive (Cutler 1983), though tends to support the suggestion that initial access does not necessarily involve decomposition: it takes no longer to recognize a complex word such as dusty than a monomorphemic one such as fancy (Manelis & Tharp 1977). But when subjects were presented with two words, of which one was suffixed and the other not (e.g. printer, slander), the task took longer than when the words both had suffixes, or both lacked them. This possibly indi-
cates that suffix-stripping is an optional strategy which can be utilized if necessary. A variety of studies have shown close relationships between the base word and derived forms, such as select and selection, destroy and destruction, (e.g. Stanners et al. 1979), but none have produced definitive evidence for (de)composition at an early stage. An authoritative survey (Henderson 1989) comments on priming studies that “findings suggest that priming by morphological relatives occurs at a stage beyond that of the word detectors” (Henderson 1989: 377). In all, evidence suggests that initial access involves whole-word look-up, but that wordformation and word-parsing routines are available if required. But full listing of words with derivational suffixes is not generalizable to all languages. In Turkish and other languages with a massive amount of productive agglutination, there must at least be some online parsing, though it may be misleading to refer to this as ‘suffix-stripping’ (Hankamer 1989).
7.
Conclusion
Three models were outlined at the beginning of this article: (a) whole-word look-up: the full-listing hypothesis; (b) on-the-spot (dis)assemblage: the decompositional hypothesis; (c) mixed processing. Empirical and experimental studies suggest that English inflectional suffixes are (de)composed in the course of speech, but that established complex words have their prefixes and suffixes firmly attached, though word parsing is an available back-up option which enables new words to be formed. The mixed processing model possibly represents a mixture of stages: whole-word look-up (initial access) may precede and overlap parsing and word formation. These findings are specific to English, and are not necessarily generalizable to other languages. However, two points at least need stressing about these conclusions. First, when morphological processing was first discussed, studies of the mental lexicon were in their infancy. Since then, it has become clear that the mind is capable of massive parallel processing (see summary in Aitchison 32003). It may be that considerably
1786
XIX. Psycholinguistische Perspektiven
more covert parsing activity is going on in parallel than we realise (cf. Bergman et al. 1988). Second, there is clearly an interrelationship between transparency and segmentability of morphemes, and productivity. It may be more revealing to look at the interaction between these than to look at a straightforward split between inflection and derivation (Hay 2002). Also an aphasic patient showed an inability to handle morphologically complex words involving suffixes that were phonologically and semantically transparent, whether inflected or derivational (Tyler et al. 1990). As researchers continue to work on an increasingly wide range of languages, and use increasingly sophisticated research tools, there is hope that many of the outstanding puzzles will soon be solved.
8.
References
Aitchison, Jean (1983⫺1984), “The Mental Representation of Prefixes”. Osmania Papers in Linguistics 9⫺10, 61⫺72 (Nirmala Memorial Volume) Aitchison, Jean (1987), “Reproductive Furniture and Extinguished Professors”. In: Steele, Ross & Treadgold, Terry (eds.), Language Topics: Essays in Honour of Michael Halliday. Amsterdam: Benjamins, 3⫺14 Aitchison, Jean (32003), Words in the Mind: An Introduction to the Mental Lexicon. Oxford: Blackwell Aronoff, Mark (1976), Word Formation in Generative Grammar. Cambridge/MA: MIT Press (Linguistic Inquiry Monograph 1) Baayen, R. Harald & Dijkstra, Ton & Schreuder, Robert (1997), “Singulars and Plurals in Dutch: Evidence for a Parallel Dual-Route Model”. Journal of Memory and Language 37, 94⫺117 Badecker, William & Hillis, Argye & Caramazza, Alfonso (1990), “Lexical Morphology and Its Role in the Writing Process: Evidence from a Case of Acquired Dysgraphia”. Cognition 35.3, 205⫺243 Bauer, Laurie (2001), Morphological Productivity. Cambridge: Cambridge University Press Bergman, M. W. & Hudson, P. T. W. & Eling, P. (1988), “How Simple Complex Words Can Be: Morphological Processing and Word Representations”. The Quarterly Journal of Experimental Psychology 40.A, 41⫺72 Bock, Kay & Levelt, Willem (1994), “Language Production: Grammatical Encoding”. In: Gerns-
bacher, Morton A. (ed.), Handbook of Psycholinguistics. San Diego/CA: Academic Press, 945⫺984 Burani, Cristina & Caramazza, Alfonso (1987), “Representation and Processing of Derived Words”. Language and Cognitive Processes 2, 217⫺227 Burani, Cristina & Laudanna, Alessandro (1992), “Units of Representation of Derived Words in the Lexicon”. In: Frost, Ram & Katz, Leonard (eds.), Orthography, Phonology, Morphology, and Meaning. Amsterdam: Elsevier, 361⫺376 Butterworth, Brian (1983), “Lexical Representation”. In: Butterworth, Brian (ed.), Language Production, Vol. II. London: Academic Press, 257⫺294 Butterworth, Brian & Howard, David (1987), “Paragrammatisms”. Cognition 26, 1⫺37 Bybee, Joan L. (1985), Morphology: A Study of the Relation between Meaning and Form. Amsterdam: Benjamins Bybee, Joan L. (1995 a), “Regular Morphology and the Lexicon”. Language and Cognitive Processes 10, 425⫺455 Bybee, Joan L. (1995 b), “Diachronic and Typological Properties of Morphology and Their Implications for Representation”. In: Feldman (ed.), 225⫺246 Bybee, Joan L. & Moder, Carol Lynn (1983), “Morphological Classes as Natural Categories”. Language 59, 251⫺270 Bybee, Joan L. & Slobin, Dan I. (1982), “Rules and Schemas in the Development and Use of the English Past Tense”. Language 58, 265⫺289 Caramazza, Alfonso (1997), “How Many Levels of Processing Are There in Lexical Access?”. Cognitive Neuropsychology 14.10, 177⫺208 Chialant, Dorian & Caramazza, Alfonso (1995), “Where is Morphology and How Is It Processed? The Case of Written Word Recognition”. In: Feldman (ed.), 55⫺76 Chomksy, Noam & Halle, Morris (1968), The Sound Pattern of English. New York: Harper and Row Comrie, Bernard (21989), Language Universals and Linguistic Typology. Oxford: Blackwell Cutler, Anne (1983), “Lexical Complexity and Sentence Processing”. In: Flores d’Arcais & Jarvella (eds.), 43⫺80 Cutler, Anne & Hawkins, John A. & Gilligan, Gary (1985), “The Suffixing Preference: a Processing Explanation”. Linguistics 23, 723⫺758 Ellis, Andrew W. (1985, ed.), Progress in the Psychology of Language, Vol. 1⫺2. London: Erlbaum
163. Speech production and perception Fay, D. (1977), “Prefix Errors”. Paper presented at the 4th Salzburg International Linguistics Meeting, August 1977 Feldman, Laurie B. (1995, ed.), Morphological Aspects of Language Processing. Hove: Erlbaum Feldman, Laurie B. & Fowler, Carol A. (1987), “The Inflected Noun System in Serbo-Croatian: Lexical Representation of Morphological Structure”. Memory and Cognition 15, 1⫺12 Flores d’Arcais, Giovanni B. & Jarvella, Robert J. (1983, eds.), The Process of Language Understanding. Chichester: Wiley Forster, Kenneth I. & Davis, Carol (1984), “Repetition Priming and Frequency Attenuation in Lexical Access”. Journal of Experimental Psychology: Language, Memory and Cognition 10, 680⫺698 Freud, Sigmund (1953), On Aphasia, translated by Erwin Stengel. New York: International University Press [11891] Gibson, Eleanor J. & Guinet, Lynn (1971), “Perception of Inflections in Brief Visual Presentations of Words”. Journal of Verbal Learning and Verbal Behavior 11, 182⫺189 Giegerich, Heinz J. (1999), Lexical Strata in English: Morphological Causes, Phonological Effects. Cambridge: Cambridge University Press Hankamer, Jorge (1989), “Morphological Parsing and the Lexicon”. In: Marslen-Wilson (ed.), 392⫺ 408 Hawkins, John A. & Cutler, Anne (1988), “Psycholinguistic Factors in Morphological Asymmetry”. In: Hawkins, John A. (ed.), Explaining Language Universals. Oxford: Blackwell, 280⫺317 Hay, Jennifer (2002), “From Speech Perception to Morphology: Affix Ordering Revisited”. Language 78, 527⫺555 Henderson, Leslie (1985), “Towards a Psychology of Morphemes”. In: Ellis (ed.), Vol. I, 15⫺72 Henderson, Leslie (1989), “On Mental Representation of Morphology and Its Diagnosis by Measures of Visual Access Speed”. In: Marslen-Wilson (ed.), 357⫺391 Henderson, Leslie & Wallis, Julie & Knight, Denise (1984), “Morphemic Structure and Lexical Access”. In: Bouma, Herman & Bouwhuis, Don G. (eds.), Attention and Performance, Vol. X. London, Hillsdale/NJ: Erlbaum, 211⫺226 Jaeger, Jeri J. & Lockwood, Alan H. & Kemmerer, David L. & Valin Jr., Robert D. van & Murphy, Brian W. & Khalak, Hanif G. (1996), “A Positron Emission Tomographic Study of Regular and Irregular Verb Morphology in English”. Language 72, 451⫺497
1787 Jarvella, R. & Job, R. & Sandström, G. & Schreuder, R. (1987), “Morphological Constraints on Word Recognition”. In: Allport, Alan & Mackay, Donald G. & Prinz, Wolfgang & Scheerer, E. (eds.), Language, Perception and Production: Relationships between Listening, Speaking, Reading and Writing. London: Academic Press, 245⫺262 Jarvella, Robert J. & Meijers, Guust (1983), “Recognizing Morphemes in Spoken Words: Some Evidence for a Stem-Organized Mental Lexicon”. In: Flores d’Arcais & Jarvella (eds.), 81⫺112 Laudanna, Alessandro & Burani, Cristina (1985), “Address Mechanisms to Decomposed Lexical Entries”. Linguistics 23, 775⫺792 Lukatela, Georgije & Gligorijevic, B. & Kostic, Aleksandar & Turvey, Michael T. (1980), “Representation of Inflected Nouns in the Internal Lexicon”. Memory and Cognition 8, 415⫺423 Lukatela, Georgije & Carello, Claudia & Turvey, Michael T. (1987), “Lexical Representation of Regular and Irregular Inflected Nouns”. Language and Cognitive Processes 2.1, 1⫺17 Manelis, Leon & Tharp, David A. (1977), “The Processing of Affixed Words”. Memory and Cognition 5, 190⫺195 Marslen-Wilson, William (1989, ed.), Lexical Representation and Process. Cambridge/MA: MIT Press Marslen-Wilson, William (1989), “Access and Integration: Projecting Sound onto Meaning”. In: Marslen-Wilson (ed.), 3⫺24 Marslen-Wilson, William (1993), “Issues of Process and Representation in Lexical Access”. In: Altmann, Gerry & Shillcock, Richard (eds.), Cognitive Models of Speech Processing: The Second Sperlonga Meeting. Hove: Erlbaum, 187⫺210 Marslen-Wilson, William & Tyler, Loraine K. (1998), “Rules, Representations, and the English Past Tense”. Cognitive Sciences 2.11, 428⫺435 Marslen-Wilson, William & Zhou, Xiaolin (1999), “Abstractness, Allomorphy, and Lexical Architecture”. Language and Cognitive Processes 14.4, 321⫺352 Monsell, Stephen (1985), “Repetition and the Lexicon”. In: Ellis (ed.), Vol. II, 147⫺195 Napps, Shirley E. (1989), “Morphemic Relationships in the Lexicon: Are They Distinct from Semantic and Formal Relationships?”. Memory and Cognition 17.6, 729⫺739 Napps, Shirley E. & Fowler, Carol A. (1987), “Formal Relationships among Words and the Organization of the Mental Lexicon”. Journal of Psycholinguistic Research 16, 257⫺272
1788
XIX. Psycholinguistische Perspektiven
Pillon, Agnesa & Partz, Marie-Pierre de & Raison, Anne-Marie & Seron, Xavier (1991), “L’orange, c’est le fruitier de l’orangine: A Case of Morphological Impairment?”. Language and Cognitive Processes 6, 137⫺167 Pinker, Steven (1999), Words and Rules: The Ingredients of Language. London: Weidenfeld & Nicolson Plag, Ingo (2003), Word-Formation in English. Cambridge: Cambridge University Press Sandra, Dominiek (1994), “The Morphology of the Mental Lexicon: Internal Word Structure Viewed from a Psycholinguistic Perspective”. Language and Cognitive Processing 9.3, 269⫺327 Schreuder, Robert & Baayen, R. Harald (1995), “Modelling Morphological Processing”. In: Feldman (ed.), 131⫺154 Segui, Juan & Zubizarreta, Maria-Luisa (1985), “Mental Representation of Morphologically Complex Words and Lexical Access”. Linguistics 23, 759⫺774 Seidenberg, Mark S. & McClelland, James L. (1989), “A Distributed, Developmental Model of Word Recognition and Naming”. Psychological Review 96, 523⫺568 Spencer, Andrew (1991), Morphological Theory. Oxford: Blackwell Stanners, Robert F. & Neiser, James J. & Painton, Scot (1979), “Memory Representation for Prefixed Words”. Journal of Verbal Learning and Verbal Behavior 18, 733⫺743
Taft, Marcus (1981), “Prefix Stripping Revisited”. Journal of Verbal Learning and Verbal Behavior 20, 289⫺297 Taft, Marcus (1994), “Interactive-Activation as a Framework for Understanding Morphological Processing”. In: Sandra, Dominiek & Taft, Marcus (eds.), Morphological Structure, Lexical Representation and Lexical Access. Hove: Erlbaum, 271⫺ 294 Taft, Marcus & Forster, Kenneth I. (1975), “Lexical Storage and Retrieval of Prefixed Words”. Journal of Verbal Learning and Verbal Behavior 15, 607⫺620 Taft, Marcus & Hambly, Gail & Kinoshita, Sachiko (1986), “Visual and Auditory Recognition of Prefixed Words”. Quarterly Journal of Experimental Psychology 38A, 351⫺366 Tyler, Lorraine K. & Behrens, Susan & Cobb, Howard & Marslen-Wilson, William (1990), “Processing Distinctions between Stems and Affixes: Evidence from a Non-Fluent Aphasic”. Cognition 36, 129⫺153 Tyler, Lorraine K. & Cobb, Howard (1987), “Processing Bound Grammatical Morphemes in Context: The Case of an Aphasic Patient”. Language and Cognitive Processes 2, 245⫺262 Tyler, Lorraine K. & Marslen-Wilson, William D. & Rentoul, James & Hanney, Peter (1988), “Continuous and Discontinuous Access in Spoken Word-Recognition: The Role of Derivational Prefixes”. Journal of Memory and Language 27, 368⫺ 381
Jean Aitchison, Oxford (Great Britain)
164. Speech errors 1. 2. 3. 4. 5. 6.
Introduction Speech errors and linguistic theory The structure of the lexicon Rules and accommodations Conclusion References
this grammar in speech production and comprehension (cf. Bierwisch 1982; Cutler 1988; Fromkin 1971; 1973, ed.; 1980, ed.; 1988; Garrett 1988; Stemberger 1982). Speech errors have been defined as involuntary deviations from intended utterances, as exemplified in the following (T ⫽ target; E ⫽ error):
1.
Introduction
(1) (a) Feature T: Cedars of Lebanon E: Cedars of Lemadon
A basic assumption of all linguistic speech errors research is that an analysis of these data can reveal something about both the mental representation of language knowledge ⫺ the grammar ⫺ and the ways in which we access
(b) Segment T: speech production E: preach seduction
1789
164. Speech errors
(c) Word T: tend to turn out E: turn to tend out (d) Stem T: The cost of flying E: The fly of costing (stem) (e) Inflectional suffix morpheme T: all the phones rang E: all the phone rangs (f) Derivational prefix T: imprecise E: disprecise (g) Syntactic constituents T: my sister went to the Grand Canyon E: the Grand Canyon went to my sister More recently, the deviant language errors produced by post-brain damaged aphasic patients have also been used as additional evidence for grammatical and production models (cf. Buckingham 1980; Dressler & Stark 1988, eds.). There are two major kinds of errors. The first category involves linguistic units, e.g. disordering, deletion, or addition of intended features, segments, morphemes, words, syntactic phrases, which show what units must be represented or derived in the grammar, as illustrated in examples (1 a⫺g) above. There are also errors which involve grammatical rules, for example, the application of a rule which should not apply as in (2): (2) T: I know where they are E: I know where they’re (wrong contraction) or failure to apply a rule which should be applied as in (3): (3) T: But when will you leave? E: But when you will leave? (Failure to invert subject and auxiliary) or a possible wrong application of a derivational or inflectional morphological rule resulting in non-occurring lexical items as in (4): (4) (a) T: E: (b) T: E:
he’s a New Yorker he’s a New Yorkan he swam in the pool he swimmed in the pool
Speech errors suggest that rules which have been posited as part of the grammar (i.e. as part of linguistic competence (Chomsky
1965)) may also be applied in linguistic performance. As will be discussed further below, some rules seem to be applied in order to prevent a proliferation of errors, referred to as accommodation (Fromkin 1971; Garrett 1980 b) illustrated in (5). (5) T: cow tracks (/s/ plural) E: track cows (/z/ plural) (Note that /kaws/ is phonologically wellformed as shown by the word mouse /maws/ but if the intended /s/ was not “changed” to /z/ the result would have been morphologically ill-formed). The majority of the examples are taken from the English UCLA corpus, but such errors occur in all languages obeying the same constraints.
2.
Speech errors and linguistic theory
The attempt to understand (and explain) the kinds of normal and aphasic speech errors that occur would not be possible without general linguistic concepts. All theories of grammar (however they may differ in detail, concepts or goals) posit a number of basic components: phonology, morphology, the lexicon, syntax, and semantics. Questions of interest concern the autonomy of these components, the ways in which they interact, their units and forms of their representation, and the rules or principles which constrain the wellformedness of structures. It is interesting that spontaneously produced speech errors reveal deviations in the units and rules of all these components, occurring at different stages of the production of an utterance (which parallel but may not be identical to the stages in a derivation of a sentence). Speech error data have proved insightful regarding various linguistic hypotheses concerning all aspects of the grammar; morphology ⫺ the structure of words and the organization of the lexicon is no exception. Morphology occupied a major role in pregenerative linguistics (cf. e.g. Bloomfield 1933; 1939; Marchand 1969; Nida 1949). In the earliest versions of generative grammar (Chomsky 1957), however, the grammar contained no morphological component and no lexicon, and all word-formation, including inflectional and derivational morphology and compounding, resulted from syntactic transformations. With the publication of Aspects of the Theory of Syntax (Chomsky 1965) a
1790
XIX. Psycholinguistische Perspektiven
lexicon was included as a separate component of the grammar but word formation (compounding and affixation) was still considered a syntactic process. With the extension of Chomsky’s (1970) lexicalist hypothesis by Jackendoff (1972), Halle’s (1973) seminal Prolegomena to a Theory of Word Formation, and Aronoff’s (1976) detailed model of word formation, morphology became an important area of study and theoretical development within the generative framework (Mohanan 1986; Kiparsky 1982; Siegel 1979; Pesetsky 1985; Anderson 1982; 1988). Psycholinguistic models have also been concerned with the processing and accessing of lexical information which in many cases require assumptions on morphological representation and rules. Some of the attempts to address these issues have included analyses of speech errors; what follows is a discussion of some of these studies.
3.
The structure of the lexicon
Traditionally, morphology concerned the structure of words and the rules of word formation. The nature of the lexicon itself was not part of morphology but rather the concern of lexicographers. With the inclusion of a lexicon as a component of the grammar, it was no longer possible to discuss morphological processes separate from the structure of the lexicon and the nature of lexical representation. Thus for example, one now needs to know whether lexical morphemes and/or words as well as derivational and inflectional morphemes are included in the lexicon, or in separate lexicons, whether complex words are derived by rule or listed in lexical paradigms, and how one should represent morphologically complex lexical entries. Analyses of speech errors have been used to address such questions. 3.1. Sub-lexicons Most psycholinguistic models of the lexicon partition entries into sub-components or modules (cf. Allport & Funnell 1981; Forster 1976; Fromkin 1985; Morton 1969). Thus, each lexical entry will have its phonological, orthographic, syntactic, and semantic representations in separate but intermeshed sublexicons. These particular sub-components have not been discussed in theories of morphology, although their inclusion does not seem to present any problems for any of the
current theories. Word substitution errors and blends provide some evidence for such partitioning, particularly for processing models. Although an isomorphism between a grammatical model and a psycholinguistic model is an open question, when such a parallel can be maintained the elegant solution is to do so. Examples (6 a) through (6 c) illustrate that in word substitution errors, the substituting words often fall into the same semantic class as the substituted words in that they share semantic features (including the polar feature values of antonyms). This may be due to their having semantic ‘addresses’ close to each other in the semantic sub-lexicon, or due to their being activated because of their shared features. (6) (a) T: E: (b) T: E: (c) T: E:
don’t burn your fingers don’t burn your toes my boss’s wife my boss’s husband a horse of another color a horse of another race
Examples (7 a⫺b) show that substituted words may instead be phonologically similar to the subtituting words, again suggesting either listings according to phonology in a phonological sub-lexicon or activation of words with shared phonology. (7) (a) T: E: (b) T: E:
white Anglo-Saxon protestant white Anglo-Saxon prostitute Universal City University City
It is interesting that word substitution errors made by aphasics and acquired dyslexics also show the same kind of semantic and/or phonological similarity (Fromkin 1987). The possible causal effect of phonological similarity is discussed by Butterworth (1981) in an analysis of word blends. 51 of 65 blends that he analyzed involved pairs of presumed words which are phonologically similar in shared phonemes, syllable structure, initial segment, or stress pattern as in the following: (8) (a) (b) (c) (d) (e)
trying/striving draft/breeze terrible/horrible grizzly/ghastly slick/slippery
> > > > >
strying dreeze herrible grastly slickery
Fay & Cutler (1977) in their study of malapropisms ⫺ word substitution errors in which the substituted word is phonologically but
164. Speech errors
not semantically similar to the intended word ⫺ conclude that: “(a) There is a single dictionary used for production and comprehension. (b) In this dictionary, words are arranged by phonemic structure, in a left-to-right manner, and based on a distinctive feature system. (c) The major partitioning of the dictionary, however, seems to be by number of syllables, with stress pattern as a second categorization within syllable categories. (d) Words may also be arranged by syntactic category.” (Fay & Cutler 1977: 515 f.)
Hurford (1981), after analyzing the same set of malapropisms, concluded instead that: “a. Malapropisms tend significantly to resemble their targets at both extremities of the lexical item ... b. Items listed in the mental lexicon are whole words, with suffixes already attached to stems.” (Hurford 1981: 423)
Hurford argues for this position on the basis of the similarity between target and error to the right of the first wrong segment which he concludes “is due … almost entirely to a coincidence of such derivational suffixes” (Hurford 1981: 420). There is, however, other evidence to suggest that in addition to whole words being listed in the lexicon (Halle 1973; Aronoff 1976), stems and/or morphemes must also be listed together with productive rules of word formation which may be incorrectly applied as illustrated by examples (4 a) and (4 b) above and the derivational errors shown in (9 a⫺b) and inflectional errors in (9 c⫺d). If the lexicon contained only actual words there would be no way to account for the possible but non-occurring words which are produced in speech errors. (9) (a) T: E: (b) T: E: (c) T: E: (d) T: E:
we often go to the movies we oftenly go to the movies motivate motify I already took a bath I already tooken a bath I meant to say I meaned to say
As Fay & Cutler (1977) and Hurford (1981) point out, there is strong evidence that words are listed in the lexicon according to various phonological criteria. This however does not contradict the need for also including a semantic sub-lexicon in which words are listed by semantic classes and/or features. There is also evidence for an orthographic sub-lexicon which is beyond the scope of this article.
1791 3.2. Two lexicons While the arguments for sub-lexicons revolve solely on processing considerations, there have been proposals for the bifurcation of words and/or morphemes into two separate lexicons based on grammatical and theoretical concerns (Stockwell et al. 1972). Golston (1991) presents evidence from speech errors in support of his claim that the grammar includes two lexicons, one consisting of content words and derivational affixes, and the second (which he calls a phrasicon) consisting of function words and inflectional affixes. In his model which he posits as being both a model of the grammar and a speech production processing model, he suggests that different kinds of speech errors occur at specific stages in production and further that other kinds of errors cannot occur. His model is similar to but not identical with Garrett’s (1980 a; 1980 b) model which also makes claims on the basis of the kinds of errors which occur at different stages or levels. There is indeed much evidence that the linguistic processing mechanisms differentiate between affixes and lexical content words and/or stems. Berg (1987) summarizes the distinction (cf. also Fromkin 1973, ed.; Garrett 1975; MacKay 1979; Nooteboom 1969): “Stems and affixes do not interact in tongue slips ... Prefixes and suffixes do not communicate with each other ... Inflectional and derivational suffixes, in turn, stay within their respective subcategorizations … Lexical morpheme errors respect word class membership to a lesser extent than do independent word errors.” (Berg 1987: 13 f.)
Golston points out that errors which consist solely of inappropriately selected function words can be accounted for by a two lexicon model in which the selection of such words occurs after the selection of content words as in the following errors from the UCLA corpus (Fromkin 1988): (10) (a) T: John’s going, isn’t he? E: John’s going, isn’t it? (b) T: That’s what we’re going to be doing today. E: That’s what we’re going to be do today. (c) T: I wouldn’t be surprised, would you? (d) E: I wouldn’t be surprised, do you? (e) E: when everyone had left T: when everyone was left
1792
XIX. Psycholinguistische Perspektiven
3.2.1. Word and stem errors Garrett (1980 b) points out that affixes and lexical stems exhibit different patterns in speech errors further supporting the notion of separate lexicons for each and specific stages at which errors involving each occur. Word stems (‘open class’ lexical content morphemes) participate in exchanges (spoonerisms or reversals) often leaving their inflectional morphemes behind (referred to as stranding) as shown in the following: (11) (a) T: E: (b) T: E: (c) T: E: (d) T: E: (e) T: E:
the cost of the purchase the purchase of the cost your keys are in my purse my keys are in your purse does Jack smoke? does smoke Jack? sweeping streets streeting sweeps wine racks rine (Rhine?) wacks
(Note that word exchanges do not always respect syntactic categories) If pronouns are included in the first lexicon (that containing lexical formatives, and derivational morphemes, but not grammatical words or inflectional morphemes) it is not surprising to find that pronouns are involved in exchanges such as in (11 b) above. 3.2.2. Derivational exchanges There are also errors involving exchanges of derivational affixes: (12) (a) T: E: (b) T: E: (c) T: E:
thinking and cognition thinktion and cogniting colder and windy coldy and winder bloody students bloodent studies (/studiz/)
There are however no attested errors involving the exchange of inflectional affixes (when they function as bound morphemes, not as phonological units such as syllables). That is, errors such as (13) do not seem to occur: (13) T: the girls are playing E: *the girling are plays 3.2.3. Stranding errors Errors such as (12 b) and (12 c) above are called stranding errors in that inflectional affixes are left behind or ‘stranded’ when the lexical stems exchange. Other examples are given in (14 a) to (14 c).
(14) (a) T: E: (b) T: E: (c) T: E:
rules of word-formation words of rule formation my check cashed my cash checked a sweeter flute a fluter sweet
Derivational morphemes rarely are stranded but such stranding errors do also occur as shown in (15). (15) T: return your call E: recall your turn Garrett’s model accounts for such stranding errors by suggesting that lexical stems are selected and inserted into syntactic planning structures or frames which contain these syntactic elements. 3.2.4. Shifts The shift or movement of affixes from the intended place in the utterance with attachment to another word is also observed, as in the following: (16) (a) T: E: (b) T: E: (c) T: E:
print outs prints out Larry Hyman’s paper Larry’s Hyman paper when someone comes up to me when someones come up to me
Again it should be noted that while derivational affixes seldom shift, grammatical words and minor grammatical categories (e.g. adverbs, intensifiers, determiners), like inflectional morphemes often do: (17) (a) T: E: (b) T: E: (c) T: E: (d) T: (e) E:
4.
I can hear her I can her hear he’s been around a long time he’s been a long around time Why do we have to? Why do have we to? I really hate to correct exams I hate to really correct exams
Rules and accommodations
It has often been pointed out that children’s errors during the acquisition stage ⫺ mouses for mice or bringed for brought for example ⫺ reveal more about their grammar construction and the rules they are forming than do their well formed utterances (cf. Art. 165). This is also true of adult speech as shown by the errors they produce, as in (18 a⫺e):
1793
164. Speech errors
(18) (a) T: E: (b) T: E: (c) T: E: (d) T: E: (e) T: E:
the last I knew about it the last I knowed about it he had to have it he haved to have it I meant to say I meaned to say Rosa always dated shrinks Rose always date shranks they’d be breastfed they’d be breastfeeded
Such possible forms (non-existent in the normal lexicon) suggest that a speaker may utilize regular inflectional rules applying them incorrectly at times to exceptional forms (although it is still possible that there may also be, in the normal situation, the selection of the correctly specified inflected form). Such rules also surface in the process known as accommodation described and illustrated in (5) above and in the following: (19) (a) T: I don’t know that I’d know one if I heard it E: I don’t know that I’d hear one if I knew it (b) T: an eating marathon E: a meating arathon (c) T: the museum that we saw E: the museum that saw us ⫺ we saw (d) T: I thought I was finishing your beer E: I thought you were finishing my beer In the Swedish example (20) the plural suffix -ar is changed to -er to accommodate to the morpholexical rules of the language.
(a) The lexicon appears to be subdivided into sub-lexicons, each containing a different set of representations. Thus a lexical entry like “coat” might be represented as in Tab. 164.1: Sub-Lexicon
Representation
phonological orthographic syntactic
/kot/ coat/COAT [N coat] ⫹ sub-categorization, etc. ‘coat’
semantic
Tab. 164.1: Sub-lexicons
(b) In the phonological sub-lexicon, words and/or morphemes appear to be listed according to initial phonemes, number of syllables, and stress patterns. (c) Lexical and grammatical morphemes, and derivational and inflectional affixes function differentially and at different stages in production. This argues for, at the very least, distinct specifications and possibly two separate lexicons which are accessed at different stages in production. (d) Both derivational and inflectional word formation rules exist which when misapplied result in possible but non-occurring lexical items. Speech error data cannot determine whether in addition to the application of such rules, derived and inflected words also occur in the one or two lexicons and their sub-components, although it appears that they do.
(20) Swedish (Linell 1982) T: tva˚ dagar i rad E: tva˚ rader i dag
Further cross-linguistic analyses of speech errors should make further contributions to questions concerning the nature and access of the lexicon and morphological representations and rules.
5.
6.
Conclusion
The discussion above aimed at showing how questions concerned with the nature and structure of the lexicon and morphological representation and rules have been investigated through an analysis of spontaneously produced errors of speech. While many questions remain to be answered and some only through normal linguistic analysis, performance data such as speech errors can contribute to our understanding of some of the issues. Just from those examined here we can draw some tentative conclusions.
References
Allport, D. A. & Funnell, E. (1981), “Components of the Mental Lexicon”. Philosophical Transactions of the Royal Society of London B 295, 183⫺196 Anderson, Stephen R. (1982), “Where’s Morphology”. Linguistic Inquiry 13, 571⫺612 Anderson, Stephen R. (1988), “Morphological Theory”. In: Newmeyer, Frederick (ed.), Linguistics: The Cambridge Survey I. Cambridge: Cambridge University Press, 146⫺191 Aronoff, Mark (1976), Word Formation in Generative Grammar. Cambridge/MA: MIT Press
1794 Berg, Thomas (1987), A Cross-linguistic Comparison of Slips of the Tongue. Bloomington, Indiana: Indiana University Linguistics Club Bierwisch, Manfred (1982), “Linguistics and Language Error”. In: Cutler, Anne (ed.), Slips of the Tongue and Language Production. Berlin: Mouton de Gruyter, 29⫺72 Bloomfield, Leonard (1933), Language. New York: Holt Bloomfield, Leonard (1939), “Menomini Morphophonemics”. Travaux du Cercle Linguistique de Prague, 105⫺115 Buckingham, Henry W. Jr. (1980), “On Correlating Aphasic Errors with Slips of the Tongue”. Applied Psycholinguistics 1, 199⫺220 Butterworth, Brian (1981), “Speech Errors: Old Data in Search of New Theories”. Linguistics 19, 627⫺662 Chomsky, Noam (1957), Syntactic Structures. The Hague: Mouton Chomsky, Noam (1965), Aspects of the Theory of Syntax. Cambridge/MA: MIT Press Chomsky, Noam (1970), “Remarks on Nominalization”. In: Jacobs, Roderick A. & Rosenbaum, Peter S. (eds.), Readings in English Transformational Grammar. Waltham/MA: Blaisdell, 184⫺221 Cutler, Anne (1988), “The Perfect Speech Error”. In: Hyman, Larry M. & Li, Charles N. (ed.), Language, Speech and Mind: Studies in Honour of Victoria A. Fromkin. London, New York: Routledge, 209⫺223 Dressler, Wolfgang U. & Stark, John A. (1988, eds.), Linguistic Analyses of Aphasic Language. New York: Springer-Verlag Fay, Donald & Cutler, Anne (1977), “Mala-propisms and the Structure of the Mental Lexicon”. Linguistic Inquiry 8, 505⫺520 Forster, Keith I. (1976), “Accessing the Mental Lexicon”. In: Wales, Roger J. & Walker, Edward (eds.), New Approaches to Language Mechanisms. Amsterdam, New York: North Holland Publishing Co., 257⫺288 Fromkin, Victoria A. (1971), “The Non-Anomalous Nature of Anomalous Utterances”. Language 47, 27⫺52 Fromkin, Victoria A. (1973, ed.), Speech Errors as Linguistic Evidence. The Hague: Mouton Fromkin, Victoria A. (1980, ed.), Errors in Linguistic Performance: Slips of the Tongue, Ear, Pen, and Hand. New York: Academic Press Fromkin, Victoria A. (1985), “Evidence in Linguistics”. In: Robins, Robert H. & Fromkin, Victoria
XIX. Psycholinguistische Perspektiven A. (eds.), Linguistics and Linguistic Evidence. Newcastle upon Tyne: Grevatt & Grevatt, 18⫺38 Fromkin, Victoria A. (1987), “The Lexicon: Evidence from Acquired Dyslexia”. Language 63, 1⫺ 22 Fromkin, Victoria A. (1988), “Grammatical Aspects of Speech Errors”. In: Newmeyer, Frederick J. (ed.), Linguistics: The Cambridge Survey II. Cambridge: Cambridge University Press, 117⫺138 Garrett, Merill F. (1975), “The Analysis of Sentence Production”. In: Bower, Gordon H. (ed.), The Psychology of Learning and Motivation, Vol. IX. New York: Academic Press, 133⫺177 Garrett, Merill F. (1980 a), “The Limits of Accommodation”. In: Fromkin, Victoria A. (ed.), Errors in Linguistic Performance. New York: Academic Press, 263⫺271 Garrett, Merill F. (1980 b), “Levels of Processing in Sentence Production”. In: Butterwort, Brian (ed.), Language Production, Vol. I. New York: Academic Press, 177⫺220 Garrett, Merill F. (1988), “Processes in Language Production”. In: Newmeyer, Frederick J. (ed.), Linguistics: The Cambridge Survey III. Cambridge: Cambridge University Press, 69⫺96 Golston, Chris (1991), Both Lexicons. Ph.D. dissertation, UCLA Halle, Morris (1973), “Prolegomena to a Theory of Word-Formation”. Linguistic Inquiry 4, 3⫺16 Hurford, Jim R. (1981), “Malaproprisms, Left-toRight Listing, and Lexicalisms”. Linguistic Inquiry 12, 419⫺423 Jackendoff, Ray (1972), Semantic Interpretation in Generative Grammar. Cambridge/MA: MIT Press Kiparsky, Paul (1982), Explanation in Phonology. Dordrecht: Foris Linell, Per (1982), “Speech Errors and the Grammatical Planning of Utterances”. In: Koch, Wolfgang & Platzack, Christer & Tottie, Gunnel (eds.), Textstrategier i tal och skrift. Stockholm: Almqvist & Wiksell International, 134⫺151 MacKay, Donald G. (1979), “Lexical Insertion, Inflection, and Derivation: Creative Processes in Word Production”. Journal of Psycholinguistic Research 8, 477⫺498 Marchand, Hans (1969), The Categories and Types of Present-Day English Word-Formation. Munich: Beck Mohanan, Karuvannur (1986), The Theory of Lexical Phonology. Dordrecht: Reidel Morton, John (1969), “The Interaction of Information in Word Recognition”. Psychological Review 76, 165⫺178
165. First language acquisition Nida, Eugene (1949), Morphology: The Descriptive Analysis of Words. Ann Arbor: University of Michigan Press Nooteboom, Sieb G. (1969), “The Tongue Slips into Patterns”. In: Sciarone, A. G. & Essen, A. J. van & Raad, A. A. van (eds.), Nomen Society: Leyden Studies in Linguistics and Phonetics. The Hague: Mouton, 114⫺132 Pesetsky, David (1985), “Morphology and Logical Form”. Linguistic Inquiry 16, 193⫺246
1795 Siegel, Dorothy (1979), Topics in English Morphology. New York: Garland Stemberger, Joseph Paul (1982), “The Nature of Segments in the Lexicon: Evidence from Speech Errors”. Lingua 56, 235⫺259 Stockwell, Robert P. & Schachter, Paul & Partee, Barbara H. (1972), The Major Syntactic Structures of English. New York: Holt, Rinehart & Winston, Inc.
Victoria A. Fromkin †, Los Angeles (U.S.A.)
165. First language acquisition 1. 2. 3. 4. 5.
Acquiring morphology Inflection Derivation Summary References
1.
Acquiring morphology
Morphology has been studied for several different reasons by researchers interested in first language acquisition. First, the acquisition of inflections has been used to assess children’s stage of language development. It is well established that, across languages, children at first make no productive use of closed class items like clitic pronouns, particles, and adpositions or of inflections marking grammatical categories such as tense, aspect, case, number, or gender. They then begin to add inflections and other grammatical markers, and their words become more complex (measured in morphemes), so that their utterances are longer. Computing the average length of children’s utterances in morphemes yields a measure of early grammatical development that is more sensitive than length-inwords (e.g., Brown 1973; Dromi & Berman 1982). Morphological development has also provided an important domain for testing hypotheses about how children learn. For example, do they learn inflections by rote, one word at a time, or do they abstract over words to come up with generalisations in the form of schemas or rules to apply to unfamiliar instances? If children can add past tense inflections to unfamiliar verbs, they must be relying on some more general principle than
rote learning. This finding is even more robust if children can be shown to inflect nonsense words, items that are non-existent in the established lexicon of their language. Such generalisations can be expressed as rules or schemas; they emerge early in the process of acquisition and show up in both spontaneous over-regularisations and elicited productions (e.g., Berko 1958; Bybee & Slobin 1982). The acquisition of morphology also offers a testing ground for crosslinguistic comparisons where some of the contributions of typology can be assessed. Do children learn morphological rules or schemas in the same way, with the same ease, in synthetic, agglutinating, and isolating languages? What aspects of typology make for error-free versus error-full paths during acquisition? And what effect does typology have on the acquisition of specific grammatical systems such as tense-aspect or passivisation on the verb, or accusative versus ergative case-marking on the noun? (See, e.g., Berman 1990; Pye 1990; Slobin 1985, ed.; 1992, ed.; 1997, ed.) Finally, the internal structure of words can be connected to conceptual and semantic development on the one hand, and to vocabulary acquisition on the other. When do children begin to analyse words into component parts? When do they start to make active use of derivational morphology in coining words? Are children sensitive to the productivity of the options available when they construct new words? These are just a few of the questions researchers have explored for derivational morphology and word formation more generally.
1796 When children begin to talk, they typically produce one word at a time. These single words may consist of bare stems only, as in English, where the stem is normally a word too; or they may be “frozen” inflected forms. For example, a child may always produce a particular noun with the same case ending on it. Or they may always produce a verb in one particular form with tense, person, number, and gender marking (e.g., Berman & ArmonLotem 1996). These initially invariant forms of words and their subsequent treatment by children discovering the relations of stems to inflectional affixes show which elements children focus on in the speech they hear. They consistently attend to stressed syllables and final syllables, for example, and tend to produce those first. Only later do they add other elements to their early words. It takes time for children to identify words or stems and their modulations ⫺ the different additions licensed by each word or word-type in a paradigm (Anderson 1985 a; 1985 b). Identification and analysis represent the first step (Peters & Menn 1993; Veneziano & Sinclair 2000). Only once that has been taken can children be credited with knowledge about the forms and meanings of inflectional paradigms or derivational patterns. And only then is it possible for researchers to generalise about children’s acquisition of such inflectional systems as case, number, gender, or tense. A major source of information about children’s morphological development consists of the errors they make en route to the adult system (see also Art. 164). The errors may be errors of omission, where children fail to add the expected inflectional or derivational morpheme to a word-stem. These errors offer indirect information about what children at each stage find difficult and therefore omit from their utterances. Other errors are errors of commission. These provide a more direct source of evidence about how children are organising their morphological paradigms. When children regularise specific wordforms, they reveal both how they have analysed and how they have organised the relevant inflections or options for word formation so far. In what follows, we look first at some of the findings for inflectional morphology (cf. 2) and then at derivational morphology (cf. 3) in acquisition.
XIX. Psycholinguistische Perspektiven
2.
Inflection
Children begin to add their first inflections and to use them in contrasting contexts sometime between 1;6 (one year, six months) and 2;0 (two years). By 2;6 to 3;0, they usually control some case endings, number marking on nouns and verbs, some person marking on verbs as well as some marking of aspect and tense, as well as gender and classifier systems in languages that mark nouns (and their associated modifiers) for these categories (see Art. 97⫺98). Just which inflections children have acquired by age three varies to a large extent with the typology of the language and with the complexity of the paradigms to be learnt. Cognitive complexity is one major pace-setter in acquisition (children must have established a conceptual category before they look for a means for talking about it), but formal complexity also plays an important role in determining when children master the forms available (Slobin 1973). If a language marks three genders as well as number and case on nouns, with different patterns of case endings for each gender and number, the inflectional system will take longer to learn than a language that marks neither gender nor case. Inflections that combine several elements of meaning (e.g., the Hebrew suffix -ot for plural number and feminine gender together) are acquired later than those that bear a clear one-to-one correspondence to a specific grammatical category (e.g., the Hebrew suffix -a for feminine gender in the default instance of singular number). Our discussion focusses on questions of complexity in the acquisition of inflections (2.1), rote learning versus rule application (2.2), agreement marking (2.3), and the linear ordering of morphemes (2.4). 2.1. Added complexity When children begin to add inflections, they modulate the meanings of the bases or stems the inflections are added to. These modulations mark additions to the basic or core lexical meanings of the pertinent terms. In general, “complexity in thought is reflected in complexity of expression” (Clark & Clark 1977: 523; Greenberg 1966). Each inflectional modulation adds complexity, whether in the forms of noun affixes to mark number and case, say, or in the forms of verb affixes to mark aspect, tense, person, and number (Bybee 1985). The added morphemes will and -ed explicitly mark added complexity of meaning
165. First language acquisition
in the contrasts between sleep and will sleep, or jump and jumped, just as the plural morpheme -s marks the added complexity of ‘more than one’ over ‘one’ in cats versus cat or the ordinal -th marks ordinal rank versus cardinality in seventh versus seven. But added complexity of meaning is not always visible morphologically, so the various possibilities are: full form/meaning transparency where a given modulation is marked by a single additional morpheme (e.g., the English plural -s on nouns or the regular past tense on verbs); added complexity of meaning with no obvious morphological addition (e.g., the English plural noun sheep or the past tense form cut); or more than one modulation of meaning packed into a single morpheme (e.g., the Hebrew feminine plural on nouns, or Italian first person imperfect past on verbs). The general prediction for acquisition, then, is that the more transparent the modulation, the earlier it should be acquired by children. Relative semantic complexity appears to account quite closely for order of acquisition in those domains where it is possible to make the relevant comparisons. It accounts for the order in which children acquire verb inflections in English (Brown 1973; see also Berman 1985 and Levy 1983 on Hebrew). But it is difficult to apply such a measure since few inflectional domains have been carefully analysed with respect to meaning, and even fewer have been compared across languages. Complexity of meaning is not the only kind of complexity children encounter during the acquisition of inflections. They must also master complexities of form ⫺ the number and kinds of changes required to mark a particular meaning in a language. Complexity of form can be analysed from different points of view. One depends on which categories are morphologically marked in a language ⫺ case, gender, tense, number, aspect, negation, and so on (Anderson 1985 a). A second factor that may increase formal complexity is the way the available structures map the relevant meanings ⫺ the number of different elements (compare the noun case systems of Hungarian and German, say); the number of case forms that carry more than one meaning within the singular or plural paradigm; or the morphophonemic rules and the extent to which the constituent morphemes in a word form remain readily analysable in the surface of each word. In Finnish, for instance, morpheme boundaries within words are often obscured,
1797 whereas in English they often remain visible. A third factor is the consistency of the formto-meaning mapping across semantic domains (Slobin 1985, ed.; 1992, ed.; 1997, ed.). Languages may differ greatly in these respects so that parts of the inflectional morphology within a language may be hard for children to acquire (e.g., noun plurals in Arabic, locative marking in Serbo-Croatian) while other inflectional paradigms may be much easier (e.g., plurals in English, locative marking in Hungarian) (Mikes 1967; Omar 1973; Ple´h 1998; Slobin 1973). Complexity of meaning and of form must be taken into account when children are acquiring morphology. 2.2. Rote or rule? Do children learn inflections for each word one by one? Or do they make generalisations based on some number of examples that they then apply to unfamiliar forms? That is, do they learn the form each member of every paradigm can have, and add each one to the paradigm as it is acquired? Such a task would demand rote-learning on a massive scale for a language like Turkish, for example, where the possible forms for a stem may be counted in the thousands. The alternative to such rote learning ⫺ where children could only learn the specific forms they happened to be exposed to ⫺ is that they analyse some sets of forms and extract from them systematic schemas that they can apply to any candidate stem to construct the desired form. One of the first studies to tackle this question was carried out by Berko (1958). She presented five- and seven-year-old children with nonsense words used in specific syntactic contexts, followed by an elicitation frame, e.g., “This man likes to biff every day. Yesterday, he __”, or “This is a wug. Now there are two of them. There are two __”, accompanied by line drawings of the scenes or objects supposedly denoted by the nonsense words. Children were asked in this way to supply several inflections including the plural and possessive on nouns, and -ing and -ed on verbs. The results showed that children could supply appropriate inflections for ‘nouns’ and ‘verbs’ they had never heard before. They could not, therefore, be relying on any rote learning of the forms for such words. Rather, they must have been using some type of rule or schema to supply the requisite ending to express a specific meaning (Bybee & Slobin 1982). This finding is also consistent
1798 with children’s spontaneous errors, where production of such regularised forms as English runned (for ran) or foots (for feet) also give strong evidence for rule application and against simple rote learning. At the same time, children are also influenced by the familiarity of the words they are asked to inflect in elicitation tasks. They more readily add inflections to familiar than to unfamiliar words. In Berko’s study, for instance, children succeeded in adding the appropriate allomorph of the plural /-ız/ to the noun glass, for glasses, but failed to do so with nonsense words ending in sibilants. In addition, children do better in elicitation tasks with nonsense words, when the situation itself calls for an inflected form (because it is pragmatically appropriate) than when the setting appears more test-like. So a child may fail to supply a plural inflection when asked about wugs (“Now there are two. There are two __”) but succeed when asked which pictures should be put away next (“Which pictures go next?”). Pragmatically appropriate tasks tend to offer a fuller picture of children’s abilities than test settings can (Hecht 1983; see also Menn & Bernstein Ratner 2000). As children learn the regular and irregular patterns of inflection in their language, they are tacitly building paradigms ⫺ sets of modulations that apply to particular classes of nouns or verbs. When they begin to identify such patterns, they must keep track of which affix-type goes on which stem-type. For example, in the Romance languages, children must classify verbs by conjugation in order to know which inflections apply. One might imagine that the most frequently heard forms would offer the first model for children, so if certain irregular verbs were the most frequent ones they heard, children should begin by generalising irregular forms. However, children attend not only to token frequencies but also to type frequencies, and their earliest generalisations appear to be based on the most frequent types (Guillaume 1927). In Romance, children consistently begin by choosing the first conjugation as their model when they add inflections to a verb, and so make errors of over-regularisation on verbs from other conjugations, as well on irregular verbs (Clark 1985). And in Semitic, children have to attend to the consonantal structure of the root in order to know which allomorph of an inflection to choose (e.g., Berman 1985).
XIX. Psycholinguistische Perspektiven
Children must also store in memory what they hear from other speakers. This enables them to recognise words and phrases when they hear them on subsequent occasions and from different speakers. As a result, they store many conventional forms for comprehension that they themselves may not yet produce. The fact that they have some stored representation of such forms means that these are available when, for instance, children are asked questions about irregular verb forms. In fact, children aged three and older are willing to make judgements about who might say a particular form, adult or child, when presented with bringed versus brought. They could have both forms stored since they hear language spoken by their peers (users of bringed) as well as by adults (brought). But they also hear conventional brought more often and from many more speakers. Child judgements about which form is “okay” at a certain point begin to favour adult over child usage (Kuczaj 1978; Platt & MacWhinney 1983). A major issue here, as in other areas of language learning, is how children come to give up their own over-regularisations in favour of the conventional irregular forms (Bowerman 1988; Clark 1993; Clark & Chouinard 2000). Frequency of exposure is one factor here. Since children hear conventional forms more frequently than any non-conventional ones, from the largest number of speakers, those should be the forms that they judge as more adult-like and that they themselves will eventually favour in production over earlier erroneous forms. Indeed, as children get older, it has been argued, their regularisation errors may largely reflect problems they are having in retrieving the appropriate (conventional) form from memory: A child who produces buyed in lieu of bought has failed to retrieve bought and so constructs a past tense form using the regular morpheme -ed (see Marcus et al. 1992). Speakers could store irregular forms in memory ⫺ these must be learnt by rote ⫺ but construct all regular forms when they are needed on the basis of an abstract rule or schema that captures the generalisation, (for English, here) that the past tense is marked by the addition of the morpheme -ed. However, other processing considerations have led researchers to argue for storage of some (not necessarily all) regular forms as well as irregular ones in addition to being able to construct regular forms by rule, rather than proposing two quite distinct
165. First language acquisition
types of processing with separate domains (cf. Frauenfelder & Schreuder 1991; Maratsos 2000). 2.3. Agreement One major function of inflectional morphology is to mark agreement in a language, to indicate which elements (or groups of elements) belong together and which play a specific grammatical role. For example, gender marking on nouns, articles, and adjectives can indicate which of these elements belong together within an utterance. Case marking on nouns typically distinguishes the grammatical functions of noun phrases as subject, direct object, or indirect (dative) object of the verb. And number agreement on nouns and verbs, along with person agreement, can also indicate which noun phrase is the subject. Languages differ considerably in the extent to which they rely on inflections for these purposes, and hence in the patterns of agreement they exhibit. Acquisition of the ways inflections function syntactically in marking grammatical concord has been rather less studied than the acquisition of isolated forms. Available data suggest that children begin working on these systems (e.g., articlenoun agreement in French or German, subject-verb and noun-adjective agreement in Hebrew) before age two though children may make many errors both of omission and commission on the way to adult mastery. In case-marked languages, children’s earliest words are case-marked, but they do not vary in form from one context to the next. In other words, the case inflection is treated initially as if it were part of the word. Very soon, however, children begin to contrast two (or more) case-markings, and to use them consistently. In Russian for example, children begin to use accusative marking on nouns following transitive verbs before age two (Gvozdev 1961). Nominative and accusative marking are usually the first two cases to be distinguished (MacWhinney 1978; Smoczynska 1985). Early case marking often interacts with word order, another device for marking grammatical relations in addition to pragmatic discourse structure. But although it is possible that children might rely on word order to mark grammatical relations prior to acquisition of case-marking, the data so far suggest that case marking emerges at least as early as consistent word order (Slobin & Bever 1982).
1799 Case on its own, though, does not provide evidence for agreement. What is needed is subject-case combined with an appropriate form of the verb, agreeing in person and number. This form of agreement, like others, begins to appear between age two and three (e.g., Clahsen 1986; Meisel 1986; Tracy 1984; Vihman 1999, ed.). In Hebrew, for example, children make a number of errors in nounadjective agreement in the first few months after age two, and often produce masculine singular adjectives with feminine singular nouns. These agreement errors then decrease in number and have typically vanished by age three. What appears critical in the mastery of agreement here is children’s prior acquisition of plural inflections for nouns. Once those are in place, Hebrew-speaking children can use the same plural ending on noun and adjective to mark agreement, e.g., ha-xaruz-im ha-gdol-im nofl-im ‘the-bead-pl the-big-pl fall-pl (the big beads are falling)’ with a masculine plural head noun, or ha-kubiy-ot hagdol-ot nofl-ot ‘the-cube-pl the-big-pl fall-pl (the big cubes are falling)’, with a feminine plural head noun. That is, children rely on the phonological similarity of the inflections across word classes in marking agreement (Berman 1985; Levy 1983). Children learning other languages also seem to rely on phonological cues to gender and number marking where possible. In Bantu languages like Sesotho, children must learn that every noun belongs in one of 14 classes, each marked by a prefix. Class prefixes (some of them identical in form to the noun prefixes) must also be added to verbs for both subject and object agreement, to adjectives, and to demonstrative, personal, possessive, and relative pronouns. Noun class markers emerge rather slowly from age two on, with both omissions and some over-use of class 9/10 prefixes (probably because many nouns used at this age come from this class). Adjectives and demonstrative, possessive, and personal pronouns are in use by age two, but often appear with the wrong class marker. By age three, though, children mark agreement with considerable consistency and accuracy. They are aided in this, it has been suggested, by the phonological similarity of class markers to the markers for each noun class (Demuth 1988). In Romance languages like French, children again make some errors in article-noun agreement once they begin to produce articles, but by age three such errors no longer
1800 occur. It takes French-speaking children slightly longer to master agreement for gender on adjectives and on pronouns. But by about age six, they are consistent in making articles and nouns agree, regardless of the natural gender of the referents. They appear to rely primarily on phonological correlations of word form (in the noun) and article in gender agreement (Clark 1985; KarmiloffSmith 1979). Reliance on phonological rather than semantic cues also seems to be the norm for Slavic (Gvozdev 1961; Smoczynska 1985) and for German (Mills 1985). Use of inflections to mark agreement emerges as soon as children begin to produce the pertinent morphology and the relevant syntactic combinations (i.e., subject-verb, article-noun, noun-adjective, etc.). This suggests that they have already represented those parts of the system in memory, and have done so on the basis of the language spoken to them in the community. It also suggests that children do not first use word order, for example, to mark grammatical relations and then switch to inflections. Rather, they will use the agreement patterns conventional in the language they hear, from the start. They may mark them imperfectly, but as they produce more of the relevant inflections, these appear to emerge without significant error. In fact, reliance on case versus word order as clues to grammatical relations varies with the typology of the target language (see Bates et al. 1984; Slobin 1985, ed.; 1992, ed.; 1997, ed.; Slobin & Bever 1982). 2.4. Linear ordering of morphemes Children appear to make no errors in the ordering of bound morphemes within words. When they add inflections to a base or stem, they do so in the right order. At first they add only single inflections, so the only error possible would be to add suffixal inflections as prefixes or prefixal ones as suffixes. But children apparently never make this type of error. Suffixes are added to the end, prefixes to the beginning, in the appropriate order. The only morpheme order error attested for English occurs in a setting that adult speakers tend to avoid: the formation of the comparative of adverbial forms like prickly or dirtily. Instead of the adult solution of using more, as in more prickly, young English-speaking children sometimes try out forms like prickerly where they add the comparative degree marker before the adverbial -ly, and thus appear to order an inflection be-
XIX. Psycholinguistische Perspektiven
fore a derivational morpheme. (However, there is some dispute about whether degree -er is an inflection at all.) Many languages license strings of inflectional morphemes, on the end or the beginning of the stem form. Here children have many more opportunities for errors in order. Yet again, studies of agglutinating and polysynthetic languages such as Turkish, Mohawk, K’iche Maya, Greenlandic Eskimo, and Walpiri report no instances of order errors (cf. Aksu-Koc¸ & Slobin 1985, ed.; Mithun 1988; Slobin 1985, ed.; 1992, ed.; 1997, ed.). This striking absence of error within words is consistent with the operating principle: “Keep the order of morphemes within a word constant across the various environments in which that word can occur” (Slobin 1985: 1231). Children’s adherence to morpheme order within words contrasts to some extent with the freedom they use in ordering their words. Use of word order to mark grammatical relations, for example, emerges only slowly. Initially, one- and even two-yearolds typically use word order as a pragmatic device to mark information as given versus new, and only later learn that word order may, in some languages, mark such relations as ‘subject-of’ (e.g., Bates 1976). Although children rarely make errors in their ordering of constituents after age three to four, they may still make word-order errors within constituents, as when they combine a quantifier like English some with a demonstrative and a noun (e.g., these beads) to produce *these some beads instead of some of these beads. Children acquire languages of quite different morphological types in similar fashion. In English, they learn a relatively lean system of inflections combined with considerable irregularity in the established vocabulary, but in Turkish they master a rich bound morphology that is highly regular in form and transparent in meaning. And in Hebrew, they learn a synthetic system with rich inflectional marking plus an elaborate system of agreement. In all three languages, children’s acquisition of inflections plays a central part in early grammatical development and simple clause structure. What they must add subsequently is mastery of subregularities or minor lexical schemas that are limited to specific sets of items, conjugations, or declensions, as well as lexical exceptions ⫺ the unpredictable, idiosyncratic irregularities of the lexicon.
165. First language acquisition
3.
Derivation
The acquisition of derivational morphology has, by comparison, been relatively neglected even though it is relevant both to grammatical form and to lexical structure. Nonetheless, there are extensive corpora of novel word formations available for English, French, German, Hebrew, Hungarian, and Polish (Clark 1993). Innovative words that rely on derivational morphemes are an important source of evidence that children have carried out some morphological analysis and are not simply using ready-made forms from the conventional lexicon. The data available consist of spontaneous innovations in the lexicon, elicited innovations, and elicited glosses of the possible meanings of novel words. Several domains have been studied in detail for the association of a derived morpheme with a specific meaning, e.g., agent and instrument nouns, compound nouns and other nominals, and denominal verbs. 3.1. Conventionality A major task in language acquisition is to master the conventional lexicon. But children often wish to talk about things for which they have not yet acquired the conventional terms. When this occurs, they may coin a word on the spot to serve their purpose (cf. Art. 80). The conventional lexicon plays a critical role in this because it displays the range of possible word-structures that children can make use of in building word forms to carry new meanings. Children fill the gaps in their lexicon as needed. For adults, such gaps may be momentary (when speakers fail to retrieve a word on the tip of the tongue and so fill in with a coinage instead) or permanent (when speakers come across a true gap in the lexicon where there is no conventional word for the pertinent meaning). The same distinction exists for children, but their still-limited knowledge of the lexicon may lead them to coin many words where conventional terms already exist. These coinages are effectively ones that, for adult speakers, would be preempted by the conventional terms. Conventional terms take priority over innovative ones. Speakers expect that if there is a conventional term for the expression of a particular meaning, it must be used for that meaning. If it is not, then the person talking must have some other meaning in mind. As a result, innovative words cannot simply be
1801 produced in place of conventional ones. To be used at all, they must contrast with them in meaning (Clark 1990; 1993; Clark & Clark 1979). For children, there are many more gaps to fill because their conventional vocabulary is still so small. The result is a large number of illegitimate coinages ⫺ innovations that are pre-empted (or blocked) by the existence of conventional words with just those meanings ⫺ as well as many legitimate ones. Both types of innovation allow us to look in some detail at what children at different stages know about word formation in their first language. 3.2. Structural options for word formation What do children know, at each stage, about possible form-meaning combinations in their language? Word formation is a critical domain for answering this question. The major options available in a language typically include derivation (with or without affixes) and compounding (Anderson 1985 b). When children innovate, they show us what they know about the forms of possible words in the language. This holds true both at the phonotactic level (where children never attempt to impose impossible consonant clusters or syllable structures) and at the level of morpheme combination. That is, children take account of typological features of their first language from a very early point in acquisition, and are sensitive to the kinds of word formation options favoured in the language they are acquiring. English-speaking children, for example, rely heavily on zero-derivation for forming new verbs from as young as age two (e.g., [to] scale for ‘weigh’, [to] oar for ‘row’; cf. Clark 1982), but Hebrew-speaking children do not. Children acquiring Germanic languages rely extensively on compounding to form new nouns (e.g., English car-smoke ‘exhaust’, cup-egg ‘boiled egg’), again from age two on, but children acquiring Romance languages do not. Derivational affixes (e.g., caker ‘someone who makes cakes’, sweeper ‘broom’; Clark 1993) tend to emerge slightly later, so in languages that favour derivation over compounding, children may not produce many coinages until age three or later. In general, children acquire suffixes before prefixes (Kuczaj 1979), for both inflection and derivation. Other options such as blending or clipping emerge later still, and are usually favoured more by adults than children (Ravid 1990; 1995). Compounding combined
1802 with derivation emerges later than compounding alone. Just as with inflections, children begin with the stem and then add the modulation of meaning, whether through affixation, compounding, or some other means. Children’s earliest choices of word-forms for new meanings appear to be governed by two main factors in addition to typology: (a) the simplicity of the form used, with a premium on the fewest changes possible in the base form, and (b) the transparency of the meaning expressed, in relation to the meaning of the base. The youngest children favour coinages with zero-derivation (as in the English innovative verbs [to] sand ‘grind into small pieces’ or [to] rug ‘vacuum a rug’, both from two-year-olds) or simple noun ⫹ noun compounding as in plate-egg ‘fried egg’ or baby-towel ‘face-cloth’ (also from two-yearolds). The compounds are also transparent to the degree that the head noun denotes the kind of category being talked about (here an egg and a towel). As children analyse the meanings of affixes, their coinages become increasingly transparent: they add the appropriate affix to a base, and use English -er, for example, to mark the agent or actor (as in rainer ‘(imaginary) person who makes rain go away’ or crayoner ‘someone who makes pictures with crayons’) (Clark & Hecht 1982; Clark & Berman 1984). In short, children’s lexical innovations show that they know a lot about the regularities of word formation in their language, even though they may not yet know the conventional terms for the referents they talk about. 3.3. Productivity Once children have acquired several options related in meaning for forming new words, they have to choose which to use on particular occasions. Their choices appear to be guided by which option is the most productive for adult speakers in the same community (cf. Art. 33). Children rely more frequently on the device adults favour than on other options also available in the language. For example, in English, four- and five-year-olds favour agentive -er over -ist and -ian, and, in Hebrew, children favour suffixal -an to mark agents over the other options available (Clark & Berman 1984; Clark & Cohen 1984). Previous research also shows that when adult preferences for one option over others change, children’s choices of forms follow the
XIX. Psycholinguistische Perspektiven
adults’. Children must therefore be tracking the frequencies of different form types in the input they hear. Shifts in what is productive for adult speakers are paralleled by shifts in children’s preferences in the forms chosen for lexical innovations (Chmura-Klekotowa 1970). Productivity in word formation presents a rather different face to productivity in inflection. In word-formation, some slots in a paradigm may be taken by existing words with the appropriate meanings, so use of the otherwise regular form is pre-empted by the existence of another form with the pertinent meaning. The degree of regularity overall is therefore somewhat lower than in inflectional paradigms. At the same time, the forms of innovative lexical items, coined to contrast with existing form-meaning combinations, offer strong evidence for the relative productivity of specific affixes or compounding patterns. In word formation, then, new formations, whether nonce uses or words that eventually enter the conventional lexicon, provide critical evidence for productivity (Berman 1987; 2000). 3.4. Order in word formation When children coin words, they often make errors, but they rarely, if ever, make mistakes in how they order the morphemes within a new word. That is, derivational morphemes are consistently added to the base or stem before children add any inflections. And where they need to add more than one derivational or inflectional affix to a single stem, these too are placed in the correct order, as in sweeper-s (verb ⫹ er ⫹ s) for the machines used to sweep streets. This attention to morphemeorder within the word is all the more striking in the face of order errors elsewhere. Children do make frequent order errors in compounds in some languages. These errors appear to be largely predictable on typological grounds and occur where there is a mismatch of head-modifier ordering in different constructions. For example, in English, the head of the verb phrase is the phrase-initial verb (e.g., throw in throw the ball). But in compound nouns, the head is in final position (e.g., climber in cliff-climber). Where the compound contains a verb stem (climb), children go through a stage of mis-ordering the modifier and head to produce compounds like a climb-cliff or a climber-cliff for ‘someone who climbs cliffs’. But in languages with the same head-modifier order across con-
1803
165. First language acquisition
structions, children never make such errors (Clark 1993; Clark & Berman 1987). Children are consistent in adhering to the ordering of morphemes within words. They begin with the base or stem form, add the relevant derivational affixes, and then add any inflections needed. Within words, children rarely if ever make errors in the linear ordering of affixal morphemes.
4.
Summary
After initial acquisition of single words, children aged two to three gain command of the syntax of simple clauses. They typically master case-marking with prepositions first, and then inflectional marking of agreement for number, gender, and person. They begin to add the modulations of case, tense, and other inflections before their second birthday, but make systematic errors along the way. Derived forms may also begin to appear between age two and three as children turn to coinage to fill gaps in their vocabulary. But the extent of such early reliance on derivation is limited. This is particularly clear in a language like Hebrew, where nearly all the word-formation processes require use of affixal morphology (Berman 2000). In languages like English, where zero derivation (without affixes) is also an option for word formation, children take advantage of it as young as age two to coin verbs from nouns. They may also begin to use derivational affixes that do not demand any change in the shape of the word stem (e.g., English -er in agent nouns like jumper) soon afterwards. However, derivational affixes appear to enter in large numbers only after age three. In both inflection and derivation, children regularise. They treat irregular forms as if they were regular, adopting those paradigms with the largest number of members (types) as their templates. The learning of lexical exceptions, whether inflectional or derivational, occurs only after mastery of regular forms. The next year, from age three to four, is critical in word formation. It is then that children analyse the internal structure of words and become capable of coining new words with derivational affixes. They do this with a range of form-types consistent with the options in the target language (Clark 1993). Indeed, the ease with which children master morphological structure depends heavily on the typology of the language, for example
whether it is synthetic, polysynthetic, isolating, or agglutinating. It also depends to a large extent on the nature of the meaning-toform mappings for inflectional and derivational paradigms. Finally, some lexical options in word formation are more productive than others, and the more productive they are, the earlier they tend to be acquired. In English and German, for instance, compounding appears much earlier than in Hebrew or French. Other options may be acquired late because they are formally very complex, as is the case for vowel-laxing and stress-shifts with derivation in the Latinate vocabulary of English, or for passive forms and certain derived nominals in Hebrew. Finally, some word-formation types may be used only in formal or written registers, and there, the patterns of usage account for late acquisitions in word-formation (see Jisa et al. 2002).
5.
References
Aksu-Koc¸, Ayhan & Slobin, Dan I. (1985), “The Acquisition of Turkish”. In: Slobin (ed.), Vol. I, 839⫺878 Anderson, Stephen R. (1985 a), “Inflectional Morphology”. In: Shopen (ed.), 150⫺201 Anderson, Stephen R. (1985 b), “Typological Distinctions in Word Formation”. In: Shopen (ed.), 3⫺56 Bates, Elizabeth (1976), Language and Context. New York: Academic Press Bates, Elizabeth & MacWhinney, Brian & Caselli, Cristina & Devescove, Antonella & Natale, Francesco & Venza, Valeria (1984), “Crosslinguistic Study of the Development of Sentence Interpretation Strategies”. Child Development 55, 341⫺354 Berko, Jean (1958), “The Child’s Learning of English Morphology”. Word 14, 150⫺177 Berman, Ruth A. (1985), “Acquisition of Hebrew”. In: Slobin (ed.), Vol. I, 255⫺371 Berman, Ruth A. (1987), “Productivity in the Lexicon: New-word Formation in Modern Hebrew”. Folia Linguistica 21, 425⫺461 Berman, Ruth A. (1990), “Acquiring an (S)VO Language: Subjectless Sentences in Children’s Hebrew”. Linguistics 28, 1135⫺1166 Berman, Ruth A. (2000), “Children’s Innovative Verbs versus Nouns: Structured Elicitations and Spontaneous Coinages”. In: Menn & Bernstein Ratner (eds.), 69⫺93
1804 Berman, Ruth A. & Armon-Lotem, Sharon (1996), “How Grammatical are Early Verbs?”. In: Martinot, Claire (ed.), Actes du Colloque Internationale sur l’Acquisition de la Syntaxe en Langue Maternelle et en Langue Etrange`re. Annales Litte´raires de l’Universite´ de Franche-Compte´: No. 631, 17⫺60 Bowerman, Melissa (1988), “The ‘No Negative Evidence’ Problem: How do Children Avoid Constructing an Overly General Grammar?” In: Hawkins, John A. (ed.), Explaining Language Universals. Oxford (UK): Blackwell, 73⫺101 Brown, Roger (1973), A First Language: The Early Stages. Cambridge/MA: Harvard University Press Bybee, Joan (1985), Morphology: A Study of the Relation between Meaning and Form. Amsterdam: Benjamins Bybee, Joan & Slobin, Dan I. (1982), “Rules and Schemas in the Development and Use of the English Past Tense”. Language 58, 265⫺289 Chmura-Klekotowa, Maria (1970), “Odbicie tendencji slowotwo´rczych jezyka polskiego w neologizmach dzieci”. Prace Filologiczne 20, 153⫺159 Clahsen, Harald (1986), “Verb Inflections in German Child Language: Acquisition of Agreement Markings and the Functions they Encode”. Linguistics 24, 79⫺121 Clark, Eve V. (1982), “The Young Word Maker: A Case Study of Innovation in the Child’s Lexicon” In: Wanner, Eric & Gleitman, Lila R. (eds.), Language Acquisition: The State of the Art. Cambridge (UK): Cambridge University Press, 390⫺425 Clark, Eve V. (1985), “The Acquisition of Romance, with Special Reference to French”. In: Slobin (ed.), Vol. I, 687⫺782 Clark, Eve V. (1990), “On the Pragmatics of Contrast”. Journal of Child Language 17, 417⫺431 Clark, Eve V. (1993), The Lexicon in Acquisition. Cambridge (UK): Cambridge University Press Clark, Eve V. & Berman, Ruth A. (1984), “Structure and Use in the Acquisition of Word Formation”. Language 60, 542⫺590 Clark, Eve V. & Berman, Ruth A. (1987), “Types of Linguistic Knowledge: Interpreting and Producing Compound Nouns”. Journal of Child Language 14, 547⫺567 Clark, Eve V. & Chouinard, Michelle M. (2000), “Enonce´s enfantins et reformulations adultes dans l’acquisition du langage”. Langages 140, 9⫺23 Clark, Eve V. & Clark, Herbert H. (1979), “When Nouns Surface as Verbs”. Language 55, 767⫺811 Clark, Eve V. & Cohen, Sophia R. (1984), “Productivity and Memory for Newly-Formed Words”. Journal of Child Language 11, 611⫺626
XIX. Psycholinguistische Perspektiven Clark, Eve V. & Hecht, Barbara F. (1982), “Learning to Coin Agent and Instrument Nouns”. Cognition 12, 1⫺24 Clark, Herbert H. & Clark, Eve V. (1977), Psychology and Language. New York: Harcourt Brace Jovanovich Demuth, Katherine A. (1988), “Noun Classes and Agreement in Sesotho Acquisition”. In: Barlow, Michael & Ferguson, Charles A. (eds.), Agreement in Natural Languages: Approaches, Theories, Descriptions. Stanford, California: Center for the Study of Language and Information (CSLI), 305⫺341 Dromi, Esther & Berman, Ruth A. (1982), “A Morphemic Measure of Early Language Development: Data from Modern Hebrew”. Journal of Child Language 9, 403⫺424 Frauenfelder, Uli H. & Schreuder, Robert (1991), “Constraining Psycholinguistic Models of Morphological Processing and Representation: The Role of Productivity”. In: Booij, Geert & van Marle, Jaap (eds.), Yearbook of Morphology 1991. Dordrecht: Kluwer, 165⫺183 Greenberg, Joseph H. (1966), Language Universals. The Hague: Mouton Guillaume, Paul (1927), “Le de´veloppement des e´le´ments formels dans le langage de l’enfant”. Journal de Psychologie 24, 203⫺229 Gvozdev, Aleksandr N. (1961), Voprosy Izucheniya Detskoy Rechi. Moskva: Idz. Akademii Pedagogicheskich Nauk RSFSR Hecht, Barbara F. (1983), Situations and Language: Children’s Use of Plural Allomorphs in Familiar and Unfamiliar Settings. Ph.D. dissertation, Stanford University Jisa, Harriet & Baruch, Elisheva & Berman, Ruth & Reilly, Judy & Rosado, Elisa & Verhoeven, Ludo (2002), “Passive Voice Constructions in Written Texts: A Crosslinguistic Developmental Study”. In: Berman, Ruth A. & Verhoeven, Ludo (eds.), Cross-Linguistic Perspectives on the Development of Text-Production Abilities in Speech and Writing. Part 2 (Special issue of Written Language and Literacy 5:2), 163⫺181 Karmiloff-Smith, Annette (1979), A Functional Approach to Child Language: A Study of Determiners and Reference. Cambridge (UK): Cambridge University Press Kuczaj, Stan A. (1978), “Children’s Judgements of Grammatical and Ungrammatical Irregular Past Tense Verbs”. Child Development 49, 319⫺326 Kuczaj, Stan A. (1979), “Evidence for a Language Learning Strategy: On the Relative Ease of
165. First language acquisition
1805
Acquisition of Prefixes and Suffixes”. Child Development 50, 1⫺13
Pye, Clifton (1990), “The Acquisition of Ergative Languages”. Linguistics 28, 1291⫺1330
Levy, Yonata (1983), “The Acquisition of Hebrew Plurals: The Case of the Missing Gender Category”. Journal of Child Language 10, 107⫺121
Ravid, Dorit (1990), “Internal Structure Constraints on New Word Formation Devices in Modern Hebrew”. Folia Linguistica 34, 289⫺347
MacWhinney, Brian (1978), The Acquisition of Morphophonology. Chicago/IL: University of Chicago Press (Monographs of the Society for Research in Child Development 43)
Ravid, Dorit (1995), Language Change in Child and Adult Hebrew. Oxford (UK): Oxford University Press
Maratsos, Michael P. (2000), “More Overregularizations after All: new Data and Discussion on Marcus, Pinker, Ullman, Hollander, Rosen & Xu”. Journal of Child Language 27, 183⫺212 Marcus, Gary & Ullman, Michael & Pinker, Steven & Hollander, Michelle & Rosen, T. John & Xu, Fei (1992), Overregularization. Chicago/IL: University of Chicago Press (Monographs of the Society for Research in Child Development 57) Meisel, Jürgen (1986), “Word Order and Case Marking in Early Child Language: Evidence from Simultaneous Acquisition of Two First Languages: French and German”. Linguistics 24, 123⫺183 Menn, Lise & Bernstein Ratner, Nan (2000, eds.), Methods for Studying Language Production. Mahwah/NJ: Erlbaum Mikes, Melanie (1967), “Acquisition des cate´gories grammaticales dans le langage de l’enfant”. Enfance 20, 289⫺298 Mills, Anne (1985), “The Acquisition of German”. In: Slobin (ed.), Vol. I, 141⫺254 Mithun, Marianne (1988), “The Acquisition of Polysynthesis”. Journal of Child Language 16, 285⫺ 312 Omar, Margaret (1973), The Acquisition of Egyptian Arabic as a Native Language. The Hague: Mouton Peters, Ann M. & Menn, Lise (1993) “False Starts and Filler Syllables: Ways to Learn Grammatical Morphemes”. Language 69, 742⫺777 Platt, Carol B. & MacWhinney, Brian (1983), “Error Assimilation as a Mechanism in Language Learning”. Journal of Child Language 10, 401⫺414 Ple´h, Csaba (1998), “Early Spatial Case Markers in Hungarian Children”. In: Clark, Eve V. (ed.), The Proceedings of the 29th Child Language Research Forum [1997]. Stanford/CA: Center for the Study of Language and Information (CSLI), 211⫺219
Shopen, Timothy (1985, ed.), Language Typology and Syntactic Description, Vol. III: Grammatical Categories and the Lexicon. Cambridge (UK): Cambridge University Press Slobin, Dan I. (1973), “Cognitive Prerequisites for the Acquisition of Grammar”. In: Ferguson, Charles A. & Slobin, Dan I. (eds.), Studies of Child Language Development. New York: Holt, Rinehart & Winston, 173⫺208 Slobin, Dan I. (1985, ed.), The Crosslinguistic Study of Language Acquisition, Vol. I⫺II. Hillsdale/ NJ: Lawrence Erlbaum Associates Slobin, Dan I. (1992, ed.), The Crosslinguistic Study of Language Acquisition. Vol. III. Hillsdale/ NJ: Lawrence Erlbaum Associates Slobin, Dan I. (1997, ed.), The Crosslinguistic Study of Language Acquisition, Vol. IV⫺V. Hillsdale/NJ: Lawrence Erlbaum Associates Slobin, Dan I. & Bever, Thomas G. (1982), “Children Use Canonical Sentence Schemas: A Crosslinguistic Study of Word Order and Inflections”. Cognition 12, 229⫺265 Smoczynska, Magdalena (1985), “The Acquisition of Polish”. In: Slobin (ed.), Vol. I, 595⫺686 Tracy, Rosemarie (1984), “Fallstudien: Überlegungen zum Erwerb von Kasuskategorie und Kasusmarkierung”. In: Czepluch, Hartmut & Janssen, Hero (eds.), Syntaktische Struktur und Kasusrelation. Tübingen: Gunter Narr, 271⫺313 Veneziano, Edy & Sinclair, Hermine (2000), “The Changing Status of ‘Filler Syllables’ on the Way to Grammatical Morphemes”. Journal of Child Language 27, 461⫺500 Vihman, Marilyn M. (1999, ed.), First Steps in Morphological and Syntactic Development: CrossLinguistic Evidence. Special issue, International Journal of Bilingualism 3
Eve V. Clark, Stanford (U.S.A.) Ruth A. Berman, Tel Aviv (Israel)
1806
XIX. Psycholinguistische Perspektiven
166. Second language acquisition 1. 2. 3. 4. 5. 6. 7.
1.
Morpheme-acquisition order Inherent inflectional morphology Contextual inflection Case marking Derivation Compounding References
-ing plural copula
auxiliary article
Morpheme-acquisition order
The first developmental studies in secondlanguage (L2) acquisition dealt with the acquisition of a given set of morphemes in L2 English. In particular, it was studied in which order these morphemes were acquired. The reason for carrying out these studies was a discussion on the nature of L2 acquisition in the mid-seventies. The central issue of this discussion was the role of the mother tongue (L1). Dulay & Burt (1974; 1975) argued that L2 acquisition, like L1 acquisition, was essentially a process of “creative construction”. Given their assumption that L2 learners relying on their L1 could only do so if they were applying old habits, they compared L1 and L2 acquisition in order to demonstrate that L1 was not a relevant factor. In order to test their claim on the irrelevance of L1, Dulay & Burt (1974; 1975) compared morpheme acquisition order in the L2 English of children with differing language backgrounds. It was found that the order of acquisition was independent of the L1. Therefore, since old habits did not play a role in L2 acquisition, it was concluded that L2 acquisition, just like L1 acquisition, was a process of creative construction. A similar conclusion was reached by Bailey et al. (1974) leading to what Krashen (1977) called the ‘Natural Order’ for L2 English represented in Fig. 166.1. The studies by Dulay & Burt and Krashen were the first of many similar studies on the acquisition of morphemes in L2 English by native speakers of differing L1s (e.g. LarsenFreeman 1975 and Andersen 1978). In addition, processes of morpheme acquisition in an L2 were compared with morpheme acquisition in other acquisitional and developmental contexts. For example, L2 acquisition was compared with processes of pidginization and creolization as well as diachronic change in language. The results of all the morpheme studies point to a common or nat-
irregular past
regular past 3rd singular possessive Fig. 166.1: Krashen’s (1977) ‘Natural Order’ for L2 English
ural order of acquisition which is seen as evidence for L2 learning as a process of creative construction. Most of these studies, however, have not paid a great deal of attention to what mental mechanisms could explain the particular orders found.
2.
Inherent inflectional morphology
As proposed in Booij (1994), two types of inflection should be distinguished, inherent and contextual inflection. Inherent inflectional morphology as opposed to contextual inflectional morphology serves to express a particular semantic content. Examples of inherent inflection are tense or aspect with verbs and number with nouns. 2.1. Past tense The acquisition of past tense morphology in English has become an area of heated debate between adherents of connectionist and symbolic models of language acquisition. The question is whether or not there is an essential difference between regular and irregular forms with respect to acquisition, storage, and production. Within a symbolic framework, linguistic knowledge of the regularly formed past tense form in English is ade-
166. Second language acquisition
quately represented by a morphological rule, whereas irregular forms are stored as associative knowledge (Pinker & Prince 1988). Within the connectionist view, on the other hand, there is no principled difference in linguistic status between regular and irregular forms. All types of past morphology are based on associative relations (Rumelhart & McClelland 1986). Notions such as rules and overgeneralizations are only to be seen as epiphenomena (Bates et al. 1983). L2 acquisition of regular and irregular verbs was studied by Lalleman et al. (1997). They wanted to find out whether or not it is possible for advanced learners of L2 Dutch to implicitly acquire the same kind of unconscious linguistic knowledge as L1 speakers. Based on the idea that production by rule is not affected by frequency of occurrence, whereas lexical retrieval is, they hypothesized that production time for irregular verbs is faster with familiar forms than with unfamiliar forms. In order to investigate this hypothesis they carried out a production time experiment for past tense formation. The results from this experiment were similar to those from the experiments on L1 English. First, differences in production time were found between regular and irregular past tense forms. Furthermore, high-frequency irregular past tense forms were produced faster than their low-frequency counterparts, while with regular verbs there is no effect of past tense frequency on the production time, neither for L1 nor for L2 speakers. Obviously, native speakers were faster than L2 learners and differences in production time between regular and irregular forms were larger with L2 learners. Lalleman et al. (1997) also compared production time of prototypical and atypical regular verbs. Prototypical regular verbs are those with a short vowel (except i, e) plus a geminate consonant in the infinitive form. They are prototypically regular because, as far as their morphology is concerned, there are no irregular counterparts. Examples are Dutch knallen ‘to bang’, bukken ‘to stoop’, klappen ‘to clap’ (an exception is vallen ‘to fall’). Prototypical regular verbs are produced more quickly and with fewer errors than other regular verbs. Comparing the percentages of irregularization errors in 11 prototypical regular verbs such as knallen ‘to bang’, bukken ‘to stoop’, klappen ‘to clap’, and in 11 atypical regular verbs such as
1807 kleven ‘to stick’ and huilen ‘to cry’, they found that L2 learners irregularize past tense forms in prototypical regular verbs less often (0,3 percent) than in atypical regular verbs (12,0 percent). In this respect L2 learners behave differently from native speakers. This L2 learner behaviour is probably due to the way in which morphological rule formation takes place. Rule formation in morphology depends on type frequency, i.e. the frequency with which different lexical items share the same morphological properties. Since there are classes of high-frequency irregular verbs which share morphological properties with regular verbs, L2 learners are likely to infer that these verbs are to be related on the basis of a particular morphological rule. As long as learners do not know that this apparent regularity is just an epiphenomenon of the lexical properties of an arbitrary list of verbs, L2 learners may treat regular verbs with similar morphology the same way. 2.2. Aspect In child language acquisition, studies by Bronckart & Sinclair (1973), Antinucci & Miller (1976), and Bloom et al. (1980) started a debate on whether past morphology is initially used for coding past tense or perfective aspect. In L1 acquisition this issue has been addressed in connection with what is called the Defective Tense Hypothesis (Weist 1986). In studies on second language acquisition the same question of ‘aspect before tense’ or ‘tense before aspect’ became relevant with respect to the development of interlanguage morphology. With respect to the acquisition of L2 Spanish by native speakers of English, Andersen (1991) found a developmental sequence for encoding tense and aspect with past morphology. Given a semantic classification of verbs into the four verb categories ‘states’, ‘activities’, ‘telic events’, and ‘punctual events’, Andersen discovered that L2 learners approach the L2 system by linking aspectual marking and lexical meaning: “[t]he Spanish Preterit and Imperfect are initially interpreted as [redundant] markers of inherent aspect” (Andersen 1991: 319). This results in a developmental process in which the Imperfect is first used for states (tenı´a ‘had’) and then also for activities (jugaba ‘played’), whereas the Preterit is first used for punctual events (se partio´ ‘broke in two’) and later for telic events (ensen˜o ‘taught x to y’) as well (see Fig. 166.2).
1808
XIX. Psycholinguistische Perspektiven
imperfect J states tenı´a ‘had’
I preterit activities jugaba ‘played’
telic events ensen˜o ‘taught x to y’
punctual events se partio´ ‘broke in two’
Fig. 166.2: Aspectual/lexical meaning in L2 Spanish (Andersen 1991: 314)
states
activities
telic events
punctual events
tenı´a tuvo ‘had’
jugaba jugo´ ‘played’
ensen˜o ensen˜aba ‘taught x to y’
se partio´ se partı´a ‘broke in two’
Fig. 166.3: Grammatical aspect in Spanish (Andersen 1991: 314)
Subsequently, learners gradually also acquire the use of the Imperfect with telic and punctual events and the use of the Preterit with activities and states (see Fig. 166.3). This means that in the end they will have acquired the variable use of past tense morphology as a device for marking grammatical aspect according to the perspective the speaker may take with respect to the temporal course of the situation. Giacalone Ramat (1995) observed a systematic opposition between simple present and past participle morphology in the initial stages of L2 Italian of native speakers of Chinese, Eritrean, Iranian, English, German, and French. An example from Giacalone Ramat (1995) is given in (1). (1) Poi aperta la botiglia ... poi then open:past.part the bottle then bevi vino. drink:3.sg wine ‘After he opened the bottle, he drinks wine.’ Giacalone Ramat argued that this opposition has to be interpreted in terms of aspect: the simple present form or the infinitive encodes imperfective aspect, whereas the bare past participle indicates perfect aspect. Another observation by Giacalone Ramat indicates that L2 learners of Italian initially encode the aspectual opposition imperfective aspect vs. perfect aspect lexically. This explains why in many cases punctual and telic verbs are marked with a past participle form (such as arrivato ‘arrived’, lasciato ‘left’, trovato ‘found’), whereas durative and stative
verbs (such as stare ‘stay’) receive imperfect marking. An example is given in (2). (2) Pensi cosi ⫹⫹⫹ quanto think:1.sg so if trovato molto bene ⫹ stare find:past.part very well stay:inf qui ⫹ va bene. here go:3.sg well ‘I think so: If I feel O.K. in Italy, I’ll stay here.’ With respect to the acquisition of L2 German, Dutch, and English, it has been noted that, independently of the L1, tense is acquired before aspect. This results from research by Dietrich et al. (1995) on L2 German, by Housen (1993) on L2 Dutch and by Vogel (1987) on L2 English. These authors draw the conclusion that their L2 learners first distinguish between past and non-past. Andersen (1991), however, concludes from a number of studies on the acquisition of L2 English, that here past inflection is first used only with punctual verbs (left), then also with telic verbs (taught), subsequently for activities too (ran), and finally also with verbs referring to states (had). 2.3. Plural For L2 learners, plural marking in German is difficult to acquire due to the fact that the German plural system consists of nine plural markers. Studies dealing with L2 learners’ plural marking in German are Köpcke (1987) and Wegener (1994; 1995). In Köpcke (1987), plural marking on pseudowords is experimentally investigated in 40 adult native
1809
166. Second language acquisition
speakers. The data are compared with L2 learners’ data from 80 English speaking American students of German who were studied by Phillips & Bouma (1980). Wegener collected data from two groups of children between 7 and 9 years of age: 4 Turkish children participating in bilingual classes and 4 Aussiedler children (3 Polish and 1 Russian) going to transition or regular classes. Both studies focused on processes of rule formation and thus on processes underlying overgeneralization. First, Wegener found that regularizations are formed with -Ø, -e, -(e)n, -s, and “practically never with -er and the Umlaut” (Wegener 1994: 277). Furthermore, “the -(e)n plural is the most frequently overgeneralized plural” (Wegener 1994: 279). With respect to the order in which these overgeneralizations were acquired, it was found that first -(e), then -(e)n, and finally -s was learnt. Köpcke’s analysis of the L2 learners’ pseudoword production leads to similar findings. The plural morphemes most frequently used were -(e)n and -e. In one-syllable words with no particular morphological properties, -(e)n and -e are used most frequently with no obvious gender bias. To explain the fact that -en and -e are overgeneralized, Wegener assumes that in dealing with the input data learners may choose between two strategies: a holistic-formulaic or a cognitive-analytic strategy (cf. Art. 163). If learners first acquire plural forms as unstructured wholes and analyze them later, the holistic-formulaic strategy is applied. If they are geared towards the identification of plural markers and rules for plural formation, learners apply a cognitive-analytical strategy. In her data Wegener has found evidence that rule formation is aided by high type frequency. High type frequency means that many different nouns have the same plural marker. High type frequency furthers rule formation because the more heterogeneous the nouns with a particular plural marker, the more productive or applicable this plural marker is. The fact that it is first the plural -e that is regularized most frequently is explained by its higher token frequency in the basic lexicon. Later in the acquisition process it is substituted by -(e)n, which has the highest type frequency. Furthermore, whereas in the target grammar, use of a particular plural marker depends on markedness and gender of the noun, in the L2 learners’ grammar it correlates with word endings: nouns ending
with -e have plural -n (as in Jungen ‘boys’), nouns ending with a consonant have plural -e (as in Türe ‘doors’, Elefante ‘elephants’), and nouns with pseudosuffix -el have -Ø or -n (as in Kugel or Kugeln ‘bullets’).
3.
Contextual inflection
Inherent inflectional morphology serves to express a particular semantic content independently of its syntactic context. Contextual inflectional morphology, however, depends on particular properties of syntactic context cf. Booij (1994). In the following sections, two examples of the L2 acquisition of contextual inflection will be discussed: the acquisition of finiteness in L2 German and L2 Dutch and adjectival morphology in L2 Dutch. 3.1. Finiteness In L2 German and Dutch the acquisition of finiteness constitutes an important target. Verb inflection in German and Dutch is context-dependent in the sense that morphology is determined by the person and number of the subject. It is obvious, however, as illustrated in (3), (4), and (5), that use of inflectional morphology also correlates with positional properties. (3) (a) Karl gibt Hans ein Buch. Karel geeft Hans een boek. Charles give:3.sg John a book ‘Charles gives John a book.’ (b) Gibt Karl Hans ein Buch? Geeft Karel Hans een boek? give:3.sg Charles John a book ‘Does Charles give John a book?’ (4) Ich sehe, daß Karl Hans ein Ik zie dat Karel Hans een I see:1.sg that Charles John a Buch gibt. boek geeft. book give:3.sg ‘I see that Charles gives John a book.’ (5) Karl möchte Hans ein Buch Karel wil Hans een boek Charles want:3.sg John a book geben. geven. give:inf ‘Charles wants to give John a book.’ Verb forms showing agreement (i.e. finite) morphology, occur in second (3 a) or first po-
1810 sition (3 b) in main clauses, whereas in subordinate clauses they occur in sentence-final position (4). Verb forms with no inflectional morphology always appear in sentence final position (5). In the initial stages of the acquisition of L2 German by L1 Italian speakers, learners apparently assume that verbs should occur in first or second position. This is what they know from their L1. Then, (as evidenced by Klein & Carroll 1992: 157 ff.) at some point they learn that lexical verbs can occur both in non-final and final position, without noticing that these positional options correlate with finite vs. non-finite morphology. What they do seem to acquire, however, is that this distributional difference correlates with the absence or presence of an auxiliary. Illustrations of this particular stage of acquisition are utterances from Gina, which were collected within the framework of the European Science Foundation project as reported in Klein & Carroll (1992). (6) gucken in eine bäckerei look:inf in a bakery ‘to look into a bakery’ (7) und dann habe eine brot and then have:1.sg a bread gebringen bring:past.part ‘and then I have brought a loaf of bread’ As Gina is able to learn the positional properties of lexical verbs independently of the relevant functional property of finiteness, it is not at all obvious how she will eventually be able to do so. What seems to be crucial to the acquisition of finiteness by learners such as Gina, is the acquisition of particle verbs with a separable particle. An example is zurückkommen (come back) in er kommt morgen zurück ‘he come:3.sg tomorrow back (he comes back tomorrow)’. The acquisition of separable particle verbs focusses learners on the variability of the position and the form of the lexical verb with respect to the particle. See e.g. the position of kommt and gekommen in (8). (8) kommt zurück vs. ist zurückgekommen The acquisition of finiteness in L2 Dutch by L1 speakers of Turkish differs from the acquisition of L2 German by Romance speakers. While the particular properties of the target languages are the same, the developmental processes differ.
XIX. Psycholinguistische Perspektiven
It is typical of Turkish learners of Dutch (see Coenen & Klein 1992; Jordens 1988) that in the initial stages they use most of the lexical verbs with non-finite morphology in final position, as in (9). Only a small group of verbs occurs with finite morphology in initial, i.e. first, second or third position, as in (10). The examples in (9) and (10) are taken from Mahmut. (9)
En dan die politie gezien. and then that police seen:past.part ‘And then he saw the police.’
(10) Ja, dan politie komt naast. yes then police come:3.sg next ‘Yes, then the police comes next.’ Ergün is a good example of a Turkish learner who has reached a more advanced developmental stage. When he begins to use infinitives in final position with modal verbs as in (11) and past participles in final position with or without auxiliaries as in (12), he also uses non-finite verbs in non-final position most of the time as in (13). (11) Hij zegt, “ik wil niet gaan he say:3.sg I want:1.sg not go:inf buiten”. outside” “He says, ‘I do not want to go outside’.” (12) Ja dan is terug naar achter yes then be:3.sg back to after politie/politie gaan. police/police go:past.part ‘Yes, then he went after the police again.’ (13) Hij gaan werk of zo. he go:past.part work or so ‘He goes to work probably.’ This opposition between (11) and (12), on the one hand, and (13), on the other, indicates that in final position lexical verbs are used to describe the ‘non here-and-now’, whereas in non-final position lexical verbs are used to refer to the immediate ‘here-and-now’ situation. At the point at which Turkish learners discover that for a description of here-and-now situations the lexical verb must be placed in non-final position, they will notice that the target model does not have -en morphology. This means that now, faced with positive evidence, they are able to learn the target-like, correct forms.
1811
166. Second language acquisition
3.2. Adjectival morphology Adjectival morphology in languages such as French, Dutch, and German depends on relations of agreement. In French, for example, depending on grammatical gender and number of the noun, adjectives vary with respect to their morphological endings: la grande hotel ‘the big hotel’, le grand cafe´ ‘the big cafe´’, les grandes hotels ‘the big hotels’. In Dutch, adjectival morphology also depends on whether the determiner is definite or indefinite: het dikke boek ‘the thick book’, een dik boek ‘a thick book’. Furthermore, in German adjectives vary according to grammatical case as well. For example, in the case of a singular noun with masculine gender, adjectives have different morphology according to whether the noun is nominative or accusative: der dicke Baum (nominative) ‘the big tree’, den dicken Baum (accusative) ‘the big tree’. An interesting case for the L2 acquisition of adjectival morphology is Dutch. Here, the target system is relatively simple. Only with nouns that have neuter gender (evident from the use of the article het) can adjectives have or not have e-morphology depending on whether the noun is definite, as in het dikke boek ‘the thick book’, or indefinite, as in een dik boek ‘a thick book’. Masculine/feminine nouns (with the article de), on the other hand, always have e-morphology regardless of whether they are definite, as in de dikke boom ‘the big tree’, or indefinite, as in een dikke boom ‘a big tree’. With regard to language use there are two ways in which this type of knowledge may be represented. One possibility, as illustrated in (14), is that the rules of adjectival morphology are stored as a property of the noun. (14) Nde with with Nhet with with
definite article: adj. ⫺e indefinite article: adj. -e definite article: adj. -e indefinite article: adj. -Ø
Another possibility is that the use of -e or -Ø is stored not as a property of the noun but as a property of the determiner de or het, such that de can be substituted by the options given in (15 a) and het by those in (15 b): (15) (a) de de [adj]e (b) het het [adj]e
een [adj]e een [adj]
If the use of adjectival morphology is based on a knowledge system as represented in (15), language learners can be expected to produce
errors such as: *de (grote) bos, *een grote bos, *het (dikke) boom, *een dik boom. However, one would not expect to find errors such as *de/*het dik boom or *de/*het groot bos. Analysis of a set of production errors (Jordens 1992) leads to the conclusion that this is indeed the case. L2 learners of Dutch make many errors as in (16). Here, the errors are due to incorrect gender. (16) Gelukkig mijn vader woont in een rustig wijk. (correct: rustige) ‘Fortunately, my father lives in a quiet neighbourhood.’ Hij is een betrouwbare mens. (correct: betrouwbaar) ‘He is a reliable person.’ Errors based on the use of an indefinite adjectival form in a definite context (*de/*het dik boom, *de/*het groot bos) do not occur. Moreover, there is evidence that learners using e-morphology with the adjective in an indefinite noun phrase use de as the definite article, while learners using Ø-morphology with the adjective use het. Hence, relations of adjectival agreement are learnt as syntagmatic patterns which constitute the morphological properties of either de- or het-nouns.
4.
Case marking
Structural case is usually seen as an example of contextual inflectional morphology, because it is the verb that determines both the number and the kind of case positions that are required. German, however, also allows a choice between dative and accusative morphology, i.e. between ihm and ihn in sentences as in (17). (17) Der Hund beißt ihm/ihn in the dog bite:3.sg 3.sg.m.dat/acc in das Bein. the leg ‘The dog bites him in the leg.’ Using experimental and corpus data, Zubin (1975; 1977; 1979) has argued that the use of dative vs. accusative morphology is determined by what he calls ‘degree of contribution’. The notion ‘degree of contribution’ is defined by Zubin as “the degree to which a participant is ‘agent-like’ in an event” (Zubin 1975: 174 f.). With respect to subject selection in particular, Ertel (1974; 1977) has shown in several experiments that it is the egocentric bias
1812
XIX. Psycholinguistische Perspektiven
which is the determining variable. This means that “symptoms of relative closeness between the subject element and the speaker’s ego” (Ertel 1977: 147) decide which participant is chosen as the subject of the sentence. In order to explain the use of case morphology in the L2 German of Dutch learners, I assumed that here too principles of egonearness and degree of contribution played a role (Jordens 1986; 1992). Use of these principles by L2 learners of German explains why in (18) the indirect object is incorrectly marked with nominative case, and why in (19) the grammatical subject is incorrectly marked with accusative case. (18) *Er wurde von uns geholfen. he was by us help:past.part (correct: Ihm) ‘We helped him.’ ist mir (19) *Den Sieg the victory:acc is me entgangen. escape:past.part (correct: Der Sieg) ‘Victory deluded me.’ Second, according to the principles underlying subject selection, sentences with a rhematic subject are processed as if they are subjectless. Hence, it can be understood why it is that with this type of subject L2 learners of German typically use incorrect accusative case. An example is given in (20). einen (20) *es wurde there become:past.3.sg indef:acc.sg.m Fall erwähnt case mentioned (correct: ein Fall) ‘a case was mentioned’ Third, object topicalization is a way of expressing ego-nearness of the object. Hence, Dutch learners of German often incorrectly use nominative morphology here. An example is given in (21). (21) *Jeder Republikaner betrachtete every Republican consider:past.3.sg er als seinen persönlichen Feind. he as his personal enemy (correct: Jeden Republikaner) ‘He considered every Republican a personal enemy.’
5.
Derivation
In French there are two morphological devices to derive nouns from verbs. From transitive verb stems, nouns can be formed with
the affix -age. An example is raffiner ‘to refine’ ⫺ raffinage. From intransitive verb stems, nouns can be formed with the affix -ment. An example is embarquer ‘to embark’ ⫺ embarquement. In an experiment on the role of error correction, Carroll et al. (1992) investigated whether native speakers of English are able to learn to use these noun formation rules productively in L2 French. Knowledge of noun formation in French was tested by having subjects give the corresponding noun to a given verb, as in the following test items: Dans cette usine, on RAFFINE le sucre ‘In this factory, they refine sugar’. Donc, c’est une usine de ... ‘So, it’s a refinery’. Le champagne doit PE´TILLER ‘Champagne should bubble’ ... est bon signe ‘This bubbling is a good sign’. “When the subjects in the experimental group made an error, they were given the correct response” (Carroll et al. 1992: 180). The aim was to find out if this procedure would enable adult learners to construct morphological generalizations. However, the results of the experiment were negative. Carroll et al. had to conclude that their experimental subjects were not able to induce “the subcategorization properties of the suffixes -age and -ment” (Carroll et al. 1992: 186). Whereas in this particular experiment Carroll et al. (1992) found no evidence that L2 learners of French learn the use of -age and -ment, there is evidence showing that in other circumstances L2 learners do learn to use L2 derivational properties. This evidence comes from research on the use of compensatory strategies. Poulisse (1995: 147), for example, noted that Dutch learners of English who did not know the lexical verb to iron produced the verb ironize instead, thus showing that they learnt to use the suffix -ize productively as a device to derive verbs from adjectives and nouns, as in normalize, foreignize, synchronize, womanize. It is a matter of investigation why it is that with respect to -age and -ment L2 learners were not able to discover the underlying generalization, whereas in the case of the suffix -ize they were. It seems reasonable to hypothesize that transfer plays a role here. While the difference between -age and -ment is a property of French which does not exist in English, there is a Dutch equivalent to the English suffix -ize. Therefore, as soon as Dutch learners of English discover that the L1 derivational morpheme -iseren can have the same function as -ize in English (as, for example,
1813
166. Second language acquisition
in normaliseren/normalize, verbaliseren/verbalize), it may be used as productively as the L1 equivalent.
6.
Compounding
In compounding, typological properties with respect to head-modifier relations become manifest (see Art. 165). In Dutch, for example, compounds have modifier-head structure. However, in comparing compounds such as afdelingshoofd ‘department-head’ with their [noun phrase ⫹ prepositional phrase] equivalent hoofd van de afdeling ‘head of the department’, it is obvious that, with regard to the structure of complex noun phrases, input data for L2 learners must be confusingly ambiguous: the same meaning can be expressed either by a modifier-head or a head-modifier structure. Interestingly, the same holds for nominal possessive constructions. A nominal possessive with a preposed genitive as in mijn vaders broer ‘my father’s brother’ can also be expressed as a [noun phrase ⫹ prepositional phrase] construction such as de broer van mijn vader ‘the brother of my father’. Particularly interesting with respect to the acquisition of compounding is a comparison of the acquisition of L2 Dutch by native speakers of Turkish and Moroccan. This is because typologically Turkish is structured in terms of modifier-head relations, whereas Moroccan has underlying head-modifier structure. Since both Turkish and Moroccan learners of Dutch encounter structures which are typologically similar to the way in which the head-modifier relations of their L1 are instantiated, there is reason for each of these learner groups to assume that Dutch is structured like their native language. Evidence from these learners’ data on the acquisition of compounds and nominal possessives shows that this is precisely what is found (Broeder 1991: 171 f.). As illustrated by the examples (22) and (23), Turkish informants prefer to use compounds or complex noun phrases with head-modifier structure, while the Moroccan subjects choose the [noun phrase ⫹ prepositional phrase] equivalent with modifier-head structure. (22) L1 Turkish: sigaretten-winkel cigarette-shop ‘cigarette shop’
L1 Moroccan: winkel van sigaret shop of cigarette ‘cigarette shop’ (23) L1 Turkish vader zus father sister ‘father’s sister (aunt)’ L1 Moroccan: die dochter van ander vrouw the daughter of other woman ‘the daughter of the other women’ An interesting example discussed recently by Clahsen (1991) is the acquisition of compounding in L2 German. In German, there is a relation between plural formation and compounding: plurals with -Ø (Lehrer ‘teachers’), -e (Schweine ‘pigs’), -e and Umlaut (Mäuse ‘mice’), -er (Kinder ‘children’), -er and Umlaut (Kräuter ‘herbs’), and -(e)n (Schrauben ‘screws’) are accessible to compounding, while plurals with the inflectional -s marker (Autos ‘cars’), are not. Examples are: Lehrermangel ‘teacher shortage’, Schweinefleisch ‘pork’, Mäusemännchen ‘little mouseman’, Kindergarten ‘kindergarten’, Kräuterkuchen ‘spicecake’, Schraubenzieher ‘screwdriver’, Autofriedhof ‘car cemetery’. Clahsen observed that native speakers of Spanish, Portuguese, and Italian initially used -e, -er, and -en as plural markers for compounding in L2 German (24 a). Furthermore, in some cases -en was left out in the early stages (24 b), while -s was never used (24 c). (24) (a) schweinefleisch ‘pork’ kindergarten ‘kindergarten’ schraubenzieher ‘screwdriver’ (b) *krankkasse (correct: Krankenkasse ‘health insurance’) *tapetefabrik (correct: Tapetenfabrik ‘wallpaper factory’) (c) kontonummer ‘bank account number’ Clahsen assumes innate level ordering to explain why the -s plural ⫺ which he regards as the default plural form ⫺ does not occur in L2 German compounding and why, at an intermediate stage of the acquisition of German at which -en is seen as the regular plural ending, -en does not occur in compounds either. Lardiere (1995) investigated the acquisition of compounding in the L2 English of native speakers of Spanish and Chinese. Sub-
1814 jects were asked to produce novel compounds in response to questions such as: “What could you call a person who cleans shoes/protects children/wears pants, etc.”. The aim of the experiment was to see whether subjects would produce plurals inside deverbal (synthetic) compounds. In the test questions, subjects were confronted with three types of plurals: regular plurals (babies, cats), irregular plurals (children, mice), and pluralia tantum (pants, people). Most of the 15 Spanish subjects used both regular and irregular plurals inside these compounds, there were only 2 subjects who correctly produced irregular plurals in compounds while omitting regular -s plurals. The Chinese speakers, on the other hand, while tending to omit the regular plural -s in compounds, also often used singular forms for irregular plurals and even for pluralia tantum. From this Lardiere concludes that “there is no evidence of the engaging of a universal constraint prohibiting regular plurals in compounds” (Lardiere 1995: 42). Furthermore she notes that given the fact that Spanish has plural objects in compounds, as in un lavaplatos ‘dishwasher’ and un abrelatas ‘can-opener’, whereas Chinese has no plural inflection at all, “the subjects’ L1 clearly seems to be having some kind of influence on their responses” (Lardiere 1995: 42). One may wonder how it is possible that Clahsen and Lardiere come to such contradictory conclusions. In order to understand why this is the case, one has to acknowledge that English and German are linguistically different as far as compounding is concerned. First, in English, the -s plural is the default while in German, contrary to Clahsen’s beliefs, it is not. As has been pointed out by Wegener (1994: 270), the -s plural in German has marked status, whereas due to high type frequency -e and -en are the default plural markers. Second, in German the -s occurring in compounds such as Arbeitsamt ‘employment service’ and Brechungsindex ‘refractive index’ is a linking morpheme and not a plural marker. Third, contrary to English, German never has the plural -s in compounds, whether in deverbal, as in Autoverkauf ‘car sale’ or in root compounds (see German Trambahn ‘tramway’ vs. English weapon(s) analysis). As far as second language acquisition is concerned, in L2 English learners are confronted with the regular plural -s in root
XIX. Psycholinguistische Perspektiven
compounds. Generalization from plurals in root compounding explains why L2 learners may incorrectly use -s in deverbal compounds as well. In L2 German, the plural -s is marked. Given the status of the plural -s in German, the differences between Spanish and Italian learners of German can be explained (see Clahsen 1991: 59). Due to their L1 only Spanish and Portuguese learners use -s initially incorrectly as in *grupps ‘groups’ (correct: Gruppen), *saches ‘things’ (correct: Sachen). However, as soon as they are confronted with more and richer input, these splurals die out quickly and are never used in compounds. Clahsen (1991: 62) also observed that at a certain point in L2 development the -n is sometimes left out in compounds like *krankkasse ‘health insurance’, *tapetefabrik ‘wallpaper factory’. Taking into account that Clahsen only found the examples given, these errors seem to be generalization errors in compounding due to the productive use of the singular form, and the plural -e and -en morphology. Overgeneralization of -e in compounds may be due to the high type frequency of -e plurals in isolation compared to their occurrence in compounds as well as, in the case of native speakers of Italian, to L1 influence.
7.
References
Andersen, Roger W. (1978), “An Implicational Model for Second Language Research”. Language Learning 28, 221⫺282 Andersen, Roger W. (1991), “Developmental Sequences: The Emergence of Aspect Marking in Second Language Acquisition”. In: Huebner, Thom & Ferguson, Charles A. (eds.), Crosscurrents in Second Language Acquisition and Linguistic Theories. Amsterdam: John Benjamins, 305⫺324 Antinucci, Francesco & Miller, Ruth (1976), “How Children Talk about What Happened”. Journal of Child Language 3, 169⫺189 Bailey, Nathalie & Madden, Carolyn & Krashen, Stephen D. (1974), “Is There a Natural Sequence in Adult Second Language Learning?”. Language Learning 24, 235⫺243 Bates, Elizabeth & MacWhinney, Brian & Smith, Stan (1983), “Pragmatics and Syntax in Psycholinguistic Research”. In: Felix, Sascha W. & Wode, Henning (eds.), Language Development at the Crossroads. Tübingen: Narr, 11⫺30 Bloom, Lois & Lifter, Karin & Hafitz, Jeremie (1980), “Semantics of Verbs and the Development
166. Second language acquisition of Verb Inflection in Child Language”. Language 56, 386⫺412 Booij, Geert E. (1994), “Against Split Morphology”. In: Booij, Geert E. & van Marle, Jaap (eds.), Yearbook of Morphology 1993. Dordrecht: Kluwer, 27⫺50 Broeder, Peter J. F. J. (1991), Talking about People: A Multiple Case Study on Adult Language Acquisition. Amsterdam, Lisse: Swets & Zeitlinger Bronckart, Jean-Paul & Sinclair, Hermine (1973), “Time, Tense, and Aspect”. Cognition 2, 107⫺130 Carroll, Susanne & Swain, Merrill & Roberge, Yves (1992), “The Role of Feedback in Adult Second Language Acquisition: Error Correction and Morphological Generalizations”. Applied Psycholinguistics 13, 173⫺198 Clahsen, Harald (1991), “German Plurals in Adult Second Language Development”. In: Theorie des Lexikons: Arbeiten des Sonderforschungsbereichs 282, Nr. 4 (Universal Grammar in Second Language Acquisition), 48⫺66 Coenen, Jose´e & Klein, Wolfgang (1992), “The Acquisition of Dutch”. In: Klein, Wolfgang & Perdue, Clive (eds.), Utterance Structure: Developing Grammars Again. Amsterdam: John Benjamins, 189⫺224 Dietrich, Rainer & Klein, Wolfgang & Noyau, Colette (1995), The Acquisition of Temporality in a Second Language Acquisition. Amsterdam: John Benjamins Dulay, Heidi C. & Burt, Marina K. (1974), “Natural Sequences in Child Second Language Acquisition”. Language Learning 24, 37⫺53 Dulay, Heidi C. & Burt, Marina K. (1975), “Creative Construction in Second Language Learning and Teaching”. In: Burt, Marina K. & Dulay, Heidi C. (eds.), New Directions in Second Language Learning, Teaching, and Bilingual Education. Washington, D.C.: TESOL, 21⫺32 Ertel, Suitbert (1974), “Satzsubjekt und Ich-Perspektive”. In: Eckensberger, Lutz H. & Eckensberger, Ute H. (eds.), Bericht über den 28. Kongreß der deutschen Gesellschaft für Psychologie in Saarbrücken. Vol. I: Wissenschaftstheorie und Psycholinguistik. Göttingen: Verlag für Psychologie, Hogrefe, 129⫺139 Ertel, Suitbert (1977), “Where Do the Subjects of Sentences Come From?”. In: Rosenberg, Sheldon (ed.), Sentence Production: Developments in Research and Theory. Hillsdale/NJ: Erlbaum, 141⫺ 167 Giacalone Ramat, Anna (1995), “Tense and Aspect in Learner Italian”. In: Bertinetto, Pier M. & Bi-
1815 anchi, Valentina & Dahl, Östen & Squartini, Mario (eds.), Temporal Reference, Aspect and Actionality. Vol. II: Typological Perspectives. Torino: Rosenberg & Sellier, 289⫺309 Housen, Alex (1993), “L2 Acquisition of Verb Morphology: A Case Study”. In: Kettemann, Bernhard & Wieden, Wilfried (eds.), Current Research in European Second Language Acquisition Research. Tübingen: Gunter Narr, 195⫺212 Jordens, Peter (1986), “Production Rules in Interlanguage: Evidence from Case Errors in L2 German”. In: Kellerman, Eric & Sharwood Smith, Michael (eds.), Crosslinguistic Influence in Second Language Acquisition. New York: Pergamon, 91⫺ 109 Jordens, Peter (1988), “The Acquisition of Word Order in L2 Dutch and German”. In: Jordens, Peter & Lalleman, Josine (eds.), Language Development. Dordrecht: Foris, 149⫺180 Jordens, Peter (1992), “The Cognitive Function of Case Marking”. In: Gass, Susan M. & Selinker, Larry (eds.), Language Transfer in Language Learning. Amsterdam: John Benjamins, 138⫺175 Klein, Wolfgang & Carroll, Mary (1992), “The Acquisition of German”. In: Klein, Wolfgang & Perdue, Clive (eds.), Utterance Structure: Developing Grammars Again. Amsterdam: John Benjamins, 123⫺188 Köpcke, Klaus-Michael (1987), “Die Beherrschung der deutschen Pluralmorphologie durch muttersprachliche Sprecher und L2 Lerner mit englischer Muttersprache”. Linguistische Berichte 107, 23⫺43 Krashen, Stephen D. (1977), “Some Issues Relating to the Monitor Model”. In: Brown, H. Douglas & Yorio, Carlos A. & Crymes, Ruth H. (eds.), Teaching and Learning English as a Second Language: Trends in Research and Practice. On TESOL ‘77. Washington D.C.: TESOL, 144⫺158 Lalleman, Josine A. & van Santen, Ariane J. & van Heuven, Vincent J. (1997), “L2 Processing of Dutch Regular and Irregular Verbs”. ITL, Review of Applied Linguistics 115⫺116, 1⫺26 Lardiere, Donna (1995), “L2 Acquisition of English Synthetic Compounding is Not Constrained by Level-Ordering (and Neither, Probably, is L1)”. Second Language Research 11, 20⫺56 Larsen-Freeman, Diane (1975), “The Acquisition of Grammatical Morphemes by Adult ESL Students”. TESOL Quarterly 9, 409⫺430 Phillips, Betty S. & Bouma, Lowell (1980), “The Acquisition of German Plurals in Native Children and Non-native Adults”. International Review of Applied Linguistics 18, 21⫺29
1816
XIX. Psycholinguistische Perspektiven
Pinker, Steven & Prince, Alan (1988), “On Language and Connectionism: Analysis of a Parallel Distributed Processing Model of Language Acquisition”. Cognition 28, 73⫺193 Poulisse, Nanda (1995). “Strategies”. In: Jordens, Peter & Lalleman, Josine (eds.), Investigating Second Language Acquisition. Berlin: Mouton de Gruyter, 135⫺163 Rumelhart, David E. & McClelland, James L. (1986), “On Learning the Past Tenses in English Verbs”. In: McClelland, James L. & Rumelhart, David E. & the PDP Research Group (eds.), Parallel Distributed Processing: Explorations in the Microstructure of Cognition. Vol. II: Psychological and Biological Models. Cambridge Mass.: Bradford Books, MIT Press, 216⫺271 Vogel, Thomas (1987), Die Entwicklung von Tempus und Aspekt im natürlichen Zweitsprachenerwerb. Universität Kiel: Englisches Seminar (Arbeitspapiere zum Spracherwerb 33) Wegener, Heide (1994), “Variation in the Acquisition of German Plural Morphology by Second Language Learners”. In: Tracy, Rosemarie & Lattey, Elsa (eds.), How Tolerant is Universal Grammar? Essays on Language Learnability and Language Variation. Tübingen: Max Niemeyer, 267⫺ 294
Wegener, Heide (1995), “The German Plural and Its Acquisition in the Light of Markedness Theory”. In: Pishwa, Hanna & Maroldt, Karl (eds.), The Development of Morphological Systematicity: A Cross-linguistic Perspective. Tübingen: Gunter Narr, 245⫺261 Weist, Richard M. (1986), “Tense and Aspect”. In: Fletcher, Paul & Garman, Michael (eds.), Language Acquisition: Studies in First Language Development. Cambridge Mass.: Cambridge University Press, 356⫺373 Zubin, David A. (1975), “On the Distributional Properties of Surface Morphology and their Consequences for Semantic Analysis”. Columbia University Working Papers on Linguistics 2, 189⫺218 Zubin, David A. (1977), “The Semantic Basis of Case Alternation in German”. In: Fasold, Ralph W. & Shuy, Roger W. (eds.), Studies in Language Variation. Washington D.C.: Georgetown University Press, 88⫺99 Zubin, David A. (1979), “Discourse Function of Morphology: The Focus System in German”. In: Givo´n, Talmy (ed.), Syntax and Semantics. Vol. XII: Discourse and Syntax. New York: Academic Press, 469⫺504
Peter Jordens, Amsterdam (The Netherlands)
167. Sprachstörungen 1. 2. 3. 4.
Einleitung Aphasie Sprachentwicklungsauffälligkeiten Zitierte Literatur
1.
Einleitung
Schon Francis Bacon hat erkannt, daß die Untersuchung des Pathologischen das Studium des Normalen fördert. In diesem Sinne können Sprachstörungen von Erwachsenen sowie Sprachentwicklungsauffälligkeiten bei Kindern psycho- und neurolinguistische Einblicke in das Wesen der Morphologie bieten, insbesondere in Prinzipien ihrer Speicherung und/oder Verarbeitung. Doch können Daten von Sprachstörungen keine direkte Evidenz für oder gegen Hypothesen eines linguistischen Modells liefern. Dazu müssen solche Daten erst im Lichte verschiedener, aber miteinander zu verbin-
dender theoretischer Modellbereiche interpretiert werden: (a) eines neuropsychologischen bzw. neuropsycholinguistischen Modells der entsprechenden sprachlichen Störung bzw. Entwicklungsauffälligkeit, (b) eines psycholinguistischen Modells normaler Sprachverarbeitung und -speicherung und unauffälligen Spracherwerbs, (c) eines linguistischen (insbesondere: morphologischen) Modells, (d) einer Brückentheorie, welche die Bereiche a) bis c) kombiniert bzw. interdisziplinär integriert (cf. allgemein Botha 1979). Einen (inzwischen überholten) klassischen Ansatz dieser Art stellen die Arbeiten von Luria dar (Luria 1970 und später). Da man aber in sonstigen Arbeiten derartig vollständig integrierte Ansätze selten findet, bleiben Schlußfolgerungen aus Sprachstörungen für
167. Sprachstörungen
morphologietheoretische Fragestellungen oft hypothetisch. Theoretische Fragen, auf welche sich ein Morphologiemodell Antworten erwarten kann, betreffen den differentiellen, nichtchaotischen Charakter spezifischer Sprachstörungen und Sprachentwicklungsauffälligkeiten. Diese können oder könnten erwartungsgemäß einzelne Bereiche ausschließlich oder vorwiegend (oder zumindest stärker) betreffen, z. B. a) die Grammatik, aber nicht Pragmatik oder Diskurs bzw. umgekehrt (s. 2.1), b) die Syntax oder die Morphologie oder das Lexikon (s. 2.2, 2.4, 3.1, 3.2) bzw. Morpheme gegenüber Wörtern, c) innerhalb der Morphologie die Flexion oder die Wortbildung (s. 2.4, 3.1), d) auch abstraktere Entitäten eines Morphologiemodells (s. 2.2, 3.1.2). In diesem Beitrag beschränken wir uns auf die wesentlichen, in der Literatur am besten dokumentierten Sprachstörungen und Sprachentwicklungsauffälligkeiten mit Schwerpunkten in der Morphologie. Erwähnenswert wären überdies u. a. Tiefendyslexie (cf. Bayer & De Bleser 1989; Coltheart et al. 1980; Lecours et al. 1989); Demenz (cf. Blanken 1986; Hamanaka et al. 1985; Schwartz 1990, Hrsg.); Schizophrenie (cf. Kraus & Mundt 1991, Hrsg.; Lecours 1993). Eine allgemeine Bibliographie geben Dittmann & Tesak (1993), einen allgemeinen Überblick Blanken et al. (1993, Hrsg.).
2.
Aphasie
2.1. Einleitung Aphasie ist eine erworbene, differentielle Sprachstörung bedingt durch eine zentrale Störung des sog. Sprachzentrums, d. h. der sprachrelevanten Verarbeitungs- und Speicherungsareale im Gehirn. Diese liegen in der linken Hirnhälfte – außer bei manchen Linkshändern, bei denen die rechte Hirnhälfte die sprachdominante ist. Läsionen dieser Areale der sprachdominanten Hirnhälfte bewirken eine Aphasie, entsprechende Läsionen der nicht-sprachdominanten Hirnhälfte rufen nach gängiger Meinung Störungen der Pragmatik hervor. Die Fragestellung, wie die Grammatik und Pragmatik verbindende Morphopragmatik, etwa der Diminutive (cf. Dressler & Merlini Barbaresi 1994), bei Hirnläsionen gestört ist, hat in diesem Zusammenhang einen eigenen Stellenwert. Morphologie als wesentlicher Teil der Grammatik
1817 stand stets im Zentrum der Interessen aphasiologischer Forschung (cf. die Überblicke in Cholewa 1993 und Badecker & Caramazza 1993). Bei der immer wieder umstrittenen Klassifizierung aphasischer Syndrome sind zumindest folgende Unterscheidungen wichtig: (a) Die Ätiologie der Aphasie, etwa Traumata (z. B. Schußwunden ⫺ in ihrer Wirkung oft eng eingegrenzt ⫺ spielten besonders nach dem Ersten Weltkrieg eine große Rolle in der klassischen deutschen Aphasielehre, die sich bei Jakobson (1941) niederschlug, oder nach dem Zweiten Weltkrieg bei Luria (1970); Schlaganfälle (vaskuläre Ätiologie, mit meist weniger umgrenzten Ausfällen); Tumore. (b) Die neurologische Lokalisierung der Läsion (z. B. im Frontallappen bei BrocaAphasie, im Temporallappen bei Wernicke-Aphasie) spielt heute wegen der ⫺ in einer Brückentheorie unterzubringenden ⫺ schwierigen und mehrdeutigen Verbindung von Lokalisation und anderen Datenbereichen eine geringere Rolle als (c) die Symptomatik. So unterscheidet man aufgrund der Flüssigkeit der Sprachproduktion allgemein zwischen flüssiger Aphasie (fluent aphasia) und nicht-flüssiger Aphasie (non-fluent aphasia). Unter den flüssigen Aphasien ist die Wernickeoder sensorische Aphasie mit ihrem Paragrammatismus in unserem Zusammenhang am wichtigsten (cf. 2.3). Kennzeichnend für die nicht-flüssige Broca- oder motorische Aphasie ist der Agrammatismus, benannt nach den dominanten Beeinträchtigungen in der grammatischen Produktion (cf. 2.2). (d) Einen weiteren Klassifizierungsfaktor bildet der Schweregrad der Störung. Zur Abgrenzung “pathologischer” von “normaler” Sprache sind neben Vergleichen von Aphasikern mit korrespondierenden (Alter, Bildung, Geschlecht, etc.) ungestörten Kontrollpersonen auch Vergleiche von aphasischen Fehlern mit alltäglichen sprachlichen Fehlleistungen (speech errors; s. Art. 164) wichtig, um festzustellen, ob sie sich nur quantitativ oder auch qualitativ unterscheiden (cf. Dressler 1987; Stemberger 1984). 2.2. Agrammatismus Nach dem ursprünglichen Konzept der motorischen Aphasie handelte es sich dabei um ein reines Produktionsdefizit ohne Störung des
1818 Sprachverständnisses (cf. z. B. Miceli et al. 1983; Nespoulous et al. 1988; Tesak 1990). Die Sprachproduktion ist durch einen sog. Telegrammstil oder Agrammatismus gekennzeichnet. Zu den klinischen Charakteristika des Agrammatismus gehören: Die langsame, stockende Produktion von Kurzsätzen, das Fehlen satzgrammatischer Funktionswörter (Artikel, Präpositionen, Auxiliarverben, nicht aber satzverbindender Konjunktionen) und das Fehlen sowie der Ersatz von Flexionsmarkern (cf. Menn & Obler 1990, Hrsg.), wie im folgenden Beispiel: “ja! ich äh immer äh äh denken äh ihr? ihr Buch? aha! richtig! dann immer denken! und äh äh sprechen ... und äh und jemand äh zu Hause mein Mann und ich Einladung äh eingeladen.” (aus Peuser 1978: 127)
Inzwischen ist allgemein sichergestellt, daß die Mehrheit ⫺ aber eben nicht alle – der Patienten mit Agrammatismus parallele Grammatikdefizite auch beim Verständnis aufweisen (Ausnahmefälle ohne Verständnisprobleme in Kolk et al. 1985; Miceli et al. 1983; Nespoulous et al. 1988). Traditionell wurde der Agrammatismus als eine zentrale Störung der Syntax interpretiert. Inzwischen variieren die Meinungen über die Natur der zugrundeliegenden Störung sehr. Einerseits bestehen Divergenzen über den primär betroffenen sprachlichen Bereich: (a) die Störung betrifft sowohl Syntax als auch Morphologie (z. B. Miceli et al. 1983); (b) die Störung betrifft nur die Morphologie, indem der Zugriff zu den morphologischen Funktionswörtern und Affixen geschädigt ist (z. B. Bradley et al. 1980), was von Kean (1979) auf (c) phonologische Gründe zurückgeführt wird (cf. Nespoulous & Dordain 1990); (d) die primäre Störung liegt am Schnittpunkt von Morphologie und Syntax, nämlich bei der Kongruenz, besonders wenn sie über die Nominalphrase hinausgeht (z. B. Bayer et al. 1987); (e) die primäre Störung betrifft die Syntax (nur sekundär die Morphologie) und besteht in einem Defizit der funktionalen Kategorien: Es fehlen die Kategorien selbst oder ihre syntaktisch-semantischen Merkmale, etwa Finitheit, Tempus, etc. (z. B. Caplan 1985; Grodzinsky 1984; Ouhalla 1993; s. auch Hackl 1996). Andererseits siedeln immer mehr Ansätze die zugrundeliegende Störung nicht im sprachlichen Wissenssystem sondern in der Verarbeitung an. Dafür spricht die Tatsache, daß Agrammatiker bei Grammatikalitätsur-
XIX. Psycholinguistische Perspektiven
teilen ⫺ auch in der Morphologie ⫺ wesentlich besser abschneiden als in Produktion und Verständnis (cf. Caramazza & Hillis 1989; Linebarger et al. 1983; Nespoulous et al. 1988). Wichtige Informationen müssen demnach erhalten geblieben sein. Linebarger et al. (1983; s. auch Schwartz et al. 1987) schlagen eine Störung in der Übertragung syntaktischer Funktionen auf thematische Rollen vor (“mapping”-Hypothese). Für Friederici (z. B. 1988) ist Agrammatismus die Folge eines pathologischen Verfalls der syntaktischen Information: Zwar ist diese vorhanden aber bereits nicht mehr parat im Zeitrahmen, in dem die morphologische Verarbeitung folgen sollte. Auch Kolk und Kollegen (z. B. Haarmann & Kolk 1994; Hofstede & Kolk 1994; Kolk et al. 1985; Kolk & Heeschen 1990) gehen in ihrer “Adaptionstheorie” und “strategic variation”-Hypothese (cf. Isserlin 1922) von Beschränkungen in der Verarbeitung aus, an die sich Agrammatiker aufgabenspezifisch adaptieren ⫺ durch Vereinfachungen und Verwendung unterschiedlicher Register (Bastiaanse 1995). 2.3. Paragrammatismus Paragrammatismus charakterisiert die Produktion bei der sensorischen oder WernickeAphasie. Sie ist im Gegensatz zum Agrammatismus flüssig und in ihrer syntaktischen Komplexität vergleichbar mit normaler Sprache. Allerdings kommt es gehäuft zu Verschränkungen im Satz, die aus einem Zusammenbruch des Monitoring und aus anderen Defiziten in der Sprachverarbeitung resultieren dürften. Eine paradigmatische Selektionsstörung ist als Ursache für die typisch auftretenden Vertauschungen vorgeschlagen worden (Jakobson 1964). Ähnlich wie in Syntax und Lexikon werden auch Flexionsformen ⫺ ohne Rücksicht auf die syntaktische Struktur – miteinander vertauscht, und zwar sowohl in der Spontansprache als auch in Tests aller Art. So wie Wortvertauschungen zumeist im selben Wortfeld verbleiben, so gehören die miteinander vertauschten Affixe gewöhnlich derselben Klasse an, z. B. Suffixe werden nicht mit Präfixen vertauscht, verbale nicht mit nominalen Flexionsaffixen, und selten Wortbildungs- mit Flexionsaffixen, wie z. B. in be-schlagen für ge-schlagen (cf. Bayer et al. 1987; Butterworth & Howard 1987; Dressler & Denes 1987). Hingegen sind Substitutionen zwischen Suffixen derselben Klasse gut bezeugt, z. B. zwischen Flexionssuffixen in ge-fang-t statt dem Partizipium
167. Sprachstörungen
des starken Verbs ge-fang-en. WernickeAphasiker produzieren auch oft morphologische Neologismen, d. h. nichtexistierende, aber meist potentiell mögliche (legale) Wörter, mit Hilfe von Wortbildungsregeln oder Analogie (cf. Buckingham 1993). Auch Kontaminationen kommen vor, wie z. B. plötzartig I plötzlich & blitzartig. Obwohl solche Paraphasien auch bei Broca-Aphasie auftreten (z. B. legal patriarch-isch statt patriarch-alisch, illegal Hochschul-schaft für Hochschüler-schaft), sind sie doch typischer für Wernicke-Aphasiker. Eine Sonderform der sensorischen Aphasie ist die Jargonaphasie, bei der ⫺ jargonartig ⫺ viele Pseudowörter erzeugt werden, die in ihrer phonologischen Gestalt aktuellen Wörtern ähneln, aber zu diesen weder in morphologischen Beziehungen stehen, noch durch phonologische Substitutionen aus ihnen abgeleitet werden können. Diese Jargonwörter können allerdings korrekt flektiert werden (cf. Buckingham 1993; Badecker & Caramazza 1993: 184), z. B. Engl. Plural chpicters. Auch produktive Anwendung von Wortbildungsregeln ist bezeugt, z. B. in ital. fratellismo ‘Bruderschaft’ als Abstrakta-Bildung von fratello ‘Bruder’ (cf. Panzeri et al. 1990). 2.4. Differentielle Störungen der Morphologie Bei morphologischen Paraphasien als typischen Substitutionsfehlern bei Aphasie stellt sich die Frage, ob es sich um Substitutionen ganzer Wörter bzw. Wortformen oder einzelner Morpheme handelt (cf. Badecker & Caramazza 1993). Hier bietet die unterschiedliche Fehleranfälligkeit verschiedener morphologischer Konstruktionen gute Anhaltspunkte. Zu den Faktoren, die bei der variablen Fehleranfälligkeit eine Rolle spielen, gehört einerseits die Vorkommensfrequenz (token frequency), aber eben nicht nur diejenige von Wörtern bzw. Wortformen, sondern auch die von Morphemen (cf. Badecker & Caramazza 1993: 182). Andererseits spielt auch die morphosemantische und morphosyntaktische Transparenz komplexer Wörter eine Rolle (cf. Dressler & Denes 1987). Z. B. waren italienische Broca- und Wernicke-Aphasiker deutlich besser beim Verstehen und der korrekten Identifizierung von morphosemantisch transparenten Komposita des Typs porta-lettere ‘Briefträger’ als von opaken des Typs mangia-preti ‘scharfer Antiklerikaler’ (wörtlich ‘Priesterfresser’). Ebenso war die Produktion von morphotaktisch transparen-
1819 ten Derivationen wie vend-i-tore ‘Verkäufer’ (zu vendere ‘verkaufen’) deutlich besser als von opaken wie scrittore ‘Schriftsteller’ (zu scrivere ‘schreiben’). Diese Auswirkungen sowohl von Frequenz als auch von Transparenz setzen Morphemzerlegung, d. h. die Verarbeitung der einzelnen Morpheme, voraus (cf. Cholewa 1993: 106⫺110). Erwähnenswert sind auch differentielle Störungen von Wurzeln und Affixen (cf. Miceli 1995; Tyler et al. 1990) sowie von Flexion und Derivation (cf. Badecker & Caramazza 1993: 183 f.; Cholewa 1993: 114⫺118). So ist bei Broca-Aphasie die Flexion, nicht aber die Derivation beeinträchtigt (cf. Cholewa & De Bleser 1995; Laine et al. 1994; Menn & Obler 1990, Hrsg.; Tyler & Cobb 1987). Hingegen kommen bei der Wernicke-Aphasie und der amnestischen Aphasie (anomia, charakterisiert durch Wortfindungsstörungen) viel mehr Fehler in der Wortbildung vor (cf. Kudo 1992). Die isolierte Bewahrung der Morphologie einschließlich der Morphosyntax ohne Zugang zur lexikalischen Semantik bei sensorisch-transkortikaler Aphasie ist in Dogil et al. (1995) dokumentiert. 2.5. Verschiedene Sprachen ⫺ verschiedene Störungen? Die Tatsache, daß Agrammatismus signifikante Störungen der Flexionsmorphologie hervorruft, hat zur Folge, daß Sprachen mit reicher und komplexer Flexionsmorphologie, wie die flektierenden (z. B. Polnisch) oder die introflektierenden (z. B. Hebräisch) bei Agrammatismus ganz offensichtlich stärker betroffen sind als solche mit geringer Flexionsmorphologie, wie z. B. Chinesisch, aber auch Englisch. Sprachvergleichende Daten und deren Analyse zeigen solche Unterschiede (cf. Friederici et al. 1991 (Vergleich Deutsch, Französisch, Niederländisch); Haverkort 1993 (Japanisch, Englisch); Menn & Obler 1990, Hrsg. (Daten aus 14 Sprachen); Sonderheft 2 des Nordic Journal of Linguistics 16 (1993) zu skandinavischen Sprachen). Aufschlußreich sind auch Untersuchungen bilingualer Patienten (z. B. Laine et al. 1994). Typologische Unterschiede in der einzelsprachlichen Flexionsmorphologie erklären auch Unterschiede in der relativen Häufigkeit von Auslassung versus Substitution von Flexionsendungen, da eine generelle Tendenz darin besteht, markierte durch unmarkierte Flexionsformen zu ersetzen. Je nachdem welche Form der unmarkierten Flexion vorliegt (Nullform/Grundformflexion versus eine spe-
1820
XIX. Psycholinguistische Perspektiven
zifische flektierte Form) kommt es entweder zu Auslassungen bzw. zu Substitutionen (z. B. die Auslassung der Verbendung -s bei der ‘3.sg.pr‰s’ im Englischen versus der Ersatz der markierten (“starken”) Partizipform gefang-en durch die weniger markierte (“schwache”) Form ge-fang-t, nicht aber durch die endungslose Form ge-fang; cf. Stark & Dressler 1990: 367; Mimouni & Jarema 1995). Oberflächlich vergleichbare morphologische und morphonologische Phänomene, welche in typologisch unterschiedlichen Sprachen eine verschiedene Rolle spielen, sind erwartungsgemäß verschieden gestört: Z. B. ist der deutsche Umlaut in Flexion und Wortbildung bei Agrammatismus stark gestört, die ungarische Vokalharmonie, weil regelmäßig, so gut wie gar nicht (cf. Dressler et al. 1996).
3.
Sprachentwicklungsauffälligkeiten
3.1. Entwicklungsdysphasie 3.1.1. Deskriptive Charakterisierung Die Bezeichnung Entwicklungsdysphasie (Spezifische Sprachentwicklungsstörung ⴚ SSES developmental dysphasia, specific language impairment ⫺ SLI, developmental language disorder) wird in der Literatur auf eine heterogene Gruppe von Kindern angewandt, bei denen isoliert die Sprachentwicklung verspätet einsetzt und mit einer Beeinträchtigung in der Entwicklung der Sprachproduktion und/ oder des Sprachverständnisses einhergeht. Eine mentale Retardierung, eine Hörbehinderung, psychische Störungen wie z. B. Autismus, psychosoziale Beeinträchtigungen oder eine nachweisbare neurologische Störung liegen dabei nicht vor. Nach ihrem Erscheinungsbild lassen sich homogenere Untergruppen unterscheiden: Einerseits eine vorwiegend expressive vs. eine expressiv-rezeptive Form; andererseits (a) eine rein morphologische/syntaktische Beeinträchtigung (Bishop 1994; Leonard 1989; van der Lely 1993; 1996) vs. (b) eine Form mit zusätzlichen artikulatorischen und/oder phonologischen Defiziten (Gopnik & Crago 1991) und (c) eine Form mit zusätzlichen semantisch-pragmatischen Auffälligkeiten (cf. Bishop & Adams 1989; Mills et al. 1992). Bis heute ist ungeklärt, ob es sich bei diesen Erscheinungsformen um unterschiedliche Beeinträchtigungen mit unterschiedlichen Ätiologien oder bloß um eine einzige Beeinträchtigung in ihren verschiedenen Gradausprägungen bzw. zu unterschiedlichen Entwicklungszeitpunkten
handelt (cf. Bishop et al. 1995). Auch wenn die Ätiologie/n für die Entwicklungsdysphasie bis jetzt unbekannt ist/sind, konnte zumindest in bestimmten Fällen eine erbliche Komponente nachgewiesen werden (cf. Tomblin 1997; in Zwillingsuntersuchungen, cf. Bishop et al. 1995; in Familien mit mehreren dysphasischen Generationen, cf. Gopnik & Crago 1991; Plante 1991; Tallal et al. 1989; 1991; Tomblin 1989; zu dysphasischen Familien siehe speziell Matthews 1994, Hrsg., Paradis 1997, Hrsg.; zur Dysphasie im Allgemeinen siehe auch Leonard 1998, Kamhi 1993, Menyuk 1993). Als Hauptmerkmal aller Formen der Entwicklungsdysphasie (und deshalb auch zentraler Forschungsgegenstand) werden in der Literatur vorwiegend speziell ausgeprägte Probleme mit Flexionsmorphologie, grammatischen Funktionswörtern und Morphosyntax genannt (cf. Watkins & Rice 1994, Hrsg. für eine Übersicht der diesbezüglichen Literatur). Diese äußern sich in einem nichtaltersadäquaten, eingeschränkten Gebrauch von Flexionsmarkern und Funktionswörtern, die häufig nicht vollständig erworben werden. “Fehlertypen” und Erwerbsabfolgen sind in der Regel mit denen von jüngeren unauffälligen Kindern vergleichbar. Konkret werden in der Literatur folgende Charakteristika der Sprache von dysphasischen Kindern allgemein hervorgehoben: (a) Auslassungen und die Verwendung von unmarkierten Formen sind typisch; (b) Übergeneralisierungsfehler (z. B. durch Anwendung einer regelmäßigen Regel) treten laut den meisten Autoren nie auf (u. a. Gopnik 1994), laut anderen gibt es sie wohl, jedoch kommt es nie zu kategorieübergreifenden Fehlern (z. B. Leonard & Bortolini et al. 1992; Rice et al. 1995; Rothweiler & Clahsen 1994; SmithLock 1993); (c) frequente sowie unregelmäßige Formen werden relativ besser beherrscht; (d) neben extrem betroffenen Bereichen ⫺ wie z. B. Subjekt-Verb-Kongruenz, Auxiliare, Kopula, Artikel sowie Kasus- und Kongruenzmarkern in der Nominalphrase ⫺ zeigen sich andere morphologische Aspekte vergleichsweise weniger betroffen und dem jeweiligen Sprachentwicklungsstand des Kindes zumindest angemessen ⫺ wie etwa die Pluralbildung von Nomina, die Vergangenheitsformen von unregelmäßigen Verben, die Partizip-
167. Sprachstörungen
bildung, der durative Aspekt auf -ing bei englischen Verben (cf. u. a. Bartke 1998; Clahsen et al. 1992; Goad 1998; Gopnik 1990; Gopnik & Crago 1991; Leonard 1989; Oetting & Rice 1993; Rice 1994; Rothweiler & Clahsen 1994; Ullman & Gopnik 1994). Wie in 2.5 bei den Aphasien bereits angesprochenen, gibt es auch beim Erscheinungsbild der Dysphasie eine Abhängigkeit von typologischen Unterschieden in der einzelsprachlichen Flexionsmorphologie. So konnten z. B. die fürs Englische allgemein betonten, massiven Schwierigkeiten mit der Subjekt-VerbKongruenz bzw. der Verbflexion (d. h. anstelle von flektierten, finiten Verbformen werden vorwiegend Infinitivformen oder eventuell Verbstämme verwendet) zwar für viele andere ⫺ aber nicht für alle ⫺ untersuchten Sprachen bestätigt werden (cf. Miller & Leonard 1998): z. B. für Deutsch (u. a. Clahsen 1988; Rice et al. 1997), für Englisch (u. a. Rice et al. 1995), für Französisch (Gopnik 1990), für Inuktitut (Crago & Allen 1994), für Japanisch (Fukuda & Fukuda 1994), für Niederländisch (Bol & Kuiken 1990), für Schwedisch (Hansson 1992); nicht aber für Italienisch (Leonard & Bortolini et al. 1992; Bortolini & Leonard 1996) und Hebräisch (Dromi et al. 1993; Leonard & Dromi 1994). Diese sprachspezifischen Unterschiede dekken sich mit entsprechenden sprachvergleichenden Daten von unauffälligen Kindern (cf. dazu auch Wexler 1994). Leonard (1998) widmet dem Thema “SLI across languages” sogar ein eigenes Kapitel. Spezifische Auffälligkeiten in der Wortbildung wurden in der Literatur zur Dysphasie bislang negiert bzw. nicht erwähnt. Einzelne Untersuchungen berichten hier jedoch über ähnliche Probleme wie in der Flexionsmorphologie und vermuten demnach bei der Dysphasie ein allgemeines Unvermögen “to abstract bound morphemes (be they affixes or roots) and represent them as such” (Dalalakis & Gopnik 1995: 202). 3.1.2. Hypothesen über das der Dysphasie zugrundeliegende Defizit Hypothesen über das der Dysphasie ursächlich zugrundeliegende Defizit gibt es in der Literatur viele und divergierende, zum Teil allerdings durchaus komplementäre (einen detaillierteren Überblick und Gegenüberstellung bieten u. a. Bishop 1992; 1997; Crago & Allen 1994; Gopnik 1994; Leonard 1998): Ei-
1821 nerseits werden unterschiedliche, spezifische Probleme beim Erwerb des grammatischen Systems, andererseits mehr allgemeine Lernund/oder Verarbeitungsdefizite postuliert. Grammatikspezifische Ansätze erklären die morphologischen Auffälligkeiten der Dysphasie entweder als ein Defizit in der grammatischen Kongruenz (“grammatical agreement deficit”-Hypothesen) oder als ein morphologisches Musterbildungsdefizit. Das Kongruenzdefizit wird unterschiedlich bewertet: Es ist ein generelles (Clahsen 1988; Rothweiler & Clahsen 1994; Clahsen & Hansen 1997) oder besteht darin, daß syntaktisch-semantische Merkmale (wie Finitheit, Tempus, Aspekt etc.) fehlen und daher nicht markiert werden können (“feature deficit”Hypothese, Gopnik 1990; Gopnik & Crago 1991). Merkmale mögen zwar vorhanden sein, aber ihre Abhängigkeiten – entweder insgesamt oder partiell ⫺ werden unzureichend verwaltet (“feature checking deficit”Hypothese, Ullman & Gopnik 1994; ausschließlich “Spec-Head”-Kongruenz: Rice 1994; van der Lely 1994; 1996), oder Merkmale werden fälschlich als optional behandelt (“extended optional infinitive”-Hypothese, Rice et al. 1995; 1997). Dem morphologischen Musterbildungsdefizit wird entweder ein “paradigm formation”-Defizit zugrunde gelegt, oder es wird dafür eine “missing-rule”Hypothese bzw. eine “impaired morphological rule-construction”-Hypothese vorgeschlagen (Gopnik 1992; 1994; Gopnik & Crago 1991; Ullman & Gopnik 1994; Gopnik et al. 1997). Demnach werden morphologisch komplexe Wörter unanalysiert gespeichert, keine morphologischen Regeln entwickelt und morphologische Defizite durch konzeptuell basierte Rekategorisierungen kompensiert. Hypothesen, die die dysphasischen Probleme auf allgemeine Lern- und/oder Verarbeitungsdefizite zurückführen, nehmen einerseits ein nicht-sprachspezifisches, allgemeines Muster- und Regelbildungsdefizit an, woraus ein “unique language learning” (Connell 1987; Swisher et al. 1995) bzw. ein abweichendes sprachliches System resultiert (Restrepo et al. 1992). Andererseits werden spezifische Defizite in der auditiv-perzeptuellen Verarbeitung, Speicherung und Wiederabrufbarkeit vermutet: Zu geringe phonetische Substanz der morphologischen Marker (in der sog. “surface”-Hypothese, u. a. Leonard 1989; 1992; Leonard et al. 1987; Leonard & McGregor & Allen 1992; Leonard & Bortolini et al. 1992; cf. auch Fellbaum et al. 1995);
1822 zu beschränktes phonologisches Gedächtnis (Gathercole & Baddeley 1990) oder Schwierigkeiten in der prosodischen Verarbeitung (Bortolini & Leonard 1996) werden als mögliche Ursachen für die Defizite genannt. In ihrer ursprünglichen Formulierung hat sich Leonards “surface”-Hypothese inzwischen ⫺ speziell auch im Sprachvergleich ⫺ als unhaltbar erwiesen (s. Leonard 1998: 246 ff. für eine kritische Auseinandersetzung damit). Die im Sprachvergleich beobachteten zwischensprachlichen Unterschiede bei specific language impairment werden neuerdings im Rahmen von sog. “morphological richness” Ansätzen dahingehend erklärt, daß “[...] children with SLI devote their limited processing resources to a dominant property of the language and by necessity, allow other aspects of the language to falter” (Leonard 1998: 257). 3.2. Mentale Retardierung Zahlreiche Studien sprachunauffälliger Kinder (u. a. Schaner-Wolles & Haider 1987) und mental retardierter Personen liefern starke empirische Evidenz für die Modularitätsthese (cf. Fodor 1983; Chomsky 1982), d. h. für eine Unabhängigkeit bzw. Dissoziation von grammatischen und allgemein-kognitiven Fähigkeiten. 3.2.1. Sprachliche Sonderbegabung bei mentaler Retardierung Trotz mentaler Retardierung können ausgesprochen fortgeschrittene grammatische Fähigkeiten ⫺ auch morphologische ⫺ erworben werden, wie eine Reihe von Fallstudien sog. “hyperverbaler” Personen (engl. hyperlinguistic) in der Literatur belegen: z. B. Laura (⫽ Marta), Rick und Antony (cf. u. a. Curtiss 1988; Yamada 1990); Christopher (Smith & Tsimpli 1995); das Mädchen D. H. (Cromer 1994); Personen mit Williams-Beuren-Syndrom (cf. u. a. Bellugi & Marks et al. 1988; Clahsen & Almazan 1998; Levy & Hermon 2000; und den Literaturüberblick in Schaner-Wolles 2000a); Franc¸oise mit DownSyndrom (Rondal 1994; 1995). Besondere Beachtung in morphologischer Hinsicht verdient Christopher, ein sog. “linguistic savant”, der im Alter von 33 Jahren trotz mentaler Behinderung neben Englisch als seiner Erstsprache zumindest 18 weitere Sprachen (genetisch und typologisch gestreut) erworben hatte (darunter eine zu Testzwecken eigens entwickelte Kunstsprache Epun mit SVO Struktur, aber einem äußerst komplexen Kongruenzsystem). Diese besondere Bega-
XIX. Psycholinguistische Perspektiven
bung charakterisieren Smith und Tsimpli wie folgt: “One aspect of Christopher’s exceptionality resides precisely in his sensitivity to and learning of complex morphology” (Smith & Tsimpli 1995: 124). Mehr oder weniger ausgeprägte Diskrepanzen zwischen unterschiedlichen grammatischen Bereichen dürften ⫺ wie auch bei der Entwicklungsdysphasie ⫺ bei mentalen Retardierungen typisch sein. Anders als bei Christopher ist bei einer solchen disharmonischen Entwicklung die Morphologie häufig selektiv beeinträchtigt: z. B. fortgeschrittene Syntax versus flexionsmorphologische Probleme bei Rick (cf. auch Cromer 1988); gute flexionsmorphologische Leistungen in der Produktion versus schwache bei Verständnistests bei Laura (Yamada 1990: 35⫺39); ausgeprägte morphologische Schwächen auch beim Down-Syndrom (cf. 3.2.2). 3.2.2. Down-Syndrom Abgesehen von wenigen Ausnahmefällen wie Franc¸oise, einer damals 34-jährigen Frau mit Down-Syndrom (DS), deren grammatische (auch flexionsmorphologische) Fähigkeiten Rondal (1994; 1995) als normal oder quasinormal bezeichnet, ist die Sprachentwicklung bei Down-Syndrom verzögert und erreicht auch im Erwachsenenalter höchstens den Entwicklungsstand eines 6⫺7-jährigen unauffälligen Kindes. Es zeigen sich Disharmonien zwischen den einzelnen sprachlichen Bereichen, wobei morphologische und speziell morphosyntaktische Fähigkeiten am stärksten zurückbleiben (cf. z. B. Fowler 1988; 1990; Fowler et al. 1994; Rondal 1994; 1995; Schaner-Wolles 1992; 1994; 2000 a; 2000 b). In der Entwicklung der einzelnen sprachlichen Fähigkeiten unterscheiden sich Personen mit Down-Syndrom jedoch qualitativ nicht von unauffälligen jüngeren Kindern. Dies gilt auch für die Morphologie (cf. Ergebnisse aus Flexionsmorphologie, Morphosyntax und Derivationsmorphologie von 82 deutschsprachigen Kindern, Jugendlichen und Erwachsenen mit Down-Syndrom in Schaner-Wolles 1989; 1992; 1994; SchanerWolles & Dressler 1986). 3.3. Hörbehinderung Beim Erwerb oraler Sprache kommt es aufgrund des auditiven Wahrnehmungsdefizits generell zu einer zeitlichen Verzögerung. Speziell der Erwerb von unbetonten Flexionsendungen ist davon betroffen, da sie auditiv besonders schlecht oder überhaupt nicht wahr-
167. Sprachstörungen
genommen werden können (cf. Cantwell & Baker 1987; de Villiers et al. 1994; MogfordBevan 1993; Wisch 1990.) Aber sogar gehörlosen Kindern ist es möglich, die wichtigsten morphologischen Strukturen ihrer Muttersprache vollständig zu erwerben, wobei weniger der Grad des Hörverlustes als vielmehr die Frühförderung den Erfolg bestimmt (cf. Peltzer-Karpf 1994). Beim erstsprachlichen Erwerb einer Gebärdensprache (s. Art. 143) verläuft die morphologische Entwicklung parallel zu der von oralen Sprachen bei nicht-hörbehinderten, unauffälligen Kindern (cf. Newport & Meier 1985; Bellugi & van Hoek et al. 1988).
4.
Zitierte Literatur
Badecker, William & Caramazza, Alfonso (1993), “Disorders of Lexical Morphology in Aphasia”. In: Blanken et al. (Hrsg.), 181⫺186 Bartke, Susanne (1998), Experimentelle Studien zur Flexion und Wortbildung: Pluralmorphologie und lexikalische Komposition im unauffälligen Spracherwerb und im Dysgrammatismus. Tübingen: Niemeyer Bastiaanse, Roelien (1995), “Broca’s Aphasia: A Syntactic and/or a Morphological Disorder? A Case Study”. Brain and Language 48, 1⫺32 Bayer, Josef (1987, Hrsg.), Grammatik und Kognition: Psycholinguistische Untersuchungen. Opladen: Westdeutscher Verlag [Linguistische Berichte, Sonderheft 1] Bayer, Josef & Bleser, Ria De (1989), “Lexikalische Morphologie und Tiefendyslexie: Eine Fallbesprechung”. In: Günther (Hrsg.), 123⫺154 Bayer, Josef & Bleser, Ria De & Dronsek, Claudia (1987), “Form und Funktion von Kasus bei Agrammatismus”. In: Bayer (Hrsg.), 118⫺163 Bellugi, Ursula & Hoek, Karen van & Lillo-Martin, Diane & O’Grady, Lucinda (1988), “The Acquisition of Syntax and Space in Young Deaf Signers”. In: Bishop & Mogford (Hrsg.), 132⫺149 Bellugi, Ursula & Marks, Shelly & Bihrle, Amy & Sabo, Helene (1988), “Dissociation between Language and Cognitive Functions in Williams Syndrome”. In: Bishop & Mogford (Hrsg.), 177⫺189 Bishop, Dorothy V. M. (1992), “The Underlying Nature of Specific Language Impairment”. Journal of Child Psychology and Psychiatry 33, 3⫺66 Bishop, Dorothy V. M. (1994), “Grammatical Errors in Specific Language Impairment: Competence or Performance Limitations?”. Applied Psycholinguistics 15, 507⫺550
1823 Bishop, Dorothy V. M. (1997), Uncommon Understanding: Development and Disorders of Language Comprehension in Children. Hove: Psychology Press Bishop, Dorothy V. M. & Adams, Catherine (1989), “Conversational Characteristics of Children with Semantic-Pragmatic Disorder. 2: What Features Lead to a Judgement of Inappropriacy?”. British Journal of Disorders of Communication 24, 241⫺ 264 Bishop, Dorothy & Mogford, Kay (1988, Hrsg.), Language Development in Exceptional Circumstances. Edinburgh: Churchill Livingstone [⫽ 1993 Hove, Hillsdale: Lawrence Erlbaum] Bishop, Dorothy V. M. & North, T. & Donlan, C. (1995), “Genetic Basis of Specific Language Impairment: Evidence from a Twin Study”. Developmental Medicine and Child Neurology 37, 56⫺71 Blanken, Gerhard (1986), Gestörtes Sprachverhalten bei seniler Demenz und Aphasie. Dissertation Universität Freiburg Blanken, Gerhard & Dittmann, Jürgen & Grimm, Hannelore & Marshall, John C. & Wallesch, ClausW. (1993, Hrsg.), Linguistic Disorders and Pathologies. Berlin: de Gruyter Bol, Gerard & Kuiken, Folkert (1990), “Grammatical Analysis of Developmental Language Disorders: a Study of the Morphosyntax of Children with Specific Language Disorders, with Hearing Impairment and Down’s Syndrome”. Clinical Linguistics & Phonetics 4, 77⫺86 Bortolini, Umberta & Leonard, Laurence B. (1996), “Phonology and Grammatical Morphology in Specific Language Impairment: Accounting for Individual Variation in English and Italian”. Applied Psycholinguistics 17, 85⫺104 Botha, Rudolf P. (1979), “External Evidence in the Validation of Mentalistic Theories: a Chomskyan Paradox”. Stellenbosch Papers in Linguistics 2, 1⫺38 Bradley, Dianne C. & Garrett, Merrill F. & Zurif, Edgar B. (1980), “Syntactic Deficits in Broca’s Aphasia”. In: Caplan, David (Hrsg.), Biological Studies of Mental Processes. Cambridge/MA: MIT Press, 269⫺286 Buckingham, Hugh W. (1993), “Disorders of Word-Form Processing in Aphasia”. In: Blanken et al. (Hrsg.), 187⫺197 Butterworth, Brian & Howard, David (1987), “Paragrammatism”. Cognition 26, 1⫺37 Cantwell, Dennis P. & Baker, Lorian (1987), Developmental Speech and Language Disorders. New York, London: Guilford Press
1824 Caplan, David (1985), “Syntactic and Semantic Structures in Agrammatism”. In: Kean (Hrsg.), 125⫺152 Caramazza, Alfonso & Hillis, Argye E. (1989), “The Disruption of Sentence Production: Some Dissociations”. Brain and Language 36, 625⫺650 Cholewa, Jürgen (1993), “Störung der lexikalischmorphologischen Wortverarbeitung bei Aphasie: Ein Literaturüberblick”. Neurolinguistik 7, 105⫺ 134 Cholewa, Jürgen & Bleser, Ria De (1995), “Neurolinguistische Evidenz für die Unterscheidung morphologischer Wortbildungsprozesse: Dissoziationen zwischen Flexion, Derivation und Komposition”. Linguistische Berichte 158, 259⫺297 Chomsky, Noam (1982), Some Concepts and Consequences of the Theory of Government and Binding. Cambridge/MA: MIT Press Clahsen, Harald (1988), Normale und gestörte Kindersprache: Linguistische Untersuchungen zu Syntax und Morphologie. Amsterdam: Benjamins Clahsen, Harald & Almazan, M. (1998), “Syntax and Morphology in Williams Syndrome”. Cognition 68, 167⫺198 Clahsen, Harald & Hansen, Detlef (1997), “The Grammatical Agreement Deficit in Specific Language Impairment: Evidence from Therapy Experiments”. In: Gopnik (Hrsg.), 141⫺160 Clahsen, Harald & Rothweiler, Monika & Woest, Andreas & Marcus, Gary F. (1992), “Regular versus Irregular Inflection in the Acquisition of German Noun Plurals”. Cognition 45, 225⫺255 Coltheart, Max & Patterson, Karalyn E. & Marshall, John C. (1980), Deep Dyslexia. London: Routledge Connell, Phil J. (1987), “An Effect of Modeling and Imitation Teaching Procedures on Children with and without Specific Language Impairment”. Journal of Speech and Hearing Research 30, 105⫺ 113 Crago, Martha B. & Allen, Shanley E. M. (1994), “Morphemes Gone Askew: Linguistic Impairment in Inuktitut”. In: Matthews (Hrsg.), 206⫺215 Cromer, Richard (1988), “Differentiating Language and Cognition”. In: Schiefelbusch, Richard & Lloyd, Lyle (Hrsg.), Language Perspectives: Acquisition, Retardation and Intervention. Austin/TX: Pro-Ed, 128⫺156 Cromer, Richard (1994), “A Case Study of Dissociations between Language and Cognition”. In: Tager-Flusberg (Hrsg.), 141⫺153 Curtiss, Susan (1988), “The Special Talent of Grammar Acquisition”. In: Obler, Loraine & Fein,
XIX. Psycholinguistische Perspektiven Deborah (Hrsg.), The Exceptional Brain. New York: Guilford Press, 364⫺386 Dalalakis, Jenny & Gopnik, Myrna (1995), “Lexical Representation of Greek Compounds: Evidence from Greek Developmentally Language Impaired Individuals”. In: MacLaughlin & McEwen (Hrsg.), 192⫺203 Dittmann, Jürgen & Tesak, Jürgen (1993), Neurolinguistik. Heidelberg: Groos (Studienbibliographien Sprachwissenschaft 8) Dogil, Grzegorz & Haider, Hubert & Schaner-Wolles, Chris & Husmann, Renate (1995), “Radical Autonomy of Syntax: Evidence from Sensory Transcortical Aphasia”. Aphasiology 6, 577⫺602 Dressler, Wolfgang U. (1987), “A Linguistic Classification of Phonological Paraphasias”. In: Dressler & Stark (Hrsg.), 1⫺22 Dressler, Wolfgang U. & Denes, Gianfranco (1987), “Word Formation in Italian-Speaking Wernicke’s and Broca’s Aphasics”. In: Dressler & Stark (Hrsg.), 69⫺81 Dressler, Wolfgang U. & Merlini Barbaresi, Lavinia (1994), Morphopragmatics. Berlin: Mouton de Gruyter Dressler, Wolfgang U. & Stark, Jacqueline A. (1987, Hrsg.), Linguistic Analyses of Aphasic Language. New York: Springer Dressler, Wolfgang U. & Stark, Jacqueline A. & Pons, Christiane & Kiefer, Ferenc & Kiss, Katalin & Me´sza´ros, E´va (1996), “Cross Language Analysis of German- and Hungarian-Speaking Broca’s Aphasics’ Processing of Selected Morphonological and Morphological Features: A Pilot Study”. Working Papers, Hungarian Institute for Linguistics, Hungarian Academy of Sciences 3 Dromi, Esther & Leonard, Laurence B. & Shteiman, Michal (1993), “The Grammatical Morphology of Hebrew-Speaking Children with Specific Language Impairment: Some Competing Hypotheses”. Journal of Speech and Hearing Research 36, 760⫺771 Fellbaum, Christiane & Miller, Steven & Curtiss, Susan & Tallal, Paula (1995), “An Auditory Processing Deficit as a Possible Source of SLI”. In: MacLaughlin & McEwen (Hrsg.), 204⫺215 Fodor, Jerry A. (1983), The Modularity of Mind. Cambridge/MA: MIT Press Fowler, Anne E. (1988), “Determinants of Rate of Language Growth in Children with Down Syndrome”. In: Nadel, Lynn (Hrsg.), The Psychobiology of Down Syndrome. Cambridge/MA: MIT Press, 217⫺245
167. Sprachstörungen Fowler, Anne E. (1990), “Language Abilities in Children with Down Syndrome: Evidence for a Specific Syntactic Delay”. In: Cicchetti, Dante & Beeghly, Marjorie (Hrsg.), Children with Down Syndrome: A Developmental Perspective. Cambridge: Cambridge University Press, 302⫺328 Fowler, Anne E. & Gelman, Rochel & Gleitman, Lila R. (1994), “The Course of Language Learning in Children with Down Syndrome: Longitudinal and Language Level Comparisons with Young Normally Developing Children”. In: Tager-Flusberg (Hrsg.), 91⫺140 Friederici, Angela D. (1988), “Agrammatic Comprehension: Picture of a Computational Mismatch”. Aphasiology 2, 279⫺284 Friederici, Angela D. & Weissenborn, Jürgen & Kail, Miche`le (1991), “Pronoun Comprehension in Aphasia: a Comparison of Three Languages”. Brain and Language 41, 289⫺310 Fukuda, Shinji & Fukuda, Suzy (1994), “Familial Language Impairment in Japanese: A Linguistic Investigation”. In: Matthews (Hrsg.), 150⫺177 Gathercole, Susan E. & Baddeley, Alan D. (1990), “Phonological Memory Deficits in Language Disordered Children: Is there a Causal Connection?”. Journal of Memory and Language 29, 336⫺360 Goad, Heather (1998), “Plurals in SLI: Prosodic Deficit or Morphological Deficit?”. Language Acquisition 7, 247⫺284 Gopnik, Myrna (1990), “Feature Blindness: A Case Study”. Language Acquisition 1, 139⫺164 Gopnik, Myrna (1992), “Theoretical Implications of Inherited Dysphasia”. In: Levy, Yonata (Hrsg.), Other Children, other Languages. Hillsdale/NJ: Erlbaum, 331⫺381 Gopnik, Myrna (1994), “Impairments of Syntactic Tense in a Familial Language Disorder”. In: Matthews (Hrsg.), 67⫺80
1825 Haarmann, Henk J. & Kolk, Herman H. J. (1994), “On-line Sensitivity to Subject-Verb Agreement Violations in Broca’s Aphasics: The Role of Syntactic Complexity and Time”. Brain and Language 46, 493⫺516 Hackl, Martin (1996), “Verb Second and Agrammatism: Evidence from Two Case Studies”. Wiener Linguistische Gazette 53⫺54, 23⫺62 Hamanaka, Toshihiko & Kanemoto, Kosuke & Ohigashi, Yoshitaka & Hadano, Kazuo (1985), “Paraphasia and Related Disorders in Primary Degenerative Dementia”. Studia Phonologica 19, 11⫺17 Hansson, Kristina (1992), “Swedish Verb Morphology and Problems with its Acquisition in Language Impaired Children”. Scandinavian Journal of Logopedics & Phoniatrics 17 (1), 23⫺29 Haverkort, Marco (1993), “Universalgrammatik, Parameter und Agrammatismus”. Linguistische Berichte 143, 60⫺80 Hofstede, Ben T. M. & Kolk, Herman H. J. (1994), “The Effects of Task Variation on the Production of Grammatical Morphology in Broca’s Aphasia: a Multiple Case Study”. Brain and Language 46, 493⫺516 Howell, S. Catherine & Fish, Sarah A. & KeithLucas, Thea (2000, Hrsg.), Proceedings of the 24th Annual Boston University Conference on Language Development, Bd. II. Somerville/MA: Cascadilla Press Isserlin, M. (1922), “Über Agrammatismus”. Zeitschrift für die Gesamte Neurologie und Psychiatrie 75, 322⫺410 Jakobson, Roman (1941), Kindersprache, Aphasie und allgemeine Lautgesetze. Uppsale: Almqvist Jakobson, Roman (1964), “Towards a Linguistic Typology of Aphasic Impairments”. In: de Reuck, Anthony V. S. & O’Connor, Maere (Hrsg.), Disorders of Language. London: Churchill, 21⫺42
Gopnik, Myrna (1997, Hrsg.), The Inheritance and Innateness of Grammars. New York: Oxford University Press
Kamhi, Alan G. (1993), “Children with Specific Language Impairment (Developmental Dysphasia): Perceptual and Cognitive Aspects”. In: Blanken et al. (Hrsg.), 625⫺640
Gopnik, Myrna & Crago, Martha B. (1991), “Familial Aggregation of Developmental Language Disorder”. Cognition 39, 1⫺50
Kean, Mary-Louise (1979), “Agrammatism: a Phonological Deficit?”. Cognition 7, 69⫺83
Gopnik, Myrna & Dalalakis, Jenny & Fukuda, Suzy E. & Fukuda, Shinji (1997), “Familial Language Impairment”. In: Gopnik (Hrsg.), 111⫺140
Kean, Mary-Louise (1985, Hrsg.), Agrammatism. New York: Academic Press
Grodzinsky, Yosef (1984), “The Syntactic Characterization of Agrammatism”. Cognition 16, 99⫺120
Kolk, Herman H. J. & Grunsven, Marianne M. F. van & Keyser, Antoine (1985), “On Parallelism between Production and Comprehension in Agrammatism”. In: Kean (Hrsg.), 165⫺208
Günther, Harmut (1989, Hrsg.), Experimentelle Studien zur deutschen Flexionsmorphologie. Hamburg: Buske
Kolk, Herman H. J. & Heeschen, Claus (1990), “Adaption Symptoms and Impairment Symptoms in Broca’s Aphasia”. Aphasiology 4, 221⫺232
1826
XIX. Psycholinguistische Perspektiven
Kudo, Takayaki (1992), “Word Formation in Aphasia: Evidence from Japanese kanji Words”. Journal of Neurolinguistics 7, 197⫺216
Leonard, Laurence B. & McGregor, Karla K. & Allen, Georg D. (1992), “Grammatical Morphology and Speech Perception in Children with Specific Language Acquisition”. Journal of Speech and Hearing Research 35, 1076⫺1085
Laine, Matti & Niemi, Jussi & Koivuselkä-Sallinen, P. & Ahlse´n, Elisabeth & Hyönä, J. (1994), “A Neurolinguistic Analysis of Morphological Deficits in a Finnish-Swedish Bilingual Aphasics”. Clinical Linguistics & Phonetics 8, 177⫺200
Leonard, Laurence B. & Sabbadini, Letizia & Leonard, Jeanette S. & Volterra, Virginia (1987), “Specific Language Impairment in Children: A CrossLinguistic Study”. Brain and Language 32, 233⫺ 252
Kraus, Alfred & Mundt, Christoph (1991, Hrsg.), Schizophrenie und Sprache. Stuttgart: Thieme
Lecours, Andre´ Roch (1993), “Glossomania and Glossolalia in Schizophasia and their Linguistic Kinships to the Jargonaphasias”. In: Blanken et al. (Hrsg.), 543⫺549 Lecours, Andre´ Roch & Lupien, Sonia & Bub, Daniel (1989), “Semic Extraction Behavior in Deep Dyslexia: Morphological Errors”. Tapuscrits 22 (Centre du Recherche du Centre Hospitalier Coˆtede-Neiges, Montre´al), 42⫺51
Levy, Yonata & Hermon, Shula (2000), “Morphology in Children with Williams Syndrome: Evidence from Hebrew”. In: Howell et al. (Hrsg.), 498⫺509 Linebarger, Marcia C. & Schwartz, Myrna F. & Saffran, Eleonor M. (1983), “Sensitivity to Grammatical Structure in so-called Agrammatic Aphasics”. Cognition 13, 361⫺392 Luria, Aleksander R. (1970), Traumatic Aphasia. The Hague: Mouton
Lely, Heather K. J. van der (1993), “Clinical and Research Congruence in Identifying Children with Specific Language Impairment”. Journal of Speech and Hearing Research 36, 580⫺591
MacLaughlin, Dawn & McEwen, Susan (1995, Hrsg.), Proceedings of the 19th Annual Boston University Conference on Language Development, Bd. I. Somerville/MA: Cascadilla Press
Lely, Heather K. J. van der (1994), “Canonical Linking Rules: Forward versus Reverse Linking in Normally Developing and Specifically LanguageImpaired Children”. Cognition 51, 29⫺72
Matthews, John (1994, Hrsg.), Linguistic Aspects of Familial Language Impairment. Montreal: McGill University [Special Issue of the McGill Working Papers in Linguistics 10]
Lely, Heather K. J. van der (1996), “Specifically Language Impaired and Normally Developing Children: Verbal Passive vs. Adjectival Passive Sentence Interpretation”. Lingua 98, 243⫺272 Leonard, Laurence B. (1989), “Language Learnability and Specific Language Impairment in Children”. Applied Psycholinguistics 10, 179⫺202 Leonard, Laurence B. (1992), “Specific Language Impairment in Three Languages: Some Cross-Linguistic Evidence”. In: Fletcher, Paul & Hall, David (Hrsg.), Specific Speech & Language Disorders in Children. London: Whurr, 118⫺126 Leonard, Laurence B. (1998), Children with Specific Language Impairment. Cambridge/MA: MIT Press Leonard, Laurence B. & Bortolini, Umberta & Caselli, M. Cristina & McGregor, Karla K. & Sabbadini, Letizia (1992), “Morphological Deficits in Children with Specific Language Impairment: The Status of Features in the Underlying Grammar”. Language Acquisition 2, 151⫺179 Leonard, Laurence B. & Dromi, Esther (1994), “The Use of Hebrew Verb Morphology by Children with Specific Language Impairment and Children Developing Language Normally”. First Language 14, 283⫺304
Menn, Lise & Obler, Loraine (1990, Hrsg.), Agrammatic Aphasia. Amsterdam: Benjamins Menyuk, Paula (1993), “Children with Specific Language Impairment (Developmental Dysphasia): Linguistic Aspects”. In: Blanken et al. (Hrsg.), 606⫺625 Miceli, Gabriele (1995), “The Summation Hypothesis and Its Implications for Output Morphological Processes”. In: Dressler, Wolfgang U. & Burani, Cristina (Hrsg.), Crossdisciplinary Approaches to Morphology. Vienna: Verlag der Österreichischen Akademie der Wissenschaften, 71⫺89 Miceli, Gabriele & Mazzucchi, Anna & Menn, Lise & Goodglass, Harold (1983), “Contrasting Cases of Italian Agrammatic Aphasia without Comprehension Disorder”. Brain and Language 19, 65⫺97 Miller, Carol & Leonard, Laurence (1998), “Deficits in Finite Verb Morphology: Some Assumptions in Recent Accounts of Specific Language Impairment”. Journal of Speech, Language and Hearing Research 41, 701⫺707 Mills, Anne E. & Pulles, Astrid & Witten, Francine (1992), “Semantic and Pragmatic Problems in Specifically Language Impaired Children: One or Two Problems?”. Scandinavian Journal of Logopedics & Phoniatrics 17, 51⫺57
167. Sprachstörungen Mimouni, Zohra & Jarema, Gonia (1995), “Agrammatic Aphasia in Arabic”. Clasnet Working Papers 3 Mogford-Bevan, Kay (1993), “Language Acquisition and Development with Sensory Impairment: Hearing-Impaired Children”. In: Blanken et al. (Hrsg.), 660⫺679 Nespoulous, Jean-Luc & Dordain, Monique (1990), “Agrammatism: a Disruption of the Phonological Processing of Grammatical Morphemes?”. In: Nespoulous & Villiard (Hrsg.), 270⫺277 Nespoulous, Jean-Luc & Dordain, Monique & Perron, Ce´cile & Ska, Bernadette & Bub, Daniel & Caplan, David & Mehler, Jacques & Lecours, Andre´ Roch (1988), “Agrammatism in Sentence Production without Comprehension Deficits: Reduced Availability of Syntactic Structures and/or of Grammatical Morphemes? A Case Study”. Brain and Language 33, 273⫺295 Nespoulous, Jean-Luc & Villiard, Pierre (1990, Hrsg.), Morphology, Phonology, and Aphasia. New York: Springer Newport, Elissa L. & Meier, Richard P. (1985), “The Acquisition of American Sign Language”. In: Slobin, Dan I. (Hrsg.), The Crosslinguistic Study of Language Acquisition. Bd. I, Hillsdale/NJ: Erlbaum, 881⫺938 Oetting, Janna B. & Rice, Mabel L. (1993), “Plural Acquisition in Children with Specific Language Impairment”. Journal of Speech and Hearing Research 36, 1236⫺1248 Ouhalla, Jamal (1993), “Functional Categories, Agrammatism, and Language Acquisition”. Linguistische Berichte 143, 3⫺36 Panzeri, Maria & Semenza, Carlo & Ferreri, Tiziana & Butterworth, Brian (1990), “Free Use of Derivational Morphemes in an Italian Jargonaphasic”. In: Nespoulous & Villiard (Hrsg.), 72⫺94 Paradis, Michael (1997, Hrsg.), Genetic Dysphasia. Elsevier [Special Issue Journal of Linguistics 10 (2⫺3)] Peltzer-Karpf, Annemarie (1994), Spracherwerb bei hörenden, sehenden, hörgeschädigten, gehörlosen und blinden Kindern. Tübingen: Narr Peuser, Günter (1978), Aphasie. München: Fink Plante, Elena (1991), “MRI Findings in the Parents and Siblings of Specifically Language-Impaired Boys”. Brain and Language 41, 67⫺80 Restrepo, Maria Adelaide & Swisher, Linda & Plante, Elena & Vance, R. (1992), “Relations among Verbal and Nonverbal Cognitive Skills in Normal Language and Specifically Language-Im-
1827 paired Children”. Journal of Communication Disorders 25, 205⫺219 Rice, Mabel L. (1994), “Grammatical Categories in Children with Specific Language Impairments.” In: Watkins & Rice (Hrsg.), 69⫺89 Rice, Mabel L. & Noll, Karen Ruff & Grimm, Hannelore (1997), “An Extended Optional Infinitive Stage in German-Speaking Children with Specific Language Impairment”. Language Acquisition 6, 255⫺295 Rice, Mabel L. & Wexler, Ken & Cleave, Patricia L. (1995), “Specific Language Impairment as a Period of Extended Optional Infinitive”. Journal of Speech and Hearing Research 38, 850⫺863 Rondal, Jean A. (1994), “Exceptional Cases of Language Development in Mental Retardation: The Relative Autonomy of Language as a Cognitive System”. In: Tager-Flusberg (Hrsg.), 155⫺174 Rondal, Jean A. (1995), Exceptional Language Development in Down Syndrome. Cambridge: Cambridge University Press Rothweiler, Monika & Clahsen, Harald (1994), “Dissociations in SLI Children’s Inflectional Systems: A Study of Participle Inflection and SubjectVerb-Agreement”. Scandinavian Journal of Logopedics & Phoniatrics 18, 169⫺179 Schaner-Wolles, Chris (1989), “Plural vs. Komparativerwerb im Deutschen”. In: Günther (Hrsg.), 155⫺186 Schaner-Wolles, Chris (1992), Spracherwerb trotz Down-Syndrom: Eine Vergleichende Studie über die modulare Organisation der Grammatik. Habilitationsschrift, Universität Wien Schaner-Wolles, Chris (1994), “Intermodular Synchronization: On the Role of Morphology in the Normal and Impaired Acquisition of a Verb-Second Language”. In: Tracy, Rosemarie & Lattey, Elsa (Hrsg.), How Tolerant is Universal Grammar? Tübingen: Niemeyer, 205⫺224 Schaner-Wolles, Chris (2000 a), “Sprachentwicklung bei geistiger Retardierung: Williams-BeurenSyndrom und Down-Syndrom”. In: Grimm, Hannelore (Hrsg.), Sprachentwicklung. [Enzyklopädie der Psychologie]. Göttingen: Hogrefe, 663⫺685 Schaner-Wolles, Chris (2000 b), “Within-Language Dissociations in Mental Retardation: WilliamsBeuren and Down Syndrome”. In: Howell (Hrsg.), 633⫺644 Schaner-Wolles, Chris & Dressler, Wolfgang U. (1986), “On the Acquisition of Agent/Instrument Nouns and Comparatives by Normal Children and Children with Down’s Syndrome”. Acta Linguistica Academiae Scientiarum Hungaricae 35, 133⫺ 149
1828 Schaner-Wolles, Chris & Haider, Hubert (1987), “Spracherwerb und Kognition – Eine Studie über interpretative Relationen”. In: Bayer (Hrsg.), 41⫺ 80 Schwartz, Myrna F. (1990, Hrsg.), Modular Deficits in Alzheimer-Type Dementia. Cambridge/MA: MIT Press Schwartz, Myrna F. & Linebarger, Marcia C. & Saffran, Eleonor M. & Pate, D. S. (1987), “Syntactic Transparency and Sentence Interpretation in Aphasia”. Language and Cognitive Processes 2, 85⫺113 Smith, Neil & Tsimpli, Ianthi-Maria (1995), The Mind of a Savant. Language Learning and Modularity. Oxford: Blackwell Smith-Lock, Karen M. (1993), “Morphological Analysis and the Acquisition of Morphology and Syntax in Specifically-Language-Impaired Children”. Haskins Laboratories Status Report on Speech Research 114, 113⫺138 Stark, Jacqueline A. & Dressler, Wolfgang U. (1990), “Agrammatism in German: Two Case Studies”. In: Menn & Obler (Hrsg.), 281⫺441 & 1458⫺1547 Stemberger, Joseph P. (1984), “Structural Errors in Normal and Agrammatic Speech”. Cognitive Neuropsychology 1, 281⫺313 Swisher, Linda & Restrepo, Maria Adelaida & Plante, Elena & Lowell, Soren (1995), “Effects of Implicit and Explicit ‘Rule’ Presentation on Bound-Morpheme Generalization in Specific Language Impairment”. Journal of Speech and Hearing Research 38, 168⫺173 Tager-Flusberg, Helen (1994, Hrsg.), Constraints on Language Acquisition: Studies of Atypical Children. Hillsdale/NJ: Lawrence Erlbaum Tallal, Paula & Ross, Randal & Curtiss, Susan (1989), “Familial Aggregation in Specific Language Impairment”. Journal of Speech and Hearing Disorders 54, 167⫺173 Tallal, Paula & Townsend, Jeanne & Curtiss, Susan & Wulfeck, Beverly (1991), “Phenotypic Profi-
XIX. Psycholinguistische Perspektiven les of Language-Impaired Children Based on Genetic/Family History”. Brain and Language 41, 81⫺95 Tesak, Jürgen (1990), “Agrammatismus: Ergebnisse und Probleme der Forschung”. Neurolinguistik 4, 1⫺41 Tomblin, J. Bruce (1989), “Familial Concentration of Developmental Language Impairment”. Journal of Speech and Hearing Disorders 54, 287⫺295 Tomblin, J. Bruce (1997), “Epidemiology of Specific Language Impairment”. In: Gopnik (Hrsg.), 91⫺110 Tyler, Lorraine K. & Behrens, Susan & Cobb, Howard & Marslen-Wilson, William (1990), “Processing Distinctions between Stems and Affixes: Evidence from a Non-Fluent Aphasic Patient”. Cognition 36, 129⫺153 Tyler, Lorraine K. & Cobb, Howard (1987), “Processing Bound Morphemes in Context: the Case of an Aphasic Patient”. Language and Cognitive Processes 2, 245⫺262 Ullman, Michael & Gopnik, Myrna (1994), “Past Tense Production: Regular, Irregular and Nonsense Verbs”. In: Matthews (Hrsg.), 81⫺118 Villiers, Jill de & Villiers, Peter de & Hoban, Esme (1994), “The Central Problem of Functional Categories in the English Syntax of Oral Deaf Children”. In: Tager-Flusberg (Hrsg.), 49⫺74 Watkins, Ruth V. & Rice Mabel L. (1994, Hrsg.), Specific Language Impairment in Children. Baltimore: Brookes Wexler, Ken (1994), “Optional Infinitives, Head Movement and the Economy of Derivations”. In: Lightfoot, David & Hornstein, Norbert (Hrsg.), Verb Movement. Cambridge: Cambridge University Press, 305⫺350 Wisch, Fritz-Helmut (1990), Lautsprache und Gebärdensprache. Hamburg: Signum Verlag Yamada, Jeni E. (1990), Laura: A Case for the Modularity of Language. Cambridge/MA: MIT Press
Chris Schaner-Wolles, Wien (Österreich) Wolfgang U. Dressler, Wien (Österreich)
XX. Morphologie in der Praxis Morphology in practice 168. Field work 1. 2. 3. 4. 5. 6. 7.
Introduction Getting to the language Collecting the data Identifying the meaning Resistant problems Technical helps References
1.
Introduction
In this article methods of gathering and evaluating data will be presented. Such data serve as basis for the morphological analyses which are discussed in the other articles in this volume. Morphological analysis can be done from many points of view, and the results vary greatly according to what kinds of questions are asked of the data and in what framework the analysis is taking place. But any and all analysis has to be based on solid linguistic data, that which is phonologically correct, natural in language use, and for which the meaning of words, parts of words or other morphemes is clear. Some methods of how to arrive at such a basis in a hitherto little or unknown language will be presented. Their main merit may well be that they have been tried and found useful. Much of what has been written on the subject is found in scattered articles and in handbooks on language learning (for a partial bibliography see Goyvaerts 1986), or in introductory textbooks on linguistic analysis or translation. Not all will be mentioned here, but enough to contain the basic information in detail.
2.
Getting to the language
2.1. The physical set-up Many (but by no means all) “insufficiently studied languages” (Kibrik 1977: 1) are found in what is commonly called the Third World.
Thanks to air and road service many of them are fairly easily accessible, though travel to some of them is still an adventure. This is true not only for means of transport, but also for living conditions in general and, maybe most importantly, hygienic eating and drinking facilities. The investigator intending to spend only a few weeks in the area in which the language is spoken should therefore first investigate the physical aspect of his stay (Healey 1975, ed.) and prepare accordingly. He is wise if he strictly observes recommendations given him by authorities and other foreigners living in the area. 2.2. The language assistant(s) If possible a selection of potential language assistants (also known as “informants”, Samarin 1967) should be available. It is always good to work with more than one. The qualification of such assistants is of utmost importance. Even if the ideal persons may not readily be found, here is a partial list of their qualifications, some of them contradictory. They should be native speakers of the language, preferably not from a “mixed marriage” where each parent speaks a different language, nor having lived out of the language area for too long as this may impair a speaker’s fine sense of feeling for his language. The best assistant would be a monolingual person who never set foot outside his area. The assistants should be neither too young nor too old. Before puberty they may not yet have full command of the language. One too old might make a good story teller but be no longer capable of giving any other type of linguistic data. They should not be ashamed of their language. An educated assistant might have gone through a school system in which use of the mother tongue was punished. Perhaps the official language was learned by means
1830
XX. Morphologie in der Praxis
of punishment. Use of the language may be forbidden and involve risks. Cases are not unknown of people “hiding” their language from foreigners by purposely giving out false information. They should be reasonably bilingual in the official language of the country. If this language is foreign to both the investigator and the assistants, sources of misunderstanding abound, both because of linguistic and cultural differences. There is also the matter of remuneration: the salary of a primary school teacher or of a secretary is a good basis on which to calculate it. The amount may have to depend on the educational background of the assistant. It is good to be informed about existing regulations and practices by other foreigners living in the area. 2.3. The investigator The attitudes the investigator carries with him have a lot to do with his effectivenes in his work. In many ⫺ but by no means all ⫺ cultures the foreigner is received with a fair amount of friendship, coloured by previous experience with other foreigners. It is very important that he respects local traditional and civil authorities (as well as higher placed people in the cities). If he cannot develop a certain affinity for the people whose language he has come to study he might better change the language. His education and position will usually place him in a higher social category than many of the people speaking the language he has come to study. Paternalism or arrogance is quickly felt by them. His assistants and their families are real people with real needs ⫺ and he might be in a better position to do something about such needs than anyone else. On the other hand friendship also involves receiving. Respect and honest appreciation of the other person are often recognized keenly.
3.
Collecting the data
3.1. The intermediate language As already mentioned, data is best collected monolingually. This would mean that the investigator learn at least the rudiments of the language first, something that can be achieved superficially in a matter of weeks (thorough learning will take at least a year or more). Eliciting data monolingually enhances
the quality, for it keeps the assistant in the right context: he does not have to translate but can react from within the language in which the data is to be gathered. The use of an intermediate language could also be avoided by collecting through an interpreter who himself fully understands what data is needed, so that he can trigger it in the language under investigation. 3.2. The questionnaire Good preparation of the kind of data to be collected often includes a questionnaire in the intermediate language, to be translated into the language under investigation (Weiers 1980). Our knowledge of general traits and peculiarities of a language family allows us to prepare such elicitation tools quite well. If a questionnaire is used to elicit straightforward translation, there are liable to be many pitfalls. Any data elicited without a natural context is easily misunderstood even in the intermediate language. And if misunderstanding is not the problem, a translation of what sounds natural in the official language might yield grammatical but quite unnatural data in the language under investigation ⫺ thus the data is falsified. The best data is not elicited via translation, but by a stimulus in the language under investigation. 3.3. The transcription For a good transcription it is best not to use the orthography (provided there is one) unless it has been developed on the basis of sound phonological principles. A poor orthography might easily underdifferentiate the sounds, yielding homographs which are not homophones, causing different morphemes to appear to be identical. Thus, the investigator might look for similarities in meaning for words that have nothing to do with each other. The investigator should have some practical knowledge of phonetic transcription including ear training so that he can accurately represent on paper what he hears. This training should include tone differentiation, should he work in a tone language. He should also have some basic training in phonological analysis. For a purely phonetic transcription might lead him astray in a different direction: he might not recognize that two utterances are the same though different phonetically and end up with more morphemes than warranted.
1831
168. Field work
If the investigation is done through a translator or by a literate assistant himself, there might be no alternative to using the orthography he is familiar with. Being aware of the pitfalls might be enough to circumvent them.
4.
Identifying the meaning
No analysis is possible without being able to identify the meaning of the morphemes under investigation. One of the most important principles to remember is that the meaning changes according to context (Pike & Pike 1977: 173 f.). 4.1. Lexical meaning The lexical meaning is relatively easy to discover as it involves the first three of the basic semantic categories (Nida 1964: 62): the meaning of the nouns (for things), verbs (for events), and descriptives (for attributes) in a language. It is good to be aware of the fact that the name of any of the above is not necessarily a simple noun, it may be a derivation, a compound or a phrase, and yield all sorts of material for morphological analysis. One word of caution: even though it may be obvious that a word is morphologically complex, a compound word or even a phrase, it is by no means obvious that the assistant can give a morpheme-by-morpheme or a word-by-word translation. To him more often than not the whole expression has one meaning; the break-down has to be supplied by the investigator. Eliciting vocabulary with lexical meaning is best done by pointing and enacting on the one hand and by analogy on the other (Beekman 1968: 361 f.), like for example: Many bees are a swarm, what are many dogs? The opposite of good is bad; what else can it be? What is the opposite of new? 4.2. Grammatical meaning The last of the four basic categories, the relationships, are far more difficult to deal with. For they encode the grammar of the language (apart from word order). This category includes all sorts of affixes as well as the particles, many of them “pesky” ones (see 5.2; Grimes 1975: 93), because they are so hard to define. No direct questioning will yield them in the language, and only the rarest of assistants is able to give a translation, once they are encountered. Even then a translation is
only a starting point to find the “real” meaning or function. Who, for example, could easily fully define the use of still or yet in English? The best chances for an equivalent in the intermediate language are found in such meaning bearers as pronouns, conjunctions, adpositions (prepositions or postpositions), singular/plural differences, negatives, question words, and maybe verb tenses. But translations are at best approximations. They do yield a beginning point for the analysis which will eventually bring to light the full meaning and use of the morpheme. 4.3. A simple methodology A quite simple method for identifying the morphemic forms and a first indication as to their meaning is that of substitution of recurring particles (Elson & Pickett 1983). It consists of collecting a number of utterances (complex words, maybe) which contain the morpheme in question. Hopefully a meaning can be supplied for the complete form, maybe other forms without the morpheme can be found and their meaning ascertained. A comparison of the paradigm will yield the various forms of the morpheme(s) under investigation as in (1) from Tzeltal (Mexico): (1) k’ab ‘hand’ akan ‘leg’ lumal ‘land’ inam ‘wife’ k’op ‘language’ at’el ‘work’ (Nida 21949: 16)
hk’ab ‘my hand’ kakan ‘my leg’ alumal ‘your land’ awinam ‘your wife’ sk’op ‘his language’ yat’el ‘his work’
This tried and true method in the tradition of the guess-and-check procedures (Longacre 1964: 11, 101 f.) still serves to recognize the majority of the morphemes in any language, including their various forms which yield the morpho-phonemic rules. It is not quite so simple to identify their full meaning. For this it is necessary to find all the possible occurrences of the form under investigation and study their meaning nuances which might be caused by the presence or absence of that form.
5.
Resistant problems
5.1. Homophones A certain number of words and particles in any language sound exactly the same. Are they the same, or different morphemes? For
1832 the lexicon this question is not quite so crucial; it is not that hard to distinguish a fish from to fish, one a noun the other a verb. If one of the pair is a grammatically functioning particle or an affix (and if a lexical meaning cannot be ascertained it always is), this question becomes relevant for the analysis and gains quite a different importance. For example, in Kaingang (Brazil), are mu‘go:pl’ and mu- ‘action in progress’ the same morpheme or two different ones? Or, asked differently: in the utterance under examination, is the occurrence of mu- a verb or a particle or an affix? This difference might be all important in deciding on the “real” meaning of the utterance (as over against the general and vague translation provided by the assistant or guessed at by the investigator) and on the analysis of the structure of that utterance ⫺ be it part of the morphology or the syntax. If the language is full of such homophones, this is one of the thornier questions to resolve. Certain verbs have a tendency to become particles (or, in an earlier stage, auxiliaries). If both members of the homophonous pair are particles or affixes, the solution is none the easier. 5.2. Pesky particles Another difficult problem is that of particles for which the meaning is difficult to ascertain. The first question to decide is whether such particles belong into the syntax or the morphology. They often have some function within discourse rather than within sentence or phrase and are therefore difficult to pin down. Nevertheless their identification is important to the understanding of the grammar. There is only one way to study them: within natural texts. 5.3. Text as context All morphemes can best be studied within the framework of text analysis, for this is where they occur most naturally. In order to do this different types of texts (Longacre 1976: 197 f.) need to be collected, recorded on tape, and transcribed if possible. Such texts are best charted according to different points of view (Wiesemann et al. 21992; Seminar für Sprachmethodik 1999): certainly a basic chart needs to be made, in which the clauses are grouped around the verbs in the text. Then an information chart lists the different types of information transmitted by the verbs in several columns, including highlighted information.
XX. Morphologie in der Praxis
Texts charted this way are the best sources of contexts for affixes or particles for which the precise function is difficult to determine.
6.
Technical helps
6.1. Recorders The recording technique is one of the most useful helps in field work (Bouquiaux & Thomas 1976, eds.: 27 f.). Today analog (cassette recorders) and digital (MiniDisc, DAT, etc.) recorders are of such high quality and at the same time so small that they can be carried almost everywhere and used even in adverse circumstances, with or without electricity. A separate microphone improves the quality and is more versatile than the inbuilt one, particularly if suprasegmental sound features are to be observed. A moving coil microphone can stand difficult climates better than a condensor microphone (Wiesemann 1991: 195 f.). In order to transcribe tape recorded data it is necessary to listen to small stretches and return to the same utterance or part of the utterance several times (Gudschinsky 1967: 9). It should be possible to rewind a bit of the tape while the play button is held down. 6.2. Data books, lists It is useful to gather the data in notebooks. To keep track of the chronological development of the data-gathering each page is numbered, each session dated, and the name of each assistant written down. It is not advisable to erase any of the attempts at writing down the data; corrections are rather written in under the original entry. At first the phonetic notations will be full of errors: it takes a while for the ear to adjust to the sounds in the language. Even after a while doubts about individual sounds are bound to persist. These can be double checked with the help of lists for contrastive listening, a type of differential drill (Brewster & Brewster 1976: 309 f.). Words with what are either identical or very similar sounds are gathered in a list. The sounds to be compared should preferably appear in the same position in each word. As the assistant pronounces each word, the investigator listens to just that troubling sound. The lists may need to be rearranged several times, the assistant may be consulted as to which words, in his opinion, have the same sound and which sounds are different.
168. Field work
1833
The morphological analysis may be done in the notebook or on separate sheets. Individual sheets are not helpful for data gathering but can easily be rearranged for analysis purposes.
(Dallas/TX: Summer Inst. of Linguistics) 1.29: 1⫺ 11 [reprinted in: Healey (1975, ed.)]
6.3. Morphology files In order to keep data handy and the analysis under control, a shoe box or something similar with subdivisions and a whole set of 3 ⫻ 5⬙ very thin cardboard slips is very useful. On each slip a word is written (normally in the left hand corner), a grammatical designation (in the middle), and a gloss (in the right hand corner). In this way the lexicon can be gathered and kept in alphabetical order, the parts of words, their forms and meanings are easily retrievable. Filing with the help of punch cards (Day 1976) is even more versatile.
Brewster, E. Thomas & Brewster, Elizabeth S. (1976), Language Acquisition Made Practical: Field Methods for Language Learners. Colorado Springs: Lingua House
6.4. Computer programs PCs are small and so powerful today, that they constitute another very useful technical tool for morphological analysis in almost any field situation. Even a small data bank can take the place of the morphology file. Specially developed programs (Weber et al. 1988; Wimbish 1990; Antworth 1990) assist in morphological analysis, glossing text and pulling out the contexts of whatever morpheme is to be studied (cf. also Art. 172): ⫺ Shoebox and LinguaLinks are data base systems for field linguists and anthropologists. Among several facilities for the field worker both programs provide also lexical databases and tools for morphological analysis and interlinear transcription of words. ⫺ PC-KIMMO is designed for generation and morphological parsing of words. The program is an implementation of the model of two-level morphology (see Koskenniemi 1983) and uses a lexicon, a module for the declaration of morphonological rules, and a module for the declaration of morphotactic rules.
7.
References
Antworth, Evan L. (1990), PC-KIMMO: A TwoLevel Processor for Morphological Analysis. Dallas/ TX: Summer Inst. of Linguistics Beekman, John (1968), “Eliciting Vocabulary, Meaning and Collocations”. Notes on Translation.
Bouquiaux, Luc & Thomas, Jacqueline M. C. (1976, eds.), Enqueˆte et description des langues a` tradition orale, vol. 1⫺3. Paris: SELAF
Day, A. Colin (1976), Getting the Edge on Card Files. High Wycombe, Bucks.: Summer Inst. of Linguistics Elson, Benjamin F. & Pickett, Velma B. (1983), Beginning Morphology and Syntax. Dallas/TX: Summer Inst. of Linguistics Goyvaerts, Didier L. (1986), Language and History in Central Africa. Antwerpen: Univ. Instelling Antwerpen (Antwerp Papers in Linguistics 44) Grimes, Joseph E. (1975), The Thread of Discourse. The Hague: Mouton (Janua Linguarum Series Minor 207) Gudschinsky, Sarah C. (1967), How to Learn an Unwritten Language. New York etc.: Holt, Rinehart and Winston Healey, Alan (1975, ed.), Language Learner’s Field Guide. Ukarumpa, Papua New Guinea: Summer Inst. of Linguistics Kibrik, Alexandr E. (1977), The Methodology of Field Investigations in Linguistics. The Hague: Mouton (Janua Linguarum Series Minor 142) Koskenniemi, Kimmo (1983), Two-level Morphology: A General Computational Model for WordForm Recognition and Production. Ph.D. dissertation, University of Helsinki Longacre, Robert E. (1964), Grammar Discovery Procedures: A Field Manual. The Hague: Mouton (Janua Linguarum Series Minor 33) Longacre, Robert E. (1976), An Anatomy of Speech Notions. Lisse: de Ridder Nida, Eugene A. (21949), Morphology: The Descriptive Analysis of Words. Ann Arbor: Univ. of Michigan Press [11946] Nida, Eugene A. (1964), Toward a Science of Translating. Leiden: Brill Pike, Kenneth L. & Pike, Evelyn G. (1977), Grammatical Analysis. Dallas/TX: Summer Inst. of Linguistics (SIL Publications in Linguistics 53) Samarin, William J. (1967), Field Linguistics: A Guide to Linguistic Field Work. New York etc.: Holt, Rinehart and Winston Seminar für Sprachmethodik (1999), Textanalyse und Grammatik, ein sprachwissenschaftliches Lehrbuch. Bonn: Verlag für Kultur und Wissenschaft
1834
XX. Morphologie in der Praxis
Weber, David J. & Black, H. Andrew & McConnel, Stephen R. (1988), AMPLE: A Tool for Exploring Morphology (Version 1.0). Dallas/TX: Summer Inst. of Linguistics
Wiesemann, Ursula & Nse´me´, Cle´dor & Vallette, Rene´ (21992), Manuel d’analyse du discours. Yaounde´: Collection PROPELCA (26) [11984 with different pagination]
Weiers, Michael (1980), Linguistische Feldforschung: Ein Leitfaden. Wiesbaden: Harrassowitz
Wimbish, John S. (1990), SHOEBOX: A Data Management Program for the Field Linguist (Version 1.2). Waxhaw: JAARS Computer Services (Summer Inst. of Linguistics)
Wiesemann, Ursula (1991), Verstehen und verstanden werden: Praktisches Handbuch zum Zweitspracherwerb. Bad Liebenzell: Verlag der Liebenzeller Mission
Ursula Wiesemann, Burbach-Holzhausen (Germany)
169. Interlinear morphemic glossing 1. 2. 3. 4. 5. 6. 7.
Basic concepts Prerequisites of morphological analysis Principles of interlinear glossing Boundary symbols Typographic conventions Summary References
1.
Basic concepts
1.1. Purpose Given an object language L1 and a metalanguage L2, then an interlinear morphemic gloss (IMG) is a representation of a text in L1 by a string of elements taken from L2, where, ideally, each morph of the L1 text is rendered by a morpheme of L2 or a configuration of symbols representing its meaning, and where the sequence of the units of the gloss corresponds to the sequence of the morphs which they render. Its primary aim is to make the reader understand the grammatical structure of the L1 text by identifying aspects of the free translation with meaningful elements of the L1 text. The ultimate purpose may be to aid the reader in grasping the spirit of the language, to control the linguistic argument the author is making by means of the L1 example or to scan a corpus for a certain gloss in order to find relevant examples. (1) Latin exeg-i monumentum implement\prf-1.sg monument.n:acc.sg aer-e perennius ore.f-abl.sg lasting:cmpr:acc.sg.n ‘I have executed a monument more durable than ore’ (Hor. Od. 3, 30, 1)
(1) illustrates the typical use of an IMG. The first line of (1) contains the L1 text line; the second line contains the IMG, and the third line contains an idiomatic translation into L2. Interlinear morphemic glossing is at the intersection of different communicative purposes. On the one hand, it is a kind of translation that accompanies the original. In this sense, it is comparable to the arrangement that one finds in synoptic editions of original and translation. On the other, it is a kind of linguistic analysis. In this sense, it competes with a fragment of a grammar. Its hybrid character leads to a number of problems and to different styles in interlinear morphemic glossing. The aim of the following treatment is a standardization of an aspect of linguistic methodology on the basis of widespread usage as developed in the 20th century. To the extent that linguistics is a science, its methods are susceptible and in need of standardization. Interlinear morphemic glossing has to do with the representation of linguistic data, comparable in this with a phonetic transcription. Just as the latter has been successfully standardized by the IPA, so interlinear morphemic glossing should be standardized. This will be done in the present article in the form of a set of rules, which are listed in section 6.1. Such a standardization only concerns linguistic science. Linguistic data are often presented to a lay public, with the purpose of education, entertainment or divulgation of the achievements of our science. Here some kind of interlinear glossing may be necessary, too. However, scientific formal-
1835
169. Interlinear morphemic glossing
ism tends to damage rather than serve the good cause. An example how interlinear glossing has been handled in a book directed to a non-specialist public is quoted in the next section (Finck 1909). The present article is biased in favor of a more formalized treatment, on the assumption that it will be easier to derive a less formal representation from the proposals made here than the other way round. The treatment is, however, not fully formal, since it focuses on interlinear glossing in printed texts. In the annotation of texts by markup languages for automatic retrieval, the same conceptual problems, but very different technical problems arise which will not be dealt with here. Data are commonly quoted from sources in which they are already provided by an analysis. In linguistic publications, it has been wide-spread usage to quote data together with their IMG and their translation, even if their form or language is different from the one used in the quoting context. That is, such composite data representations have been treated as indecomposable blocks. Such scruples do not seem to be warranted. Primary data may be quoted and provided with the quoting author’s analysis and translation (cf. Bickel et al. 2004: 1). 1.2. Precursors Interlinear glossing has precursors in the descriptive tradition which link it up not with some kind of morphological representation, but with efforts to bring out the spirit of the language. The point there is not to provide a formal representation of a piece of linguistic data, but to render the language-specific construal of the world intelligible. To this end, literal translations were provided. For instance, G. Gabelentz (1901: 460), in a passage arguing that the personal verb suffixes in Semitic languages are possessive pronouns, gives the following Arabic example: “ya-kfı¯ka-hu`m er genügt dir gegen sie (eig. er-genügt-dein-ihr)”. The IMG is a late-comer in linguistics. Early grammars were intended as primers, the user was expected to work through them and learn the morphemes; so no glossing was necessary. Many scientific grammars, e.g. of Latin, Greek, Arabic etc., were meant for the initiated who needed no glossing either (not seldom even the free translations were spared). Even comparative studies, historical or typological, left the analysis of the examples of diverse languages to the reader. H. C.
Gabelentz, in the middle of a discussion of Lule, Osage and other languages, presents the following passage: “Im Dakota (meine Grammatik der DakotaSprache § 34) dient die 3 Pers. Plur. Act. dazu, das Passivum auszudrücken, sogar wenn ein Actor im Singularis hinzuzudenken ist, z. B. Jesus Jan en˜ hi q ix Jordan watpa ohna baptizapi, Jesus kam zu Johannes und sie tauften ihn (st. er wurde getauft) im Jordanfluss.” (H. C. Gabelentz 1861: 465)
Here the reader who does not have the grammar mentioned on his desk is given no chance. Pace Gabelentz, IMGs are needed when two conditions coincide: the level of analysis is above morphology, and the reader is not expected to be familiar with the languages under discussion (which is generally the case in typology, but not in descriptive or historical-comparative linguistics). W. v. Humboldt (1836 [1963]: 534) invented his own device to help the reader identify L2 meaningful elements with L1 morphemes. He gives the following example from Classical Nahuatl: 1 2 3 4 5 6 7 8 9 ni- c- chihui -lia in no- piltzin ce calli 1 3 2 4 5 6 7 8 9 ich mache es für der mein Sohn ein Haus While dispensing with the IMG proper, this method fails for L1 elements which cannot be rendered by L2 words. Beside the literal translation illustrated above, G. Gabelentz (1901) uses a variety of techniques. He also has interlinear glosses, as when he says: ‘Der Satz “Ich bin Dein Sohn” heißt im Maya: a ⫺ meχen ⫺ en. Dein Sohn ich,’ (Gabelentz 1901: 383) and occasionally (e.g. Gabelentz 1901: 400) he uses Latin as L2 in IMGs. Finck (1909) is one of the first linguistic publications that illustrate the working of a language with a sizable text provided with a free translation and an IMG. The following sentence from his Turkish text (Finck 1909: 83) illustrates his glossing style (s. Fig. 169.1). As may be seen, these forerunners have no grammatical category labels yet. Finck glosses Turkish -ın ‘gen’ by Germ. der because this word displays a morphological trace of the genitive. Similarly, Turkish -up ‘ger’ is glossed by -enderweise, maybe the closest to a gerund that German can muster. This pro-
1836
XX. Morphologie in der Praxis
Fig. 169.1
cedure is a tribute to the non-specialist readership that the booklet aims at, but necessarily falsifies the working of the language by attributing lexical meanings to its grammatical morphemes. It took a long time until interlinear morphemic glossing became firmly established. In Bloomfield’s Language, of 1933, examples abound, but they are presented like this: “Some languages have here one word, regardless of gender, as Tagalog [kapa’tid]; our brother corresponds to a Tagalog phrase [kapa’tid na la’la:ki], where the last word means ‘male’, and our sister to [kapa’tid na ba’ba:ji], with the attribute ‘female’” (Bloomfield 1933: 278). IMGs that fulfill most of the requirements set out below appear first in the sixties of the 20th century. From the eighties on, they become standard in publications dealing with languages whose knowledge is not presupposed. Editors and publishers increasingly require them even for languages like Latin, French and German that used to be wellknown to linguists. The development is towards (not only translating, but even) glossing every language except English. This is apparently a symptom of a global development in which every language except English becomes exotic. Good IMGs are relatively costly, both for the scientist and for the typesetter. Authors and publishers are therefore not too eager to produce them (well). There is at least one software on the market that aids the linguist in generating systematic IMGs for his texts, the interlinearizer that comes with the program Shoebox, from the Summer Institute of Linguistics (cf. Simons & Versaw 1988; Art. 168). Since IMGs are fairly recent in linguistics, they have seldom been treated by linguistic methodology. The first treatise of the present subject is Lehmann (1982). Subsequent work
includes Simons & Versaw (1988), Lehmann et al. (21994), Lieb & Drude (2000), Bickel et al. (2004). They have been freely made use of in the present treatment. 1.3. Levels of representation Interlinear morphemic glossing must be seen in the larger context of representation of linguistic data and, even more comprehensive, of the documentation of a language (cf. Lieb & Drude 2000). On such a background, an isolated example given in a descriptive context is a particularly constrained case of the edition and annotation (also called ‘markup’ for technical purposes) of a piece of primary linguistic data for posterity. In other words, a general-purpose edition of a linguistic corpus is a kind of maximum model, subject to the full set of rules for explicitness, detail and elaboration, from which the quotation of an isolated example in the context of some grammatical discussion represents a subset delimited by considerations of feasibility, usefulness and the like. Every linguistic representation of some piece of raw data, even if it limits itself to a phonetic transcription, involves some linguistic analysis (Lehmann 2004). Insofar, no sharp boundary is to be drawn between the sheer representation of data and their analysis. Bearing this in mind, we can speak of various levels at which linguistic data may be represented. Presupposing spoken language data, at least the following are relevant: (a) raw data recording (video or audio tape), (b) phonetic transcription of the utterance, (c) orthographic representation of the utterance, (d) morph(ophon)emic representation of the utterance, (e) IMG of the utterance, (f) free translation of the utterance into the background language,
169. Interlinear morphemic glossing
(g) descriptive and explanatory comment on pragmatic or cultural aspects of the utterance. This set may be supplemented by even more representations (cf. Lieb & Drude 2000). There may be a phonological representation distinct from both levels (b) and (d). There may be a syntactic representation, e.g. in the form of a labeled bracketing. And there may be a semantic representation instead of, or in addition to, representation (f). In such representations, the portion of linguistic analysis is probably even stronger than in the seven levels enumerated. The raw data have a temporal structure which is projected onto a spatial line in written representations. These representations are synchronized more or less closely. For instance, representation (f) generally matches L1 sentences, units of level (g) may be associated with L1 units of any size, and representation (e) may match representation (d) morpheme by morpheme. This has different consequences for the typographic layout. For instance, units of level (g) may be associated with the running text by making full use of a multidimensional display, while representation (f) may be in a lateral column at the same height as its original, as is usual in synoptic editions and also practiced in the example from Finck (1909) given in section 1.2. Other representations should be arranged in lines one of which is beneath the other and runs in parallel with it. For the purposes of descriptive and typological grammatical analysis and exemplification, the seven-level set is generally reduced to only three. What may be called the ‘canonical trilinear representation’ of linguistic examples involves: ⫺ a representation of L1 at one of the levels (b), (c) or (d), ⫺ an IMG in L2 (level e), ⫺ an idiomatic translation into L2 (level f). An IMG will seldom be paired with a phonetic representation, because this serves phonetics, while an IMG serves grammar. They therefore form an unequal pair. If both are required, they will generally be mediated by another representation, a morphophonemic or orthographic one. It makes a difference for the glossing whether L1 is rendered in a morphophonemic representation or in conventional orthography. In the former case, the rules of
1837 orthography do not apply, and the linguist may dress up the representation in such a way that a biunique mapping onto the IMG is facilitated. In the latter case, morpheme boundaries may be obscured by the orthography, and there will be delimiters such as blanks, hyphens and punctuation marks which do not necessarily represent grammatical boundaries and may interfere with the glossing. However, the choice between an orthographic and a scientific representation of a text is generally a higher-order choice which cannot depend on glossing requirements. In particular, an example may be quoted unchanged from a primary source (think of Sanscrit examples). It may then not be possible to insert boundary symbols and the like in the L1 text. Glossing conventions therefore have to be adjusted to use with orthographic representations. If the first line representing the L1 text differs too much from a morphophonemic representation, then it is advisable to expand the canonical trilinear representation by an additional morphophonemic representation. It will then be this line that the IMG refers to. The two languages involved will be called L1 and L2 throughout. However, it should be clear that the relationship between them is asymmetric: L1 is the object language, L2 is the metalanguage. The symbols occurring in an IMG have a different status from the elements of the text line that they gloss: For present purposes, the L1 text line consists of morphs, while the IMG consists of names of L2 morphemes and of grammatical categories (cf. section 3.2). There can, thus, be no question of “mirroring” the structure of the L1 expression by the sequence of the L2 elements. Instead, an element in an IMG serves as a kind of mnemonic hint to the meaning or function of its corresponding L1 element. 1.4. Delimitation The complete set of representations rendering an L1 text may be sufficient to derive a grammatical description from it (as postulated in Lieb & Drude 2000, § 1.1). However, given its inherent restrictions, an IMG cannot by itself compensate for a grammar (or just a morphology). Apart from the form of presentation, the most important substantive difference between a grammatical description and an IMG lies in the fact that the grammar treats of categories in the sense of classes, while the IMG identifies individual morphemes. For instance, a grammar treats of
1838
XX. Morphologie in der Praxis
the verbal category of aspect. An IMG contains a gloss for an individual aspect morpheme, e.g. perf, neglecting the question of whether this is actually an aspect morpheme or rather a tense morpheme, and also leaving unanswered questions concerning other members of the paradigm, let alone the construction and use of the perf morpheme. Some of these kinds of information may be given in other representations, e.g. a syntactic representation. By the same token, the IMG does not indicate the syntactic category of a word form. For instance, the IMG of Germ. laufend is ‘run:part.prs’, showing that the form contains a morpheme whose function it is to mark a present participle. The gloss is not ‘run (part.prs)’ or anything of the sort, meaning that laufend is a present participle. While the latter is true, it is not the task of an IMG to give this information. Moreover, the type of morphological unit is not an object of an IMG. Thus, concepts like ‘stem’, ‘root’, ‘prefix’ do not appear in IMGs. Such information may, to a large extent, be inferred from a proper IMG, since the gloss of a root differs typographically from the gloss of a grammatical formative. Similarly, an IMG cannot replace a lexicon. Here again, elements appearing in an IMG are but names of elements appearing in the L1 line. They are not meant to exhaust the meaning of such an element. Finally, an IMG is not meant to replace an idiomatic translation. Thus, it cannot and should not render closely the sense of an L1 item in the given context. An IMG is regularly accompanied by a free translation which fulfills precisely this purpose.
2.
Prerequisites of morphological analysis
Interlinear glossing might appear to be just an elementary form of representing data. As a matter of fact, it presupposes a morphological analysis. The following analytic problems are directly reflected by the glosses. 2.1. Unmarkedness and zero morphemes Where the L1 text contains a morph, the IMG contains an element rendering it. Where the L1 text contains nothing, the issue of rendering it is complicated by markedness theory. Germ. Herr may be glossed by ‘master’ or by ‘master(nom.sg)’. Latin mone-t
may be glossed by ‘warn-3.sg’ or by ‘warn (ind.act)-3.sg’ (according to R16). Moreover, one may believe that such forms contain zero morphemes and put thus: Herr-Ø ‘master-nom.sg’, mone-Ø-Ø-t ‘praise-ind-act3.sg’. All of these IMGs are formally correct. The choice among them is not a matter of appropriate glossing, but of morphological theory. For interlinear glossing, only the general rule R1 is relevant. 2.2. Allomorphy If the L1 representation to be glossed corresponds to standard orthography, the analyst has no decisions to make in its regard. Otherwise, a good option for the representation (as well as for any writing system) is a morphophonemic representation which steers a middle course as far as allomorphy is concerned: Phonologically conditioned allomorphy is resolved (ignored), morphologically conditioned allomorphy is not resolved (is rendered). The IMG, on the other hand, shows morphemes, not allomorphs. In order to understand what this implies, consider three examples. Modern Yucatec Maya expresses completive and incompletive aspect by suffixes on transitive and (one conjugation class of) intransitive verbs as follows: aspect valence
completive
incompletive
transitive intransitive
-ah -Ø
-ik -Vl
Tab. 169.1: Aspectual suffixes in Yucatec Maya
For instance, t-u hats’-ah ‘past-sbj.3 beatcmpl (he beat it)’. One might think that the table contains four morphemes. Actually, however, transitivity is inherent in the verb stem and conditions allomorphy in the aspect suffix. The conditioning factor should not make part of the gloss. That is, the correct gloss for -ah is not ‘tr.cmpl’, but simply ‘cmpl’. See also 4.5. Yucatec Maya also has personal clitics that precede nouns as possessive cross-reference markers and verbs as subject cross-reference markers. If the noun or verb starts with a vowel, a glide is inserted in its front. The choice between the two glides w and y is morphologically conditioned: If the pronoun is of first person singular or of second person, it
1839
169. Interlinear morphemic glossing
is w; if the pronoun is of third person, the glide is y. For instance, in watan ‘poss.1.sg Ø:wife (my wife)’, u yatan ‘poss.3.sg Ø:wife (his wife)’. It is also possible to regard the noun forms modified by the initial glide as stem allomorphs, in which case the glide would not even receive the gloss by ‘Ø’. However, in the third person, a pronominal clitic followed by the glide can be omitted. Thus, yatan by itself means ‘his wife’. (Historically, the glide is indeed a reflex of an older cross-reference marker). We therefore have u y-atan ‘poss.3 Ø-wife’ ~ y-atan ‘poss.3wife’, and we face the problem that the same element is not even a morph in one context, but a full-fledged morpheme in another. Whatever the correct morphological analysis may be, the IMG presupposes it and brings it out. Last, consider gender marking in a language such as Latin (cf. Art. 48). Puellae bonae means ‘good girls’, pueri boni ‘good boys’. Apart from motion, gender is inherent in a noun stem. It is, however, recognizable by the declension suffixes. Nevertheless, the gloss of the morph in question does not contain the conditioning category. The noun forms will be glossed ‘girl.f:nom.pl’, ‘boy.m: nom.pl’, implying that gender is a category of the stem, not of the suffix. What about the adjectives? Gender is not inherent in an adjective stem. We may therefore gloss them by ‘good:nom.pl.f’ and ‘good:nom.pl.m’. Then one and the same element would be a morpheme on adjectives, but a conditioned allomorph on nouns, and therefore it would get two different glosses. Since two different glosses for the same element are not admissible in interlinear glossing (R4), this would entail that there are two homonymous declension suffixes -ae in Latin, which is obviously undesirable. We may stop this consideration here, since the problem is obviously not one of glossing, but one of morphological analysis. R2 codifies the convention that IMG expressions represent morphemes, not allomorphs.
3.
Principles of interlinear glossing
3.1. General In the canonical trilinear representation, one L1 text line is matched by two L2 lines, the IMG and the free translation. This entails a division of labor between the two L2 representations. The free translation is the idio-
matic semantic equivalent of the L1 line, the IMG is a representation of its morphological structure. There is consequently no need for the translation to be particularly literal, just as there is no need for the IMG to repeat the morphemes that appear in the translation. For instance, a polysemous L1 item will be rendered by its contextual sense in the free translation, but by its generic meaning in the IMG (R8). Unnecessary parallelism between the two L2 lines is redundant; the trilinear canonical representation offers an occasion to provide additional information. In principle, the degree of detail displayed in an IMG depends on the purpose the example with its gloss is meant to serve. However, the author cannot foresee the purposes to which others will want to use his examples. A morphological detail that is not at stake in the current discussion may be essential for the argument another linguist may wish to base on the example. For this reason, the principle is to allow for as much precision and detail as seems tolerable (R3). The following rules specify the properties of a complete IMG. They do not exclude less detailed IMGs where they suffice. Cf. R13 and R23 for possibilities of underspecifying morphological structure. The IMG of a morpheme is some sort of name for it, a name that alludes to its meaning or function and is insofar mnemonic or, at least, more helpful to the non-specialist than the L1 morph itself. It must therefore have a certain recognition value. R4, which actually is a tightening of R1, therefore requires that given a particular L1 morpheme, its gloss will be the same in all contexts; and apart from full synonymy, no two morphemes of L1 will have the same gloss. These points will be elaborated in the following subsections. 3.2. Glossing vocabulary Glosses are taken from a language L2 that serves as a metalanguage of L1. L2 is based on a natural language ⫺ in this article, English ⫺, but with far-reaching deviations from natural language use. The glossing vocabulary consists of the following kinds of symbols: ⫺ vocables: ⫺ L2 morphemes and stems ⫺ grammatical category labels ⫺ boundary symbols. The difference between the two kinds of vocables is the following: Morphemes and stems
1840 are taken from natural L2 vocabulary and are meant to be translation equivalents (in a sense to be made precise below) of L1 items. For instance, the notation “Germ. Schreibtisch ‘write-table (desk)’” is to be interpreted thus: The German word form Schreibtisch ‘desk’ consists of two morphs, of which schreib- means ‘write’ and tisch means ‘table’. Grammatical category labels, on the other hand, are taken from scientific terminology and are meant to categorize the function of L1 items. For instance, “Germ. schreib-en ‘write-inf (write (inf.))’” is to be interpreted thus: The German word form schreiben ‘write (inf.)’ consists of two morphs, of which schreib- means ‘write’, while -en is an infinitive marker (that is, -en does not mean ‘infinitive’; it is the German word Infinitiv which means ‘infinitive’). To bring out this essential difference between the two kinds of IMG vocables, L2 morphemes and stems are written in straight orthography, while grammatical category labels are written in (small) capitals (R29). A grammatical category label represents (i.e. is the name of) the value of a grammatical category (the latter being taken, technically, as a parameter or attribute). For instance, the label ‘acc’ is the name of the value ‘accusative’ of the morphological category ‘case’. Just like a grammatical category label is a name of a value of a grammatical category, what is called ‘L2 morphemes and stems’ are actually names of L2 morphemes and stems. In the following, we will abide by the simpler way of speaking. The choice and use of vocables are treated in the following subsections; boundary symbols are treated in section 4. 3.3. Lexemes An L1 lexeme is, in principle, glossed by an L2 lexeme (R5(a)). Sometimes more than one L2 word is necessary, for instance in Germ. fabulieren ‘invent.stories’. However, profusion is to be avoided. Adjectives that do not require a copula in predicative function are often glossed by adding a copula, e.g. West Greenlandic anurli ‘windy’ is glossed ‘be. windy’ in Fortescue (1984: 65). This is only correct if a word of this class requires an attributor in attributive function. Otherwise it wrongly implies that there is no difference between adjectives and verbs, and it tends to obscure the fact that the language does not use a copula with adjectival predicates.
XX. Morphologie in der Praxis
L1 cardinal numerals are glossed by Arabic numbers. An issue arises for proper names, which are often not glossed at all. However, there is no room here for an exception to the general rule: a proper name is rendered by its counterpart in L2. Some proper names have conventional counterparts that are specific to L2; Engl. John corresponds to Germ. Hans, and Engl. Munich corresponds to Germ. München. These then appear in the IMG. Whenever there is no such languagespecific convention, the counterpart of an L1 name is usually the same word in L2. If L2 is English, no problem arises for the form in which L2 lexemes are quoted in the IMG. In other languages, lexemes have a citation form in conformity with L2 conventions. If this is an inflected form, like the nominative for nouns or the infinitive for verbs, then it is excluded from an IMG by R5 (b), and instead the bare stem must be used. The reason is that such a gloss would seem to imply that there is a nominative, or an infinitive, in the L1 line where actually just a stem is being glossed. 3.4. Grammatical formatives L1 morphs are, in principle, glossed by citation forms of L2 morphemes. However, interlinear morphemic glossing crucially revolves around grammatical properties of L1 items. These will differ between L1 and L2. Even if, in a number of cases, the L2 stem appearing in a gloss has the same grammatical properties as the L1 morph that it represents, this cannot be expected and therefore not be relied upon. For instance, Latin eum could be glossed by Engl. him, and at the typological level, they do share a number of features. However, eum is accusative and can thus not be indirect object, while him is the form for direct and indirect object. Therefore, grammatical items of L1 are generally not glossed by grammatical items of L2, but by a configuration of symbols taken from the scientific metalanguage and representing their grammatical features, i.e. by grammatical category labels (R6). Thus, Latin eum may be glossed by ‘ana:acc.sg.m’. No bound grammatical or derivational morphemes should appear in IMGs. Free grammatical morphemes may be used to render free grammatical morphemes. However, use of those in the second column of Tab. 169.2 is discouraged unless L1 happens to exhibit the same ambiguity as English:
1841
169. Interlinear morphemic glossing
word class
instead of
use
copulas, auxiliaries
be have (except to mean ‘possess, own’)
cop, pass, prog ... pf, oblg ...
prepositions
by with for as from to of
ag, erg ... inst, com, assoc ... ben, dest ... eqt, ess ... abl, del ... dat, all, dest, term, inf ... gen, assoc ...
subordinators
that if
comp, sr (, d3) int, cond.sr
relativizers
that who which
rel rel.hum.nom ... rel.nhum.nom ...
Tab. 169.2: Free grammatical morphemes
Some morphemes are extremely deeply entrenched in the semantic or pragmatic system of the language and simply have no translation equivalent in L2. Two common ways out are a) to repeat the significans of the item in the gloss, and b) to indicate the class of the item instead of its meaning. Thus, we find the German modal particle eben glossed either as ‘eben’ or as ‘ptl’. Both glosses are inadequate. If there is no translation equivalent in natural L2, then the linguist has a specialized metalanguage to describe such functions. For the sake of an IMG that is not devoted to modal particles in particular, a gloss like ‘reaff’ (reaffirmed) will be fully sufficient and more helpful than either of the aforementioned. A gloss is a proper name of an L1 morpheme. It does not give information on the grammatical class of the morpheme in question other than what is implied by the name itself. If a gloss is ‘acc’, one assumes that the morpheme belongs to the grammatical class of the case morphemes. It is the task of the grammar to clarify whether or not this implication is correct in a particular case. The gloss will not be ‘case.acc’ or anything of this sort. For the same reason, the gloss of the perfective aspect is simply ‘prfv’ and not ‘prfv.asp’, and so on. From this it follows that the gloss will not be ‘asp’ either. In the literature, one frequently encounters glosses such as ‘ptcl’ (particle), ‘agr’ (agreement), ‘art’ (article). If
L1 possesses only one particle, agreement morpheme (hardly imaginable) or article (this is possible), then these glosses are sufficient. In all other cases, this kind of gloss is not helpful because it does not give the information on the meaning or function of the morpheme that a gloss is supposed to give. Moreover, the whole glossing becomes inconsistent, as some glosses name particular morphemes, while others name the class a morpheme belongs to. More on this in section 3.9.1. 3.5. Ambiguity Each morpheme of L1 should be recognizable by its gloss. The reader is supported in this task if glosses are consistent within one publication. It will rather confuse him if Yucatec Maya k’ı`in is once glossed ‘sun’ and the next time ‘day’. Polysemy is resolved in the idiomatic translation. The gloss renders neither the contextual sense nor the full meaning range of an item. Naturally, this does not apply to homonymy. Homonymous L1 morphs represent different morphemes and therefore receive different glosses. This is stipulated by R7, which follows from R4. If the senses of an item are reducible to a Gesamtbedeutung, then this should be used in the gloss (R8). For instance, the Turkish dative/allative suffix -a is glossed by ‘dat’. The Gesamtbedeutung rather than the Grundbedeutung should appear in the gloss, because it has better chances to fit all the diverse
1842 contexts in which the item occurs. Sometimes, there is either no Gesamtbedeutung, or if there is, L2 does not have a term for it. In cases like Yucatec Maya k’ı`in ‘sun, day’, there are various alternatives. First, the Grundbedeutung may be used as the gloss; thus Yucatec Maya k’ı`in ‘sun’. However, if all the occurrences of a polysemous morpheme in a particular publication reflect the same (derived) reading, then generally no useful purpose is served if it is consistently glossed by its basic meaning. For instance, all the occurrences of Yucatec Maya k’ı`in in a particular text might mean ‘day’. Then this would be the appropriate gloss. Finally, any kind of reduction may seem misleading. Then two or even more senses may be indicated in the gloss, separated by a slash, e.g. Yucatec Maya k’ı`in ‘sun/day’. (2) illustrates the same convention. (2) Korean Toli-n⁄n kae-hako cal Toli-top dog-add often/well non-ta. play:prs-decl ‘Toli likes to play with the dog.’ Syncretism often involves extensive polysemy and/or homonymy. If it should be made explicit in an IMG, then e.g. the gloss for Lat. ancillae would have to be ‘maid.f:gen.sg/ dat.sg/nom.pl’. This may be appropriate if the discussion in the context deals with syncretism. Otherwise, only the category actually required by the context may be shown, e.g.: (3) Latin ancillae orant maid.f:nom.pl pray:3.pl ‘the maids pray’ In other words, in cases of syncretism the last two options of R8 must be resorted to. A whole paradigm of markers may be used in two clearly distinct functions. For instance, a set of cross-reference markers may combine with a verb to reference its subject, and with a noun to reference its possessor. Here again, the two alternatives mentioned are open: either gloss the verb markers by ‘sbj’ and the noun markers by ‘poss’, or gloss them by ‘sbj/poss’ in both positions (which is, actually, never done). A third alternative ⫺ one that is actually resorted to in Mayan linguistics; cf. Art. 170, section 6.1.2 ⫺ is to coin
XX. Morphologie in der Praxis
a concept and a term for a paradigm that is used in these two functions and use this in the IMG. 3.6. Features and functions As remarked in section 1.4, an IMG cannot fill the place of a grammar. In particular, the grammatical category label that represents a morpheme in the gloss cannot possibly represent the full functionality of that morpheme. It can only serve as a mnemonic identifier for the reader. We just saw that the full polysemy of an item cannot be accounted for in a gloss. The same goes for functional information associated with a morphological position. If the slot filler is a verb agreement affix or crossreference marker, then its meaning is in the sphere of person, number and gender. Consider conjugation endings as in Germ. lieb-e ‘love-sbj.1.sg’, lieb-st ‘love-sbj.2.sg’, lieb-t ‘love-sbj.3.sg’. The information that these suffixes cross-reference the subject is functional information associated with the morphological slot. It must be given in the grammar; the IMG may simply read lieb-e ‘love1.sg’ etc. The same would apply, in principle, if the verb cross-references more than one of its dependents. Here, however, it has become customary to distinguish the references of the cross-reference markers by indicating their syntactic function, as in (4). (4) Swahili ni-li-mw-ona m-toto sbj.1.sg-pst-obj.cl.1-see cl.1-child ‘I saw the/a child’ The information that the initial prefix references the subject, while the one following the tense prefix references the direct object must be contained in the grammar. The task of the gloss is to identify the particular element, not to specify the rules of its use. Insofar, adding functional information concerning the morphological slot itself ⫺ ‘sbj’ and ‘obj’ in (4) ⫺ is a service to the reader that may be useful, but that also clutters up the gloss (cf. R3). The distinction between morphological categories and syntactic or semantic functions is also relevant in the domain of case and valence. The frequent confusion among syntactic/semantic functions, cases and valence-derivational functions also manifests itself in glossing habits. One frequently encounters glosses such as Turkish ates¸-in ‘fire-
1843
169. Interlinear morphemic glossing
poss’ instead of ‘fire-gen’, ates¸-e ‘fire-io’ instead of ‘fire-dat’ or ‘... -send-dat ...’ instead of (5). The quality of the glossing reflects the quality of the morphological analysis. (5) Swahili Musa a-li-ni-andik-ia Musa sbj.cl.1-pst-obj.1.sg-send-appl barua letter ‘Musa sent me a letter’ 3.7. Derived stems The morpho-semantic structure of a derived stem may be completely regular and transparent, as in Germ. wolk-ig ‘cloudadjvr (cloudy)’, or it may be opaque, as in Germ. heil-ig ‘salvation-adjvr (holy)’. If the discussion focuses on word-formation, then both of these words will be glossed as indicated. If the internal structure of stems is of no relevance, then it will not be shown in the L1 text line, and consequently the glosses can reduce to ‘cloudy’ and ‘holy’, respectively. For opaque complex stems, morphological segmentation plus corresponding gloss often amounts more to etymology than to morphological analysis. It also unnecessarily obscures the correspondence of the gloss to the idiomatic translation. This should be borne in mind before one carries it through as a general principle in text editions. In an ideal methodological situation, an IMG is taken from a lexicon, where the gloss constitutes one of the fields in the microstructure of each lexical entry. The German lexicon may contain, e.g., the three entries Huf ‘hoof’, Eisen ‘iron’ and Hufeisen ‘horseshoe’. If the latter occurs in an L1 text, then it may either be analyzed or not. In the former case Huf and Eisen will be looked up in the lexicon and will be matched by their glosses, while in the latter case Hufeisen will be looked up and be glossed accordingly. 3.8. Submorphemic units There are two kinds of submorphemic units: parts of morphemes with a sound-symbolic value and strings of phonemes inserted between morphemes for euphonic or similar reasons. The former kind is not generally subjected to morphemic analysis and may therefore be left out of consideration here. The latter kind may be illustrated by the second element in forms such as French a-t-il ‘has he’ and Germ. Weihnacht-s-gans ‘Christmas goose’. If the submorphemic unit is not
at stake in the context, then the first choice is to abstain from an analysis by regarding the submorphemic unit as part of a stem alternant: Weihnachts-gans ‘Christmas-goose’. The second choice is to render the submorphemic unit by Ø, e.g. a-t-il ‘has-Ø-he’. A euphonic submorphemic unit may be glossed by ‘eu’ instead of ‘Ø’. 3.9. Grammatical category labels 3.9.1. General As was said in 3.4, the gloss for a grammatical item is generally not a grammatical item of L2, but a grammatical category label (R6). For instance Yucatec Maya ya`an is not rendered by ‘be’, but by ‘exist’, one of the reasons being that L2 ‘be’ is a copula, while Yucatec Maya ya`an is not. While this poses few problems for such categories for which the European grammaticographic tradition possesses terms, it does pose a problem for certain classes of semi-grammaticalized items such as function verbs and coverbs. Coverbs are words which are grammaticalized from verbs to minor parts of speech, mostly adpositions. If they function as the latter, they may express a semantic role. In Mandarin, for instance, yo`ng has the lexical meaning ‘use’ and the grammatical meaning ‘instr’, as in (6). (6) Chinese Ta¯ yo`ng sho˘u zo˘u lu`. he use/INSTR hand walk road ‘He walks on his hands.’ This kind of problem is not solved by putting the lexical meaning in upper case (‘use’), since ‘use’ is neither a grammatical concept in L2 nor a term of the grammatical metalanguage. Applying R8 in such cases would imply opting in favor of the Gesamtbedeutung of the item, which in such cases is the grammatical meaning. The gloss would then be ‘instr’ (or some more language-specific grammatical category which may suit better this particular function). The problem remains, however, that the same word can occur as the sole predicate of a clause, in the meaning ‘use’ (e.g. ta¯ yo`ng sho˘u ‘he uses his hand’). An IMG ‘instr’ would be hardly intelligible there. The alternative of only using the Grundbedeutung ⫺ ‘use’ in (6) and throughout ⫺ would be in conflict with the principle that morphological analysis must be kept distinct from etymology. Here the third alternative offered by rule R8 may be re-
1844
XX. Morphologie in der Praxis
label
intended meaning
comment
a adv agr agt art asp aux card clf clt ep
transitive subject adverb agreement agent article aspect auxiliary cardinal classifier clitic epenthetic
evid pat prep ptl tns
evidential patient preposition particle tense
in morphemic glosses, the abbreviation is ERG specify meaning specify agreement categories this is not a value of a morphological category only if it has no determinative properties specify particular aspect only if there is only one auxiliary morpheme in the language only if it is a morpheme or grammatical feature this is a word class this is neither a morphological category nor a value of one has no morphological status, should not be separated in the first place specify particular evidential this is not a value of a morphological category this is a word class this is (at best) a word class specify particular tense
Tab. 169.3: Labels to be avoided
sorted to, viz. providing both meanings in the gloss of each occurrence of the item, thus: yo`ng ‘use/instr’. An IMG identifies an L1 morpheme. It names a value, not a parameter. Mentioning the name of the generic category in the gloss instead of the specific value is nevertheless widespread usage. One finds both Japanese yom-i and yon-de glossed by ‘read-conv’ (converb), which hinders the reader in his attempt to keep the converb forms apart. One finds Onondaga wa? ha-ye? kwa-hnı´:-nu? ‘he bought tobacco’ glossed as ‘tns:he/it-tobacco:buy-asp’ (Woodbury 1975: 10), which is of no use for somebody studying the interdependence of incorporation with tense and aspect. IMGs not seldom contain labels that do not correspond to the principles introduced so far. Sometimes, elements without morphological status are separated and glossed. Sometimes, the parameter instead of the particular value of a grammatical category is identified. Sometimes, syntactic or semantic instead of morphological information is given. Here is an incomplete list of labels that have repeatedly been found in glosses but which should be avoided. 3.9.2. List of grammatical categories and their glossing labels No list of grammatical category labels can be complete. The list following in Tab. 169.4
(which incorporates the list in Lehmann et al. 2 1994) only contains the most widespread categories. When more than one abbreviation is mentioned, they are given in the order of preference. To the extent that these abbreviations are or become wide-spread, they get the status of linguistic abbreviations like ‘NP’, which need not be defined when used. If a publication uses labels not contained in the following list, it must explain them in an individual list of abbreviations. Grammatical category labels are subject to two conflicting requirements: they must be both distinct and short. The former requirement takes precedence. It is, for instance, not possible to use ‘comp’ in one and the same publication to mean both ‘completive’ and ‘complementizer’. The list in Tab. 169.4 avoids such clashes. However, in an individual publication that has nothing to do with complementation, the aspect may, of course, be abbreviated by ‘comp’ (instead of ‘cmp(l)’, as in the list). Parenthesized parts of an abbreviation are only necessary if a distinctness conflict arises. Tab. 169.4 contains only such terms which may appear in an IMG. In other publications, similar lists of terms for syntactic categories and functions and for semantic and pragmatic functions may be found. ‘Cross-reference position’ means a morphological slot, usually on a verb, occupied by pronominal elements that agree with or
1845
169. Interlinear morphemic glossing value
abbrev.
category
1st person 2nd person 3rd person abessive
person person person
ablative absolute
1 2 3 (prv) (avers) abl absl
absolutive
abs
abstract accusative action nominalizer active
abstr acc acnnr act
actor
acr
actor topic additive addressee-honorific addressee-humble adelative adessive adhortative aditive adjectiv(al)izer admonitive adverbializer adversative affirmative agent nominalizer agentive alienable
a add 2hon 2hml adel adess (hort) (all) adjr adm advr adrvs affmt agnr ag al
allative allocutive anaphoric andative animate anterior anticausative
all alloc ana and an ant acaus
antipassive aorist
apass aor
voice tense-aspect
applicative
appl
deverbal verb derivation
apprehensional assertive
appr asrt
interpropositional relation modality
associative assumed attenuative attributor auditory augmentative auxiliary benefactive cardinal caritive
ass(oc) assum atten at aud aug aux ben card (prv)
adnominal case evidential deverbal verb derivation nominal evidential denominal nominal derivation
comment
use ‘privative’ and ‘aversive’ local case nominal grammatical case or cross-reference position nominal grammatical case deverbal nominal derivation 1. voice; 2. case or cross-reference position grammatical case or cross-reference position voice case honorification honorification local case local case
‘from’ (⫽ separative) free non-incorporated form of noun in ergative system
2. in active system
use ‘hortative’ use ‘allative’ derivational or syntactic mood derivational or syntactic interpropositional relation opposite to negative deverbal nominal derivation possessive attribution morpheme local case honorification pronominal deictic tense deverbal verb derivation
case numeral
‘whereas’ normally unmarked
‘to’ kind of addressee-honorific
relative tense ⫽ deagentive, blocking of actor argument perfective past (as opposed to imperfect) subtypes may be distinguished by appl.rec, appl.inst etc. ‘lest’ subtype of declarative: high degree of commitment ‘with, a`’ links an attribute to the head if it is the only auxiliary root ‘for’ if marked grammatically use ‘privative’
1846
XX. Morphologie in der Praxis
Tab. 169.4: continued value
abbrev.
category
causative circumstantial clamative classifier
caus circ (excl) clf
deverbal verb derivation interpropositional relation
cohortative collective comitative common
(hort) coll comit comm
comparative complementizer completive conative concessive conditional
cmpr comp cmp(l) cntv conc cond
conjectural conjunctive connector, -ive consecutive construct converb continuous copula crastinal dative deagentive debitive declarative deferential definite deictic of 12 person deictic of 1st person deictic of 2nd person deictic of 3rd person delative demonstrative dependent verb form desiderative destinative
conjc conj conn consec const (ger) cont cop cras dat (acaus) (oblg) decl defr def d12 d1 d2 d3 del dem (subj)
determiner
det
deverbal verb derivation local case; also on non-finite verb forms (⫽ supine) pronominal
detransitivizer
detr
deverbal verb derivation
different subject diminutive direct direct evidential direct object directional
ds dim dr direv do dir
denominal noun derivation voice evidential cross-reference position case or verb derivation
distal distributive donative
dist distr don
des dest
nominal
case gender degree of comparison subordinator aspect mood interpropositional relation interpropositional relation; mood evidential interpropositional relation interpropositional relation nominal
comment ‘in, by’ use ‘exclamative’ followed by class identifier, e.g. hum use ‘hortative’ ‘with, in the company of’ either masc. or fem.; cf. ‘human’ and ‘animate’ ⫽ sr normally ⫽ perfective ‘although’ ‘if’; ‘would’ of non-finite predicate if there is only one ‘so that’ construct state use ‘gerund’
aspect/aktionsart tense grammatical case sentence-type honorification determination determination determination determination determination local case determination
if there is only one tomorrow use ‘anticausative’ use ‘obligative’ normally unmarked ~ speaker-humble
‘down from’ use ‘subjunctive’
determination nominal or verbal
‘to’; if typically for human destinations, use ‘benefactive’ will normally be def, indef, gnr, spec, nspec see also ‘anticausative’ and ‘introversive’ vs. inverse ‘towards’; use and and ven for deictic directionals remote from deictic center auxiliary of benefactive construction
1847
169. Interlinear morphemic glossing Tab. 169.4: continued value
abbrev.
category
comment
dual dual exclusive dual inclusive dubitative durative dynamic egressive elative emphasizer/emphatic equative
du, dl de di dub dur dyn egr elat emph eqt
number number number mood aktionsart aktionsart aktionsart local case funct. sentence perspective 1. case; 2. predicative
vs. stative
ergative
erg
essive evidential exclamative exclusive
ess evid excl (de, pe)
grammatical case or cross-reference position case verbal mood
exist(ential) experiential extrafocal
exist exper exfoc
grammatical verb aspect verbal
extraversive
extrv
deverbal verb derivation
factitive
fact
familiar feminine fientive
fam f (inch, proc) fin 12
denominal/deadjectival verb derivation pronominal gender denominal verb derivation
finite first person dual inclusive focus formal frequentative
‘out of’ e.g., class of pronoun ‘as’; feature/marker of adjective in nominal clause in ergative system ‘as’; see also ‘transformative’ if there is only one use ‘dual exclusive’, ‘plural exclusive’ status of subordinate clause of cleft-sentence transitivization by addition of undergoer A-fact NP ‘make NP A’
use ‘inchoative, processive’
verbal if treated as a quasi-singular; otherwise ‘dual inclusive’
foc frm freq
funct. sentence perspective mood aktionsart
future generic genitive gerund gerundive habitual habitual-generic habitual-past hesitative hesternal hodiernal future hodiernal past honorific hortative human humble
fut gnr gen ger (oblg) habit
tense determination grammatical case verbal use ‘obligative’ aktionsart
hesit hest hodfut hodpst hon hort hum hml
funct. sentence perspective tense tense tense honorification mood honorification
hypocoristic hypothetical illative immediate
hcr hyp ill imm
affect mood local case tense
multiple times on several occasions
verbal adverb or converb ~ customary use ‘habitual’, ‘generic’ use ‘habitual’, ‘past’ yesterday’s past today’s future today’s past 1st person imperative comprises ‘speaker-humble, addressee-humble, referent-humble’ ‘into’ specifier of other tenses
1848
XX. Morphologie in der Praxis
Tab. 169.4: continued value
abbrev.
category
comment
immediate/imminent future immediate past imperative imperfect imperfective impersonal
immfut
tense
(recpst) imp impf ipfv impr
mood tense-aspect aspect
imperfective past; vs. aorist
impersonal passive
ips
voice
inactive
inact
inalienable
inal
grammatical case or cross-reference position nominal
only if formally distinct from the specific persons passive without promotion to subject in active system
inanimate inceptive inchoative inclusive
inan (ingr) inch (di, pi)
denominal verb derivation
incompletive, noncompletive inconsequential indefinite independent indicative indirect object inessive inferential infinitive ingressive injunctive instructive instrument nominalizer instrumental intensive interrogative
incmp(l)
aspect
incons indef indep ind io iness infr inf ingr inj (man) instnr
interpropositional relation determination mood mood cross-reference position local case mood or evidential verbal aktionsart mood
inst(r) ints int
case verbal sentence type
intransitive
intr
verbal
intransitive subject
s
cross-reference position
introversive inverse invisible irrealis iterative jussive
intrv inv invs irr iter juss
deverbal verb derivation usually verbal determination mood aktionsart mood
lative ligature linker
lat lig lnk
local case nominal nominal
locative locative topic manner nominalizer masculine masculine personal
loc lt mannr m mhum
local case voice deverbal nominal derivation gender gender
use ‘recent past’
possessive attribution morpheme or feature use ‘ingressive’ N/A-inch ‘become N/A’ use ‘dual inclusive’, ‘plural inclusive’ normally ⫽ imperfective
only if distinct from indicative ‘inside’
use ‘manner’ deverbal nominal derivation often aktionsart particle or morphological category morpheme or grammatical category only if opposed to both a and p; use sbj otherwise blocking of undergoer argument vs. direct several times on one occasion 3rd ps. imperative or dependent mood ‘to ~ from ~ via’ links subconstituents of a phrase, typically an NP; properly includes ‘attributor’
1849
169. Interlinear morphemic glossing Tab. 169.4: continued value
abbrev.
category
logophoric malefactive manner medial
log mal man med
pronominal or verbal deverbal verb derivation case determination
medial mediative mediopassive middle motivative narrative near future negative neuter nominalizer
medv medt medp mid mtv narr nrfut neg n nr
verbal case voice voice case tense tense
nominative nonnon-finite non-future non-human non-masculine personal non-past non-plural non-singular
nom n nfin nfut nhum nm npst npl nsg
non-specific non-visual non-volitional noun class n object obligative oblique obviative optative ordinal participle (marker) partitive passive past patient nominalizer patient topic paucal pejorative perfect perfective pergressive perlative place nominalizer pluperfect plural plural exclusive plural inclusive pluritive polite positional positive possessive
nspec nvis nvol cln obj oblg obl obv opt ord ptcp prtv pass pst patnr pt pau pej p(r)f pfv (perl) perl locnr plup pl pe pi (pl) (frm) posit (affm) poss
gender deverbal nominal derivation or syntactic subordination grammatical case
comment
also on non-finite verbs medial distance from deictic center verb form in a chain ‘between, among; by means of’ excludes passive ‘by’; sometimes called ‘causal’ after ‘immediate future’ see also the more specific ones e.g. npst
verbal tense gender gender tense number number determination evidential verbal
<3 > 1; only if there is a plural for > 2 non-eye-witness where n is a number or a feature
cross-reference position mood case person mood numeral verbal case voice tense deverbal nominal derivation voice number affect tense-aspect aspect local case deverbal nominal derivation tense number number number
vs. proximate
use ‘perlative’ ‘through’ past or perfect of a past
plural of a singulative; use ‘plural’ use ‘formal’ verbal possessive adjective, pronoun and cross-reference position
use ‘affirmative’ not for an adnominal case relation; that is gen or at
1850
XX. Morphologie in der Praxis
Tab. 169.4: continued value
abbrev.
category
comment
postcrastinal postelative posterior postessive post-hodiernal potential precative predicative present preterite pre-hesternal primary object privative processive, -ual progressive prohibitive prolative proprietive prospective proximal proximate punctual purposive quality nominalizer quotative realis recent past reciprocal reduplicative referent-honorific referent-humble referentive reflexive reinforcement relational(izer) relative
pocras postel post postess pohod pot prec pred prs (pst) prhest po pr(i)v proc prog proh prolat propr prosp prox prx pnct (dest) qualnr quot rls recpst rec(p)
tense local case relative tense local case tense mood mood nominal tense
future after tomorrow ‘from behind’
3hon 3hml rfr r(e)fl (intns) rell rel
honorification honorification case voice or pronominal
relative remote remote past repetitive reportative resultative reversive same subject secondary object semelfactive sensory separative sequential simultaneous singular singulative sociative speaker-honorific speaker-humble specific speculative stative
(rfr) (dist) rempst rep rprt res rvrs ss so smlf sens (abl) seq sim sg sgt soc 1hon 1hml spec specl stat
tense cross-reference position case denominal verb derivation aspect mood local case case or denominal derivation tense-aspect determination person aspect or aktionsart
‘behind’ future after today for requesting predicative form use ‘past’ past before yesterday ‘without’ negative imperative ‘along, by (way of)’ ‘having, provided with’ ‘going to’; opposite of perfect near the deictic center vs. obviative use ‘destinative’
deverbal nominal derivation mood tense voice or pronominal
marking indirect speech vs. irrealis ⫽ immediate past gloss by function ‘about’ use ‘intensive’
nominal subordinative and/or pronominal
in relative clause use ‘referentive’ use ‘distal’
tense aktionsart evidential aspect or aktionsart aktionsart
only if distinct from iterative
cross-reference position aktionsart evidential interpropositional relation interpropositional relation number nominal verbal honorification honorification determination evidential aktionsart
use ‘ablative’ vs. simultaneous vs. sequential restricted vs. collective ‘together’
1851
169. Interlinear morphemic glossing Tab. 169.4: continued value
abbrev.
category
comment
subelative subessive subject subjunctive sublative subordinator
subel subess sbj subj subl sr
local case local case cross-reference position mood local case interpropositional relation
‘from under’ ‘under’
superdirective superelative superessive superlative super-lative terminative topic transformative
(supl) supel supess sup supl term top trnsf
local case local case degree of comparison local case local case or aktionsart funct. sentence perspective case
transitive
tr
verbal
transitive patient
p
cross-reference position
transitive subject
a
cross-reference position
transitivizer translative trial undergoer unrestricted unspecified
trr trnsl trl ugr (pl) unspec
deverbal verb derivation local case number cross-reference position
validator venitive verbalizer visible visual vocative volitional, volitive zero
ven vr, vbz vs vis voc vol Ø
deictic verb derivation determination evidential case verbal
person
‘to under’ only for the single universal subordinator (‘that’) use super-lative ‘from above’ ‘above’ ‘to above’ ‘up to’ ‘becoming’; dynamic counterpart of essive morpheme or grammatical category only if opposed to both s and a; use obj otherwise only if opposed to both s and p; use erg otherwise ‘across’ only if distinct from paucal use ‘plural’ unspecified argument of relational base use ‘assertive’, ‘declarative’
eyewitness
making no contribution to sentence meaning
Tab. 169.4: Grammatical category labels
refer to a dependent in a specific syntactic function. ‘Case’ means a case relator that may take the form of a case affix or an adposition. Verb derivational morphemes get these glosses only if they are homonymous with nominal case relators.
4.
Boundary symbols
4.1. Basic rules Rules R1 and R4 guarantee correspondence between units in the L1 text and in the IMG.
They do not, however, insure that the vertical alignment works in a mechanical way. This is desirable in certain contexts such as automatic parsing. It can be guaranteed in a fully formalized representation, which would then take the form of a table (s. Lieb & Drude 2000). In less formal situations, it cannot be fully guaranteed because there may be good reasons not to insert morpheme boundaries in the L1 text while still representing each morph by a separate gloss (cf. R13). Correspondence of boundary symbols in the L1
1852 and the IMG lines is therefore not generally an equivalence, but only an implication: boundary symbols in the L1 line are matched by corresponding boundary symbols in the IMG (R9). We will review the kinds of boundaries and their delimiters in turn. The word boundary is shown by a blank in L1. This is repeated in the IMG, and conversely there is a blank in an IMG only if there is a corresponding blank in the L1 line. This particular rule (R10) is therefore stricter than R9. R10 prohibits two situations: a word being rendered by a sequence of two words; and a sequence of two words being rendered by one word. The first situation will be discussed in section 4.5. Sometimes a sequence of two L1 units (words or morphemes) corresponds to one L2 unit. In principle, this situation should not arise in the IMG because each of the L1 units should have its own gloss. However, it is possible that either the L1 units have no meaning in isolation or else mean something totally different than their combination, the latter being idiomaticized. In such cases, glossing them separately might give a misleading impression of the workings of the grammar. When the bisected L1 unit forms an orthographic unit (e.g. a compound), one may simply dispense with the analysis (cf. section 3.7). For instance, instead of Germ. be-komm-en ‘appl-come-inf’, one can write bekomm-en ‘get-inf’. If the orthography requires a boundary, as in Yucatec Maya le kah ‘when’, the first choice is to gloss the items separately (in this case, ‘def sr’) and to leave the semantic interpretation to the idiomatic translation. The second choice is to indicate the semantic unity of the two L1 items typographically by replacing the blank by a boundary symbol that does not interfere with the orthography, e.g. by an underscore: le_ kah ‘when’ (R11). If L1 orthography links the two items by another symbol that is also an IMG boundary symbol, as in Engl. vis-a`-vis ‘facing’, no satisfactory solution is known. Apart from special cases to be noted, the morpheme boundary is shown by a hyphen in L1 (R12). This is repeated in the IMG; and here again the converse applies, too. Apart from the vis-a`-vis type exception, this does not pose any problems. It does, however, happen that the L1 text contains a combination of two morphemes, but no boundary is shown between them. Various motivations for this are conceivable, be it that two morphemes are fused in a portmanteau morph,
XX. Morphologie in der Praxis
be it that the position of the boundary is not clear or irrelevant, be it that the analyst does not want to disfigure L1 orthography with boundary symbols. In such cases, a colon in the IMG is a hint at a morpheme boundary existing, but not shown in the L1 line (R13). The purpose of R13 is to allow the analyst to forgo a segmentation while still saving R1 and insuring biuniqueness of the other boundary symbols. Several examples may be seen in (1). The colon is also used to render a portmanteau morph, e.g. French au ‘dat: def’. More on this in section 4.5. Special symbols may be introduced to distinguish kinds of morpheme boundaries. For instance, the use of the plus sign to signal a boundary in compounding, as in German Weihnachts⫹gans ‘Christmas⫹goose’ is rather widespread; and occasionally it is also found in derivation, as in German wolk⫹ig ‘cloud⫹ adjvr (cloudy)’ (R14). No orthography distinguishes clitic boundaries from word and morpheme boundaries. If L1 is represented in conventional orthography, then the simplest solution for an IMG is not to distinguish them either. Thus French je le sais ‘I know it’ will be glossed as ‘sbj.1.sg do.3.sg.m know.sg’, while Latin itaque ‘and so’ will be glossed by ‘so:and’. If clisis is important or the L1 representation is non-orthographic, then the clitic boundary will be shown by an equal sign both in the L1 text and in the IMG, thus: ita⫽que ‘so⫽and’ (R15). If a zero morph or morpheme is represented in L1 by Ø (cf. section 2.1), no special measures need be taken. If it is not there represented, then its gloss is enclosed in parentheses (R16), like this: Lat. timor ‘fear.m (nom.sg)’. In this example, a stem is accompanied by two (complexes of) grammatical category labels, ‘m’ and ‘nom.sg’. The first is separated by a period because it corresponds to an inherent feature of the stem. The second is enclosed in parentheses because it corresponds to a separate morpheme. 4.2. Discontinuity Discontinuous units ⫺ words or morphemes ⫺ are like bisected units in that one semantic unit is represented by two expression units. However, they present the added difficulty that their parts are not adjacent, so the IMG has to make it explicit what belongs together. For a discontinuous stem or affix, diverse so-
1853
169. Interlinear morphemic glossing
lutions have been proposed in the literature. Among them is the proposal (Bickel et al. 2004) to repeat the same gloss under each part of the discontinuous item. However, this seems misleading, as the syntagmatic cooccurrence of synonymous L1 items is not at all rare ⫺ e.g. in hypercharacterization ⫺ and must be distinguished from discontinuity. An unambiguous solution for a circumfix is to set it off by angled brackets, like this: Germ. ge>lauf<en ‘<part.prf>run’ (run (part. prf.))’ (R17). Discontinuous words are rare. The first choice is to try and gloss each part independently, as done for the German circumposition um … willen ‘for’ in (7). (7) German um unser-es Heil-es willen for our-gen.sg salvation-gen.sg sake ‘for (the sake of) our salvation’ The second choice is to treat them by the same formalism as for circumfixes. Consider the case of preverbs. In several Indo-European languages, they may be distantiated from their host verb to yield a discontinuous verb stem. There are two options for glossing such discontinuous compounds: If the compounding is relatively transparent, one may prefer to provide the preverb and the base each with its gloss. If the compound is completely lexicalized, this might be misleading, and so it may be preferable to treat it as a discontinuous morpheme in the gloss, as in (8). (8) German es hör>t jetzt3.sg now ‘it stops now’ Infixes, too, require a special boundary symbol in order to insure that the root bisected by them is perceived as a unit. This is achieved enclosing them in angled brackets as shown in (9)⫺(10) (R18). (9)
Latin vic-o conquer<prs>-1.sg ‘I conquer’
(10) Indonesian t<el>unjukpoint ‘forefinger’
The gloss of a left-peripheral infix precedes the gloss of its host, the gloss of a right-peripheral infix follows it (Bickel et al. 2004). 4.3. Reduplication Reduplicative segments may have the same kinds of grammatical functions as affixes, and sometimes they are formally not easily distinguished from affixes. Therefore they must be glossed just like affixes, but at the same time they must be formally distinguished from affixes. This is achieved by providing the same kind of gloss for them as for grammatical formatives, but separating them by a tilde (R19; Bickel et al. 2004), as in (11)⫺(12). (11) Ancient Greek ge´~graph-a prf~write-1.sg ‘I have written’ (12) Yucatec Maya k’a´a~k’as intns~bad ‘wicked’ 4.4. Other morphological processes Morphological processes not covered by the above conventions comprise transfixation, internal modification, metathesis, subtraction and suprasegmental processes (cf. ch. VIII). These are like infixation in not being peripheral to the base, but they differ from it in that the grammatical meaning in question is not associated with a single string of segments which, if subtracted, leaves the base. The notation recommended here distinguishes them from the other morphological processes, but not from each other. Such a morpheme can hardly be signaled in the L1 representation. In the IMG, its gloss follows the gloss of the base, separated by a backslash (R20). An example of transfixation is the Arabic broken plural, as in buju¯t ‘house\pl (houses)’. Apophony, metaphony, e.g. German säng-e ‘sing\irr-1/3.sg (I/ he would sing)’, and tone shift, as in Yucatec Maya. ha`ats’ ‘beat\introv (beat (unspec. object))’ are treated in the same way. 4.5. Semantic and grammatical features The gloss of a grammatical morph often consists of a set of symbols. They are separated by a period, as in Germ. Tisch-es ‘tablegen.sg’ (R21). The same rule applies in the
1854 situation mentioned in section 3.3, where an L1 lexeme is glossed by more than one L2 words. These, too, are separated by a period, as in Germ. fabulier-en ‘invent.stories-inf’. Lexical stems fall into grammatical classes. Noun stems, for instance, have gender; verb stems have valence. If such grammatical categories are covert, this information is not deducible from (the gloss of) the lexical meaning. It therefore makes sense to represent it in the gloss of the stem. The Latin example puellae ‘girl.f:nom.pl’ of section 2.1 shows how this may be done for gender. The same would be possible with transitivity. Instead of Yucatec Maya hats’-ah ‘beat-cmpl’ as shown in section 2.2, we might put ‘beat.tr-cmpl’. It does not seem necessary to have a rule here beyond R3 and R21. The period between values of different morphological categories cumulated in one morpheme is dispensable between person, gender and number, provided the resulting letter sequence is unambiguous. Thus, Latin lauda-mus may be glossed as ‘praise (prs.ind)1.pl’ or ‘praise (prs.ind)-1pl’. Sometimes the period is used as a generalpurpose symbol to hide the lack of an analysis, including the function of the colon as regulated by R13. This is not recommendable if ⫺ as is usually the case ⫺ the period is also used in the function regulated by R21. Given R21, the notation Lat. orant ‘pray.3.pl’ would imply that orant consists of a single morph. An IMG should at least make the distinction between a morph and a grammatical feature of a morph. In other words, if the author knows the number and order of morphs in an L1 form, then he should indicate them. If the author does not even know so much, he should probably not use the example. Still, in emergency situations, R23 may be viable, which allows for linking IMG elements by an underscore without any implications for L1 morphological structure. This would allow for putting orant ‘they_pray’. 4.6. Composite categories Two cross-reference categories may share a morphological slot, as in (13). (13) Mayali Kamak kan-bolk-bukka-n good sbj.2&obj.1-country-show-npst ke. your ‘It is good that you will show me your country.’ (Evans 1997: 400)
XX. Morphologie in der Praxis
In principle, the case is analogous to one declension suffix showing both number and case. However, when actor and undergoer cross-reference is cumulated in one morpheme, sticking to R21 would lead to obscurity. Instead, information on the two dependents should be separated by ‘&’ or by ‘>’ (R22). The ‘greater than’ sign has two advantages here: it is iconic, and it dispenses with the use of function labels such as ‘sbj, obj, acr, ugr’ (simply ‘2>1’ in (13)). It has the disadvantage that the same symbol is used for discontinuous and infixed material, which may lead to conflicts. This case must be kept distinct from a portmanteau morph, viz. when two cross-reference categories that generally each have their own morphological slot fuse in one morph occasionally. There R13 applies. 4.7. Constituency The IMG abides at the level of morphology. The text may be represented at other levels in addition, if this is desired. Still, IMGs are used most frequently in publications on syntax, where not only morphological, but also syntactic properties of the examples are at stake. Very often it suffices to identify one constituent in the example, for instance the prepositional phrase or the relative clause that is the subject of analysis. Then no harm is done, but on the contrary the reader is helped in scanning the example, if constituency is shown by brackets. Thus in (14), the relative clause is identified by the bracketing. (14) Yucatec Maya le ma´ak chowak u ho’l-e’ def person [long poss.3 head]-d3 ‘the person who has long hair’ In principle, this may be done either in the L1 line or in the IMG (it need not be repeated in both). However, since the IMG line is the one that contains the grammatical analysis, the bracketing seems more natural there (R24). In principle, an IMG may even be combined with a labeled bracketing; but above some rudimentary level, this will soon lead to illegibility.
5.
Typographic conventions
IMGs obey a number of typographic conventions all of which aim at facilitating the reader’s task. First, if there are more lines of linguistic representation (cf. section 1.3), for instance one of syntactic constituency or lines
1855
169. Interlinear morphemic glossing
that show syntactic, semantic or pragmatic functions of the construction, then these follow the IMG, as stipulated in R25. Second, words (neither larger nor smaller units) of L1 are left-aligned with their glosses (R27). Further, since IMGs are generally longer than the L1 text they render, they are printed in a smaller type-face (R28), and grammatical category labels are abbreviated (R29). Here is an example of a publication which does not observe these rules (MonodBecquelin 1976: 138 on Trumai): sˇysˇyk letsi k’ate sˇy hai-ts sˇysˇy-ka-ke “avec du piment, je rends le poisson piquant (regarde)” // piment / avec / poisson / actualis. / 1e`re pers. erg. / piquant-causatif-marque d’adjectivisation // Furthermore, since IMG lines are not sentences, the relevant orthographic rules of punctuation, initial capitalization and syllabification do not apply (R30⫺R32).
6.
Summary
Instead of a prose summary, a list of the rules and symbols proposed follows: 6.1. Rules 6.1.1. Glossing principles R1.
R2.
R3.
R4. R5. R6.
With the exceptions specified below, there is a symbol or a configuration of symbols in the IMG if and only if there is a morph in the L1 text that it corresponds to. The IMG represents morphemes, not allomorphs. Therefore, the gloss of a grammatically conditioned allomorph does not contain the grammatical category conditioning it. An IMG should be as precise and detailed as tolerable. The limits of precision and detail are defined by practical considerations of complexity and intelligibility. There is a biunique mapping of individual L1 morphemes onto glosses. (a) An L1 lexeme is glossed by L2 lexemes. (b) L1 stems are glossed by L2 stems. The gloss of a grammatical morph is a configuration of grammatical category labels each of which represents the
R7. R8.
value of a grammatical category. A grammatical morph should not be glossed by an L2 bound morpheme. It may be glossed by an L2 word if that has the same function as the L1 morph. Homonymy is resolved in the IMG, polysemy is preferably not. The gloss of a polysemous L1 item should represent, in the order of decreasing preference, ⫺ its Gesamtbedeutung, ⫺ its Grundbedeutung, ⫺ the set of its senses, ⫺ its contextual sense.
6.1.2. Boundary symbols R9. Apart from R30, there is a boundary symbol of a certain type in the IMG if there is a corresponding boundary symbol in the L1 text. More strictly, there is a blank, hyphen, plus, equal sign, angled bracket and tilde in an IMG if and only if there is an identical symbol in the L1 text corresponding to it. R10. A word boundary is shown binniquely by a blank ( ). R11. Two successive orthographic L1 words which must be glossed by one L2 word are linked by an underscore (_). R12. A morpheme boundary is generally shown by a hyphen (-). R13. A morpheme boundary not shown in the L1 text is indicated by a colon (:) in the IMG. This applies also to portmanteau morphs. R14. A boundary in a compound stem, and possibly also in a derived stem, may be shown by a plus sign (⫹). R15. A clitic boundary may be shown by an equal sign (⫽). R16. A gloss of a zero morpheme or allomorph is enclosed in round parentheses (( )). R17. The string enclosed in a discontinuous L1 item P1 ... P2 is enclosed in inverted angled brackets (P1> ...
1856
R20.
R21. R22.
R23.
R24.
XX. Morphologie in der Praxis
grammatical category labels) and separated from its source by a tilde (~). A grammatical meaning expressed by a non-segmentable morphological process (transfixation, internal modification, metathesis, subtraction, suprasegmental process) is not set off in the L1 representation. Its gloss follows the gloss of the base, separated by a backslash (\). Elements of an IMG that represent components of one L1 morph are separated by a period (.). As a special case of R21, components of one L1 cross-reference morph that have distinct reference are separated by the ampersand (‘&’) or, where no conflict with R17 and R18 arises, by the greater-than sign (‘>’) for actor and undergoer cross-reference. An L1 word form whose morphological structure is not represented in the IMG may be represented by a set of symbols whose status as representing morphs or features is ignored and whose sequence has no implications as to L1. Such symbols that jointly correspond to an L1 word form are joined by an underscore (_). If constituent structure is to be displayed, square brackets ([ ]) can be inserted in the IMG.
6.1.3. Typographic conventions R25. The IMG is in the line immediately below the corresponding L1 text line. R26. The distance between an L1 text line and the line immediately preceding it is greater than that between it and the IMG line belonging to it. R27. Each L1 word form is left-flush with the L2 word or complex of symbols rendering it. If such an arrangement is impossible, the following is a minimum requirement: If there is, in an IMG, an equivalent to an element of an L1 text line, it is contained in the line immediately below that line. R28. The IMG is printed in a smaller typeface than the L1 text. If this is impossible, then at least grammatical category labels are in small capitals. R29. Grammatical terms appearing in IMGs are abbreviated, without a period at the end, and set in (small) capitals. R30. There is no punctuation in an IMG. Parentheses including optional material in the L1 line are not repeated in the IMG, either (cf. R16). R31. There is no sentence-initial uppercase in an IMG. R32. There is no syllabication either in the L1 line or in the IMG.
6.2. Symbols L1
IMG
meaning
xy x_y z x⫺y x⫹y x⫽y z xy
xy z x_y x⫺y x⫹y x⫽y x/y x:y (x) ab<x> <xy>a x\y x.y x&y (x>y) [x] [x]Y
word boundary between x and y x and y are two orthographic words, but one lexical word x and y jointly render z without morphological analysis morpheme boundary between x and y x and y form a compound or a derivative stem x and y are joined by clisis x and y are alternative meanings of ambiguous z morpheme boundary between x and y not shown in the L1 text x does not have a significans in the L1 text x is an infix in ab xy is a circumfix around a y is a non-segmentable morphological process on lexeme x x and y are semantic or grammatical components of z x and y are grammatical components of z cross-referencing two different dependents x is a syntactic constituent x is a syntactic constituent of category Y
a<x>b x>a
1857
170. Grammaticography
7.
References
7.2. Sources of examples
7.1. Specialized literature
Bloomfield, Leonard (1933), Language. New York etc.: Holt, Rinehart & Winston
Bickel, Balthasar & Comrie, Bernard & Haspelmath, Martin (2004), The Leipzig Glossing Rules. Conventions for Interlinear Morpheme by Morpheme Glosses. Leipzig: Max-Planck-Institut für Evolutionäre Anthropologie
Evans, Nicholas (1997), “Role or Cast? Noun Incorporation and Complex Predicates in Mayali”. In: Alsina, Alex & Bresnan, Joan & Sells, Peter (eds.), Complex Predicates. Stanford: CSLI Publications, 379⫺430
Lehmann, Christian (1982), “Directions for Interlinear Morphemic Translations”. Folia Linguistica 16, 199⫺224
Finck, Franz Nikolaus (1909), Die Haupttypen des Sprachbaus. Leipzig: B. G. Teubner [Nachdr. d. 3., unveränd. Aufl.: Darmstadt: Wiss. Buchgesellschaft, 1965]
Lehmann, Christian (2004), “Data in Linguistics”. Linguistic Review 21.3/4, 175⫺210 Lehmann, Christian & Bakker, Dik & Dahl, Östen & Siewierska, Anna (21994), EUROTYP Guidelines. Strasbourg: Fondation Europe´enne de la Science (EUROTYP Working Papers)
Fortescue, Michael (1984), West Greenlandic. London etc.: Croom Helm (Croom Helm Descriptive Grammars) Gabelentz, Hans Conon von der (1861), “Über das Passivum. Eine sprachvergleichende Abhandlung”. Abhandlungen der philologisch-historischen Classe der Königlich-Sächsischen Gesellschaft der Wissenschaften 8, 449⫺546
Lieb, Hans-Heinrich & Drude, Sebastian (2000), Advanced Glossing: A Language Documentation Format. Berlin: Technische Universität (Working Paper)
Monod-Becquelin, Aurore (1976), “Classes verbales et construction ergative en trumai”. Ame´rindia 1, 117⫺143
Simons, Gary F. & Versaw, Larry (1988), How to use IT. A Guide to Interlinear Text Processing. Dallas, Tx.: Summer Institute of Linguistics (Revised edition, Version 1.1)
Woodbury, Hanni (1975), “Onondaga Noun Incorporation: Some Notes on the Interdependence of Syntax and Semantics”. International Journal of American Linguistics 41, 10⫺20
Christian Lehmann, Erfurt (Germany)
170. Grammaticography 1. 2. 3. 4. 5.
General problems Grammaticographic problems in morphology Structure of a grammar Descriptive devices References
1.
General problems
1.1. Basic concepts Like many scientific and, especially, linguistic terms, the term ‘grammar’ is ambiguous in designating both a sector of the object area of linguistics and also a scientific account of this sector. Grammar2 is, thus, at a meta-level with respect to grammar1, which is its object. The grammar1 of a language is the systematic part of the way the language maps meanings
onto expressions. The grammar2 of the linguist is a scientific representation of this part of a language ⫺ in the form of a book, a computer program or some other suitable medium ⫺, whose locus within a comprehensive presentation of a language will be characterized more precisely in 3. Grammaticography is an activity related to grammar2. This term, too, is ambiguous in the same way. At the object level, grammaticography is (the practice ⫺ experience or art ⫺ of) grammar writing. At the meta-level, it is the investigation of methodological principles that reconcile this practice with linguistic theorizing. Because of the analogous ambiguity, the term ‘lexicography’ has been coupled with the term ‘meta-lexicography’; and the same could be done for ‘grammaticogra-
1858 phy’. Metagrammaticography starts from two ends: On the one hand, it takes stock of successful grammars and distills from them the grammaticographic principles that they follow. On the other hand, it takes successful models of language description as a theoretical basis and deduces from them requirements for an adequate grammar2. Grammaticography is related to investigation of grammar just as lexicography is related to lexicology (investigation of the lexicon). Metalexicography has had a firm position in (applied) linguistics for several centuries. While grammaticography in the sense of ‘production of grammars’ goes back to antiquity, metagrammaticography is a fairly recent discipline. The earliest treatment known to us is Gabelentz (1891/1901, Zweites Buch, esp. VI. Capitel): “Die Darstellung der Einzelsprache”. The term ‘grammaticography’ does not appear much earlier than Cherubim (1973). This disproportion between lexicography and grammaticography is not easy to account for. In systematic terms, neglect of grammaticography is simply unjustified and detrimental both to the linguistic discipline and to the quality of actual grammars. In terms of the history of linguistics, there has apparently been a disequilibrium between lexicon and grammar: Research into the former started out as a practical concern, i.e. as lexicography; and only towards the end of the 19th century did lexicology in the modern sense begin. Research into grammar started out as theory of grammar with the modists of the 13th century. The beginnings of modern grammaticography with Gabelentz (1891) were taken up by O. Jespersen, but remained largely inconsequential otherwise. Theory of grammar1 got an even stronger position in the second half of the 20th century. The complete neglect of linguistic methodology during the period of the dominance of generative grammar included the complete neglect of grammar2; the concept or problem of grammaticography did not surface in linguistics. As a consequence, most grammars2 actually published, while incorporating conceptions of theories of grammar1 en vogue at their time, do not in their general organization reflect any theoretical foundation at all. In assessing the relationship between the practice of grammar writing and linguistic theorizing, we may recognize that the last third of the 20th century has brought a significant progress in grammar writing. Apart from many noteworthy monographs, the
XX. Morphologie in der Praxis
series Lingua Descriptive Studies alias Croom Helm Descriptive Grammars, Cambridge Studies in Linguistics, Mouton Grammar Library and London Oriental and African Language Library may be singled out, as they are devoted to comprehensive accounts of hitherto underdescribed languages. This progress in the quality of descriptive grammars can be attributed not so much to some particular model or theory of grammar, but, rather, to a general agreement to keep grammars comprehensible for those outside a particular framework and/or jargon. Roughly since the forties of the 20th century, a certain amount of grammatical descriptions of minority languages had appeared that applied some model of structural or ⫺ less often ⫺ transformational linguistics. It was soon recognized that these grammars were not even used by linguists, let alone by laymen. Since the seventies of the 20th century, most comprehensive grammars stick to what their authors call “traditional grammar”. It was recognized that “traditional” descriptive grammars remain up-to-date and readable much longer than any model-oriented grammars. Such grammars also do not strive for the precision sometimes attributed to formal descriptions. Neglect of precision in itself is not a virtue, but historically it must be seen as a reaction to a kind of hollow precision that served to conceal lack of insight. In theory, grammaticography and theory of grammar inform each other. In practice, mutual awareness leaves much to be desired. Still, good grammar writing does apply the achievements of linguistic theorizing. Some new concepts and approaches (or more elaborated old ones), presentational techniques, empirical domains to be looked at and included in a grammar, etc. have established themselves as part of recent grammaticographic tradition. To the extent that these innovations are picked up from various models of grammar, this approach is eclectic. Grammars have to be both consistent and comprehensive, and these goals are sometimes in conflict. The eclecticism often practiced in contemporary grammaticography may appear as a symptom of inconsistency, but is more properly seen as a consequence of the fact that none of the available theories of grammar suits all languages and all empirical domains equally well. Natural languages comprise heterogeneous phenomena which favor different models and methods. If a grammarian is able to choose the most pow-
170. Grammaticography
erful approach for each descriptive domain and to combine them in the compilation of a comprehensive grammar, this is a virtue rather than a vice. Moreover, this approach to writing grammars is actually the only way to determine the borders of applicability of theoretical concepts and models and their relations to one another, and thus to achieve theoretical progress. In this sense, metagrammaticography provides an interface between the practice of grammar writing and the theory of grammar. On the other hand, metagrammaticography can be viewed as a sort of self-reflection inside the field of descriptive linguistics, which can help to accumulate positive experiences and most suitable (technical) solutions to be applied in future grammatical descriptions. Like most disciplines concerned with the practical application of scientific insights, and like metalexicography in particular, metagrammaticography is at least as much a prescriptive discipline as it is a descriptive one. It is descriptive insofar as it is faced with actually published grammars and with the esteem these are held in, a datum which it has to take into account lest it become escapist. It is, however, prescriptive in that it converts both the principles distilled from grammaticographic practice and the theorems of grammatical theory into instructions of grammar writing (cf., e.g., Mosel 2003, Noonan to app.). Metagrammaticography is part of the methodology of linguistics. Consequently, the present treatment will account for observable practice, but will also derive recommendations from it. 1.2. Universalism and particularism The description of a language has two opposite tasks: to bring out the uniqueness of this language and to render it comparable with other languages. To achieve the former task, the description must not identify the categories and operations of the language with those of other languages, but rather describe the language in its own terms. To achieve the latter, the description has to characterize the language in terms of a linguistic type, presupposing and referring to universal parameters of human language as a background against which the peculiarities of the language stand out. To do this, it must describe the language in the same terms as other languages. The two tasks seem, thus, irreconcilable. Consequently, many existent grammatical descriptions sin by an overemphasis on either partic-
1859 ularism, which renders the language different from anything that may be familiar to the reader and, thus, unintelligible, or on universalism, which converts the language into just another instance of something well-known to the reader and, thus, uninteresting. Both extremes miss their target (cf. also 1.6.1). Many of the grammaticographic principles discussed below revolve around this problem. The purpose of a language description is not only to tell the reader the facts to any desired degree of detail, but, at a more general level, to convey to him an impression of what this language is like. This evaluation proceeds against the general background of ‘le langage’, shared by author and reader, and possibly against the comparative background of other languages, too. In the former respect, the grammar will take for granted many aspects of the language that it shares with all other languages (e.g. the sheer existence of word formation, and its most general cognitive and communicative functions). In the latter respect, many grammars, especially textbooks, even presuppose as deserving no comment those features of the language that it shares with the metalanguage of the description. Thus, few grammars of French will bother to mention that it lacks infixation or that clause structure follows the accusative model of fundamental relations. 1.3. Purpose and kinds of grammar Grammars may be classified according to a variety of parameters. The first distinction is between a comparative grammar and a grammar limited to one language. There are various kinds of comparative grammars according to the purpose and method of comparison: a general comparative grammar is a systematic survey of grammatical phenomena in the languages of the world. A historical-comparative grammar traces the evolution of the grammar of a proto-language into its daughter-languages by comparing these. A contrastive grammar compares grammatical phenomena of two languages, mostly with some practical application in mind. In contrast to comparative grammars, grammars devoted to just one language are sometimes called descriptive grammars; but of course such a grammar can also be prescriptive; see below. A grammar of one language may take the synchronic or the diachronic perspective. A grammar that does the latter is a historical grammar. In contradistinction to a language history, a historical grammar is normally not
1860 subdivided according to stages of the language, but may rather have the same structure as a synchronic grammar. A historical grammar usually defines one stage of a language ⫺ not seldom an ancient stage ⫺ as the end point of its scope and differs from a synchronic grammar in tracing the properties of that stage back to the relevant proto-language. Focusing now on non-comparative synchronic grammars, we may distinguish between what has been known traditionally as textbook vs. reference grammar. A textbook (German Sprachlehre) aims at teaching the language. It arranges grammatical material in the sense of some didactic progression; and grammatical information may be intertwined with other didactic material such as lessons, cultural information and exercises. A reference grammar, on the other hand, gives a systematic and comprehensive overview of the grammatical system or even the whole system of a language, but is limited to that. Its purpose is to provide orderly access to information on the language system rather than to help the user learn the language. It may presuppose some familiarity with the language or with linguistics. There also used to be a distinction between linguists’ vs. learners’ (“scientific” vs. “pedagogical”) grammars. This has lost somewhat in significance because it has been recognized that writing for colleagues is compatible with the requirement of user-friendliness (cf. also the notes on recent developments in 1.1). The orientation of a grammar may be descriptive or prescriptive. A descriptive grammar is based on data that are independent from the analyst and describes these. Its purpose is to ascertain and systematize the facts in some empirical domain. A prescriptive grammar represents what the author considers correct. Its purpose is to orient the reader towards the norm of the language. While the distinction between descriptive and prescriptive linguistics is clear in principle, there are some factors that tend to blur it in practice. Firstly, no grammar (descriptive or other) can account for all the variation that occurs in its object area. It will exclude at least speech errors, false starts, hesitation phenomena etc.; and it may exclude youth language or archaic ritual language from its scope. This delimitation amounts to the identification of a norm. This does not necessarily coincide with the highest norm in the entire speech community; it simply means
XX. Morphologie in der Praxis
the identification of a prototypical manifestation of the variety under study and the marginalization of other varieties. The exclusion of some facts from the variety being described is similar to the decisions made in a prescriptive grammar. Secondly, most users of a grammar, even of a descriptive grammar, require reliable information on what is grammatical in the language or, at least, what is normally said. They want to be given rules. To satisfy this need, the author has to distinguish manifestations of the norm from deviations from it and has to formulate rules for grammatical constructions. This, again, is what a prescriptive grammarian does. In this article we will concentrate on descriptive reference grammars. However, a good portion of the following metagrammaticography applies to other types of grammar as well. 1.4. Data and variation Data are the empirical basis of a grammar (see Lehmann 2004 a). There are different kinds of them, and they play different roles in grammars. This is determined, first and foremost, by the orientation of the grammar in the sense of 1.3. In a descriptive grammar, data function as the empirical basis and the object of research. Therefore the data must exist independently from the production of the grammar. A descriptive grammar is therefore based on a corpus of data. In a prescriptive grammar, on the other hand, data have the function of illustrating the norm. They may be taken from a corpus of texts that is conceived as representing the norm ⫺ for instance the works of classical writers; but if the author considers herself master of the norm, then she may simply make up illustrative examples, which are then no data in the strict sense. Some descriptive grammars, it is true, are also based on material produced by the author of the grammar. In this case, however, the data are not independent from the scientist, which means they are not reliable data by scientific standards. The corpus on which a descriptive grammar is based may essentially be of two kinds. It either exists prior to the grammar as a body of texts available in the speech community; or it is collected by the linguist in fieldwork (in the broadest sense). In morphology as opposed to syntax and discourse structure, many data may be systematically elicited in informant work, rather than drawn from texts (see Art. 168).
170. Grammaticography
The object of a grammatical description is a certain language, more precisely, some particular variety of a language. The first task of the linguist is to delimit this variety in synchronic and diachronic terms. However narrowly she may delimit it, her empirical data will necessarily be only a sample of the object of description. With a corpus collected in a small village of the Mexican state of Quintana Roo, one may publish a grammar of the speech of those male peasants of that village who were born between 1940 and 1950; or one may publish a grammar of Yucatec Maya. It commonly corresponds both to the purpose of the author and to the interest of the readership to situate the variety described at a rather high level of generality. It is, then, the responsibility of the author to guarantee that her data actually represent her object, i.e. that they are valid. This presupposes some homogeneity of the object. The description of Yucatec Maya based on that sample will be valid to the extent that an observationally similar description could have been produced with data from Valladolid (Yucatan). Mutatis mutandis, the same goes for a grammar of French or Latin. The actual sources of the examples ⫺ informants or texts ⫺ have to be identified (the former, if they agree to it). If their representativity could be an issue, it has to be demonstrated that they do represent the object of description. For instance, many a linguistic description is based on a corpus of narrative texts, and it is by no means to be taken for granted that such a corpus does truly represent the language in question. Again, young fieldworkers tend to cooperate with young informants, whose speech usually represents just one sociolect in the community. Therefore, the sociolinguistic coordinates of the data have to be indicated. It is essential that the data sources be made entirely explicit. The usage observed in traditional grammars of giving examples without indication of the source is no longer acceptable. To put it differently, an example without indication of its source will be regarded as made up by the author of the description and insofar unreliable (no matter whether she is a native speaker or not). 1.5. Topicality Scientists are all but resigned to seeing new theories and analyses pop up every day only to be superseded by even newer theories and analyses, so that most of scientific work be-
1861 comes irrelevant not only to the general public but even to the scientific community itself. Things are different for a grammar (and likewise for a dictionary). First of all, for most languages of the world the publication of a grammar is an event that is not likely to be repeated within a generation. For all those languages that are threatened by extinction, a comprehensive description is the only chance for them to be revived and the only way to protect them from falling into oblivion after their extinction. Secondly, a grammar is a scientific product of potential importance for the general public, since it may and must be the basis of the development of a norm and the composition of primers. All of this heightens the responsibility of the grammarian. The main virtue of the grammarian is (alas) not originality and inspiration but consistency and reliability. To the extent that grammarians have become aware of these conditions, they have refrained from tying their work to some fashionable model and from concentrating too much on trendy issues. The grammarian has the unenviable task of keeping herself informed on progress in general comparative linguistics but free of ephemeral trends. Fortunately this task has been facilitated in the past decades by comprehensive and up-todate surveys of general comparative grammar such as Shopen (1985, ed.) and Payne (1997). 1.6. Concepts and terms At least two issues arise with respect to the use of concepts and terms in grammatical descriptions: (a) What is the appropriate attitude towards conservatism and neology? (b) What degree of familiarity with concepts and terms should be presupposed? 1.6.1. Conservatism and neology Originality is a value in science that does not include terminology. Established terminology must be used. Most phenomena in the languages of the world, whether or not they happen to be known to the analyst or to the particular philological or linguistic school that she has been raised in, are actually known in general linguistics and have a term to them which, at least sometimes, is appropriate and well-established. For instance, the semantic relation often called purposive had been called ‘destinative’ by Finno-Ugricists half a century earlier. It is among the duties of the analyst to get informed about estab-
1862 lished terminology before she coins her own terms. Sometimes, it is true, established terminology is inappropriate. For instance, the term ‘possessive classifier’ was deemed inappropriate even by its coiner, but for want of a better term continues to be used. It is most important that distinct concepts be designated by distinct terms and that these be unequivocal; questions of beauty of terms have less priority. Secondly, the question regularly arises whether the particular phenomenon in the language being described is simply an instance of something known from other languages. Here, a middle course between two extremes must be steered (cf. 1.2). One is particularism. From the concept of the ‘langue’ which is both historically and systematically unique, and of the language sign which is arbitrary and unlike anything to be found in other languages, this position deduces that every category of the language must be named by a term of its own in order to avoid misleading identification with something known from other languages. This position has been cherished to the extreme in Russian and American structuralism. It renders a description unusable and blocks comparison of the language with other languages. The other extreme is universalism. It assumes a universal grammar and subsumes every phenomenon of the language under pre-established concepts and terms. For instance, the concept of ‘case’ is stretched to cover (non-morphological) syntactic properties of noun phrases; and then sometimes what used to be called case is renamed ‘surface case’. This position is well-represented in generative grammar, but not alien to some brands of functionalism either. The appropriate middle course is characterized by two propositions: (a) descriptive concepts belong to the level of linguistic types; (b) descriptive concepts are prototypical in nature. Take the concept of ‘dative’ as an example. For it to be a typological concept means that it is not a universal property of human languages, that some languages have a dative and others don’t, that ‘dative’ is a type of which there are concrete instantiations in the languages and that there are partial functional equivalents to the dative both inside a language that has it and across languages. Being a typological concept for a (value of) a grammatical category, i.e. for a (kind of) linguistic sign, ‘dative’ is characterized both by its meaning or function
XX. Morphologie in der Praxis
(roughly ‘case whose basic meaning (Grundbedeutung) is the recipient and whose generic meaning (Gesamtbedeutung) is a participant related indirectly to the situation’) and by its structure or form (which for the dative is the same as for ‘case’, i.e. a grammatical morpheme bound to a nominal expression that marks the latter’s syntactic or semantic function in the clause). The prototypicality of such concepts relates to both sides of the linguistic sign. For the meaning or function, it necessitates the Jakobsonian distinction between basic meaning and generic meaning, as just exemplified for ‘dative’. For the structure or form, it may involve the identification of a focal instance on a scale of grammaticalization, with some range of variation at both sides. For instance, a case is typically marked by an agglutinative affix, but may be marked by a highly general adposition, on the one hand (as in Japanese or Hawaiian), or by more fusional morphological means, on the other hand (as in Sanskrit). Of course, such definitions are not always available for grammatical concepts, and often it is the task of the analyst to decide whether an established term covers her particular phenomenon. Extensions of traditional usage are allowed and necessary, because in the last analysis, most of these terms were originally conceived for Greek and Latin and must necessarily be extended in their meaning if they are to be applied to any other language at all. If this were not so, we could never speak of conjugation or of passive formation in English. What is essential here is that such extensions be controlled, i.e. that a general definition (at the typological level) of the kind alluded to be provided, from which it may be deduced that the term is indeed applicable to the particular phenomenon under analysis. If terms are used in this way, then neology in a grammatical description may be kept to a minimum. The most strongly grammaticalized categories are, at the same time, the most deeply entrenched in a particular language system and consequently the most arbitrary and language-specific ones. Here the question of whether such a language-specific phenomenon should be subsumed under a category known in typology can become a real issue. An example is provided by the German verb category whose values are traditionally called ‘subjunctive I’ or ‘present subjunctive’ (e.g. singest) and ‘subjunctive II / past subjunctive’ (e.g. sängest). These terms are actually
170. Grammaticography
not helpful as far as the functions of these categories are concerned. If German happened to be an exotic language first described in the second half of the twentieth century, then these categories would probably be called ‘hearsay evidential’ and ‘irrealis’, respectively. Consider the two sets of crossreference markers in Mayan languages as another example. Members of the first set precede the finite verb, members of the second set (displayed in Tab. 170.9 below) follow it. Mayan linguistics, which originates in American structuralism, calls these two paradigms by purely arbitrary labels ‘set A’ and ‘set B’, in order to avoid any functional implications. The two sets do combine heterogeneous functions. In Yucatec Maya, set A cross-references the transitive subject and the subject of an intransitive verb in one aspect-mood category, while set B cross-references the subject of a nominal clause, the transitive object and the subject of an intransitive verb in another aspect-mood category. Relating this to certain ergativity-splits, one may say that set A cross-references the subject, while set B crossreferences the absolutive. However, this is odd, because a transitive verb would then be flanked by a subject and an absolutive index. Moreover, set A also precedes nouns, crossreferencing the possessor. This function can certainly not be subsumed under the denominator ‘subject’. In other words, concepts provided (so far) by typology do not help here; language-specific concepts and terms have to be coined. The only question that remains is whether they must be as empty as ‘set A/B’. 1.6.2. Definitions In principle, the explicit introduction of concepts and terms of general linguistics, especially of analytic concepts such as ‘infix’ or ‘completive aspect’, is not the task of a grammatical description, but of genres such as terminological dictionaries and studies in general-comparative grammar. If one could presuppose a systematic methodological organization of linguistic science, a descriptive grammar should only have to categorize its phenomena by the concepts of general-comparative grammar and designate them by the corresponding terms. However, such rigor is not practicable. On the one hand, the extent to which familiarity with established concepts and terms may be presupposed obviously depends on the intended readership of the description. If non-specialists are addressed, technical terminology either has to be intro-
1863 duced explicitly or must not be used at all. On the other hand, established usage may be insufficient with respect to the language being described. Then modifications to it must be made explicit. Still, a grammarian must sometimes be reminded that she is not writing an introduction to linguistics. Specific suggestions on morphological terminology are found in Art. 169, section 3.9.2. 1.7. Description and argumentation A grammar is not the kind of treatise that has a point. The data are not used to argue for a particular analysis, the analyses are not arranged in a way to convey a certain theoretical insight or to demonstrate a certain method. All of this is necessary and justifiable in other kinds of scientific treatises. A grammar is, in the first place, a systematic encyclopedia of the grammatical functions and structures of the language. Consider the French clitic personal pronouns as an example. The grammarian, of course, has a position on the issue of whether they are affixes or clitics. In an article devoted to the subject, it would be appropriate to assemble the evidence for her thesis and argue for it. In a grammar, these pronouns will just be called clitic pronouns (or personal affixes), their various properties will be described, both those like their fixed relative order which are more affix-like and those like their alternative prethematic or postthematic position which are more clitic-like. It is not necessary to turn the description into a persuasory discourse, as is done in some grammars like Cole (1982) and McGregor (1990). Neither is a grammar the place for a linguist to teach linguistic methodology or to demonstrate that she has applied it. What matters is that her analytical decisions be transparent. This is chiefly done by means of examples that exhibit the phenomenon some descriptive statement attributes to the language. It is normally also not necessary to back statements on obligatory rules with ungrammatical examples of constructions that violate them. What is important is that it be made explicit under which conditions the rule applies. True, there will always be cases where the grammarian has not been able to ascertain these conditions. In order not to mislead the reader, she should admit this, for instance by modifying a descriptive statement by the phrase ‘under unclear conditions’. Several grammaticographic treatments (e.g. Comrie & Smith 1977; Noonan to app.)
1864
XX. Morphologie in der Praxis
require that the analyst note those properties that the language does not possess. In a comparative grammar, such statements are natural by-products of the comparison. In a monolingual grammar, such statements involve an implicit comparison. Moreover, there is no limit to the things that a language does not have. A feasible form of complying with the requirement is to limit such negative statements to such features that might be expected on genetic, areal or typological grounds and to relate them to the description of some feature the language does possess. Thus, in the section on counting of an onomasiological grammar of German it would be mentioned that the nouns Stück ‘piece’ and Mann ‘man’ can replace the counted noun generically and anaphorically under certain conditions. This may naturally be rounded off by a statement to the effect that this is the closest to numeral classifiers that the language can muster.
2.
Grammaticographic problems in morphology
2.1. Tasks of a morphological description Basically, the core task of the morphological part of a grammar is to describe: A. (i) the internal structure of word forms (grammatical words) (morphological structure), and (ii) the form and the meaning of grammatical items (non-lexical morphemes). These aspects of a morphological description overlap to the extent that (a) the grammatical items of a language are bound morphemes, and (b) the internal structure of word forms in this language can be described in terms of these morphemes (typically as a construction of a lexical stem and a chain of bound grammatical items). Since the domain covered by these two conditions is fairly large in many languages, the usual practice of grammarians is not to separate tasks A(i) and A(ii) from one another for descriptive purposes. As a result, those sub-domains where one of these conditions does not hold commonly come out in the periphery of a morphological description (as represented by sections on, e.g., prepositions or compounding). Generally, preference is given to what may be termed word-oriented morphology (as in A(i)), that is, the word is taken as a starting
point for the description. Grammatical items thus fall into two strictly distinct groups which are described quite differently, viz. those which are constituent elements of word structure and those which are words themselves. Within this approach, the very existence of a bound grammatical item constitutes a descriptive statement about word structure. An alternative approach, which may be termed morpheme-oriented morphology, would be to start from the morpheme, so that the class of words as well as word classes emerge as a result of the distributional analysis of morphemes. No grammar takes this approach exclusively. However, many grammars contain sections which consist of an annotated list of grammatical formatives. For instance, McGregor (1990), in the relevant sections of his chapter 3, has various lists of grammatical items for each of which he provides the significans, a functional label and a list of functions fulfilled in various constructions. Each item of the latter list contains a reference to other sections where the item is treated in its relevant context. The word-oriented approach implies a distinction between the morphological description proper (as outlined in A) and two “classificatory” tasks: B. (i) delimiting the class of linguistic units to be referred to as words (both in the sense of ‘word form’ and of ‘lexeme’; cf. Art. 26), and (ii) a classification of lexemes into parts of speech (cf. Art. 70) and further grammatically relevant sub-classes, at least inasmuch as morphological patterns vary by word class. Quite commonly, classificatory solutions are just incorporated into the general structural outline of a grammar, rather than constituting a subject of description in its own right; in other words, the classes of linguistic units dealt with in morphology are presented as established a priori. There are obvious theoretical grounds for this well-established tradition: First, these tasks are scarcely solved by purely morphological criteria; syntax and semantics play quite a role in an appropriate classification of linguistic units. In this sense, this classification can be regarded as “external” with respect to morphology proper (tasks A). Secondly, this way of presentation overcomes a certain circularity which is apparently inherent in grammatical reasoning,
170. Grammaticography
whereby relevant classes are distinguished on the basis of distribution and associated morphological categories, the latter being, in their turn, defined with respect to and in terms of pre-established grammatical classes. However this problem may be solved in the process of grammatical analysis, the usual descriptive practice is to take both the level of the word and the word classes for granted. On the other hand, classificatory solutions adopted in a given descriptive grammar do constitute autonomous descriptive claims and are even argued for in case they deviate from a “traditional” one, be it a languagespecific or the general-linguistic tradition. In addition, minor (sub-)classifications which are relatively or exclusively language-specific (as, e.g., inflectional classes) are more likely to be presented as a subject of morphological description than the major ones. That is, many grammarians take the classificatory tasks to have (at least to some extent) a default solution which can be applied without special discussion. This default solution is in fact very close to the most traditional parts of speech. By accepting the default solution, a grammarian virtually refrains from a language-specific classification of linguistic units; that is, it is not considered a task of a particular descriptive grammar, but rather a universal ready-made structural template for describing languages. However, the impact of classificatory solutions on the quality of grammatical description can hardly be overestimated: (a) They play a major role in structuring the morphological description and thus determine the general outline of the grammar (cf. 2.2). (b) They strongly affect the descriptive solutions (tasks A), insofar as morphological patterns and categories can be defined only with respect to the word classes; once a classification is established, lots of further descriptive solutions are predetermined. (c) They serve as a basis for the morphologysyntax interface, inasmuch as syntactic constructions often make reference to a word class and/or a further morphological sub-class of the elements involved. (d) Last not least, these solutions establish the grammar-lexicon interface, in the sense that they determine a system of grammatical indices to be provided for items listed in a lexicon (cf. 3.1).
1865 This means that if a language does not fit the default classificatory solution, the consequences of adopting this solution can turn out harmful for the whole enterprise. On the other hand, the default solution is, in a sense, the most user-friendly one, in that it facilitates obtaining information from the grammar, since the reader is faced with a familiar structural outline: the key words (commonly constituting chapter headings, cf. 2.2) tend to be exactly those he expects to find. Thus, a grammarian normally has to weigh the default solution against its descriptive cost in order to find an appropriate compromise. This problem, although often not pronounced as such, seems to be fundamental for descriptive morphology. A classification into word classes as represented in the morphological part of a grammar is often far more detailed than is needed for morphology proper. The usual descriptive practice is to introduce all more or less grammatically relevant classifications of words by structuring a chapter where the morphological patterns associated with some of these classes are described. In this sense, a description of the internal structure of word forms and a grammar-oriented classification of lexemes are not distinguished in the practice of grammarians, although these ⫺ obviously overlapping ⫺ descriptive domains presumably never coincide exactly. This practice has the advantage of highlighting the correlation between classifications based on distinct parameters (morphological, syntactic, semantic), the descriptive significance of which is beyond doubt. To sum up: the practice of descriptive morphology combines three distinct domains of language structure: the word structure, the grammatical items, and the grammar-oriented classification of lexemes, the latter constituting the skeleton of a morphological description. As will be seen below, this combination, however well-motivated, involves some descriptive problems which could be avoided otherwise. 2.2. Structuring a morphological description The well-established tradition of structuring morphological descriptions is based on a simple and elegant idea of converting grammaroriented classifications of linguistic units into the section headings of a grammar. This idea makes it possible to combine two descriptive domains (word classes and word structure) in
1866
XX. Morphologie in der Praxis
1. Open word classes with rich morphology: 1.1. (Optional:) a general overview of morphological structure (e.g., in terms of a general structural template comprising slots for constituent morphemes). 1.2. Further (functional) sub-classification, if relevant for morphology (e.g., transitive vs. intransitive verbs). 1.3. Inflectional paradigm: 1.3.1. The structure of the paradigm(s), i.e., the set of grammatical categories associated with the word class under description and their possible values. 1.3.2. Further (formal) sub-classification into inflection classes (if any) and how the items of the paradigm are constructed for each inflection class. 1.3.3. Exceptions, defective paradigms etc. 1.3.4. (Optional:) semantics and usage of inflectional categories can be picked up as a special issue. 1.4. Word formation (sections organized first by the category of the output of a process, second by the category of the base). 1.4.1. Compounding 1.4.2. Derivation: 1.4.2.1. Regular (productive) derivation; the internal structure is close to a semantically ordered morphemic lexicon (a morphological item ⫹ its meaning and distribution). 1.4.2.2. Irregular (non-productive) derivation. 1.4.3. Reduplication, conversion, etc. 1.5. (Optional:) some important semantic and/or derivational sub-classes (e.g., reflexive verbs). 2. Open word classes with little or no morphology: 2.1. Semantic sub-classes (e.g., spatial adverbs, temporal adverbs, etc.). 2.2. Correlated structural features (if any) and word formation (often including etymology, grammaticalization notes, etc.). 3. Closed classes: 3.1. A general semantic and syntactic overview. 3.2. (Optional:) exhaustive listing of items with their meanings and distribution, ordered according to functional (e.g., interrogative pronouns, indefinite pronouns etc.) or syntactic properties (e.g. postpositions with dative, postpositions with genitive etc.). 3.2.1. If the items are inflected, the paradigm is given for each. 3.2.2. (Optional:) distinctive structural features of each group (if any) and word formation techniques (e.g., conversion, grammaticalization, etc.). Tab. 170.1: Types of morphological chapters with respect to their internal structure
a rather natural fashion, inasmuch as it can be assumed that different word (sub-)classes are associated with distinct morphological patterns; hence, the morphological information is appropriately distributed over these (sub)-classes. The internal structure of chapters varies according to the word class. Roughly, they fall into three groups: open classes with rich morphology, open classes with no or little morphology, and closed classes. A schematic representation of how the respective chapters are most commonly structured is given in Tab. 170.1. As shown by this scheme, the morphological parts of grammars contain rather heterogeneous chapters: some are morphological in the word-oriented sense, since they deal with the internal structure of words (cf. the dis-
tinction introduced in 2.1); others are morphological in the morpheme-oriented sense, since they describe the form and the functions of closed-class (⫽ grammatical) items; and, finally, there are chapters which are concerned almost exclusively with (non-morphological) sub-classifications of a word class. This may be seen as a direct consequence of the fact that the three descriptive domains combined in morphology (cf. 2.1) remain essentially distinct. Not only do the respective chapters differ in their internal structure, they are bound to take distinct descriptive approaches. In some cases, a semantic classification seems to appear in a grammar just because there is nothing morphological to say about a class. The well-established tradition of providing a semantic classification of
170. Grammaticography
adverbs independently of its grammatical significance is a case in point. Also, it seems that if a language has no nominal morphology, the chapter on nouns is much more likely to contain their semantic sub-classification. In fact, a chapter subdivision is the form of introducing a comprehensive classification, but it seems bizarre to leave a section thus generated empty. A more consequential drawback of the structuring scheme outlined above is that it provides no natural way of accounting for morphological items shared by distinct classes (e.g., for person/number paradigms attached to both verbs and nouns, as, for example, in Yucatec Maya). Another instance of the same problem is represented by grammatical items which function both as a free morpheme (hence, have to be described in a separate chapter of the closed-class type) and as a bound morpheme (hence, a constituent element of some morphological pattern related to a certain word class); cf., e.g., locative prepositions vs. applicative suffixes on verbs in Rwanda (Art. 141, section 3.5.1). Various ways to overcome this drawback of the traditional approach have been applied in different grammatical descriptions. The options are: (a) Introduce generalized word classes comprising several word classes sharing some morphological categories (e.g., “nominals”). This solution has a limited domain of applicability because of its hierarchical nature. If the distribution of morphological patterns in a given language does not follow any hierarchical classification (e.g., a sub-class of nouns has an adjectival paradigm, as in Russian), this strategy does not work. (b) Select a “locus” for the description of an item ⫺ a morphological category, a single morpheme, or whatever ⫺ (for example, the noun for the category of case), and refer to this description in all other relevant chapters (in this case, for instance in the section on pronouns). This solution may work if the semantic and formal properties of the item in question do not vary by word class. If comparable items are even slightly different, a reference will not suffice. (c) Link the descriptions of similar items by cross-references. This is a solution widely
1867 applied, obviously necessary, but not sufficient, since it does not provide a way to describe the item with several instantiations as a whole. (d) Separate word formation (for all word classes) from the main body of description and treat it in a special chapter. This compromise solution can overcome the drawbacks of the word-oriented approach at least for derivational morphemes. (e) Separate morphological semantics from the main body of description and treat it in a special chapter. This is a very strong, but rarely applied solution, presumably because of its “ambitious” flavor. (f) Create a morphemicon as part of the morphological description as described below (3.5). The last three structuring solutions tend to be applied in one or another form in recent descriptive grammars. They are, in fact, a concession to the morpheme-oriented approach to grammar. In any case, an appropriate structure for a grammatical description can be achieved only by means of an appropriate combination of various solutions. 2.3. Identification of grammatical items The task of describing the morphological structure of a word form (cf. A(i) in 2.2) is almost never addressed directly in descriptive grammars. Instead, it is decomposed into several distinct sub-tasks. The most widely applied decomposition is based on the distinction between inflection and word formation (in particular, derivation; see Art. 38). Its controversial theoretical status notwithstanding, this distinction is widely applied in morphological descriptions. Not only are these types of phenomena often rigidly distinguished (cf. 2.2); these sub-domains of morphology are commonly described in quite disparate fashion: (a) To describe the inflection of a language means to describe each word class in terms of the associated paradigms, that is, the inflectional patterns are considered properties of pre-established word classes. To describe the derivation is to describe the semantic impact and distribution of each single derivational morpheme; hence, a word sub-class to which some derivational morpheme applies is
1868 viewed as a property of that morpheme (this includes the issues of productivity, regularity, etc.). (b) Inflectional forms of different words can be identified as representing the same item of a paradigm independently of their formal similarity; the identity of a paradigm is grasped in terms of certain semantic-functional labels (even if these only imply an approximation of the respective meanings), not in terms of formal identity. In contrast with this, the derivational sub-task involves a semasiological description of particular morphemes, which usually presupposes formal similarity of its various allomorphs. (c) The semantics of inflectional items is assumed by default to be independent of the semantic properties of the word; they are allowed to have a few, usually context-dependent functions. Conversely, a derivational morpheme can have a variety of meanings depending on individual stems it is applied to. These properties can be traced back to the concepts of “prototypical” inflection and derivation and the role they are assumed to play in the grammar and, in particular, in the grammar-lexicon interface (Art. 36). The distinction between inflection and derivation thus determines the choice of a descriptive perspective. That is, there is a considerable difference in how inflectional and derivational items are treated in descriptive grammars, from the identity of an item up to the types of information to be provided. In particular, the concept of inflection implies a more function-oriented identification procedure, that is, both formal identity of items even in case of clearly related meanings, and an obvious formal contrast are very easily disregarded, as far as the inflection is concerned (hence, multiple non-distinctions within paradigms; see Plank (1992) for an overview of this problem, and Zaliznyak (1967: 19⫺34, 129⫺148) for case paradigms). Here is an example: Russian is an aspect-dominated language in the sense that the grammatical category of tense works differently for perfective and imperfective verbs. Tense allows for two different ways of description, schematically represented in Tab. 170.2 by the tense forms of the verb pair delat’ ‘make/ do (impfv.)’ and sdelat’ ‘make/do (prfv.)’:
XX. Morphologie in der Praxis
(a) form-based description ⫹Past ⫺Past Imperf. Future
dela-l, sdela-l dela-et, sdela-et budet delat’
(b) function-based description
Past Present Future
imperfective verbs
perfective verbs
dela-l dela-et budet delat’
sdela-l ⫺ sdela-et
Tab. 170.2: Russian tense: alternate paradigms
Basically, the two accounts differ in how they treat forms like sdela-et ‘pfv:make-fut/ npast.3.sg (will make, will have made)’. Variant (a) highlights the formal identity between this item and the Imperfective Non-Past delaet. They are identified as instantiating the same item of the tense paradigm and get the same functional label (Non-Past) (see Comrie 1978: 66 f.). Yet these forms are clearly distinct semantically: the form of perfective verbs can refer only to the future, while with imperfective verbs it is employed basically for reference to the present (although the future meaning is not excluded). Thus, the descriptive cost of this solution is a dependency of the range of functions of an inflectional item on the verb class. By contrast, variant (b) disregards the formal identity between the items under discussion. Instead, it highlights their functional distinction and, accordingly, the functional similarity between the Perfective Non-Past (Future) of the perfective verbs and the Imperfective (analytical) Future. The latter pair of forms are assigned the same position in the paradigm and get the same functional label (Future), despite the obvious formal contrast. The dependency on the aspectual meaning of the verb stem is thereby built into the structure of the paradigm, instead of constituting an independent descriptive statement. The descriptive cost of this solution is (i) the formal heterogeneity of the Future and (ii) the unmotivated formal identity of the Perfective Future and the Imperfective Present (which would show up recurrently in the description of multiple formal variants of this marking, only one of which is represented in Tab. 170.2). Variant (b) is apparently preferred in existent descriptions of Russian (even though
1869
170. Grammaticography
some of them do suggest variant (a), most often the formal identity between the Imperfective Present and the Perfective Future is not mentioned). In other words, the paradigm is defined in such a way as to simplify the semantic description at the cost of somewhat artificial formal complexity. Formal identity of this type is never neglected as easily for derivational formatives; to continue with Russian examples, its verbal postfix -s’a/-s’ expresses an extremely wide variety of meanings in combination with different verb stems (middle, reflexive, reciprocal, anticausative, dispersive, to mention only some of them). Yet these are commonly treated as instances of the same linguistic unit. In view of this difference in descriptive perspective, the well-known problem of delimiting inflection and derivation gets a new kind of relevance: how a grammatical item is identified and described depends on its affiliation with inflection vs. derivation, at least above a certain degree of discrepancy between form and meaning. It might be the case that these two sub-domains of morphology are in fact delimited in descriptive grammars depending on which approach seems more “suitable” for a given item. On the other hand, the choice of one or another perspective can affect the very properties of the items identified (as they would be described in the grammar), inasmuch as the prototypical features of inflection vs. derivation are embodied in the identification procedure.
3.
Structure of a grammar
3.1. Comprehensive presentation of a language No grammar can be complete. Given limitations of every kind, the author has to set priorities. These follow from the main purpose of the grammar and from external conditions. Needless to say, on the basis of available publications, one may decide to produce a partial description. However, the aim may be to produce a comprehensive presentation of the language. Assume a field structured in a hierarchical fashion. For an account of it to be comprehensive means that it is balanced in terms of the amount of detail provided for each of the sections at a given level of the hierarchy. In this sense, a grammar such as McQuown (1990) (of Totonac), which comprises 64 pages on phonology, 107 pages on
morphology and 11 pages on syntax is unbalanced and therefore not comprehensive. A comprehensive account of a language is articulated on the three levels shown in Tab. 170.3. level 1 level 2 level 3
the documentation of the language, which is a corpus representing (analyzed) primary data; the description of the language, whose object are the data of level 1; a methodological reflection on the description, whose object is the account of level 2.
Tab. 170.3: Levels of presentation of a language
The levels are, thus, in a meta-relation to each other. The reflection of level 3 has the function of accounting for the purpose of the description, the conditions under which it was carried out, including the achievements of previous scholarship, its theoretical and methodological prerequisites, the many choices and decisions that the author has made; and as mentioned in 1.4, it reflects on the character and limitations of the data base. Most of this part commonly takes the form of an introduction to the description. Some of it may be relegated to an appendix. The corpus on which the grammar is based (level 1) will normally not be reproduced in full. Good modern presentations of a language contain an appendix that presents some specimina of representative texts (see Mosel 2002, § 6.2 for the scientific importance of such a collection). These are provided in the form of video or audio recordings and rendered in the canonical trilinear representation as explained in Art. 169 (more on the relationship between the corpus and examples in the running description in 4.5). The descriptive part of a comprehensive account of a language (level 2) consists of two subparts, the system of the language and the setting of the language. Here we may be brief on the latter. The setting of the language comprises an explanation of relevant glossonyms, the genetic affiliation of the language and its dialects, its ethnographic situation (i.e. the situation of its speech-community), its cultural situation (including, importantly, its written tradition) and its sociolinguistic situation, i.e. its internal stratification, its status in the speech community and its areal relations. This part of the description provides a referential background for paramet-
1870 ricization by diachronic, diatopic, diastratic and diaphasic variables that will prove necessary in the account of the language system. It is, at the same time, the one part of a comprehensive presentation of a language that is most likely to take a historical perspective even if the rest is purely synchronic. The language system has two main parts, the expressive subsystems and the significative subsystems (corresponding to Martinet’s second and first articulation, respectively). The primary expressive subsystem is the phonology (with the phonetics), the secondary one is normally the writing system. (This is, of course, also true for [ancient] corpus languages.) The significative subsystems are not alternative, but jointly exhaustive: the grammar and the lexicon. They are significative because their units embody a mapping of meanings onto expressions. The significative subsystems are articulated in terms of levels of complexity: At the level of the word form and below, we have morphology, which ⫺ in the form of inflection ⫺ is an essential part of the grammar and also ⫺ in the form of word formation ⫺ of the lexicon (see Art. 36). At higher levels, we have the syntax as part of the grammar, and phraseology as its counterpart inside the lexicon. Since there is, in principle, no borderline between grammar1 and lexicon, there is also overlap between a grammar2 and a dictionary of a language. The morphology describes the word-formation patterns of the language. Their products are nevertheless listed in the dictionary. Only in the limiting case of a word-formation process that applies completely regularly and productively ⫺ e.g. formation of certain verbal nouns ⫺ may one renounce to representing each of the products in a lexical entry of its own. Contrariwise, the dictionary will contain a couple of entries that result from a word-formation process that is no longer productive in the language. Mentioning it in the morphology would entail repeating the same list of items that is in the dictionary. In such cases, redundancy is preferred to parsimony. The drawbacks of redundancy are that it induces inconsistency and that it may be uneconomical. They are outweighed by the advantages, which include user-friendliness and theoretical soundness (the mental grammar and lexicon are redundant in the same way). Moreover, we are not talking about a literal repetition of material in two parts of the description, since the ordering principles and, conse-
XX. Morphologie in der Praxis
quently, the ways of accessing the information are entirely different between dictionary and grammar2. Another important relationship between the grammar and the dictionary is that the former introduces and defines the terms that appear in the cells on grammatical information of the microstructure of a dictionary entry. This concerns such categories as noun class, gender, countability, aktionsart, inflection class etc. Technically speaking, this kind of specification in a dictionary entry determines whether the item can be used in a certain construction described in the grammar. 3.2. Onomasiology and semasiology As we said in section 1.2, the fundamental problem of grammaticography is to provide a common format for descriptive grammars while at the same time taking care not to obliterate the individuality of the language being described by forcing it into a Procrustean bed. The general task of a language is to provide a mapping between meanings and expressions. The meanings have an extra-linguistic substrate (i.e. one independent either of language altogether or at least of particular languages) in cognition and social interaction (communication). The expressions have an extra-linguistic substrate in phonetics and semiotics. It is the mapping itself, achieved in the grammar and in the lexicon, that is proper to each language. There can therefore be a universal system of cognitive and communicative domains coded by languages and a universal system of expression techniques and sounds used by languages. There cannot be a universal system of grammar. A system of lexical and grammatical description that provides comparability of the object language with other languages can therefore be either based on a system of cognitive and communicative domains and then describe how the language in question manifests each facet of these in its expressions; or else it can be based on a system of expressive (structural) devices and then describe what each of them is used for in the object language. The former approach is traditionally called onomasiological (or synthetic, more recently ‘functional’), the latter semasiological (or analytic, more recently ‘structural’) (cf. first Gabelentz 1891/1901: 84⫺104 and, more recently, Lehmann 1980, 22002: § 1.2.1 and Mosel 2003: §7). Each of these approaches is in itself coherent and capable of providing a complete description. Many grammars stick
170. Grammaticography
to one of them. For instance, Jespersen (1937) is a purely semasiological grammar, while Givo´n (1993) is a purely onomasiological grammar. However, each of these approaches is one-sided, as the onomasiological approach corresponds to the viewpoint of the speaker, while the semasiological approach corresponds to the viewpoint of the hearer. Thus, an onomasiological grammar answers questions of the type “how can I express such and such a thought, or fulfill such and such a communicative function, in this language?”, while a semasiological grammar answers questions of the type “what does such and such an expression of this language mean?”. Since grammars, just as dictionaries, are generally meant to serve both the speaker and the hearer, the ideal grammar2 consists of two parts, an onomasiological and a semasiological one. It should be clear that such an arrangement is also maximally user-friendly, because whatever question a user may pose to a grammar is posed either in the speaker or in the hearer perspective. Moreover, most of the descriptive problems reviewed in 2 resolve themselves if the twofold approach is chosen. A couple of published grammars approach this ideal to some extent. Despite appearances and declarations of the author, Gabelentz (1881), a grammar of Classical Chinese, is relatively far off his own mark because the “synthetic grammar” is not really onomasiological, but just a construction-based syntax. McGregor (1990), a grammar of Gooniyandi, organizes its chapters in the familiar bottomup fashion appropriate for a semasiological account. However, some syntactic chapters and part of a final chapter on semantics take an onomasiological perspective on the items introduced in lower-level sections. The chapter on morphology in Haspelmath (1993), a grammar of Lezgian, contains sections on noun morphology and verb morphology, each of which is subdivided into a formbased and a function-based subsection; but otherwise the grammar mixes the two approaches. Lehmann (22002) is not a grammar, but just a description of possession in Yucatec Maya. It is subdivided into a chapter that introduces the relevant structures in a semasiological perspective, and three chapters that take the opposite perspective. It must be said that to this day, most grammars mix the two approaches in uncontrolled ways (cf. Lehmann 2004 b).
1871 It is nowadays standard to tie the distinction between morphology and syntax to the distinction between levels of grammatical complexity. However, it has often been associated with the distinction between an onomasiological and a semasiological grammar. From traditional grammars of the nineteenth century to our day, many a grammarian has said that her morphology deals with the system and structure of linguistic forms, while her syntax deals with their use. It might thus appear that the morphology of such a grammar is semasiological, while the syntax is onomasiological. However, such systematicity is only apparent. Most traditional morphologies contain chapters on functional categories such as person and tense ⫺ and to this extent they are onomasiological. And the typical traditional syntax has a chapter on the use of the cases, thus starting off from a structural concept introduced in the morphology and tracing its functions ⫺ and to this extent it is semasiological. Actually, the canonical level-dependent distinction between morphology and syntax can only be made in a semasiological grammar. A semasiological grammar is organized according to the hierarchy of structural complexity of linguistic units. For all those languages which possess the word form as a level of grammatical structure, the section of semasiological grammar devoted to this and lower levels will be the morphology. In the last quarter of the twentieth century, there have also been functional (i.e. onomasiological) grammars that presupposed a distinction between morphology and syntax; e.g. Comrie & Smith (1977) and the series of grammars based on it. Since this distinction cannot be made in an onomasiological grammar, it introduces inconsistency into it. For instance, the grammar of many languages manifests the concept of definiteness, but some do so at the level of morphology, while others do so at the level of syntax. In the twentieth century, many structural (thus, semasiological) grammars were published. Contrariwise, there are as yet few purely functional grammars. A theory of cognitive and communicative domains of language started to be developed only in the last quarter of the twentieth century, and for most of these domains we still lack both a solid foundation and an internal structure adequate for linguistic description. The present handbook tries to partly compensate for this shortcoming by organizing those of its
1872 chapters that are devoted to grammatical categories and operations (XII and XIV) according to an onomasiological perspective. It is inevitable that the articles of those chapters break the boundaries between morphology and syntax. The handbook on syntax in this series might have done the same with equal or greater justification. 3.3. Structure of a semasiological grammar Much of the language system is structured in terms of hierarchies of levels of complexity. This concerns, first of all, the system of grammar1 as mentioned in 3.1; but it may also include, with some imprecision, the distinction between expressive and significative systems. This yields the well-known series ‘phonology ⫺ morphology ⫺ syntax ⫺ discourse’, and inside grammar the sequence ‘stem ⫺ word form ⫺ phrase/syntagm ⫺ clause ⫺ (complex) sentence’. The history of linguistics has, cum grano salis, followed this progression. Syntax was a step-child of linguistic description up to the end of the nineteenth century, and even the 20th century has seen many “grammars” that actually boil down to a phonology plus a morphology. Discourse has come to be studied systematically only in the last third of the 20th century. The progression is deeply entrenched in the awareness of linguists, who think it must shape the organization of their grammar. Almost all grammars work essentially bottom-up (cf. Mosel 2003: § 5.2). This seems to correspond to a didactic progression which starts from elementary units and proceeds stepwise to complex units. A top-down progression, as it is prescribed in Comrie & Smith (1977) and followed in Lingua Descriptive Studies, seems unnatural because in the treatment of a given unit (e.g. the complex sentence) one is forced to appeal to constituent concepts (the clause, in this case) which have not yet been introduced. However, the picture must be modified slightly. The bottom-up approach corresponds to analytical thinking, which combines elementary units according to rules and aims at constructing a complex whole in a compositional fashion. The top-down approach corresponds to holistic thinking, which starts from a whole and understands its parts in terms of their function in the whole. As is well-known, the two approaches do not exclude, but complement each other. To give an example: A cleft-sentence is best understood if one knows what contrastive focus is and what it entails for semanto-syntac-
XX. Morphologie in der Praxis
tic structure. One can then identify the structural constituents in terms of their role in the complex construction. Contrariwise, it will be hard to construct the purport of a cleft-sentence in a bottom-up fashion, starting from a copular predication over an empty subject that combines with something that looks like a relative clause. More precisely, although the chapter on the cleft-sentence does presuppose the notions of copula clause, relative clause and complement clause, it is not the case that the grammar can build on one of these and expand it into a cleft-sentence. The external grammatical relations of a given unit ⫺ the subordinate clause in this example ⫺ are not treated in the chapter dealing with this unit, but instead in the chapter of a higher unit whose internal relations they are and of which the given unit is a constituent. This responds to the principle ⫺ last put forth in construction grammar ⫺ that the formation of complex constructions is goal-directed. The same goes for the morphology. The semasiological description of word-formation does not start from a certain derivational suffix, combine it with bases of different categories and then look what the category of the result is. Instead, there is a section on stems of a certain category, e.g. the adjective. The category has certain elementary members, i.e. adjectival roots. Next there are possibilities of forming adjective stems by various formal processes, e.g. by suffixal derivation. One of them is our derivational suffix. Finally, it is seen that it may combine with bases of various categories to yield the result at stake. The general principles of such an arrangement of a semasiological description may be formulated as follows. (a) The description works bottom-up through the hierarchy of grammatical levels. (b) For the grammatical unit of each of these levels, the categories into which it is articulated are identified. For each of the potentially complex categories, its internal syntagmatic structure is analyzed: First, a set of constructions according to kinds of syntagmatic relation is enumerated. For each of these constructions, the nature and distribution of its elements is set forth. Finally, given a certain construction of elements of two categories, one of the categories may comprise a grammatical (in particular morphological) para-
170. Grammaticography
digm of elements. Such paradigms are discussed as part of the description of the particular construction. (c) Only the internal syntagmatic structure of a given unit is part of the treatment of that unit. Any structural phenomena which concern the relation of a given element to its context are treated at the point where the including construction ⫺ the one which provides the context ⫺ is treated. To give a final example: At the level of the word form, categories such as ‘finite verb form’ are identified. There are kinds of finite verb forms according to their internal structure, e.g. periphrastic forms and synthetic forms of different kinds. One of these is the synthetic form that consists of a tensed stem and a personal ending. In the latter position, there is a paradigm of morphemes whose internal structure is treated now. It may be seen that this approach combines a bottom-up progression to ever more complex constructions with a top-down analysis of each of these constructions. 3.4. Structure of an onomasiological grammar The cognitive and communicative domains that are coded in language comprise concepts and operations a subset of which manifest themselves in grammatical structure. These are such concepts as the addressee or directed motion and such operations as abstraction or making a question. Such concepts and operations are assembled in functional domains. They are by now well-researched in functional typology, so that they can be tentatively enumerated. Since they provide the highest-level subdivision of an onomasiological grammar, this will be done here in Tab. 170.4. Onomasiological description proceeds from very general cognitive and communicative functions as indicated in the central column of Tab. 170.4 through more specific subareas such as those of the right-hand column down to the functional categories and processes of the language under description. These are finally mapped onto the structural devices and grammatical formatives introduced in the semasiology. For instance, in a grammar of German, the chapter on possession treats, among other things, possessive pronouns, the genitive attribute and its equivalents, the possessive dative, possessive predications with
1873 haben and sein etc. Most of these structural devices recur also in other chapters of the onomasiological description; but this chapter is where the question is answered how German expresses an attribution of possession as in Turkish vakt-im var ‘time-poss.1.sg exist (I have time)’. 3.5. Additional parts of a grammar The subdivision of a grammar follows mainly from the overall organization of the presentation of a language as outlined above. Since the term ‘grammar’ is not seldom used in a broad sense almost equivalent to ‘language description’, it bears repeating that, conceptually, the data corpus, on the one hand, and the methodological reflection, on the other hand, are not part of the grammar; and neither are the account of the setting of the language, the lexicon, the phonetics, phonology and orthography. What remains is simply the morphology and the syntax. Of the morphology, inflection is properly included in grammar, while word formation could, in principle, be treated either in the dictionary or in the grammar. For practical reasons, it is always included in the grammar (if it is treated at all). An important challenge taken up in many modern grammars written on a typological background is to present an epitome of what the language is like and how it works, i.e. to present the language as instantiating a linguistic type. From what has been said in 1.7 it follows that the overall presentation style of a reference grammar does not by itself fulfill this task. It is therefore widespread practice to provide an additional chapter ⫺ usually at the beginning of the grammar ⫺ that characterizes the language at the typological level. While the main body of a grammar is descriptive, this chapter provides the opportunity to compare the language with other languages and to answer the Humboldtian (1836: 417) question “auf welche Art [sie] die hauptsächlichen Fragen löst, welche aller Spracherzeugung als Aufgaben vorliegen”. Furthermore, a description of a language system contains a number of lists. The most important of these is, of course, the list of significative units (signs) of the language. It comprises the lexemes (including phraseologisms) in their citation form plus the morphemes, i.e. the roots and the grammatical and derivational formatives, including the affixes. The question is in which part of the
1874
XX. Morphologie in der Praxis
domain
basic functions
representative concepts and operations
apprehension & nomination
an entity is grasped by categorizing and individuating it; it is named by a label or a descriptive expression a concept is enriched, or an object is identified a representation is related to and delimited within the universe of discourse the relation of an entity to another one is established or inheres in one of them
categorization, types of concepts, empathy
concept modification reference
possession
spatial orientation quantification
predication participation
temporal orientation
illocution, modality, evidentiality contrast
nexion
communicative dynamism
an entity is localized in space statically or dynamically the extent of the involvement of a set of entities in a predication is delimited information is attributed to a referent a situation is articulated into an immaterial center and a set of participants and circumstants related to it and to each other a situation is designed with respect to its internal temporal structure and limits and temporally related to another situation a proposition is rendered relative to speaker, hearer and reality a concept or proposition is assessed qualitatively by comparison with similar ones a situation is expanded into a complex one, or several situations are linked together a proposition is articulated in foreground and background
attribution, apposition, relativization determination, deixis, reference tracking
possession in reference, possessive predication, external possessors reference points, local relations, spatial and gestalt properties of objects quantification in reference and in predication; counting, ordering existence, situation, characterization control & affectedness, central vs. peripheral roles, alignment of fundamental relations situation types, aspectuality, temporal relations
speech acts, obligation, volition, possibility, toning, evidentiality negation, comparison, gradation, intensification speech reproduction, complementation, interpropositional relations discourse structure, functional sentence perspective (topicalization, focusing, emphasis)
Tab. 170.4: Functional domains
overall description and in which way these lists are presented. In the ideal situation, there is a dictionary of the language beside the grammar. If so, then the dictionary contains entries for all of these items. It then properly includes a morphemicon (inventory of morphemes), i.e.
morphemes are lemmata just like stems or citation forms of words. For the sake of userfriendliness, not only morphemes, but also variants are conceded an entry, the latter reducing to a reference to the main entry. A dictionary entry refers to the grammar in two ways: (a) implicitly by the grammatical cat-
1875
170. Grammaticography
egorization and other grammatical information provided in the microstructure of each entry (see 3.1); (b) explicitly in the form of references to the relevant sections of the grammar for such entries which are grammatical formatives. In such a situation, the grammar contains no morphemicon. Instead, each grammatical morpheme is introduced in the semasiological description as part of its construction and paradigm. The distribution of information on individual grammatical items between dictionary and grammar is then a question of fine-tuning. There must be a certain amount of overlap as regards generic properties. Details on meaning and function, especially idiosyncratic properties, are provided in the lexicon, details on distribution and conditions of variation are provided in the grammar. If there is no extra dictionary, then a number of second-best solutions are available. Of these, the relative best is a glossary or vocabulary appended to the grammar that contains all the lexemes and morphemes that are mentioned in the grammar, including the examples, each coupled with its meaning or function. Again, variants are listed, too; and for the grammatical items there are references to all the places where they are discussed. The absolute minimum requirement is an alphabetical index of grammatical items at the end of the work, again with the relevant references. Among the indices there is also a subject index. While this is nothing special of books on grammar, it does assume a special role here. If the grammar is bipartite and both the semasiology and the onomasiology are organized hierarchically according to established conceptions, then the user can retrieve the information by systematic search through the table of contents if this is sufficiently detailed. Even then the grammar probably needs an index, as there are liable to be at least some concepts and terms that are language-specific or appear in unwonted contexts. But as long as grammars are not organized in such a way, an index of functional terms is necessary in a structural grammar to compensate for the lack of the opposite perspective; and vice versa for a functional grammar. Another kind of list that is often attached to a grammar ⫺ and not seldom also to a dictionary ⫺ is the list of inflection paradigms (mostly of the specimen type; see 4.3) of various word classes and their subclasses. If the language has a relatively complicated
inflectional morphology, as e.g. Ancient Greek conjugation, then it makes sense not to interrupt its systematic description by pages filled with conjugation tables and instead to relegate these to an appendix. Among the various lists that a grammar shares with other genres, such as lists of tables, figures, primary text sources, bibliographical references and the like, the list of abbreviations is of special interest. Among all the abbreviations used in a grammar, the (abbreviations of) grammatical category labels form a distinct subset. Whether or not other kinds of abbreviations are listed, the latter must certainly be listed, and they may constitute a list of its own. A separate list of grammatical category labels also helps the author to be consistent.
4. 4.1.
Descriptive devices
Representation of a single grammatical item 4.1.1. Representation of the significans Basically, one representation type is chosen for the whole grammatical description (i.e. excepting the phonetics and phonology). The primary choice is between a standard orthographic representation and a (technical) linguistic representation. The former is generally preferred because it makes the description more accessible to the speech community and other non-linguists interested in the language. Sometimes, standard orthography is supplemented by additional diacritics, e.g. to mark word stress. There are, however, situations where a linguistic representation is necessary or preferable, if a language has no orthography or if using traditional orthography would complicate comprehension of the grammar and is not in accordance with its descriptive tasks (cf., e.g., recent Chinese grammars). In this case, authors choose a basic representation trying to obtain a compromise between its theoretical ambitions, convenience for descriptive purposes and user-friendliness (especially in a learner’s grammar, the latter includes easy rules of reading). Sometimes printing (publishing) facilities prove to be a limiting factor. If the representation is based on linguistic principles, it is either a broad phonetic transcription or a phonemic or a morphophonemic representation. Use of a phonetic transcription as the basic representation in a grammar is very rare and essentially lim-
1876 ited to situations in which a grammar is published before a satisfactory phonological analysis of the language is available. In general, a morphophonemic representation is to be preferred to a phonemic one, but lowerlevel representations are, of course, necessary in those sections that deal with lower level variation. In phonetic representations, symbols are chosen from the International Phonetic Alphabet. At higher levels and if a certain phonetic distinction is irrelevant in the language, usually less technical letter symbols are preferred. The basic representation is used throughout a grammar; yet, a description of morphology often involves more or less significant deviations from this representation. In particular, an orthographic representation can prove insufficient for an adequate and sufficiently detailed morphemic segmentation, either in a language with intricate morphonology or in case the traditional orthography obscures certain morphological phenomena. Hence, some additional elements can be incorporated into representations of some forms or even phrasal examples in order to provide a transparent morphemic segmentation. For example, Russian orthography makes use of graphemes like я, ю which stand for combinations /ja/, /ju/ after a vowel or for single phonemes /a/, /u/ after a palatal consonant (very loosely, in the latter environment these graphemes are employed to signify the palatalization of the preceding consonant, in contrast with their plain equivalents а, u which occur after non-palatal consonants). This spelling convention makes the orthographic representation unsuitable for morphemic segmentation and for the representation of single morphemes, as is necessary, i.a., in combination with interlinear morphemic glossing (Art. 169). In a strictly orthographybased Russian grammar one would find numerous items like an “ending -а/-я”, a representation which suggests allomorphy, but in fact just reflects the fact that the ending /a/ occurs both after non-palatal and palatal consonants (in particular, after /j/). In order to avoid such misleading representations, a spelling convention can be employed which renders /j/ between vowels as well as palatal consonants in a more transparent fashion (Zaliznyak 1967: 201⫺294). As an example, consider the following forms, where morphemic borders are indicated as imposed by the orthographic representation:
XX. Morphologie in der Praxis
sg.nom sg.gen sg.dat
student
father-in-law hero
тuдт тuдт-а тuдт-u
тт тт-я тт-ю
грои˘ гро-я гро-ю
Tab. 170.5: Russian declension in standard orthography
The phonemic representation is:
sg.nom sg.gen sg.dat
student
father-in-law hero
stud’ent stud’ent-a stud’ent-u
t’es’t’ t’es’t’-a t’es’t’-u
g’iroj g’iroj-a g’iroj-u
Tab. 170.6: Russian declension in phonemic representation
It is now clear that the endings do not vary between these words; yet the transcription cannot help but contain deviations from the orthography which are of no relevance here and can only be misleading (cf., e.g., /i/ for in the last column). Now consider the same forms in an alternative spelling (the deviations from the established convention are shown by capital letters): student sg.nom тuдт sg.gen тuдт-а sg.dat тuдт-u
father-in-law hero тт тт-а тт-u
грои˘ ˘ -А гроИ ˘ -U гроИ
Tab. 170.7: Russian declension in adapted orthography
The orthographic representation is brought closer to the phonemic one in order to show that the three nouns follow the same inflection pattern, yet the deviations are reduced to a necessary minimum. The symbols reflecting a distinct representation type are underscored either by capitalizing or by using symbols from another alphabet (continuing the Russian example above, the symbols /’/ and ˘ ). Such /j/ might be used instead of and И mixed representations may be used as a compromise between the incompatible requirements of using a standard orthography and providing an adequate morphological analysis. Another side of the same problem is the representation of a morpheme abstracted from its morphological environments; that is, technically speaking, a unitary representation
1877
170. Grammaticography
of a set of allomorphs. Several techniques (which may also be combined) are in use: (a) Regular phonemic alternations are ignored (a technique normally favored by the use of the orthographic representation). (b) For more or less regular morphonemic phenomena, special symbols for morphonemes are employed, commonly, a capitalized phoneme symbol (for instance to abstract from vowel harmony phenomena). (c) The allomorphs are just listed, commonly separated by a slash. There seems to be no established means to represent the significans of non-segmental items (e.g. metaphony). 4.1.2. Representation of the significatum The meaning of a grammatical morpheme may be rendered in several ways, among them: ⫺ by using an appropriate functional (mnemonic) term, e.g. ‘iterative’, ‘ablative case’, etc., which may then be referred to by means of a conventional abbreviation, e.g. iter, abl etc., where upper case identifies (abbreviations of) grammatical category labels; ⫺ by translating it into the metalanguage (e.g. ‘constantly’, ‘from’). The choice of a name for a category may be determined by a language-specific terminological tradition or by a desire to keep the terminology as transparent and “speaking” as possible (see 1.6.1). Semantic information in morphology almost never reduces to naming categories appropriately (although, in some cases, a reference to the syntactic part of grammar is considered sufficient). The semantics of derivational morphemes is commonly described in terms of processes and operations. This is appropriate wherever operations apply in a regular and productive way, yet it can be misleading in the case of fossilized patterns. In the latter case, restrictions on the distribution should be carefully described (sometimes, just a list of relevant items seems the best solution, to be sure, with comments on their semantic motivation). The semantics of inflectional items is commonly described in terms of grammatical features structuring the respective paradigm (cf. 4.2).
To visualize the range of meaning of a polysemous (multifunctional) morpheme, a semantic map is a useful descriptive device (see Haspelmath 2003). This is a languageindependent n-dimensional (mostly with n ⫽ 2) arrangement of monosemic grammatical functions by their similarity. A semantic map of a particular case, e.g. the Latin dative, will cover a contiguous area that comprises a subset of case functions (recipient, addressee, experiencer etc.) appearing on the map. A similar map can be drawn for the dative of another language, e.g. Turkish, and then the semantic expansion of the dative in the two languages may be easily compared. Such a semantic space is also the locus for diachronic semantic variation. 4.2. Category and feature Rules of syntax may refer to a morphological category or grammatical feature that a constituent bears. The formulation of such rules often presupposes the specification of such categories in a parameter-value format, where a grammatical category constitutes a parameter, and its subcategories, the values (s. Art. 28). Thus, the inflectional information of a Latin verb form may be represented as follows: cantaveritis 2 m u m m u
person number tense anteriority mood voice
Fig. 170.1: Inflectional information of a Latin verb form
Another task of a morphological description is to provide an appropriate system of inflection classes to be picked up in a lexicon (cf. 2.2 and Art. 65). The following duties await the analyst here: (a) The inflection classes have to be brought into a hierarchy. Wherever a language has some complexity in this area, major inflection classes usually have minor subclasses, and these have a few totally irregular members. For instance, one of the major conjugation classes in Latin is the consonantal conjugation. One of its subclasses is constituted by a perfect formation which involves lengthening of the
1878 root vowel. This, in turn, has emo¯ ‘buy’ as a member, which is slightly irregular in having an epenthetic consonant in the perfect participle emptum ‘bought’. (b) The inflection classes have to be named by appropriate terms. Since the conditioning factor of an inflection class is often the stem final, this is suitable as a label of the class, as in the Latin a conjugation. The concept of a particular inflection class reflects its position in the taxonomy, e.g. ‘consonantal conjugation with lengthening perfect’. This, of course, leads to cumbersome terms which may be abbreviated by numbers and letters, e.g. ‘3e’ (third conjugation, subclass e). The abbreviations are needed, among other things, in the section of grammatical information of the lexical entry of such a verb. (c) For didactic purposes, those inflected forms of the paradigm are identified from which one can deduce the entire inflectional paradigm. For this reason, Latin nouns are often quoted not only in the nominative, but also the genitive, e.g. ge¯ns, gentis ‘stem’. For German, one even needs the nominative plural in addition (Hahn, Hahns, Hähne ‘rooster’). For many semiregular Latin verbs, if one knows the first person singular present indicative active, the first person singular perfect indicative active and the perfect participle, then one can conjugate the verb correctly through all of its categories. Traditional grammars and dictionaries therefore specify the forms of these categories for semiregular verbs, e.g. emo¯ ⫺ e¯mi ⫺ emptum ‘buy’. Instead of such a set of word forms, it is also possible to mention a set of allomorphs appearing in such word forms. For instance, a German apophony class may be identified by i ⫺ a ⫺ u (as in singe ⫺ sang ⫺ gesungen). 4.3. Paradigms Inflectional morphology is commonly represented in terms of paradigms arranged as one- or multidimensional tables. For each word class, the set of tables has the same structure in terms of number of columns and rows and their category labels. There are two linguistic forms of such tabular representations. In the traditional paradigm, a plain example word is represented in the forms of all the values of the categories, as in Tab. 170.8.
XX. Morphologie in der Praxis
number singular
plural
person 1 2 3
laud-o laud-as laud-at
laud-amus laud-atis laud-ant
Tab. 170.8: Specimen paradigm of Latin personal endings
This kind of representation is the original linguistic sense of the Greek term para´deigma ‘example’. It is still used in textbooks or if the morphology is so fusional that inflectional markers are hard to represent separately. This leads us to the second form of representing a paradigm, in which the cells of the table are occupied just by inflectional morphemes, as in Tab. 170.9. number singular
plural
person 1 2 3
-en -ech ⫺
-o’n -e’x -o’b
Tab. 170.9: Marker paradigm of Yucatec personal suffixes
This kind of representation is more abstract and proper to linguistic science. It is viable for agglutinative morphology. The specimen paradigm is, in a sense, a definition per ostensionem that appeals to intuitive understanding. The marker paradigm only works to the extent that the information that it presupposes is made explicit in the grammar. This concerns, in particular, syntagmatic information on the way the markers combine with the stem (see 4.4). If complex morphophonemics are involved, it is prudent and userfriendly to complement the rules by a specimen paradigm. For a given morphological category, the order in which its values are enumerated is fixed by tradition in many cases; e.g. ‘1st, 2nd, 3rd person’, ‘nominative, genitive, dative, accusative, ablative, vocative case’. A properly linguistic criterion for their arrangement is syncretism (Art. 66). Especially for semantically motivated syncretism, it is convenient if the syncretistic category values are adjacent in the table. Tab. 170.10 and Tab. 170.11 are two examples to illustrate ways of displaying syncretistic paradigms:
1879
170. Grammaticography number singular
plural
dual
deva¯s
devau
case deva devas devam devena deva¯ya deva¯t devasya deve
vocative nominative accusative instrumental dative ablative genitive locative
deva¯n devais devebhyas deva¯na¯m devesø u
deva¯bhya¯m devayos
Tab. 170.10: Declension of Old Indic a-stems (paradigm of deva- ‘god’)
rule or instruction how to construct a certain morphological form. E.g., concerning the Latin supine II like dictu¯ ‘to be said’: in order to make the supine II of a verb, form its passive participle, convert this into a stem of the u-declension and form the ablative singular of this stem. Syntagmatic relations are commonly analyzed in terms of constituency or dependency and accordingly visualized by tree diagrams or dependency stemmata, each illustrated here with an example: N
The tables representing paradigms are at most two-dimensional. For agglutinative morphology (e.g. the Turkish case paradigm), even a one-dimensional table suffices to represent the set of values that each morphological category can take. If two morphological categories are cumulated in a morpheme, as e.g. person and number in many conjugation paradigms, a two-dimensional table becomes necessary. If more than two categories are fused in a morpheme, or if inflection classes are involved in addition, a suitable combination of two-dimensional tables is chosen, as also shown in Tab. 170.11. Although such complex paradigms are not necessary for agglutinative morphology, it is prudent at least to give a couple of examples of morphologically complex word forms to render the formulas concrete. 4.4. Syntagms The accurate statement of the combination of a morphological marker (of a certain category) with its host may require specification of the category of the base, the conditions for the combination and specification of the resulting category. This may take the form of a
tense
Nzr
Vbzr
Ver-
A
N
Adr
ein
-ig
[ [ [Ver-]Vbzr [ [ein]N [-ig]Adjr ]A ]V[-ung]Nzr ]N Fig. 170.3: Dependency structure of complex stem
The valence class of a verb or some other relational word is, above all, a syntactic property. However, in many languages, valence classes coincide with inflection classes, so that specification of a valence class may be relevant for morphology, too. This may be done in the form of a construction formula, e.g. for a trivalent verb: [NP]Sbj __ [NP]dir.Obj [AdvP]. Such valence classes may also con-
past m
f
n
cˇita´l
cˇita´la
cˇita´lo
gender person
m/f/n
sg.
1 2 3 1 2 3
cˇita´ju cˇita´esˇ cˇita´et cˇita´em cˇita´ete cˇita´jut
Tab. 170.11: Russian conjugation
-ung
Fig. 170.2: Constituent structure of complex stem
present
number
pl.
V
cˇita´li
1880
XX. Morphologie in der Praxis
position category
5 negation
4 subject
3 tense/aspect
2 relative
1 object
0 stem
Fig. 170.4: Swahili verb prefixes
rank category
0 stem
1 adn/num*
2 relational0⫺2
3 modal0⫺2
4 (associating/complementizing)
Fig. 170.5: Kayardild nominal suffixes
tribute to an account of the function of a given case. There are two basic types of sequential order of elements, scope order and template order. Scope order presupposes a binary construction consisting of a head and a dependent one of which has scope over the other, and specifies the position of the dependent as against the head. Thus, the position of an adnominal relative clause is determined with respect to its head nominal, and the position of the complement NP with respect to an adposition. At the morphological level, this convention is replaced by specifying the position of bound morphemes with respect to their hosts, or of grammatical formatives with respect to a lexical root or stem. In a syntagm such as Fig. 170.3, the relative order of the suffixes -ig and -ung is a contingent consequence of the fact that each of them has scope order with respect to the operand preceding it. This may be summarized in a formula like [ [ X ]N/V -ig ]A. Template order is the relative order of elements which do not form a construction with each other. It is typical of clitics which attach to whatever their host is, in a fixed order that does not reflect any semantic scope. Therefore template order cannot be specified in the same straightforward way as scope order. Instead, a schema is set up consisting of a fixed sequence of slots, a slot being a syntagmatic position for an element. For instance, the order of the prefixes on the Swahili verb may be described by the template in Fig. 170.4. The interpretation of such a template is straightforward: Each of the five prethematic slots is reserved for an element out of the paradigm designated by the slot label. If a slot remains unoccupied, this does not concern the order of the other elements.
Some morphological systems are so complex that different ranks of complex stems may be assumed such that affixes may be added at each rank. The suffix positions of the Kayardild noun may be schematized as in Fig. 170.5 (according to Evans 1995: 122). At the first rank, the stem may be followed by any number of adnominal case and number suffixes. At rank 2, up to two relational cases, and at rank 3, up to 2 modal cases may follow. At rank 4, there may be either an associating or a complementizing suffix. Templates are necessary when the sequential position of elements is not in consonance with their semantic scope. Structuralist accounts of the middle of the 20th century often give templates even for constructions which are decomposable into binary constructions in each of which scope order obtains. An example would be ‘stem ⫺ past tense ⫺ personal ending’ for the German regular finite verb. This is helpful in visualizing a complex form as a whole, but should be preceded by a separate account of each of the elementary constructions involved. 4.5. Illustrative examples Abstract or formal statements should be supplemented by example expressions from the language being described. One of the most conspicuous differences between older and modern linguistic descriptions is the growing extent to which this principle is heeded. One might suppose this to be a trivial issue. Actually, however, descriptions that contain no examples are very often not intelligible and, to this extent, worthless. Since most linguistic statements fail in one or another respect and simply do not do justice to the real complex-
1881
170. Grammaticography
ity of the language, an example can teach more than an ever so neatly formalized rule. Above a certain degree of syntagmatic complexity, such an example should be presented in the canonical trilinear representation (cf. Art. 169), i.e. the text should be provided with an interlinear morphemic gloss and an idiomatic translation. While this was practically never done in traditional descriptions and even in structural linguistics, modern grammatical descriptions, especially those written in a functional or typological framework, often abound in analyzed illustrative examples. In this respect, there has been a real progress in descriptive linguistics. There is a certain amount of trade-off between the corpus specimina as presented in the appendix and the examples in the running descriptive text. First of all, extensive illustration in the description can be reduced and be replaced by references to the corpus. Generally, such segments from the corpus are copied as illustrative examples in the running text which represent frequent patterns and do not involve any additional complications irrelevant to the issue at hand. Second, not all of the levels of representation figuring in the edition of the text corpus need be repeated in an example of the running text. For many purposes and especially for simple examples, some representation of the significans (orthographic or morphological, i.e. phonological with morpheme boundaries marked) coupled with a translation as a shorthand semantic representation will be sufficient. Thirdly, wherever uniformity of a set of examples is required, for instance in a conjugation paradigm, examples will not (entirely) be drawn from the corpus, but will rather complement it. Normally, the illustration of a morphological phenomenon does not require a whole example sentence. Very often, a word form or, at most, a phrase will do. Where a sentence appears to be needed to adequately exemplify a grammatical phenomenon, this is a hint at the syntactic rather than morphological nature of the phenomenon. Also, while a corpus sentence may appear to be a more natural piece of language use, it may also distract, by its complexity, from the morphological phenomenon that is at stake. The situation is, of course, different in syntax. There the author bears heightened responsibility if she simplifies corpus sentences for illustrative purposes.
5.
References
Cherubim, Dieter (1973), “Grammatikographie”. In: Althaus, Hans Peter & Henne, Helmut & Wiegand, Herbert Ernst (eds.), Lexikon der germanistischen Linguistik. Tübingen: Niemeyer, 768⫺778 Cole, Peter (1982), Imbabura Quechua. Amsterdam: North-Holland (Lingua Descriptive Studies, 5) Comrie, Bernard (1978), Aspect: An Introduction to Study of Verbal Aspect and Related Problems. Cambridge: Cambridge University Press Comrie, Bernard & Smith, Norval (1977), “Lingua Descriptive Studies: Questionnaire”. Lingua 42, 1⫺72 Evans, Nicholas (1995), A Grammar of Kayardild. Berlin, New York: Mouton De Gruyter Gabelentz, Georg von der (1881), Chinesische Grammatik: mit Ausschluß des niederen Stils und der heutigen Umgangssprache. Leipzig: Weigel Gabelentz, Georg von der (1891), Die Sprachwissenschaft: Ihre Aufgaben, Methoden und bisherigen Ergebnisse. Leipzig: Weigel Nachf. [2. überarbeitete Aufl.: Leipzig: Tauchnitz, 1901] Givo´n, Talmy (1993), English Grammar: A Function-Based Introduction, Vol. 1⫺2. Amsterdam & Philadelphia: J. Benjamins Haspelmath, Martin (1993), A Grammar of Lezgian. Berlin: Mouton de Gruyter (Mouton Grammar Library, 9) Haspelmath, Martin (2003), “The Geometry of Grammatical Meaning: Semantic Maps and Crosslinguistic Comparison”. In: Tomasello, Michael (ed.), The New Psychology of Language: Cognitive and Functional Approaches to Language Structure, Vol. II. Hillsdale/NJ: L. Erlbaum, 211⫺242 Humboldt, Wilhelm von (1836), “Über die Verschiedenheit des menschlichen Sprachbaues und ihren Einfluß auf die geistige Entwicklung des Menschengeschlechtes”. Reimpr.: Humboldt, Wilhelm von (1963), Schriften zur Sprachphilosophie. [⫽ Werke in fünf Bänden, hrsg. v. A. Flitner und K. Giel, Bd. III]. Darmstadt: Wissenschaftliche Buchgesellschaft, 368⫺756 Jespersen, Otto (1937), Analytic Syntax. Kopenhagen: E. Munksgaard Lehmann, Christian (1980), “Aufbau einer Grammatik zwischen Sprachtypologie und Universalistik”. In: Brettschneider, Gunter & Lehmann, Christian (eds.), Wege zur Universalienforschung. Sprachwissenschaftliche Beiträge zum 60. Geburtstag von Hansjakob Seiler. Tübingen: Narr (Tübinger Beiträge zur Linguistik 145), 29⫺37
1882
XX. Morphologie in der Praxis
Lehmann, Christian (22002), Possession in Yucatec Maya. Erfurt: Seminar für Sprachwissenschaft der Universität (ASSidUE, 10) Lehmann, Christian (2004 a), “Data in Linguistics”. Linguistic Review 21.3/4: 175⫺210 Lehmann, Christian (2004 b), “Funktionale Grammatikographie”. In: Premper, Waldfried (ed.), Dimensionen und Kontinua: Beiträge zu Hansjakob Seilers Universalienforschung. Bochum: N. Brockmeyer (Diversitas Linguarum, 4), 147⫺166 McGregor, William B. (1990), A Functional Grammar of Gooniyandi. Amsterdam & Philadelphia: J. Benjamins (SLCS, 22) Mosel, Ulrike (2002), “Analytic and Synthetic Language Description”. In: Ezawa, Kennosuke & Kürschner, Wilfried & Rensch, Karl H. & Ringmacher, Manfred (eds.), Linguistik jenseits des Strukturalismus. Akten des II. Ost-West-Kolloquiums, Berlin 1998. Tübingen: G. Narr, 199⫺208 Mosel, Ulrike (2003), Grammaticography ⫺ the Art and Craft of Writing Grammars. Kiel: Seminar für Sprachwissenschaft der Universität McQuown, Norman A. (1990), Grama´tica de la lengua totonaca (Coatepec, Sierra Norte de Puebla).
Me´xico, D.F.: Universidad Nacional Auto´noma de Me´xico (Coleccio´n Lingüı´stica Indı´gena, 4) Noonan, Michael to app., “Grammar Writing for a Grammar-Reading Audience”. In: Payne, Thomas & Weber, David (eds.), Grammar Writing. Amsterdam & Philadelphia: J. Benjamins (Typological Studies in Language) Payne, Thomas (1997), Describing Morphosyntax: A Guide for Field Linguists. Cambridge: Cambridge University Press Plank, Frans (1992, ed.), Paradigms: The Economy of Inflection. Berlin & New York: Mouton de Gruyter (Empirical Approaches to Language Typology, 9) Shopen, Timothy (1985, ed.), Language Typology and Syntactic Description, Vol. 1⫺3. Cambridge etc.: Cambridge University Press Zaliznyak, Andrey (1967), Russkoe imennoe slovoizmenenie [Russian nominal inflection]. Moscow: Nauka
Christian Lehmann, Erfurt (Germany) Elena Maslova, Stanford (U.S.A.)
171. Lexicography 1. 2. 3. 4.
Morphology in dictionaries Morphological models Morphology and lexicography References
1.
Morphology in dictionaries
1.1. Monolingual dictionaries We begin our discussion of the contribution of morphology to lexicography with a review of the treatment of inflection (cf. Mugdan 1989⫺1990 a) and word formation in monolingual dictionaries intended for native speakers. The questions that need to be asked are: What kinds of morphological information do dictionaries contain? How do they present them? 1.1.1. English Our starting point is the second edition of the Oxford English Dictionary (Simpson & Weiner 21989, eds.). Each “main word”, i.e. headword or primary entry, begins with its “identification”, which includes:
“The inflexions, i.e. plural of substantives, and principal parts of verbs, when other than the ordinary -s, -ed.” (Simpson & Weiner 21989, eds.: xxvii)
Irregular inflections only are indicated for nouns and verbs. In the area of word formation, the OED distinguishes “combinations”, i.e. compound words, and “derivatives”, i.e. words formed by suffixation. Combinations are treated as main words if they have a specialized meaning which cannot be deduced from the meanings of their elements and therefore require a full definition, or if they have multiple senses. Otherwise they are treated as run-ons, at the end of the article for the main word constituting the first element of the combination. The category of derivatives does not include words derived by prefixation, since, because of the alphabetical ordering of the dictionary, these must be treated as main words in their appropriate place. Derivatives by suffixation are also treated as main words, unless they are of infrequent occurrence or have only one or a small number of meanings directly relatable
171. Lexicography
to the meaning of the main word from which they are derived; in which case they are treated in the article for the main word in a separate paragraph at the end. Also given main word status are prefixes, suffixes and combining forms that are especially productive in the formation of derived and compound words. The practice of the OED in the treatment of morphological information is followed in broad terms by other British English dictionaries, though with some significant variations. In the treatment of inflections, Collins English Dictionary (Hanks 1979, ed.; Treffry 4 1998, ed.), Chambers English Dictionary (Schwarz et al. 1988, eds.) and the Concise Oxford Dictionary (Allen 81990, ed.; Pearsall 10 1999, ed.) give not only those formed irregularly but also regular formations which are likely to cause confusion or difficulty, e.g. where the final consonant of the root doubles with the addition of the inflectional suffix (dragging) or where a letter is dropped from the root (arguing) or where the addition of a suffix changes the pronunciation of the root (younger). The inflections of adjectives are treated as well as those of nouns and verbs, and the Concise Oxford at least indicates which two-syllable adjectives regularly take the -er/-est comparative and superlative suffixes. These dictionaries also enter irregularly inflected forms of words as headwords, with a cross-reference to the base or citation form. The treatment of compounds and derivatives is, however, variable. Chambers treats all compounds and suffixed derivatives as runons; prefixed derivatives are headwords, except that derivatives with productive prefixes, like re-, sub-, un-, are listed without definition at the bottom of the page unless they require definition by virtue of a specialized or particularly complex meaning. Collins enters all compounds as headwords and all suffixed derivatives, unless the meaning of the derivative is clear from the root, when it is treated as a run-on under the root. It also, like Chambers, has lists of prefixed derivatives at the bottom of the page. The Concise Oxford has prefixed derivatives only as headwords, and suffixed derivatives are treated in the same way as in Collins. With compounds, however, the Concise Oxford uses a spelling rather than a meaning criterion: if the compound is written solid it is treated as a headword, otherwise it is treated as a run-on under the first element. The treatment of inflection is particularly detailed in Webster’s Third New International
1883 Dictionary (Gove 1961, ed.). All inflectional forms are indicated in the dictionary, whether irregular or regular. Regular inflections are given only by suffix, but where any inflected form of a word is irregular all the forms are given in full. If an inflected form would fall in the alphabetical sequence “more than five inches” from its root, then it is entered as a headword and cross-referenced to the root. These practices also mean that all adjectives which form their comparative and superlative by suffixation have that indicated. In addition to the inflectional information in the dictionary entries, the front matter to Webster’s Third contains two pages of detailed spelling rules for forming the plurals of nouns. It also contains two pages of detailed rules for writing compounds. All types of compound ⫺ solid, hyphenated and open ⫺ are entered as headwords, as are suffixed derivatives where they need separate definition. 1.1.2. French English has a relatively limited inflectional morphology, with a small number of irregularities, affecting only the plural morpheme in the noun and the past tense and past participle morphemes in the verb (cf. Art. 119). It is, therefore, practicable to indicate, where necessary, the irregular forms in the dictionary entries themselves. French, however, shows a more complex inflectional morphology, especially in the verb (cf. Art. 121). The Grand Larousse (Guilbert et al. 1971, eds.) contains a “Tableau des Conjugaisons”, covering several pages, to which the verb entries in the dictionary are cross-referenced. For example, acque´rir is indicated as belonging to conjugation 13. This dictionary also contains in its front matter extensive discussion of the word formation processes of prefixation, suffixation and compounding, from both a diachronic and a synchronic perspective. An alternative solution to the inflectional question is found in Le Petit Robert (Rey 1973, ed.; cf. Drivaud 1997, ed.). For nouns and adjectives, the feminine form and the plural form (when it is not the regular -s) are given near the beginning of the entry for the word (e.g. journal, aux; jumeau, elle). For verbs, a number of strategies are employed: there is an appendix containing the “Conjugaisons des Verbes Franc¸ais”; regular -er and -ir verbs have their inflections given only in the table of conjugations; verbs with minor irregularities, such as the inclusion of -e- in bougeons (1st person plural present of bouger) or the c¸ in plac¸ons,
1884 have these irregularities given in the dictionary entry for the verb; major irregular verbs are referred to a prototype, e.g. “PRODUIRE, conjug. conduire”, where the forms for that group of irregular verbs are given, as well as in the table of conjugations. Le Petit Robert includes derivational prefixes and suffixes as headwords and also contains a table of suffixes as an appendix. 1.1.3. German In German, the inflectional forms are yet more diverse and varied (cf. Art. 120). In the Deutsches Wörterbuch (Wahrig et al. 1980⫺ 1984, eds.), the front matter contains numbered tables for the declension of nouns and pronouns, for the declension of adjectives and indefinite pronouns, and for the conjugations of verbs, which are used for cross-reference from the dictionary entries. Each noun entered in the dictionary is marked for gender, and the genitive singular and the plural inflections are given; uncountable nouns (i.e. with no plural inflection) are marked as such, as are nouns occurring only in the plural form; for the adjectival and nominal declension of nouns the user is referred to the declension tables. A similar cross-reference is made for adjectives, where three tables are devoted to regular patterns and five to deviations from these. In the case of verbs, socalled “weak” verbs are regarded as the regular category, and there is no cross-reference to a conjugation table, though one is provided. In the case of “strong” or irregular verbs, the entry for the verb contains a number which cross-refers to a conjugation table. Irregular inflectional forms are entered as headwords in the dictionary, with an indication of their form and the citation form to which they belong, e.g. “brachte <1, 3 Sg Ind Prät von> bringen”. Clearly, in the case of German, and to some extent of French, an important question is how and where to show the inflectional forms of words, whether in the dictionary entries themselves or in the tables outside of the main body of the dictionary with cross-references in the entries. It is a question to which we shall return (cf. 3.2). 1.2. Learners’ dictionaries Learners’ dictionaries are a particularly English phenomenon, though a French and a German learners’ dictionary will be referred to below. They aim to be a comprehensive and detailed guide to usage for, mainly, ad-
XX. Morphologie in der Praxis
vanced foreign learners. Care is, therefore, taken to ensure that all necessary information, especially grammatical information, is included (Jackson 1988: 174⫺191). The first in the marketplace, in 1948, was A. S. Hornby’s Oxford Advanced Learner’s Dictionary, now in a much revised fifth edition (Crowther 51995, ed.). In this dictionary, irregular inflections are shown and where the addition of an inflectional suffix causes spelling changes, e.g. the doubling of a stem-final consonant. All words with irregular inflections are listed as headwords, with cross-reference to the stem form. There is a table of irregular verbs as an appendix. All adjectives forming the comparative and superlative in -er/-est are said to be marked for this, though the dictionary (4th edition, cf. Cowie 41989, ed.) is less than consistent in practice (Bolinger 1990: 135). In the area of word formation, suffixed derivatives are entered as runons after a distinctive triangle symbol, if the meaning is clear from the root word; otherwise a derivative is entered as a separate headword, as it is if the spelling connection is not clear, e.g. with satisfaction from satisfy. Compounds are usually listed as run-ons under the first element and are preceded by a distinctive square symbol; however, if the first element of a compound is not a headword, or if the compound is a loanword or has an idiomatic meaning, then it is entered separately. Affixes and combining forms are listed as headwords and accompanied by explanations and examples. The second advanced learners’ dictionary was the Longman Dictionary of Contemporary English in 1978, now in a third edition (Summers 31995, ed.). It has the same information on inflections as its Oxford competitor, except that it contains no information about adjective inflections in the entries apart from irregular ones. In word formation, suffixed derivatives are entered as run-ons if the meaning is derivable from the root, but compounds are always entered as separate headwords. The dictionary (2nd edition, 1987) contains a ten-page appendix on word formation, including alphabetical listings of “word beginnings” and “word endings”. Common affixes and combining forms are listed as headwords and cross-referenced to the appendix. The third English advanced learners’ dictionary was the Collins Cobuild English Language Dictionary (Sinclair 21995), which, in its construction ⫺ based on a large computer
171. Lexicography
corpus ⫺ and in its presentation of information, marked a significant departure from conventional lexicographic practice. Inflectional forms are given in full, at the beginning of the entry, for all words that have inflections, both regular and irregular. Each entry is for a single orthographic word; there are no separate entries for a word form belonging to more than one word class. There is one entry for play, for example, with the inflectional forms plays, playing, played; the form plays indicates both ‘noun plural’ and ‘verb 3rd person singular present tense’. In word formation, a suffixed derivative is entered as a run-on, after a distinctive diamond symbol, only if the meaning is directly derivable from the stem; otherwise, if the derivative is very common or has a different meaning from the stem, it has a separate entry. Compounds always have separate headword status, as do common prefixes and suffixes (including combining forms). A fourth advanced learners’ dictionary in English is the Cambridge International Dictionary of English (Proctor 1995, ed.). For inflection the general principle is that, in the case of nouns and verbs, only irregular inflections are given in the dictionary, but regular inflections are described in the front matter and boxed “language portraits” in the body of the dictionary add further information. In the case of adjectives, all those that can form comparative and superlative with the suffixes -er/-est are so marked in the dictionary, as are ungradable adjectives. One of the features of the Cambridge International is the multiple entry of headwords with more than one core sense. Derivatives are entered, each on a separate line, below the sense to which they relate. Additionally, there is a “language portrait”, entitled “Combining Forms”, which describes the main derivational prefixes and suffixes of English. Compound words all have headword status. English learners’ dictionaries both reflect and vary the diversity of practice in the treatment of morphological information that was evident in the monolingual dictionaries of English. The same is true of the French learner’s dictionary, Dictionnaire du Franc¸ais Contemporain (Dubois et al. 1971, eds.). It contains in its front matter tables of “Conjugaisons des Verbes” and “Suffixes et Pre´fixes”. In the dictionary entries, if a verb has an irregular conjugation, it is cross-referenced by number to the table of conjugations, e.g. “sortir (conj. 28)”. An interesting
1885 feature of this dictionary are the boxed articles in the body of the dictionary on derivational prefixes, e.g. archi-, hypo-, in-. A first German advanced learners’ dictionary appeared in 1993, Langenscheidts Großwörterbuch Deutsch als Fremdsprache (Götz et al. 1993, eds.). The inflections of all nouns (genitive and plural) and of all verbs (simple past and perfect tenses) are given; additionally the “3rd person singular present tense” is given for irregular verbs where it differs from the stem of the infinitive headword. This dictionary has paid particular attention to aiding the learner understand the word formation processes of German. A larger number of prefixes and suffixes are entered as headwords, with a label indicating how productive they still are, and with some typical examples. Common compounds are listed within the entries for words, after a K(for “Komposita”) symbol where the word is the first element of a compound, and -K where it may be the second element. 1.3. Bilingual dictionaries In bilingual dictionaries, inflectional information is normally given for headwords. In a French/English bilingual dictionary, for example, inflections are given for French words in the French-English side of the dictionary, and for English words in the English-French side; though this does not invariably apply. In the Collins-Robert French/English Dictionary (Atkins et al. 21987, eds.), the irregular or unobvious plurals of nouns are indicated in this way. However, for gender variants, the feminine and masculine forms of a French noun are given on the English-French side if they could both be the translation of an English noun, e.g. “singer .. n chanteur m, -euse f”. The feminine forms of French adjectives are also given on the English-French side, if they are not formed regularly. In the case of verb inflections, the Collins-Robert follows the respective monolingual traditions: French verbs with irregular inflections, on the French-English side of the dictionary, are given a number, which refers the user to the tables of verb conjugations in the appendix; the inflections of irregular English verbs are given at the entry for the verb in the EnglishFrench side of the dictionary, though there is also an appendix of irregular English verbs. Inflected forms of nouns and verbs that cannot be easily related to the root form are en-
1886
XX. Morphologie in der Praxis
tered as headwords, with a cross-reference to the citation form. The separate monolingual traditions are also followed in the Collins German Dictionary (Terrell et al. 1980, eds.). The genitive and plural inflections are given for all German noun headwords, on the German-English side, except for certain regular noun inflections, indicated in a table in the front matter. For English, the irregular plurals, and only these, are given for nouns on the English-German side. Similarly, the irregular, or “strong” verbs have their inflections given in the entries for the simple verb. This means that verbs which are derived by prefix or compounding from a simple irregular verb are cross-referenced to that verb for the inflectional information. German verbs that have regular inflections apart from not forming the past participle with ge- are marked with an asterisk, e.g. bedienen*, hantieren*. Irregular inflectional forms are listed as headwords and cross-referenced to the citation form. In bilingual dictionaries, the need to provide the information that a user requires at the most convenient place vies with the pressing need to keep the dictionary within manageable bounds of both volume and cost. Inflectional information would most conveniently, from the user’s point of view, be provided at the point of look-up, e.g. for German verbs on the English-German side of the dictionary. But providing information there would mean either increasing the volume of the dictionary or leaving something else out. The user must therefore look in two places to discover, for example, the past tense form of the German translation of speak: in the English-German side to discover that the equivalent in German is sprechen, and in the German-English side to discover that the past tense form is sprach. How to present inflectional ⫺ and word formation ⫺ information in an accessible way is an important question, to which we shall return below (in 3).
2.
Morphological models
2.1. Dictionary entries Before we consider which of the models of morphology is most used in describing morphological data in dictionaries, it will be useful to consider the nature of a dictionary entry. A dictionary entry consists of a headword and an accompanying article that gives
a lexical description of the headword. The lexical description includes phonological, morphological, syntactic, lexicological, semantic and pragmatic information (Jackson 1988: 35⫺47; but cf. Dubois & Dubois 1971: 90). The headwords of a dictionary usually include the “simple” words (i.e. the root forms that have not undergone any word formation processes) of the vocabulary in their citation form (see 3.1), some compound words, and some derived words. Compounds and derivatives are usually given separate headword status if their meaning has diverged from that of the root so that it can no longer be deduced directly from the form of the word itself. Otherwise, compounds and derivatives are treated as run-ons under the simple word, i.e. as part of the morphological/lexicological description of the simple word. Headwords may also include inflected forms of words (for cross-reference purposes), affixes and combining forms, abbreviations, and so on. This raises the question whether dictionaries take the word or the morpheme as their basic unit of description. The vast majority of headwords in dictionaries, for European languages at least, are words, in their citation form, though what counts as a dictionary word is not always a simple matter to decide (Dubois & Dubois 1971: 61⫺65). Other items that appear as headwords (prefixes, suffixes, and so on) are included because they are deemed to represent useful reference material for dictionary users; they are sometimes included additionally or instead in an appendix to the dictionary. Whether the word is the appropriate basic unit for highly inflecting languages like Swahili is another question (Bwenge 1989). Some of the information contained in the article that accompanies a headword may be morphological. This includes the inflectional forms of the headword, usually if these are not formed on the regular pattern for the word class concerned (Jackson 1988: 148). It also includes the words that have been formed from the headword by the word formation processes of compounding and derivation, where the resulting forms do not require separate definition. In the case of derivation, however, it is only suffixed derivatives that are included in the article; words derived by prefixation have separate headword status. This discrepancy arises from the exigencies of the alphabetical ordering of dictionary entries and the assumption that a user
1887
171. Lexicography
cannot be expected to make a morphological analysis of a word, i.e. recognise the presence of a derivational prefix, before looking it up.
morphology; it does not treat the forms that arise by regular or irregular derivational processes (Hoenigswald 1967: 106).
2.2. Word and paradigm Dictionaries for European languages operate with words as their basic unit (Zgusta 1971: 159). The implicit working model of lexicographers tends therefore to be the wordand-paradigm model of morphology (Zgusta 1971: 119). This is most obvious in the French monolingual dictionaries reviewed in 1.1.2, where the inflections of verbs are indicated in the dictionary entry merely by a number, which refers the user to a paradigm in the table of conjugations at the front or back of the dictionary. But it is equally the model implicit in the indication of irregular inflections in the dictionary entry itself. First of all, a regular paradigm of inflections is assumed ⫺ the “declension” of nouns, the “conjugation” of verbs ⫺ which is either listed in the front or back matter of the dictionary or assumed to be in the general knowledge of the dictionary user (Hoenigswald 1967: 109). Secondly, the inflectional forms are given in the articles for irregularly inflected words that enable the user to construct the appropriate paradigms for each word. For example, the “principal parts” of verbs in German and English are given (3rd person singular present tense, past tense, past participle), from which the paradigms for the various tenses can be worked out (Lyons 1977: 515). What the word-and-paradigm model does not always make explicit, because of the assumption of a regular paradigm, is that there may be gaps in the paradigm even for some regularly inflected words (Hoenigswald 1967: 109). Not all verbs in English have the full paradigm forms: stative verbs (e.g. know) do not usually form progressive tenses; modal verbs (e.g. can) do not have non-finite forms. There are nouns that have no plural forms (German Armut, English poverty), usually marked in learners’ dictionaries as “uncountable”. The word-and-paradigm model assumes a language whose major word classes inflect and where a dictionary user can either be referred to an appropriate table or be given some principal parts from which to construct the paradigms. Where a language does not have a series of inflected forms, the paradigm model is irrelevant, e.g. for Burmese (Zgusta 1971: 162). The model really only has something to say about the inflectional aspects of
2.3. Word formation One reason why the word-and-paradigm model does not apply well to derivation and other types of word formation is that it is essentially a static model. Word formation is conceived in terms of processes. General monolingual synchronic dictionaries are static descriptions; they record the products of word formation processes, rather than the processes themselves. However, there are ways in which dictionaries can indicate the process nature of word formation. Firstly, the process by which a compound or derivative comes into being may be indicated in the etymology of the word (see Art. 174). Secondly, dictionaries often list affixes and combining forms as headwords, or in an appendix, with an implication that they continue to be productive in the processes of forming new words. With compounding, that is more difficult. There is much that the practice of lexicography can learn from the linguistic investigation of morphology (Jackson 1988: 250) and vice versa; it is to this we now turn.
3.
Morphology and lexicography
3.1. Citation forms The form of a word that is entered as the headword is known as the “citation form” (Lyons 1977: 513; Jackson 1988: 9) or the “canonical form” (Zgusta 1971: 119; Landau 1989: 76). For words that undergo inflection, the citation form represents the paradigm (Landau 1989: 76; Zgusta 1971: 119). In that sense they may be regarded as “theoretical units” (Dubois & Dubois 1971: 62), since a headword (e.g. biscuit) stands for all the possible forms of the word (i.e. singular biscuit and plural biscuits). In English dictionaries, the citation form is the stem to which the inflectional suffixes are added (Lyons 1977: 513). So, for regular verbs in English, the citation form is the infinitive without to (e.g. play), to which the inflectional suffixes are added (e.g. plays, played, playing). In German and French dictionaries, the citation form is also the infinitive (e.g. jouer, spielen); but in these languages it is not the stem to which the inflectional suffixes are added. The infinitive suffix, -er and -en respectively, has to be
1888 removed to produce the stem forms (jou-, spiel-). In Der Sprach Brockhaus (see Brockhaus 81972), however, verbs are entered in the 1st person singular present tense form (e.g. “ich spiele”). In the case of Swahili, it is by no means clear what the citation form should be (Bwenge 1989: 5 f.). Nouns in Swahili, for example, whether singular or plural, always have a prefix that marks number, but which also acts as a gender/class marker: should the headword be the stem form, without prefixes, or the singular form? The former would make it difficult for a user of the dictionary to locate a word unless they knew which noun class it belonged to. The latter would group together in the dictionary all the nouns belonging to the same class, and it assumes that the user will then recognise the root and be able to deduce the plural form. What the citation forms should be in any particular instance is not always immediately obvious. The choice will be informed both by the insights of morphology and by the practical demands of the dictionary. This is why it is important, for example, for inflected forms, if they are not readily relatable to their stem, to be entered as headwords and crossreferenced to the citation form. Such a practice is especially necessary in bilingual dictionaries, where the users may not be familiar with the word that they are looking up (Landau 1989: 77). 3.2. Inflectional variants It is well-established that, as part of the idiosyncratic information about words that dictionaries contain, inflections are included, especially if they are irregular (Jackson 1985: 54). Idiosyncratic information is not predictable from the general rules of grammar; in the case of inflections, the declension of a noun or adjective or the conjugation of a verb is not predictable from the phonological form of a word (Bauer 1983: 192). The requirement on a dictionary is that it should be possible to construct the whole paradigm of a word from the information given in the entry (Zgusta 1971: 121). This is usually interpreted to mean that regular inflections need not be given, since the dictionary user is assumed to know them; Webster’s Third (Gove 1961, ed.) and Collins Cobuild (Sinclair 21995, ed.) are exceptions to this generalisation. For irregular inflections, enough information needs to be given for the complete paradigm to be deduced, as the principal parts of “strong” verbs in German allow the para-
XX. Morphologie in der Praxis
digms for all the tenses to be constructed. However, this is not always or not always obviously the case. In German dictionaries it is customary to give for nouns the genitive singular inflection and the plural inflection (e.g. Blick -(e)s, -e; Genosse -n, -n). Unless a user knows the way in which the singular paradigm for German nouns works, he will not be able to conclude that in the case of Blick the dative singular form is Blick, but in the case of Genosse it is Genossen. And if he knows that, he hardly needs the information that is given in the dictionary (Wiegand & Kucˇera 1982: 309 f.). What information does a dictionary user need to have about inflections? There are two different look-up questions in respect of inflections that a user might come to a dictionary with (Mugdan 1983: 180). The user may come with an inflected word form and want to find the corresponding headword or citation form (lemma). Alternatively, the user may come with a citation form and want to know what the inflected forms of the lemma are (cf. Bergenholtz & Mugdan 1982: 29 f.). To satisfy the first query, a dictionary needs to contain a grammatical description that includes inflectional paradigms, as well as entering irregularly inflected forms as headwords with cross-reference to the citation form (Mugdan 1983: 182 f.; cf. Cowie 1983: 101). To satisfy the second query, again the grammatical description needs to be included in the dictionary, and, if possible, an indication of irregular inflections needs to be given at the entry for the citation form (Mugdan 1983: 187 f.). There appears to be a strong argument for including a brief grammatical description in the dictionary, to which dictionary entries can be cross-referenced (cf. Bergenholtz & Mugdan 1982: 19). This is customary practice for the conjugations of verbs in French dictionaries (see 1.1.2), but because the complexities are too great to be able to handle in the entries themselves rather than for any principled reason. The inclusion of such a grammatical sketch in bilingual dictionaries may be a solution to the problem of where to indicate the inflectional information in these dictionaries (cf. 1.3; Mugdan 1983: 189). A particular problem ⫺ though it is probably not unique ⫺ is posed by the adjective inflection in English. First of all, there is an alternation in forming the comparative and superlative forms between an inflectional suffix (-er, -est) and more/most before the adjec-
171. Lexicography
tive (Jackson 1985: 54 f.). The inflectional suffix is usual with words of one syllable and with some two-syllable words; the periphrastic form is obligatory with three-syllable words and longer. Dictionaries need to indicate, at least for two-syllable words, where -er/-est is allowed and where not, as in the Cambridge International (Proctor 1995, ed.)(see 1.2). But the inflectional suffix may not be possible even for some one-syllable words, but not for any morphological reason (Bolinger 1990: 135): e.g. sourer or wronger do not “sound right”. Further linguistic investigation and guidance would greatly aid lexicographers in instances like these. 3.3. Derivatives Derivation, as a word formation process, is of great interest to dictionary compilers. They are keen to include in their works all new words that have become lexicalised (accepted into the language) since the previous edition of a dictionary. Because the number of derivatives can greatly enhance the calculation of the number of “references” included, they may be used to inflate the claimed size of a dictionary (Landau 1989: 78; Mugdan 1984: 241; Jackson 1998: 27). In their function as a record of the vocabulary of a language, dictionaries are naturally much concerned with word formation processes. Dictionaries need clear policies and practical guidance on several matters relating to the treatment of derivatives and compounds. One of the matters concerns the relationship between derivation and inflection, “a key point in the linguistic theory of word formation” (Ilson 1985: 163). The sharp boundary that is implied by the separate treatment of inflection and derivation in this article does not exist in reality (Zgusta 1971: 129); it is in many respects considered to be controversial (Lyons 1977: 521). The point can be illustrated from the present and past participles in a number of languages (Dubois & Dubois 1971: 63; Landau 1989: 91). The participles are usually considered to be inflectional forms of the verb; so they do not merit an entry in a dictionary unless they are formed irregularly. However, some participles are used more commonly than others as adjectives and, in this function, may develop a meaning that differs in some respect from the meaning of the verb to which the participle is related, e.g. French e´tonnant, English surprising. Such participles would warrant being entered as separate headwords, which implies
1889 a lexical derivation rather than a grammatical inflection. From the perspective of derivation, many derivational processes are highly regular and predictable, e.g. un- as a negative-forming prefix in English, or the -ish suffix added to adjectives, neither of which involve a change of word class of the item to which they are affixed. To that extent they are like inflections (Zgusta 1971: 127), and a dictionary of contemporary Polish did not intend to include regular derivations (Saloni et al. 1990: 7). However, derivational rules tend to be less productive than inflectional ones, and they sometimes lead to unpredictable semantic effects (Lyons 1977: 524). A lexicographer has to decide at what point to count an item as a different word and therefore deserving a separate entry in the dictionary. How then are derivatives treated in dictionaries? Prefixed derivatives, because of the alphabetical ordering of dictionary entries, are separated from the stem to which they are related, “so that the user loses the chance to grasp with a quick glance the combinatory potentialities of the given stem” (Kahane & Kahane 1967: 257). Suffixed derivatives are treated in two ways, and dictionaries differ in how individual derivatives are handled. The first way is to include a derivative in the article for its stem as a run-on entry, in bold type but usually with no, or minimal, definition. Derivatives whose meaning is directly related to that of the stem are normally handled in this way. The second way is to accord a derivative separate entry status as a headword, thus confirming the derivative as a different word. Derivatives whose meaning differs from that of the stem or whose use requires more extensive explanation are normally handled in this way (Landau 1989: 78). How a dictionary treats individual derivatives may depend on practical considerations of economy and retrievability and the weight given to these by the lexicographer (Cowie 1983: 102). The inconsistency in the treatment of prefixed as against suffixed derivatives and the implications this has for the view that a user has of derivational processes needs further attention from linguists and lexicographers. (Cf. the comments on the treatment of word formation in learners’ dictionaries by Prcic (1999). 3.4. Affixes and combining forms Should the elements of word formation processes, though they are not words themselves, be accorded headword status in the dictio-
1890 nary? The practice of doing this is variable (Bergenholtz & Mugdan 1982: 31), but their inclusion is seen as highly desirable (Benson et al. 1986: 236; Mugdan 1984: 300). The arguments for inclusion include the fact that they are meaning-bearing items (Müller 1982: 156), and that their inclusion may enable the dictionary user to decode new derivatives that are not yet lexicalised (Bergenholtz & Mugdan 1982: 32; Müller 1982: 167). For example, the inclusion of the combining form -ectomy will enable a user who knows the meaning of tonsil to decode the derivative tonsilectomy (Benson et al. 1986: 236). The information included for affixes and combining forms needs to indicate not only their meaning, but also how they form new words (Benson et al. 1986: 236, cf. Bauer 1983: 193). Arguably, this kind of explanation may be more satisfactorily handled in an appendix, where all word formation processes can be considered together, as in some French dictionaries (see 1.1.2) and learners’ dictionaries (see 1.2). To provide adequate points of access for users, it may be necessary to provide a three-fold treatment of derivatives: the inclusion of lexicalised derivatives as words; the inclusion of affixes and combining forms as headwords; and the treatment of word formation processes in an appendix. Lexicographers usually make a distinction between affixes and combining forms, though some dictionaries operate just with the labels “prefix” and “suffix”, or more unusually just with “combining form” (Proctor 1995, ed.). It is not certain that this is a distinction that all linguists would make (Ilson 1985: 164). There needs to be some clarification for lexicographers of the distinctions between inflectional affix, derivational affix, and combining form. They probably form a continuum from the obviously inflectional affix (like the English plural -s), through the inflectional affixes like English present participle -ing, through highly productive derivational affixes like the agentive -er, to the more obviously lexical re(“again”) and combining forms like -ology (“study of”) (Ilson 1985: 164). 3.5. Compounds In contrast to derivatives, one of the crucial issues in the treatment of compounds in dictionaries is deciding when a word ⫹ word or root ⫹ root combination constitutes a word or lexical item in its own right and therefore warrants entry in a dictionary. On the one hand this issue is connected with the general
XX. Morphologie in der Praxis
problem of the definition of a lexical unit, when this is not a simple word nor apparently a syntactic phrase (Dubois & Dubois 1971: 30). On the other hand, it is connected with the criteria that might be applied to decide that a combination should be regarded as a compound. Spelling is no guide in English (Cowie 1983: 103; cf. Mugdan 1984: 285 for German), though some dictionaries use that as a criterion, requiring a compound to be written “solid” or at least “hyphenated”, but not “open”. More acceptable and reliable criteria include the requirement that a compound should have a semantic unity or form a semantic unit (Dubois & Dubois 1971: 30) or have a “unity of signification” (Cowie 1983: 104), as well as the requirement that a compound should be stable (Zgusta 1971: 135) or institutionalised (Lyons 1977: 535). This last criterion excludes occasional or nonce compounds, which may be readily produced in some languages but are bound to the context in which they are coined. Only if they subsequently become generally accepted by use ⫺ institutionalised or lexicalised ⫺ do they merit inclusion in a dictionary; but the lexicographer is faced with deciding when that point has been reached (Lyons 1977: 536), and it may in any case be indeterminable (Mugdan 1984: 245). The other important issue for lexicography is how to treat compounds in the dictionary. Should they all be accorded headword status, or can they be included as run-ons under one or other of the constituent roots? From the user’s perspective, a strict alphabetical ordering of headwords is probably ideal (Mugdan 1984: 290). But economy usually dictates that the second course is followed as far as possible. In that case, are there any criteria for deciding which element of the compound it should be entered under? If, for example, it is considered desirable to indicate the relationship between a compound like fruit-picking and the construction from which it derives (e.g. “They picked fruit”), then this compound needs to be entered under the verb pick rather than the noun fruit (Cowie 1983: 99). But compounds are usually entered under the first lexical constituent. Not only may this be arbitrary, but it also implies a relationship between compound and the item (Lyons 1977: 537), when the compound may have developed a specialised meaning that is no longer relatable to either of its constituents; though that may be a reason for entering the compound as a separate headword. Another approach to this question is to take
171. Lexicography
the practical perspective of the dictionary user: where would a user expect to find a particular compound? It may be necessary to enter a compound in more than one place, or to take account of the strategies that users employ. It would appear that strategies differ from language to language (Bogaards 1990: 79 ff.); French speakers, for example, take more account of a constituent’s or word’s frequency, while Dutch speakers are more sensitive to a word’s class, tending to look for an item under a noun element before an adjective or verb element. 3.6. Conclusion That morphological information should be included in dictionaries is without dispute. Such information includes at least the inflectional properties of headwords (Lyons 1977: 517), and additionally derivational features (Saloni et al. 1990: 5). Another proposal specifies as the morphological information about lexical items: structure in terms of morphemes or alternating phonological structures; irregular morphological structures linked to particular morpho-syntactic features; partial similarities to other words (in derivatives and compounds); cliticizing properties (Hudson 1988: 311). These are the areas where linguists and lexicographers can fruitfully interact in assembling the information that needs to be included in a dictionary. More particularly lexicographical are the considerations of how to present this information, both to be true to the data and its description, and to take account of the practical needs of the dictionary user. How much of the information should be included in the dictionary entries? How extensive should the cross-referencing be? What is the place of grammatical descriptions in the front or back matter of the dictionary? Perhaps there is an argument for every dictionary to contain a brief grammatical sketch to which the entries can cross-refer (Bergenholtz & Mugdan 1982; Mugdan 1989⫺1990 b) or even for there to be a reference book that is dictionary and grammar combined (Lemmens & Wekker 1991).
4.
References
1891 Bauer, Laurie (1983), English Word Formation. Cambridge: Cambridge University Press Benson, Morton & Benson, Evelyn & Ilson, Robert (1986), Lexicographic Description of English. Amsterdam: John Benjamins Bergenholtz, Henning & Mugdan, Joachim (1982), “Grammatik im Wörterbuch ⫺ Probleme und Aufgaben”. In: Wiegand (ed.), 17⫺36 Bogaards, Paul (1990), “Ou` cherche-t-on dans le dictionnaire?”. International Journal of Lexicography 3.2, 79⫺102 Bolinger, Dwight (1990), “Review of Oxford Advanced Learner’s Dictionary, Fourth Edition”. International Journal of Lexicography 3.2, 133⫺145 Brockhaus (81972) ⫽ Der Sprach Brockhaus. Wiesbaden: F.A.Brockhaus Bwenge, C. (1989), “Lexicographical Treatment of Affixational Morphology: A Case Study of Four Swahili Dictionaries”. In: James, Gregory (ed.), Lexicographers and their Works. University of Exeter (Exeter Linguistic Studies 14), 5⫺17 Cowie, Anthony P. (1983), “On Specifying Grammatical Form and Function”. In: Hartmann, Reinhard R. K. (ed.), Lexicography: Principles and Practice. London: Academic Press, 99⫺107 Cowie, Anthony P. (41989, ed.), Oxford Advanced Learner’s Dictionary of Current English. Oxford: Oxford University Press Crowther, Jonathan (51995, ed.), Oxford Advanced Learner’s Dictionary of Current English. Oxford: Oxford University Press Drivaud, Marie-He´le`ne (1997, ed.), Le Robert Colle`ge. Paris: Dictionnaires Le Robert Dubois, Jean et al. (1971, eds.), Dictionnaire du Franc¸ais Contemporain. Paris: Librarie Larousse Dubois, Jean & Dubois, Claude (1971), Introduction a` la Lexicographie: Le Dictionnaire. Paris: Librarie Larousse Götz, Dieter & Haensch, Günther & Wellmann, Hans (1993, eds.), Langenscheidts Großwörterbuch Deutsch als Fremdsprache. Berlin: Langenscheidt Gove, Philip B. (1961, ed.), Webster’s Third New International Dictionary. Springfield, Mass: G. & C. Merriam Co. Guilbert, Louis & Lagane, Rene´ & Niobey, Georges (1971, eds.), Grand Larousse de la Langue Franc¸aise. Paris: Librarie Larousse
Allen, Robert E. (81990, ed.), The Concise Oxford Dictionary of Current English. Oxford: Clarendon Press
Hanks, Patrick (1979, ed.), Collins English Dictionary. London, Glasgow: Collins
Atkins, Beryl T. et al. (21987, eds.), Collins-Robert French-English English-French Dictionary. Glasgow: Collins
Hausmann, Franz Joseph & Reichmann, Oskar & Wiegand, Herbert Ernst & Zgusta, Ladislav (1989⫺1990, eds.), Wörterbücher, Dictionaries, Dic-
1892
XX. Morphologie in der Praxis
tionnaires ⫺ Ein internationales Handbuch zur Lexikographie. Berlin, New York: Walter de Gruyter
Müller, Wolfgang (1982), “Wortbildung und Lexikographie”. In: Wiegand (ed.), 153⫺188
Hoenigswald, Henry M. (1967), “Lexicography and Grammar”. In: Householder, Fred W. & Saporta, Sol (eds.), Problems in Lexicography. Bloomington/IN: Indiana University; The Hague: Mouton, 103⫺110
Pearsall, Judy (101999, ed.), The Concise Oxford Dictionary. Oxford: Oxford University Press
Hudson, Richard (1988), “The Linguistic Foundations for Lexical Research and Dictionary Design”. International Journal of Lexicography 1.4, 287⫺312 Ilson, Robert (1985), “The Linguistic Significance of Some Lexicographic Conventions”. Applied Linguistics 6.2, 162⫺172 Jackson, Howard (1985), “Grammar in the Dictionary”. In: Ilson, Robert (ed.), Dictionaries, Lexicography and Language Learning. Oxford: Pergamon Press (British Council ELT Documents 120), 53⫺59 Jackson, Howard (1988), Words and their Meaning. Harlow: Longman Jackson, Howard (1998), “How Many Words in YOUR Dictionary?”. English Today 55, Vol. 14.3, 27⫺28 Kahane, Henry & Kahane, Rene´e (1967), “Problems in Modern Greek Lexicography”. In: Householder, Fred W. & Saporta, Sol (eds.), Problems in Lexicography. Bloomington: Indiana University; The Hague: Mouton & Co. Landau, Sidney I. (1989), Dictionaries, The Art and Craft of Lexicography. Cambridge: Cambridge University Press Lemmens, Marcel & Wekker, Herman (1991), “On the Relationship between Lexis and Grammar in English Learners’ Dictionaries”. International Journal of Lexicography 4.1, 1⫺14 Lyons, John (1977), Semantics, Vol. II. Cambridge: Cambridge University Press Mugdan, Joachim (1983), “Grammatik im Wörterbuch: Flexion”. In: Wiegand, Herbert Ernst (ed.), Studien zur neuhochdeutschen Lexikographie III. Hildesheim etc.: Olms (Germanistische Linguistik 1⫺4.82), 179⫺237 Mugdan, Joachim (1984), “Grammatik im Wörterbuch: Wortbildung”. In: Wiegand, Herbert Ernst (ed.), Studien zur neuhochdeutschen Lexikographie IV. Hildesheim etc.: Olms (Germanistische Linguistik 1⫺3.83), 237⫺308
Prcic, Tvrtko (1999), “The Treatment of Affixes in the ‘Big Four’ EFL Dictionaries”. International Journal of Lexicography 12.4, 263⫺79 Proctor, Paul (1995, ed.), Cambridge International Dictionary of English. Cambridge: Cambridge University Press Rey, Alain (1973, ed.), Le Petit Robert: Dictionnaire Alphabe´tique et Analogique de la Langue Franc¸aise. Paris: Dictionnaires Le Robert Saloni, Zygmunt & Szpakowicz, Stanislaw & Swidzinski, Marek (1990), “The Design of a Universal Basic Dictionary of Contemporary Polish”. International Journal of Lexicography 3.1, 1⫺22 Schwarz, Catherine et al. (1988, eds.), Chambers English Dictionary. Edinburgh: W & R Chambers Sinclair, John (21995, ed.), Collins Cobuild English Language Dictionary. London & Glasgow: Collins [11987] Simpson, John A. & Weiner, Edmund S. C. (21989, eds.), Oxford English Dictionary. Oxford: Clarendon Press Summers, Della (31995, ed.) Longman Dictionary of Contemporary English. Harlow: Longman [21987; 11978] Terrell, Peter et al. (1980, eds.), Collins/Pons German Dictionary. London, Glasgow: Collins; Stuttgart: Ernst Klett Treffry, Diana (41998, ed.), Collins English Dictionary. Glasgow: HarperCollins Wahrig, Gerhard & Krämer, Hildegard & Zimmermann, Harald (1980⫺1984, eds.), Brockhaus Wahrig Deutsches Wörterbuch. Wiesbaden: F. A. Brockhaus; Stuttgart: Deutsche Verlags-Anstalt Wiegand, Herbert Ernst (1982, ed.), Studien zur neuhochdeutschen Lexikographie II. Hildesheim etc.: Olms (Germanistische Linguistik 3⫺6/80) Wiegand, Herbert Ernst & Kucˇera, Antonin (1982), “Brockhaus-Wahrig Deutsches Wörterbuch auf dem Prüfstand der praktischen Lexikologie”. In: Wiegand (ed.), 285⫺373
Mugdan, Joachim (1989⫺1990 a), “Information on Inflectional Morphology in the General Monolingual Dictionary”. In: Hausmann et al. (eds.), 518⫺525
Zgusta, Ladislav (1971), Manual of Lexicography. Prague: Academia
Mugdan, Joachim (1989⫺1990 b), “Grundzüge der Konzeption einer Wörterbuchgrammtaik”. In: Hausmann et al. (eds.), 732⫺749
Howard Jackson, Birmingham (Great Britain)
1893
172. Computational linguistics
172. Computational linguistics 1. 2. 3. 4. 5. 6.
Computational linguistics Models of morphology in computational linguistics Morphology in language technology Morphology learning Morphology tools for linguists References
1.
Computational linguistics
It is useful to distinguish in Computational Linguistics between applications and modules. Applications are geared toward a specific user-oriented goal (e.g., automatic translation, or dialogue with an information system), whereas modules are necessary in a wide range of applications. For example, syntactic parsing as a module is useful in the development of both Machine Translation systems and Dialogue Systems. Modules can be characterised by a transformation between different linguistic representation levels, e.g. from text to speech, or from a string of words to a tree structure representing the sentence’s structure. The goal of much work in Computational Linguistics is to design efficient and accurate computational models for such transformations. In developing a syntactic parser, for example, we may build a model using a grammar, a lexicon and a heuristic search procedure which together transform strings of words (the input representation) into labeled trees representing the syntactic structure (the output representation). In this article, I will treat morphology as one such module, which can operate in two directions. In morphological analysis, a complex word form is transformed into a string of morphemes with a characterisation of the structural relations between the different morphemes (possibly represented in a tree structure), and its stem or citation form. In
morphological synthesis (or generation), a stem or root form of a morphological paradigm and a set of grammatical features is taken as input, and the corresponding complex word form is generated. At a general level, all morphological processing modules will need a lexical database associating morphemes with linguistic information, a model of the combination possibilities of morphemes and the effects thereof on spelling and pronunciation (mostly in the form of rules, and a search procedure using this information to actually generate or analyze words. In this article, I will briefly discuss the kind of models that have been proposed for designing the morphological module in Computational Linguistics applications (cf. 2), go into the different application areas where morphological modules are used (cf. 3), and give an overview of tools which could help morphologists in their research (cf. 5). I will also provide pointers to some recent work on “learning of morphology” which seems to attract the attention of both linguistic and computational morphology (cf. 4). There are several good introductions to the field of Computational Linguistics in general (Jurafsky & Martin 2000; Manning & Schütze 1999; Allen 21995).
2.
Models of morphology in computational linguistics
As Computational Morphology focuses on developing models that achieve one of the morphological mappings discussed earlier (segmenting a string into its parts and disambiguating the parts in morphological analysis, construction of a string on the basis of lexical and morphological information in
morphological analysis surprisingly took
(((surprise)Verb ing)Adjective ly)Adverb (take)Verb-Past
morphological synthesis establish (past participle) clean (superlative)
established cleanest
Tab. 172.1: Morphological analysis vs. synthesis
1894 morphological synthesis), theoretical linguistic distinctions like inflection versus derivation (cf. Art. 38) do not play an important role. Formalisms developed for modeling morphology focus on the concrete construction processes involved (concatenation, Ablaut, root-and-template interleaving, suppletion, etc.; cf. Art. 53), on the spelling and phonological changes these processes produce (cf. Art. 35, 44), and on the constraints under which all this occurs (morphotactics; cf. Art. 42). Issues of productivity (cf. Art. 33) will be reflected in Computational Morphology by a particular choice of the division of labour between lexical storage and rulebased processing. We will only briefly describe the main types of models here. More technical and broader overviews of the field of Computational Morphology can be found elsewhere (Sproat 1992; 2000). 2.1. The role of lexical databases In any Computational Morphology model, there is a trade-off between the contents of the lexical database and the size of the rule set. In languages like English with a relatively poor morphology, it is feasible to construct a lexical database with complete paradigms of complex word forms rather than morphemes (a full form lexicon). The construction can be done semi-automatically (by morphological synthesis, which is easier than analysis), leading to an analysis by synthesis approach. All inflectional and most of the derivational processes can then be solved by lexical retrieval rather than computation. However, for most languages this is not a feasible solution. For example, compounding is extremely productive in languages like German and Dutch, and for languages like Turkish and Finnish, inflection and derivation cannot be solved by this “pre-computation” and lexical storage approach. Moreover, the approach fails for words (stems) not contained in the lexicon. For implementation, lexical databases tend to be represented as letter tries (Fredkin 1960). A trie is a tree data structure with nodes representing a position in a word, and arcs leaving a node representing letters that can follow the corresponding position in the word. The root node represents the start position, and individual words are paths in the trie. Leaf nodes represent the lexical information associated with the word the path of which ends at that node. This way the redundancy in the spelling of words is used to com-
XX. Morphologie in der Praxis
press the lexicon into a data structure that is small and that can be searched efficiently. Fig. 172.1 is an example of the trie structure representing the five words ask asking away be became. Note how the redundancy is removed. 2.2. Finite-state morphology The most influential model developed for morphological analysis and synthesis to date is finite-state morphology (FSM), also called two-level morphology. It was developed by Koskenniemi (Koskenniemi 1983; 1984) for Finnish and other languages, inspired by unpublished work by Martin Kay and Ron Kaplan, published only much later (Kaplan & Kay 1994). The model consists of a trie lexicon structure and a set of rules implemented as a finite-state transducer, and can perform both analysis and synthesis. The most important ideas in this approach were that all or most morphological phenomena can be described with regular expressions, and that the morphological mappings can be described with only two levels: a lexical representation, and a surface (spelling or speech) representation. Traditional linguistic descriptions like generative phonology (Chomsky & Halle 1968) made use of an ordered series of rewrite rules with intermediate representations to transform lexical representations into surface forms, where each rule works on the output of the previous rule, thereby giving rise to complex rule interactions. In addition, the expressive power of the formalism used to describe each rule (context-sensitive grammar) was obviously higher than needed, and in a computational perspective, this type of rule only works in one direction (from lexical to surface). A further advantage of finite-state approaches is that the computational machinery involved is language-independent, and not ad hoc for a specific language. In finite-state morphology, each rule is implemented as a bi-directional Finite-State Transducer linking the lexical and surface representation, making it useful for both analysis and synthesis, and all rules are applied in parallel. The base unit of a two-level rule or constraint is a pair of symbols, one from the alphabet of the lexical representation, one from the alphabet of the surface representation. E.g. the pair n : m (a lexical n corresponding to a surface m), or V : 0 (a lexical vowel deleted at the surface level). See Fig. 172.2 (a), which represents an assimila-
1895
172. Computational linguistics
Fig. 172.1: Example of a trie data structure containing the lexical items ask asking away be became
Fig. 172.2: Three representations of a finite-state morphology rule; two-level rule (a), transition network (b), and transition table (c)
tion rule: a lexical n has to be realised as a surface m when followed by a bilabial consonant (B ⫽ m, b, or p, the symbol ⫽ can refer to any symbol). Fig. 172.2 (b) and (c) show the corresponding Finite-State Transducers as a network (transitions show how states are allowed to change) and as a transition table. The automatic compilation of two-level rules into finite-state transducers is not trivial, but can be done. The strength of the approach is that two-level rules can refer to the lexical as well as to the surface representations for defining their context.
A finite-state morphology model consists of a set of rules of this type. In morphological synthesis, the rules get a series of lexical symbols as input and produce all surface forms allowed by the rules, in morphological analysis, they get a surface form as input, and produce all possible lexical representations (constrained by the lexicon). The lexicon in finitestate morphology consists of a list of stem forms, and a number of affix lexicons (lists of inflectional and derivational suffixes). Lexicon entries contain associated lexical information, including a list of pointers to contin-
1896 uation lexica (lexica which can follow this lexicon entry). The model has been applied to many languages and many morpho-phonological phenomena, and has proven to be an elegant and practical approach to Computational Morphology. Many overviews, tutorial material, extensions, and references are available (Gazdar 1985; Ritchie et. al. 1992; Antworth 1990). An extension for morphologically motivated phenomena like Umlaut is presented in Trost (1991). Despite its success, finitestate morphology is not without problems; the lexicon system leads to redundant representations, and some processes like reduplication and root-and-template morphology cannot be described without complicating the formalism. The success of finite-state approaches in Computational Morphology no doubt contributed to the investigation of the applicability of finite-state methods to syntactic analysis and in applications such as shallow semantic interpretation (Abney 1996; Karttunen et al. 1996; Roche & Schabes 1997, eds.). Together with statistical approaches, finitestate methods dominate Computational Linguistics today. 2.3. Hierarchical lexicons Another popular approach to computational morphology makes use of the concept of taxonomies and inheritance of properties to represent morphological knowledge. The basic insight here is that irregular words normally deviate only in a few characteristics from regular words. E.g., an irregular verb like run is just like a regular verb like work (compare run, runs, running to work, works, working), except that the past tense and past participle are formed in a different way. This “elsewhere condition” type of reasoning, so common in linguistic morphology, can be modeled elegantly with inheritance networks (Daelemans et al. 1992). Although there are many general-purpose knowledge representation and programming languages (of the frame-based or object-oriented type) that allow the implementation of this type of reasoning, the most important exponent of this approach is DATR (Evans & Gazdar 1996), a special-purpose programming language for lexical knowledge representation. DATR was guided in its design by formal adequacy goals (explicit declarative semantics and explicit theory of inference), and notational adequacy goals (it should be expressive enough
XX. Morphologie in der Praxis
to describe all relevant generalizations). Another concern was efficient implementation. DATR fragments have been developed for the morphology of various languages. Both approaches are complementary as the finite-state approach is limited in the mechanisms it allows for handling morphotactics, which is what most datr work focusses on. Not all models can be assigned to one of these approaches; one successful morphological analysis module in the context of speech synthesis, DECOMP (Allen et al. 1987), uses ad hoc weights to improve disambiguation. I will return to the problem of morphological ambiguity resolution in the section on morphology learning (cf. 4).
3.
Morphology in language technology
Morphological analysis is a basic component in many language technology applications. In Word Processing applications, for example, the accuracy of hyphenation, spelling checking, and grammar checking is highly dependent on the presence of some form of morphological analysis. Morphological processing can also be considered a core component in any language processing system, from speech synthesis over information retrieval to machine translation. Depending on the sophistication of the morphological processing involved, different terms have been used in language technology. The term stemming or suffix stripping is used to refer to a sort of “poor man’s” morphological analysis involving simple rules (often without a lexicon) to reduce an inflectional form of a word to its stem. A classical example of this approach, adapted to many languages is the Porter stemmer (Porter 1980). This approach is of course only feasible with some accuracy for languages like English which have a simple morphology. The term lemmatization refers to a more advanced process in which a complex word form is reduced to its lemma (or citation form) and the possible morphosyntactic classes it can have as retrieved from the lexicon and deduced from the rules. In this case the structure of the word is not further analyzed. So whereas full morphological analysis would analyze optimalizations as “((((optimal)Adjective -ize) Verb -ation)Noun -s)Noun-plural”, a lemmatizer would output “optimalizationNoun-plural”, depending on the contents of the lexicon, and the definition of the rules, of course.
172. Computational linguistics
3.1. Word processing Automatic hyphenation is the process of splitting words at the end of lines in order to minimize whitespace when right justification is used. Although the process is conventional, it is mostly based on linguistic units like syllables and morphemes. In languages like English where hyphenation is morpheme-based, morphological segmentation is required. However, given the poor morphology of English, a good computational solution is to use a set of splitting patterns found in a dictionary rather than full morphological analysis (Liang 1983). In a language like Dutch, however, where hyphenation is based on syllabification and morphological structure, the situation is much more complex. It is fairly easy to implement the basic syllabification rules for Dutch (based on the maximal onset principle and a language-specific rule avoiding syllables ending in a short vowel). It is sufficient to collect a list of possible syllable onsets. Interestingly, these rules are overridden by morphological rules. In compounds and some derivations, e.g. with the suffix -achtig (transl. -ly), the morphological boundary overrides the syllable boundary, giving rise to oppositions like groe-nig (groen ⫹ ig, ‘greeny’) versus groen-achtig (groen ⫹ achtig, ‘greeny’) where in the first case the maximal onset principle is preserved, and in the second case the morphological boundary has precedence over the syllable boundary in hyphenation. A fairly accurate morphological analysis is required to solve this problem in principle, in this case splitting patterns will necessarily be error-prone because of the high incidence of new compounds in Dutch, as can be witnessed every day in Dutch newspapers (Daelemans 1988). Spelling Checking is based on a very simple principle to detect errors. Given a list of words of the language, every word encountered in a text which does not belong to the list is a spelling error. In languages with a productive morphology, this leads to an annoying “overkill”; the software continuously flags correct words as errors. Especially compounding is a problem. Extremely productive in languages like German and Dutch, this morphological process is responsible for the creation of many complex words which are used ad hoc in a text, and never gain enough frequency to warrant their inclusion in the word list. Morphological analysis is the only way to solve this problem. Similar arguments hold for languages with an extensive deriva-
1897 tional or inflectional morphology like Finnish or Turkish, where storage of all word forms in a word list is impossible. Kukich (1992) describes the role of natural language processing in spelling correction. In Grammar Checkers, software that checks the grammatical correctness of sentences in a text, morphological analysis is essential to be able to detect agreement errors, e.g. between subject and verb in many languages, and between modifiers and nouns in languages like German and French. Processes very similar to morphological analysis are also necessary in word processing of languages like Chinese, and Japanese and Korean when written with Chinese characters, where word boundaries are not marked by spaces or other typographical means. A further good example is Vietnamese: it is written in Latin characters, but white space indicates syllable boundaries rather than word boundaries. 3.2. Module in larger systems Morphological analysis is a necessary component in complete Natural Language Processing systems (for speech recognition and synthesis, language understanding, language generation, language translation, information retrieval, etc.), mostly as a means to increase the lexical coverage of such a system. For example, a syntactic parser needs lexical information about every word in a sentence to be parsed. In case words are not in the lexicon, morphological analysis helps extract useful lexical information from these unknown word forms (often complex forms of known words), increasing the “virtual” coverage of the lexicon. In speech processing, morphological analysis is important as well. In Speech Recognition as a means to keep the recognizer lexicon small, in Speech Synthesis as a means to increase lexical coverage and solve ambiguous pronunciations like th in nothing versus anthill. With the availability of the WWW, Information Retrieval has become a ubiquitous technology. In search engines, keywords can be input to retrieve a number of documents containing them, ranked according to relevance according to statistical or heuristic measures. The reliability of an information retrieval engine is measured in terms of recall (how many of the documents relevant for my query did I get) and precision (how many of the documents returned by the system actu-
1898
XX. Morphologie in der Praxis
ally were relevant). Morphological processing is one way of increasing the recall of search engines. By expanding a keyword to all its morphologically related forms (by morphological synthesis) on the basis of its stem (found by morphological analysis), a wider range of (possibly relevant) documents is found, increasing recall.
4.
Morphology learning
The last decade, statistical approaches have started dominating the field of computational linguistics (Manning & Schütze 1999). The field has evolved from deductive to inductive. Recently, machine learning methods have been added to the tools of inductive computational linguistics. Machine Learning is a subfield of Artificial Intelligence concerned with the design of algorithms that learn from examples (Mitchell 1997). When used for building a linguistic model explaining some set of data, machine learning algorithms and linguists share the same task and purpose, which makes the approach potentially interesting for linguistics. Machine Learning algorithms can be supervised, in which case they get examples of available input and required output, or unsupervised, in which case they only get examples of available input, and have to figure out useful groupings or clusters of the data. The goal of a learning approach is to use the examples to find useful generalizations about the inputoutput mapping to be learned. In an example of supervised machine learning applied to morphological analysis (van den Bosch & Daelemans 1999), a morphological analysis system is induced based on a large set of examples of complex words and their corresponding morphological analysis from the CELEX lexical database (Baayen et al. 1993, eds.). It can reconstruct the analyses it used for training, and apply the same systematicity to previously unseen complex words with high accuracy. The approach has been applied to English, German, and Dutch, and shows that even complex spelling changes can be handled in this classification-based way. Daelemans et al. (1997) is an example of a similar approach for morphological synthesis. In this case, the output of the learning is a rule system for diminutive formation in Dutch which is very similar to linguistic solutions proposed for the problem. Supervised learning methods solve a problem which approaches like finite-state trans-
ducer leave basically unsolved: the disambiguation problem. Even in languages with a minimally complex morphology, a morphological analysis system can lead to many possible analyses (different segmentations into morphemes, different assignments of grammatical classes to morphemes), many of which are spurious. A supervised machine learning approach implicitly uses frequency information from the data it was trained on to make probabilistic disambiguation decisions. Especially in unsupervised learning of morphology, a lot of progress has been achieved the last few years. Starting from a list of complex words, several unsupervised learning techniques have been experimented with to automatically extract information like lists of stems and affixes, and of how they can be combined (Goldsmith 2001; Kazakov & Manandhar 2001; Yarowsky & Wicentowski 2000). An exciting new approach in between supervised and unsupervised learning would start from a bilingual corpus, semi-automatically aligned. Suppose a morphological analyzer exists for German and not for Dutch. Once a bilingual aligned corpus German ⫺ Dutch is collected, the analyses generated for German can be projected to the Dutch part of the corpus and used to induce a morphological analyzer for Dutch (Yarowsky & Ngai 2001).
5.
Morphology tools for linguists
Some of the approaches used in Computational Morphology are at the heart of what constitutes a linguistic approach to morphology (finding the right generalizations and morpheme inventories for describing the morphology of a language, and looking for formalisms which allow describing them in sufficient detail). It is therefore not surprising that many Computational Morphology researchers have tried to build reusable tools that will be of interest to linguists as well. In this section a (probably incomplete) overview of the most important ones is given (cf. also Art. 168). Finite-state morphology is a well-developed field which has spun off many useful tools. The Xerox Research Center in Europe (XRCE) has developed tools for morphological analysis in many languages, based on finite-state technology and is also active in research in this area. Evan Antworth of SIL
172. Computational linguistics
provides a useful two-level morphology software package called PC-KIMMO. Finite-state tools for inflectional morphological analysis and synthesis of English implemented using widely-available unix utilities are available from the University of Sussex (Minnen et al. 2001). Also many implementations for DATR have been developed. Dafydd Gibbon provides an implementation called Zdatr, and the developers of DATR (Roger Evans and Gerald Gazdar) maintain a web portal with tutorial information, references, and references to implementations. Simpler application-oriented tools like Porter stemmers are widely available as well. Many implementations of Machine Learning algorithms are available as well, but are generally not directly usable for linguistic research. Systematic evaluation is only just starting (Maxwell 2002). BOAS is a userfriendly environment for the development of morphological analyzers which makes use of Machine Learning (Oflazer et al. 2001). Interactive demonstrations of the supervised learning approach of van den Bosch & Daelemans (1999) are available as well via the demonstrations section of their website.
6.
References
Abney, Steven (1996), “Partial Parsing via FiniteState Cascades”. In: Carroll, John & Briscoe, Ted (eds.), Proceedings of the ESSLLI ‘96 Robust Parsing Workshop. Prague: ESSLLI, 8⫺15 Allen, John & Hunnicutt, M. Sharon & Klatt, Denis (1987), From Text to Speech: The MITalk System. Cambridge: Cambridge University Press Allen James (21995), Natural Language Understanding. Redwood City/CA: The Benjamin/Cummings Publishing Company Antworth, Evan L. (1990), PC-KIMMO: A TwoLevel Processor for Morphological Analysis. Dallas/ TX: Summer Institute of Linguistics (Occasional Publications in Academic Computing, 16) Baayen, R. H. & Piepenbrock, P. & Rijn, H. van (1993, eds.), The CELEX Lexical Database (CDROM). Philadelphia/PA: Linguistic Data Consortium, University of Pennsylvania
1899 Daelemans, Walter M. P. (1988), “Automatic Hyphenation: Linguistics versus Engineering”. In: Heyvaert, F. J. & Steurs, Frieda (eds.), Worlds behind Words. Leuven: Leuven University Press, 347⫺364 Daelemans, W. & Berck, P. & Gillis, S. (1997), “Data Mining as a Method for Linguistic Analysis: Dutch Diminutives”. Folia Linguistica XXXI.1⫺2, 57⫺75 Daelemans, Walter & Smedt, Koenraad De & Gazdar, Gerald (1992), “Inheritance in Natural Language Processing”. Computational Linguistics 18.2, 205⫺218 Evans, Roger & Gazdar, Gerald (1996), “DATR: A Language for Lexical Knowledge Representation”. Computational Linguistics 22.2, 167⫺216 Fredkin, E. (1960), “Trie Memory”. Communications of the ACM, 490⫺499 Gazdar, Gerald (1985), Finite State Morphology. Linguistics 23, 597⫺607 Goldsmith, John (2001), “Unsupervised Learning of the Morphology of a Natural Language”. Computational Linguistics 27.2, 153⫺198 Jurafsky, Daniel & Martin, James H. (2000), Speech and Language Processing: An Introduction to Natural Language Processing, Computational Linguistics, and Speech Recognition. New Jersey: Prentice Hall, Englewood Cliffs Kaplan, Ron & Kay, Martin (1994), “Regular Models of Phonological Rule Systems”. Computational Linguistics 20.3, 331⫺378 Karttunen, L. & Chanod, J. & Grefenstette, G. & Schiller, A. (1996), “Regular Expressions for Language Engineering”. Natural Language Engineering 2.4, 305⫺338 Kazakov, Dimitar & Manandhar, Suresh (2001), “Unsupervised Learning of Word Segmentation Rules with Genetic Algorithms and Inductive Logic Programming”. Machine Learning 43, 121⫺ 162 Koskenniemi, Kimmo (1983), “Two-Level Model for Morphological Analysis”. Proceedings of the 8th International Joint Conference on Artificial Intelligence. Los Alamos/CA: Morgan Kaufmann, 683⫺685
Bosch, A. van den & Daelemans, W. (1999), “Memory-Based Morphological Analysis”. Proceedings of the 37th Annual Meeting of the ACL. San Francisco/CA: Morgan Kaufmann, 285⫺292
Koskenniemi, Kimmo (1984), “A General Computational Model for Word-Form Recognition and Production”. Proceedings 10th International Conference on Computational Linguistics. Stanford/CA: ICCL
Chomsky, Noam & Halle, Morris (1968), The Sound Pattern of English. New York/NY: Harper & Row
Kukich, Karen (1992), “Techniques for Automatically Correcting Words in Text”. ACM Computing Surveys 24.4, 377⫺439
1900 Liang, Franklin Mark (1983), Word Hy-phen-a-tion by Com-put-er. Ph.D. dissertation, Stanford University Manning, Christopher D. & Schütze, Hinrich (1999), Foundations of Statistical Natural Language Processing. Cambridge/MA: MIT Press Maxwell, Mike (2002), “Resources for Morphology Learning and Evaluation”. In: Gonzalez Rodriguez, Manuel & Suarez Araujo, Carmen Paz (eds.), LREC 2002: Third International Conference on Language Resources and Evaluation, Vol. III. Paris: ELRA, 967⫺974 Minnen, Guido & Carroll, John & Pearce, David (2001), “Applied Morphological Processing of English”. Natural Language Engineering 7.3, 207⫺223 Mitchell, Tom M. (1997), Machine Learning. New York/NY: McGraw-Hill Oflazer, Kemal & Nirenburg, Sergei & McShan, Marjorie (2001), “Bootstrapping Morphological Analyzers by Combining Human Elicitation and Machine Learning”. Computational Linguistics 27.1, 59⫺86 Porter, M. F. (1980), “An Algorithm for Suffix Stripping”. Program 14.3, 130⫺137 Ritchie, Graeme D. & Russell, Graham J. & Black, Alan W. & Pulman, Stephen G. (1992), Computa-
XX. Morphologie in der Praxis tional Morphology: Practical Mechanisms for the English Lexicon. Cambridge/MA: MIT Press Roche, Emmanuel & Schabes, Yves (1997, eds.), Finite-State Language Processing. Cambridge/MA: MIT Press Sproat, Richard W. (1992), Morphology and Computation. Cambridge/MA: MIT Press Sproat, Richard (2000), Lexical Analysis. In: Dale, Robert & Moisl, Hermann & Somers, Harold (eds.), Handbook of Natural Language Processing. New York, Basel: Marcel Dekker, 37⫺57 Trost, Harald (1991), “X2morph: A Morphological Component Based on Augmented Two-Level Morphology. Proceedings of the 12th International Joint Conference on Artificial Intelligence. San Francisco/ CA: Morgan Kaufmann, 1024⫺1030 Yarowsky, D. & Ngai, G. (2001), “Inducing Multilingual pos Taggers and np Bracketers via Robust Projection across Aligned Corpora”. Proceedings of NAACL-2001. San Francisco/CA: Morgan Kaufmann, 200⫺207 Yarowsky, D. & Wicentowski, R. (2000), “Minimally Supervised Morphological Analysis by Multimodal Alignment”. Proceedings of ACL2000. San Francisco/CA: Morgan Kaufmann, 207⫺216
Walter Daelemans, Antwerpen (Belgium)
XXI. Morphologie und Nachbardisziplinen Morphology and related fields 173. Namenkunde 1. 2. 3. 4.
Eigennamen und ihre Funktionen Interne onymische Morphologie Externe onymische Morphologie Zitierte Literatur
1.
Eigennamen und ihre Funktionen
Eigennamen (auch Propria, Onyme) werden unter die Substantive subsumiert und erfüllen spezifische referentielle Funktionen. Im Gegensatz zu den Appellativen (Gattungsbezeichnungen) wie z. B. Mensch oder Stadt, die eine ganze Klasse von Gegenständen bezeichnen, referieren Eigennamen prototypischerweise nur auf ein einziges Denotat (Monoreferentialität), z. B. Goethe oder Frankfurt. Im Falle mehrfacher Referenz werden feste Erweiterungen vorgenommen: Bei Familiennamen kommt der Rufname hinzu (Fischer J Joschka Fischer), bzw. bei Rufnamen der Familienname, bei Toponymen weitere Spezifikationen (Frankfurt (am) Main vs. Frankfurt (an der) Oder). Während Appellative ihr semantisches Potential nutzen, um die Referenz zu leisten, kommt Eigennamen in der Regel keine Semantik zu (s. jedoch Christoph 1987). Diese idealtypische Generalisierung ist bei genauerem Hinsehen einzuschränken: So können Spitznamen durchaus ein motiviertes semantisches Potential besitzen (z. B. Schlaffi
für einen wenig dynamischen Menschen), doch müssen sie dies nicht (man kann theoretisch auch jeden anderen Gegenstand Schlaffi nennen). In manchen Sprachen wie z. B. dem Deutschen müssen Rufnamen den Sexus des Referenten anzeigen (Christian vs. Christiane/ Christine). Auch sog. Gattungseigennamen (Teutoburger Wald, Goethestraße) enthalten semantische Strukturen. Zwar können Eigennamen lexikalische Rest- oder Scheinstrukturen enthalten ⫺ was aus ihrer diachronen Abkunft aus Appellativen oder aus volksetymologischen Umformungen resultiert ⫺, doch unterstützen diese keineswegs die Identifikation: Ein Lehrer kann Fischer heißen und eine Großstadt Düsseldorf. Potentielle semantische Strukturen werden also beim Eigennamen neutralisiert. Doch kann es, gerade bei eher außergewöhnlichen Eigennamen, zu Resemantisierungen kommen, etwa wenn ein Anwalt Mörder heißt; doch geschieht dies seltener im umgekehrten Fall, also wenn ein Mörder den gewöhnlichen Namen Richter trägt. Auch in Wortspielen, Witzen und der Literatur werden transparente Strukturen (re)motiviert. In der Regel gilt jedoch: Eigennamen individualisieren, Appellative generalisieren und charakterisieren (s. Abb. 173.1). Der gestrichelte Pfeil deutet den potentiellen Restbestand lexikalisch-semantischer
Abb. 173.1: Die unterschiedlichen Referenzleistungen von Eigennamen und Appellativen
1902 Strukturen an, der jedoch unmotiviert bleibt und nicht zur Referenzleistung beiträgt (s. die unterbrochene gepunktete Linie). Die Identifikation des Objekts verläuft direkt vom Ausdruck zum Objekt (Direktreferenz; s. den durchgezogenen Pfeil). Damit können Eigennamen zwar über transparente, doch nicht über motivierte Strukturen verfügen (vom genuinen Gattungseigennamen abgesehen). Eigennamen benennen, identifizieren bzw. etikettieren ein Objekt. Ihre Funktion entfalten sie nicht auf der Basis semantischer Merkmale. Monoreferenz kann auch mit anderen Mitteln hergestellt werden: Zum einen durch definite Beschreibungen mithilfe appellativischer Ausdrücke (z. B. Deutschlands größtes Bankenzentrum für Frankfurt), wobei diese häufig onymische Bestandteile enthalten; zum anderen durch Indikatoren, d. h. Deiktika und Proformen, die jedoch auf einen Kobzw. Kontext verweisen und damit einen solchen zur Voraussetzung haben (diese Stadt hier für ‘Frankfurt’; er für ‘Goethe’) (Werner 1974). Eigennamen haben diesen Verfahren gegenüber den Vorteil der Eindeutigkeit, des kurzen Ausdrucks und der Ko(n)textunabhängigkeit. Daher leisten sich alle Sprachen den “Luxus” von Eigennamen; dabei erhalten nur relevante Objekte einen Namen, in aller Regel Personen, der menschliche Lebensraum und vom Menschen verfertigte Objekte. Gemäß der Beschaffenheit des Denotats unterscheidet man zwischen Anthroponymen (Rufnamen, Familiennamen, Pseudonymen, Spitznamen etc.), Toponymen (Städte-, Straßen-, Flur-, Gewässernamen etc.), Ergonymen (Produktnamen, Bücher-, Filmtitel etc.), Praxonymen (Kriege, Abkommen etc.) und Phänonymen (Wetterhochs und -tiefs, Taifune etc.). Dabei nimmt die Prototypik von Eigennamen nach hinten hin ab: Konkrete, distinkte Objekte wie Personen und Örtlichkeiten bilden die Objektgruppe, die am ehesten onymisch realisiert wird. Da man immer wieder mit neuen Personen, Örtlichkeiten etc. in Kontakt kommt, ist man wie bei keiner anderen Wortart lebenslang mit dem Erwerb neuer Einheiten befaßt: Eigennamen bilden kein geschlossenes Inventar. Auch wenn sie in der Regel nicht bzw. nur rudimentär in Wörterbüchern verzeichnet werden, so gehören Eigennamen dem Sprachsystem im Allgemeinen und dem Lexikon im Besonderen an. Auf der anderen Seite gehören Eigennamen zu den Wörtern, die man am ehesten wieder
XXI. Morphologie und Nachbardisziplinen
vergißt, vor allem dann, wenn man sie nicht häufig gebraucht. Die Onomastik hat sich wie kaum eine andere linguistische Disziplin bisher in fast ausschließlicher Weise mit diachronen Gesichtspunkten befaßt (Fleischer 1970). In der Regel geht es um Fragen der Etymologie (s. Art. 174), da die meisten Namen appellativischen Ursprungs sind, sich jedoch im Laufe der Zeit zu mehr oder weniger opaken Ausdrükken entwickelt haben. Daß es dabei auch zur Herausbildung interner morphologischer Strukturen kommen kann, ist erst in jüngster Zeit ins Blickfeld geraten. Prinzipiell gilt es, zwei Domänen onymischer Morphologie zu unterscheiden: interne Morphologie, die die Binnenstruktur des Onyms selbst betrifft bzw. die Markierung der Proprialität (s. 2), und externe Morphologie, die die spezifisch onymische Flexion und Wortbildung zum Gegenstand hat, z. B. die Pluralbildung von Eigennamen oder die Bildung von Adjektiven aus Eigennamen (s. 3).
2.
Interne onymische Morphologie
Da Eigennamen nur Benennungsfunktion haben, sollte erwartbar sein, daß interne Morphologie afunktional ist. Dennoch lassen sich verschiedene morphologische Strukturen erkennen, die nicht nur Relikte einstiger appellativischer Morphologie darstellen. Diese morphologischen Strukturen tragen jedoch nicht dazu bei, das Referenzobjekt zu identifizieren, sondern sie markieren die Proprialität des Wortes, also seinen Eigennamenstatus. Dabei handelt es sich gemäß Dressler (2000) um extragrammatische Morphologie. Die Erfassung interner morphologischer Strukturen ist bisher sowohl von der Onomastik als auch von der Morphologie vernachlässigt worden. 2.1. Der Eigenname im Spannungsfeld zwischen Motivierbarkeit, Transparenz und Opakheit Das “Dilemma” von Eigennamen besteht darin, daß sie zwar eine grundlegend andere Funktion als Appellative ausüben, ihre diachrone Quelle jedoch in aller Regel in der Appellativik besteht. Dies schließt nicht aus, daß es auch Phantasienamen gibt (z. B. Spitznamen, Pseudonyme, Tiernamen) oder arbiträre Zahlen- und/oder Buchstabenkombinationen (z. B. Autokennzeichen, Identitätsnummern), doch ist dies ein wegen seiner ho-
173. Namenkunde
hen kognitiven Kosten realiter selten verwendeter Namentyp: Eine arbiträre und womöglich lange Zahlen- oder Buchstabenfolge stößt schnell an die Grenzen der Memorierbarkeit. Der prototypische Eigenname rekrutiert sich in seiner Entstehungsphase aus motivierter appellativischer Lexik (Eichstätt, Homburg < (zer) hohen Burg, Becker); dabei kann es sich auch um Fremdlexik handeln (Köln < lat. Colonia Agrippinensis ‘Kolonie des Agrippa’). Um der Verwechslungsgefahr mit einem Appellativ zu entgehen, sollte der Eigenname idealerweise möglichst weit entfernt vom Appellativ sein. Dem kommen gerade opake Namen wie Köln, Tolsterglope, Golodkowski nahe. Doch haben solche Namen, vor allem wenn sie mit Länge gekoppelt sind, hohe kognitive Kosten: Ein Familienname wie Becker ist leichter memorierbar als Golodkowski. Ist es ein weitgehend opaker Name, so bildet man Merkhilfen, d. h. man schafft sich sekundär transparente Strukturen. Weder totale Opakheit noch potentielle Motivierbarkeit können als propriale Ideale betrachtet werden. Üblicherweise bewegen sich die Eigennamen zwischen diesen beiden Polen, was sich wie folgt skizzieren läßt: Man geht davon aus, daß zu germanischer Zeit die zweigliedrigen Rufnamen motiviert waren im Sinne sog. Heilswünsche, z. B. Theoderich bzw. Dietrich ‘im Volke mächtig’. Dabei herrschten semantisch, morphologisch, syllabisch und euphonisch gesteuerte Abfolgeprinzipien und Restriktionen (Seibicke 1985; Greule 1996). Eine solche motivierte Rufnamenvergabe findet sich heute z. B. in China. Viele Eigennamen enthalten volltransparente Strukturen, ohne daß Motivierbarkeit, ein sinnvoller Bezug zum Objekt, möglich wäre: Familiennamen wie Beckenbauer, Neumann, ebenso Städtenamen wie Mannheim, Würzburg bestehen aus bekannten lexikalischen Bausteinen, ohne damit eine Identifikation des Referenzobjekts vollziehen zu können. Das schwedische Familiennamensystem hat diese Strategie zum Prinzip erhoben: Namenkomposita wie Sjöberg ‘Seeberg’, Sjögren ‘Seeast’, Lindström ‘Lindenstrom’ etc. gehören nicht nur zu den häufigsten Namentypen, sondern sie werden bei dem in Schweden einfachen Familiennamenwechsel sogar explizit empfohlen (Nübling 1997). Partielle Transparenz besteht, wenn appellativische Bestandteile leichte formale Veränderungen erfahren: Schmitt/Schmidt/Schmitz
1903 zu Schmied, Eichstätt zu Eiche und Stadt/ Stätte. Hierzu gehören auch Abweichungen in der Akzentposition (Ebersfe´lde), Apokopen (Ulm), Erweiterungen (Schilfa, Oberhausen), Abweichungen in der Graphie () etc. Solche formal leicht dissoziierten und damit noch identifizierbaren Morphe bezeichnet man als Pseudomorphe (Nübling 2000: 130) oder Paläomorphe (Tournier 1985: 87 f.). Auch Semitransparenz bildet eine vielgenutzte Kompromißlösung zwischen motivierten und opaken Strukturen: Hier ist mindestens ein Glied opak und ein Glied transparent (meist das letzte): Klinsmann, Kullmann bzw. Hamburg, Homburg etc. Vor dem Endstadium totaler Opakheit sind solche Eigennamen anzusiedeln, die noch morphologische Strukturen erkennen lassen, doch deren Einzelmorphe opak sind. So interpretiert man das Toponym Tolsterglope als Kompositum (als Tolster- ⫹ -glope), ohne daß der Sprecher einen dieser Teile an bekannte Lexik oder Onymik anschließen könnte. Eigennamen sind morphologisch erstarrt. Morphologische Operationen, wie sie bei Appellativkomposita möglich sind, sind bei Onymen nicht mehr gegeben, wie etwa Koordinierbarkeit: *Morgen fahre ich erst nach Ham- und dann nach Homburg. Das Erstglied solcher zweigliedrigen Komposita steht in der Regel der Appellativik ferner und weist eine höhere morphologische Bandbreite auf, während das Letztglied eher an die Appellativik anschließt, einem begrenzteren Inventar angehört und eher Rückschlüsse auf den Eigennamentyp zuläßt. Ähnlich wie in der Wortbildung scheint es auch hier das Letztglied zu sein, das den kategorialen Status markiert. Häufig sind gerade bei deutschen Toponymen Erweiterungen zur Dreigliedrigkeit, wobei dann meist ein transparentes, oft sogar motiviertes Erstglied präponiert wird: Oberjosbach ⫺ Niederjosbach. Prinzipiell scheint im Deutschen die Peripherie eines morphologisch komplexen Toponyms eher der Appellativik nahezustehen, während das Zentrum (-jos-) opak ist. Auf einer Skala von (potentieller) Motivierbarkeit (Koch, Neustadt) bis hin zu totaler Opakheit (Stratz, Tolsterglope) sind es diese drei dazwischenliegenden Verfahren der vollen, der partiellen und der Semitransparenz, die am meisten genutzt werden. Hierdurch entgehen die Eigennamen potentieller
1904 Verwechslungsgefahr mit den Appellativen, nutzen jedoch gleichzeitig deren Bekanntheit, was der Memorierbarkeit zugute kommt. 2.2. Integration proprialen Materials Eine häufig genutzte Technik der Eigennamenbildung besteht in der Nutzung bereits bestehender Eigennamen: Toponyme verwenden anderweitiges toponymisches oder anthroponymisches Material; entsprechend nutzen auch Anthroponyme anderes anthroponymisches oder toponymisches Material: Rhein J Rheinhausen, Rheinfelden etc.; Ludwig J Ludwigsburg, Ludwigshafen; Peter J Petersen, Peters; Hessen J Hesse, Hess etc. (s. 3.2). Bisher wurde noch nicht hinreichend untersucht, in welche Richtung diese onymischen Wander- und Entlehnungsbewegungen verlaufen. 2.3. Spezifisch onymische Morphe Von besonderem Interesse sind spezifisch onymische (auch: propriale oder onomastische) Morphe. Diese haben keinen synchronen Bezug zur Appellativik. Indem sie ausschließlich Proprialität markieren, erfüllen sie extrem kategoriale Funktionen. Daher werden diese Verfahren der internen Morphologie zugerechnet; allerdings ergeben sich hier Übergänge zur externen onymischen Morphologie (zu den Techniken s. 3.3). In manchen Sprachen sind Rufnamen produktiv aus Appellativen ableitbar, d. h. sie bilden kein festes Inventar (Onomastikon), aus dem geschöpft wird. Im Kinyarwanda (s. Art. 141), einer Klassensprache, besteht ein Appellativ üblicherweise aus einem Präpräfix, einem Präfix und einem Stamm, während der Eigenname durch die Abwesenheit des Präpräfixes gekennzeichnet wird: umuhuungu bedeutet ‘Junge’, segmentierbar als umu-huungu; Muhuungu dagegen ist ein Eigenname; ebenso amabuye ‘Steine’ (a-ma-buye) vs. Mabuye als Eigenname. Eine umgekehrte Markierung findet sich in einer anderen Klassensprache, dem Zulu: Hier erhält der Eigenname ein sekundäres Präfix, u-, wobei der anlautende Vokal des ursprünglichen (appellativischen) Suffixes schwindet: intombi ‘Mädchen’ vs. uNtombi als Eigenname (Kuhn & Serzisko 1982). Onymische Affixe zeigen nicht nur Proprialität an, sondern sie können auch Auskunft über die Art des Eigennamens und damit ⫺ indirekt ⫺ des Referenzobjekts erteilen. Die polnischen Familiennamen machen besonders intensiven Gebrauch von onymi-
XXI. Morphologie und Nachbardisziplinen
schen Suffixen: Zu den häufigsten zählen -ska (für weibliche Träger) und -ski (für männliche Träger), die ursprünglich aus Beinamen von Adligen hervorgingen und deren Besitz anzeigten (z. B. Tarno´w J Tarnowska, Tarnowski). Damit hatten sie ein Toponym als Basis. Später breitete sich dieses Suffix auf Toponyme aus, die die Herkunft (nicht mehr den Besitz) der betreffenden Person bezeichneten, und schließlich entwickelte es sich zum häufigsten aller polnischen Familiennamensuffixe, indem es sich ab dem 16. Jh. auch mit Rufnamen und Appellativen verband, z. B. Wis´niewski aus wis´nia ‘Sauerkirsche’ (Szczepaniak 2002). Auch das Schwedische kultiviert mit -son ein Familiennamensuffix, dessen appellativische Herkunft zwar noch ersichtlich ist, doch das sich formal davon dissoziiert hat (keine Homophonie mehr mit dem Simplex son ‘Sohn’). Suffigiertes -son in Gustavsson bedeutete ursprünglich ‘Sohn von’ und ist als Onym längst auch auf Frauen beziehbar. Mit dieser Desemantisierung ist es zu einem Indikator für Familiennamen geworden. In jüngster Zeit wird es sogar produktiv gemacht, indem es auch in Verbindung ganz anderer Erstglieder auftreten kann (Lindson, Balson). Das Deutsche liefert gerade bei den Toponymen viele Beispiele für onymische Morphe: Bei Eigennamen auf -ach, -itz, -a, -ow [o:], -wang(en), -holt, in der Schweiz -ikon, kann es sich nur um Siedlungsnamen handeln (allenfalls um sekundäre Familiennamen). Das bekannteste toponymische Suffix dürfte -ing(en) sein (Tübingen, Gundelfingen, Freising), das sich aus einem Zugehörigkeitssuffix entwickelt hat und ursprünglich mit (männlichen) Rufnamen verband: Gundolf ⫹ -ingen (> Gundelfingen) bedeutete ursprünglich ‘die Leute/die Siedlung des Gundolf’. Daß onymische Suffixe morphologischen Status besitzen, erweist sich bei volksetymologischen Prozessen: Rellingen ist eine sekundäre Neubildung aus in Reinlage, Ording aus (in) Urden. Heida und Wiesa (aus Heide, Wiese) wurden sekundär neugebildet nach Toponymen wie Vechta (s. 2.4). Bei der Entlehnung und Integration von Toponymen aus fremdsprachlichem Material ins Deutsche sind auffällige Strukturbildungen beobachtet worden: Sowohl bei Entlehnungen aus dem (Gallo-) Romanischen am Südwestrand Deutschlands als auch aus dem Slawischen im Osten Deutschlands kommt es zu strukturellen Übereinstimmungen, besonders bzgl. des Letztglieds: So enden viele To-
1905
173. Namenkunde
ponyme trotz unterschiedlicher Etyma auf -atsch (z. B. Fillatsch im Westen, Trebatsch im Osten), -itsch, -an, -aun, -un etc. Dies läßt auf eine gewisse “Gestaltung des onymischen Systems des Deutschen” (Eichler 2001: 165) schließen. Auffällig ist auch, daß sich viele Ortsnamen trotz ihrer fremdsprachlichen Herkunft in zwei Glieder segmentieren lassen, d. h. daß sekundär pseudomorphologische Strukturen mit meist bekannten onymischen Morphen entstehen (z. B. -en und -in wie in Wilthen und Berlin).
die späteren (sekundären) morphologischen Zäsuren vorgeben. Wie das Beispiel Venusberg < Vennsberg zeigt, können auch sekundäre onymische Strukturen entstehen. Damit werden sämtliche Verfahren der Markierung von Proprialität durch volksetymologische Umformungen produziert. Dies läßt den Schluß zu, daß es sich um funktionale Strukturen handelt.
2.4. Sekundäre Herstellung von Transparenz und onymischer Morphologie: Volksetymologie In ihrer theoretischen Tragweite unterschätzt wurden bisher die gerade im onymischen Bereich ausgesprochen häufig auftretenden Prozesse sog. Volksetymologie. Unzutreffend ist der alternative Terminus der sekundären Motivierung. Wenn ⫺ meist opake ⫺ Eigennamen volksetymologisch umgeformt werden, so entstehen dabei keineswegs adäquatere Beschreibungen des Referenzobjekts, sondern es entstehen nur Strukturen, die an Bekanntes anschließen und ⫺ hierin dürfte der Sinn liegen ⫺ die Memorierbarkeit erleichtern. Beispiele aus der deutschen Toponymik sind Hiddensee < dän. Hiddens-ø, eigentlich ‘Hiddens-Insel’, Sauerland < Suderland, Venusberg < Vennsberg, Meerholz < Me¯rold etc., aus der Anthroponymik Augstein < Augustin, Wohlrabe < Wallraff etc. (Bach 1953; Leys 1966; Olschansky 1996; Vennemann 1999). In keinem der (beliebig zu vermehrenden) Beispiele entsteht erhöhte referenzsemantische Adäquatheit im Sinne motivierter Strukturen, sondern sie verharren auf der Stufe der formalen Transparenz. Koch (1963) geht so weit, gerade den Mangel an Sinn als bestes Indiz volksetymologischer Umbildung zu bewerten: “Eben das Widersinnige und Ungereimte eines ON. [Ortsnamens] ist das beste Kennzeichen der Volksetymologie. Deren Wesen muß also woanders liegen als beim Sinngeben oder Sinngebenwollen” (Koch 1963: 165). Auch er sieht die Funktion in mnemotechnischer Erleichterung. Zu diesem Komplex besteht noch großer Forschungsbedarf. Stichprobenhafte Untersuchungen zeigen, daß die opake Vorlage meist mehrere Silben umfaßt (was die Anforderungen an die Memorierbarkeit und damit den Bedarf an Herstellung vermehrter Transparenz erhöht), und daß es meist die Silbengrenzen sind, die
Eigennamen können Basis für unterschiedliche morphologische Prozesse sein. Solche Prozesse können (1) Flexionsformen ein und desselben Namens (Goethe-s Genitiv Singular) hervorbringen (s. 3.1) oder durch die Wortbildungstechniken der Derivation (Personenname Leverkus J Ortsname Leverkusen) und Komposition (Rhein-hausen) zu (2) Eigennamen einer andern Kategorie (tsch. Hrabal-ova´ movierter Familienname) und (3) Appellativen einer gleichen oder andern Wortart (Benz-in, boykott-ieren) führen (s. 3.2). Ein Sonderfall von (1) sind onymische Pluraliatantum (die Azoren), aus denen kein Singular zurückgebildet werden kann (*eine Azore). Lexikalische Techniken der Bedeutungs-Konversion wie bei Pegnitz (Flußname J Ortsname) zu (2) oder Manchester (Ortsname J Stoffbezeichnung) zu (3) werden ebenfalls berücksichtigt, auch wenn hier keine explizit morphologischen Mittel eingesetzt werden. Für die explizite deonymische Morphologie spielen solche Vorgänge eine Rolle, wenn die zu Appellativen umgedeuteten Eigennamen Basen morphologischer Prozesse sind: Personenname Krösus J Appellativ Krösus ‘reicher Mann’ (Singular) J Krösuss-e (Plural). Ob “Personengruppennamen” wirklich onymisch sind, ist umstritten (s. 1). Der propriale Status von Ethnonymen (Kamerun-er ‘Bewohner des Landes Kamerun’) ist fragwürdiger als der von Kollektivbezeichnungen für die ‘Mitglieder einer Familie’ ((die) Buddenbrook-s).
3.
Externe onymische Morphologie
3.1. Flexion Eigennamen teilen als Unterklasse der Substantive deren grammatische Kategorien. Invariant ist die Person (immer 3.), variant sind Numerus, Kasus und Genus. Plural und oblique Kasus werden in der Regel morphologisch overt symbolisiert. Genus ist inhärentes Merkmal, kann aber ausdrucksseitig z. B. in typischen Endungen reflektiert sein (Michael-
1906 a ‘feminin’ zu Michael; Claudi-us/Claudi-a ‘maskulin/feminin’). Morphosyntaktische Auswirkungen haben die namensubstantivischen grammatischen Kategorien insofern, als sie die Deklination ihrer Begleiter bestimmen (ewig-es Rom). Numerus als eine stark semantische Flexionskategorie zum Ausdruck natürlicher Ein- bzw. Mehrzahligkeit (s. Art. 100) kollidiert bei Eigennamen in pluralischer Form mit dem für Propria konstitutiven Faktum der Monoreferentialität. Dennoch bleibt ein Plural wie (die) Mann-s (zum Familiennamen Mann) ebenso Eigenname (‘Kollektiv, das von den zu dieser Familie Mann gehörenden Personen gebildet wird’) wie ein geographisches Namen-Pluraletantum die Azor-en. In Redeweisen wie die zwei Deutschlands (BRD, DDR) dagegen handelt es sich um appellativischen Gebrauch. Im Deutschen kommt Kasus von Eigennamen fast nur in Form des attributiven Genitivs vor (Goethe-s; Polen-s). Vom Verb regierte syntaktische Kasus sind anders als etwa im Polnischen (Go¢a˛b-owi/-a Dativ/Akkusativ Singular) archaisch (Goethe-n Dativ/ Akkusativ Singular). Es gibt die Tendenz, Eigennamen möglichst in ihrer Nennform und dadurch unflektiert zu gebrauchen, d. h. ⫺ den Kasus betreffend ⫺ den Nominativ auch da stehen zu lassen, wo syntaktisch ein anderer Kasus gefordert wäre: (in einem Bericht) des “Neuer Tag” statt des “Neuen Tags”. Nicht eindeutig der Kasus- oder Numerusflexion zuzuordnen sind wegen der Homonymie der Suffixe die Bildungen auf -s vom Typ Fischer-s Fritz ‘Fritz Fischer’. Die Paraphrasen ‘Fritz aus der Familie der Fischer-s’ oder umgangssprachlich ‘den Fischer-s ihr Fritz’ sprechen für pluralisches -s, die syntaktische Konstruktion und die Paraphrase ‘(Vater) Fischer-s Sohn Fritz’ dagegen für genitivisches -s. Möglicherweise ist aber dieses Suffix bei Eigennamen weder numeralisch noch kasuell, sondern konstituiert eine eigene (Misch-)Kategorie ‘kollektiv-genealogisch’ o. ä., wobei ‘kollektiv’ nur Affinität zu ‘pluralisch’ und ‘genealogisch’ ebenfalls nur Affinität zu ‘genitivisch’ (‘herkunftsmäßig’) aufweist. Eigennamen mit flexivischer Kennzeichnung weichen gegenüber der Flexion von Appellativen oft ab. Das wird besonders deutlich, wenn es zu Eigennamen homonyme Appellative gibt (von denen jene z. T. herkommen): vgl. proprial/appellativisch die Mann-s / die Männ-er; die Herz-en-s (mit phonotaktisch bedingtem -en- zwischen den Sibilan-
XXI. Morphologie und Nachbardisziplinen
ten) / die Herz-en; des Karl Rabe / des Rabe-n. Zusätzlich fällt die allomorphische Armut der Namenflexive gegenüber den appellativischen Flexiven auf: proprial die Mann(-s), Rabe(-s), Zahn(-s) vs. appellativisch die Männ-er, Rabe-n, Zähn-e. Das s-Suffix ist hier wie in andern Spezialbereichen des Wortschatzes, wo die formale Identität des Lexems möglichst unangetastet bleiben soll (etwa bei Fremdwörtern), das probate Mittel. Es erhöht die Silbenzahl nicht, führt nicht zu möglichen Hiatusproblemen zwischen Stammauslaut und Suffix, ist nicht mit konkomitanten Änderungen des Basismorphems verbunden (wie etwa -er mit Umlaut des Stammvokals bei Männ-er) usw. Entsprechenden Appellativen nachgebildete Plurale wie die Liebermänn-er sind spielerischer Natur. Daran, daß pluralische Formen wie Mann-s (kommen) oder genitivische Formen wie (Thomas) Mann-s (Werk) artikellos gebraucht werden können bzw. müssen und bei Artikelgebrauch ein nichtsymbolisierter Plural die Mann-Ø (kommen) möglich bzw. ein nichtsymbolisierter Genitiv des Thomas Mann-Ø zwingend ist, sieht man, daß im Deutschen bei Phrasen mit Namen ebenso “Monoflexion” (Admoni 41982) anzutreffen ist wie bei Phrasen aus Artikel und Adjektiv (d-er groß-e vs. (ein-Ø) groß-er). Genus (s. Art. 98) kann ausdrucksseitig mit (Pseudo-)Suffixen korrelieren, die zu bestimmten, vom Genus mitkonstituierten flexivischen Klassen passen, z. B. Ann-e ‘feminin’ (vgl. appellativisch Nicht-e, Kann-e ‘feminin’; s. 2 zur “internen” onymischen Morphologie). Genus ist ausdrucksseitig auch mit derivativischen Suffixen verbunden, die Genus determinieren wie bei Lott-chen/Fritzchen ‘neutrum’ (vgl. appellativisch Töchterchen/Söhn-chen, Lämp-chen/Stühl-chen ‘neutrum’) oder Sexus symbolisieren und mit dem passenden Genus einhergehen wie bei tsch. Hrabal-ova´, dt. archaisch (Luise) Müller-in, dt. Michael-a, alle ‘feminin’ (s. 3.2 zur “externen” onymischen Wortbildungsmorphologie). 3.2. Wortbildung Eigennamen unterliegen als Basen für Wortbildungsprozesse keinen prinzipiell anderen Beschränkungen als die appellativischen Substantive, d. h. sie können explizit und implizit abgeleitet sowie zu Kompositionen verwendet werden. Die Wortbildungsprodukte können wiederum Eigennamen sein oder Appellative
1907
173. Namenkunde
Derivation
Konversion
Komposition
Eigennamen
(Hermann) Bayreuth-er Personenname
Bayreuth Eigenname Synonym für die Wagner-Festspiele
Neu-Bayreuth Eigenname Synonym für die Nachkriegs-Festspiele
Appellative
(alle/ein) Bayreuth-er Einwohnerbezeichnung
(es gibt inzwischen viele) Bayreuth-s ‘Festspiele mit Merkmalen, die ursprünglich nur die Bayreuther Festspiele aufwiesen’
Bayreuth-Besucher ‘Besucher der Stadt oder der Festspiele’
Bayreuth-er (Studenten) bayreuth-ische (Lande) Relationsadjektive
Tab. 173.1: Der Ortsname Bayreuth als Basis für die Bildung von anderen Eigennamen und Appellativen
werden. Beispielbasis für die folgenden deonymischen Bildungen ist der Ortsname Bayreuth: Derivationen von Eigennamen aus Eigennamen dienen u. a. den Zwecken der Movierung (tsch. Hrabal-ova´), der Personennamenbildung aus Ortsnamen (Oppenheim-er), der Ortsnamenbildung aus Personennamen (Leverkus-en), der Namen-Diminution (Kosename Ton-i aus Anton; Änn-chen aus Ann-e). Die derivativischen Mittel zur deonymischen Bildung von neuen Eigennamen sind spezifisch proprial bei Hrabal-ova´, Ton-i, Leverkus-en, homomorph mit solchen zur Bildung von appellativischen Substantiven bei Oppenheim-er, Änn-chen (vgl. Städt-er, Töchterchen). Derivierte Eigennamen können zu Appellativen konvertiert werden (Hein-i ‘Klein Heinrich’ J ‘Mensch mit seltsamem Verhalten’). Ziel der meisten deonymischen Derivationen (s. Art. 89) sind Appellative. Eine große Gruppe machen hier die Einwohnerbezeichnungen aus. Im paradigmatischen Verhältnis der Ableitungen zu ihrer toponymischen Basis läßt sich auch deren mögliche morphologische Binnengliederung erkennen. (1) Ortsname Bayreuth Veron-a Damask-us Ländername Schweiz Israel Chin-a Marokk-o
J J J J
Einwohnerbezeichnung Bayreuth-er Veron-es-e Damasz-en-e(r)
J J J J J
Einwohnerbezeichnung Schweiz-er Israel-i(-s) Chin-es-e(-n) Marokk-an-er
Einwohnerbezeichnungen müssen nicht de(top)onymisch gebildet sein. Sie können auch
primär vorliegen und ihrerseits Basis für die Ableitung von Orts- und Ländernamen sein: Einwohnerbezeichnung [bei den] Franke-n (Dativ Plural) J Orts-/Ländername Frank-en (vgl. auch den Typ [der] Grieche-n Land J Griechen-land). Die hier angezeigten Ableitungsrichtungen entsprechen der diachronischen Entwicklung, sind aber auch synchronisch gerechtfertigt, wenn man die längere Form als Wortbildungsprodukt annimmt, die kürzere als Wortbildungsbasis. Eine rein auf Proportionalanalogie gegründete Morphologie sieht aber auch synchronisch rückbildende deonymische Ableitungen kürzerer Ethnonyme aus längeren Ländernamen vor: (2) Afghan-istan Usbek-istan
: :
Finn-land Lett-land
: :
Afghan-en ⫽ x x ⫽ Usbek-en Finn-en ⫽ x x ⫽ Lett-en
Eine große, kaum blockierte Gruppe deonymisch abgeleiteter Appellative sind die einwohnerbezeichnenden Substantive (Bayreuther Singular/Plural, Chin-es-e(n)) und die aus den Ländernamen abgeleiteten Adjektive (Bayreuth-er/bayreuth-isch, chin-es-isch). Die Beziehung zwischen adjektivischem und (letztem) ethnonymischem Suffix ist hochgradig regulär (-er, -e, -i, wird durch -isch ersetzt vice versa): s. zu obigen Einwohnerbezeichnungen die Adjektive bayreuth-isch, veron-esisch, damasz-en-isch, chin-es-isch, marokk-anisch, israel-isch (schweiz-er-isch ist eine der wenigen Ausnahmen). Die Basen für die Ableitung der Bewohnerbezeichnungen und der Adjektive aus Toponymen sind also in der Regel identisch und können intern noch ein-
1908 mal aus toponymischem Stamm und einer Endung bestehen: Der intern strukturierte Derivationsstamm chin-es (zu Chin-a) ist Basis für Chin-es-e und chin-es-isch (Fuhrhop 1998: 141 ff.). Allomorphie des toponymischen Stamms kommt vor (Damask- vs. damasz-). Eine andere Gruppe von deonymischen Derivaten bilden aus Personennamen abgeleitete appellativische Substantive wie Benz-in (zu Benz) oder Verben wie mendel-n (auf Basis der Grundform Mendel), röntg-en (auf Basis des Stamms Röntg von Röntg-en). Detoponymisch ist etwa finnland-is-ier-en. Konversionen (s. Art. 90) aus Eigennamen können intra-onymisch vor sich gehen, so wenn aus Personennamen Firmen- und Markennamen gebildet werden (Ford “), aus Ortsnamen Vereinsnamen (Liverpool), aus Flußnamen Ortsnamen (Pegnitz), und sie können zu Appellativen werden wie der Markenname Uhu“ für ‘Flüssigklebstoff’. Oft ist schwer zu entscheiden, ob ein Name direkt konvertiert wurde oder das Ergebnis einer Rückbildung aus Komposita mit diesem Namen ist: Diesel für (und aus?) appellativisch Diesel-motor, Diesel-kraftstoff. Auch Komposition (s. Art. 87) unter Beteiligung von Eigennamen kann entweder neue Eigennamen hervorbringen oder Appellative. Beispiele für Komposita aus mehreren Namen sind kopulative Bildungen wie Leuthäuser-Schnarrenberger (Familienname ⫹ Familienname), Castrop-Rauxel (Ortsname ⫹ Ortsname), Frankenpfalz (Ländername ⫹ Ländername; oberfränkisch-oberpfälzisches Grenzgebiet) und determinative Bildungen wie Karl-Rüdiger (in einer Reihe mit Wolf-/ Heinz-Rüdiger usw.) zur progressiven Differenzierung von Rüdiger oder wie Müller-Lüdenscheid (Personenname ⫹ Ortsname) oder Baden-Baden (Ortsname ⫹ Ländername) zur regressiven Differenzierung der vielen Personen namens Müller bzw. mehrerer Orte namens Baden. Beispiele für Komposition unter Beteiligung von Appellativen sind determinative Bildungen wie Ober-/Unterammergau (Adjektiv ⫹ Ortsname J Ortsname), Kleinhans (Adjektiv ⫹ Familienname J Familienname), Karl-s-ruhe (Personenname ⫹ Fuge ⫹ Substantiv J Ortsname) zur progressiven Differenzierung eines Eigennamens (Kleinhans) bzw. eines Appellativs durch einen Eigennamen (Karl-s-ruhe) oder Bildungen wie Davos-Dorf, Davos-Platz zur regressiven Differenzierung. Aus ursprünglich regressiv-determinierend erscheinenden Reihenbildungen
XXI. Morphologie und Nachbardisziplinen
wie Hans-dieter, Hans-veit, Hans-jürgen entwickelt sich eine Art Halbpräfix Hans-. Dreigliedrige weibliche Familiennamenkomposita wie Greiner-Petter-Memm sind kopulativ aus den unmittelbaren Konstituenten GreinerPetter und Memm aufgebaut, wobei GreinerPetter selber eine Struktur aufweist, in der Petter den im Thüringischen häufigen Familiennamen Greiner regressiv determiniert. Beispiele für Komposition unter Eigennamenbeteiligung, die Appellative hervorbringt, sind etwa Otto-Motor mit der Struktur Personenname ⫹ Substantiv oder Grüß-August mit der Struktur Verb ⫹ Personenname, wobei August schon vorher zum Appellativ konvertiert gewesen sein kann (Vorname J ‘unbeholfener Mensch’). 3.3. Morphologische Typen deonymischer Prozesse Im Prinzip unterscheiden sich deonymische Prozesse morphologisch nicht von deappellativischen. Hier wie dort sind davon die Stammposition (Mailand/Mailänd-er wie Land/ Länd-er) und die Suffixposition betroffen, an der es additive, modulatorische und ⫺ mit den selben Einschränkungen wie in der appellativischen Morphologie ⫺ subtraktive Prozesse gibt. Was den Sprachbau betrifft, sind die additiven und subtraktiven Prozesse dem agglutinierenden Typ und dem Grundform-Prinzip zuzuordnen, die modulatorischen dem alternierenden Typ und dem Stamm-Prinzip (Harnisch 2001). Die in Tab. 173.2 zusammengestellten Prozesse operieren mit morphologischen Einheiten, wenn auch z. B. das Jo- von Joseph allenfalls sekundär morphologisch interpretiert sein kann: s. die scherzhafte proportionalanalogische Reihe Sepp/ Jo-seph, Hans/Jo-hann, Kurt/x; x ⫽ Jo-kurt. Bei -seph/Sepp oder Urs-/Usch- (über Urschmit Sibilantenassimilation) liegt Stammallomorphie vor. Andere Prozesse an Eigennamen operieren mit rein lautlichen Mitteln. Die lautliche Entstellung der Urformen von Eigennamen ist ein automatischer Prozeß, der mit den semantischen Besonderheiten und den besonderen Gebrauchsbedingungen der Eigennamen zu tun hat (s. 1). Motiviert dagegen sind lautliche Manipulationen an Eigennamen, wie sie v. a. bei der Herstellung von Koseformen zu Personennamen vorgenommen werden. Neben den Vorteil, den gekürzte Formen bei häufigem Gebrauch im kommunikativen Nahbereich haben, tritt das Prinzip der laut-
1909
173. Namenkunde
additiv
modulatorisch
subtraktiv
Hans Hans-i
Ägypt-en Ägypt-er
Mark-us Marc
additiv-modulatorisch
additiv-subtraktiv
modulatorisch-subtraktiv
Chin-a Chin-es-e
Jo-seph Sepp-i
Urs-ul-a Usch-i
Tab. 173.2: Suffixmorphologische Prozeßtypologie
lichen Verringerung als Ikon der semantischen Verniedlichung. Das hindert nicht, daß morphologische Substanz, die von der Lautsymbolik her geeignet ist, ‘Kleinheit’ anzuzeigen, wie das -i, wieder hinzutritt: Joseph J Sepp J Sepp-i. Ideale Grundeinheit rein lautlicher Verkürzungen ohne Rücksicht auf eine eventuell vorliegende interne morphologische Struktur sind offene Silben, die entweder ohnehin vorliegen wie bei Fe.li.ci.tas J Fe.li (Silbentilgung), Lie.se.lo.tte J Li.lo (Wahl auseinanderliegender betonter Silben), E.li.sa.beth J Li.li (Reduplikation) oder erst hergestellt werden wie bei Hans-Joachim J *Hansjo J Hajo). Offensichtlich werden Zweisilbler mit vokalischem Auslaut angestrebt, s. auch Fe.li.ci.tas J Zi.ta, wo das s lautlich und nicht als *-s-Suffix getilgt wird. Wo ein typischer Namenausgang durch die Kürzung nicht schon angeboten wird wie bei Rü.di.ger J Rü.di, wird er hergestellt: Ru.dolf J Ru.di. Diese vokalischen Wortausgänge werden dann morphologisch reanalysiert und sekundär motiviert (s. 2.4), hier als typisch weibliches Vornamensuffix -a (Zita J Zit-a wie Paul-a zu Paul), als typisch diminuierendes Kosesuffix -i (Feli/Rudi J Fel-i/Rud-i wie Hans-i zu Hans), als untypisch weibliches oder typisch männliches, aber immer reihenbildendes Suffix -o (Lilo J Lil-o in der Reihe [Ca.ro.lin J] Caro J Car-o usw.; Hajo J Haj-o in der Reihe Brun-o, Ud-o, usw.). Daß solche (pseudo)suffixischen Muster wahrgenommen werden, zeigen Namenmoden wie die gleichzeitige Beliebtheit von Namen auf -ian (Christ-ian, Sebast-ian, Flor-ian usw.).
4.
Zitierte Literatur
Christoph, Ernst-Michael (1987), “Studien zur Semantik von Eigennamen”. Namenkundliche Informationen, Beiheft 10 Doleschal, Ursula & Thornton, Anna M. (2000, Hrsg.), Extragrammatical and Marginal Morphology. München: Lincom Dressler, Wolfgang U. (2000), “Extragrammatical vs. Marginal Morphology”. In: Doleschal & Thornton (Hrsg.), 1⫺10 Eichler, Ernst (2001), ‘‘Germania Romana und Germania Slavica im toponymischen Vergleich”. In: Bentzinger, Rudolf & Nübling, Damaris & Steffens, Rudolf (Hrsg.), Sprachgeschichte ⫺ Dialektologie ⫺ Onomastik ⫺ Volkskunde. Stuttgart: Steiner, 159⫺167 Eichler, Ernst & Hilty, Gerold & Löffler, Heinrich & Steger, Hugo & Zgusta, Ladislav (1995⫺ 1996, Hrsg.), Namenforschung: Ein internationales Handbuch zur Onomastik. Bd. 1⫺3. Berlin, New York: Walter de Gruyter Fleischer, Wolfgang (1970), “Onomastische Strukturen in der deutschen Sprache der Gegenwart”. In: Onomastica Slavogermanica V, Abhandlungen der Sächsischen Akademie der Wissenschaften zu Leipzig, Phil.-hist. Klasse 62.2, 35⫺44 Fuhrhop, Nanna (1998), Grenzfälle morphologischer Einheiten. Tübingen: Staufenburg Greule, Albrecht (1996), “Morphologie und Wortbildung der Vornamen: Germanisch”. In: Eichler et al. (Hrsg.), Bd. II, 1183⫺1187 Harnisch, Rüdiger (2001), Grundform- und StammPrinzip in der Substantivmorphologie des Deutschen: Synchronische und diachronische Untersuchung eines typologischen Parameters. Heidelberg: Winter
Admoni, Wladimir (41982), Der deutsche Sprachbau. München: Beck
Koch, Max (1963), “Volksetymologie und ihre Zusammenhänge”. Beiträge zur Namenforschung 14, 162⫺168
Bach, Adolf (1953), Deutsche Namenkunde. Heidelberg: Winter
Kuhn, Wilfried & Serzisko, Fritz (1982), “Eigennamen im Rahmen der Dimension der Apprehen-
1910
XXI. Morphologie und Nachbardisziplinen
sion”. In: Seiler, Hansjakob & Lehmann, Christian (Hrsg.), Apprehension. Teil 1: Bereich und Ordnung der Phänomene. Tübingen: Narr, 277⫺293 Leys, Odo (1966), “Der Eigenname in seinem formalen Verhältnis zum Appellativ”. Beiträge zur Namenforschung, Neue Folge 1, 113⫺123 Nübling, Damaris (1997), “Deutsch-schwedische Divergenzen in Entstehung und Struktur der Familiennamen: Ein Beitrag zur kontrastiven Onomastik”. Beiträge zur Namenforschung 32.2, 141⫺173 Nübling, Damaris (2000), “The Semiotic and Morphological Structure of German Toponyms: Different Strategies for Indicating Propriality”. In: Doleschal & Thornton (Hrsg.), 127⫺137 Olschansky, Heike (1996), Volksetymologie. Tübingen: Niemeyer Seibicke, Wilfried (1985), “Überblick über Geschichte und Typen der deutschen Personennamen”. In: Besch, Werner & Betten, Anne & Reichmann, Oskar & Sonderegger, Stefan (Hrsg.), Sprachgeschichte: Ein Handbuch zur Geschichte der deutschen Sprache und ihrer Erforschung, Bd. II. Berlin, New York: Walter de Gruyter, 2148⫺2163
Szczepaniak, Renata (2002), Onymische Suffixe als Signal der Proprialität: Das Polnische als Paradebeispiel [Vortrag, gehalten am 20.08.2002 beim 21. Internationalen Kongress für Namenforschung in Uppsala; wird veröffentlicht] Tournier, Jean (1985), Introduction descriptive a` la lexicoge´ne´tique de l’anglais contemporain. Paris: Champion-Slatkine Vennemann, Theo (1999), “Volksetymologie und Ortsnamenforschung: Begriffsbestimmung und Anwendung auf ausgewählte, überwiegend bayerische Toponyme”. Beiträge zur Namenforschung, Neue Folge 34, 269⫺322 Werner, Otmar (1974), “Wie kann man mit einem begrenzten Inventar über unbegrenzt viele Gegenstände sprechen?” In: Heilmann, Luigi (Hrsg.), Proceedings of the 11th International Congress of Linguistics 1972, Bd. II. Bologna: Il Mulino, 173⫺ 187
Rüdiger Harnisch, Passau (Deutschland) Damaris Nübling, Mainz (Deutschland)
174. Etymologie 1. 2. 3. 4. 5. 6. 7.
Was ist Etymologie? Etymologie und Wortbildungslehre Etymologie und Wortgeschichte Die Wortprägung Theoretische Voraussetzungen Erschließungsmethoden Zitierte Literatur
1.
Was ist Etymologie?
Die Etymologie ist der Teil der Sprachwissenschaft, in dem Entstehung (und Geschichte) der Wörter einer Sprache untersucht werden. Im strengen Sinn beschreibt die Etymologie nur die Entstehung der Wörter; doch verlangt eine verwertbare Beschreibung der Entstehung eine Ergänzung durch die Behandlung der darauffolgenden Wortgeschichte mindestens in dem Umfang, in dem der Zugang zu der Bildung des Wortes von der betrachteten Sprachstufe aus beeinträchtigt wird. Soll also etwa die Etymologie des neuhochdeutschen Wortes Kiefer (⫽ Nadelbaum) beschrieben werden, so ist einerseits auf die Bildung des althochdeutschen Wortes kien-foraha einzugehen: Es ist ein Kompositum aus foraha
‘Föhre’ und kien ‘Kienspan’ (‘die Föhre, aus der Kienspäne vor allem als Beleuchtungsmittel gewonnen werden’); dann müssen aber, um diese Erklärung der Bildung mit dem heutigen Wort zu verbinden, die besondere Lautentwicklung, die morphologische Verdunkelung und die Veränderung von der Benennung einer besonderen Art Föhre zu einem undurchsichtigen Pflanzennamen beschrieben werden, also die Wortgeschichte (oder zumindest ein Teil von ihr). In einem solchen Fall wird das Material zur Beurteilung der Etymologie, die Vorform, aus der Überlieferung der betreffenden Sprache selbst gewonnen, in anderen Fällen wird die Beurteilungsgrundlage durch den Vergleich von Wörtern gleichen Ursprungs in verwandten Sprachen (im Rahmen einer Wortgleichung) erst erschlossen. Während sich die traditionelle etymologische Wissenschaft fast ausschließlich um die Wortentstehung gekümmert hat, wurde vor allem in der romanistischen Etymologie nicht ohne Grund die Wichtigkeit der Geschichte des Wortes betont; denn die tiefgreifenden Entwicklungen der französischen Sprache lassen bei früh ent-
174. Etymologie
standenen Wörtern eine bloße HerkunftsEtymologie in der Tat als unbefriedigend erscheinen. Der Gegensatz zwischen den beiden Interessenschwerpunkten wurde dabei schlagwortartig ausgedrückt durch e´tymologie-origine und e´tymologie-histoire du mot (Baldinger 1959 [1977: 219]). Diese Umorientierung führte in der romanistischen Sprachwissenschaft dazu, daß nun umgekehrt die Wortentstehung vernachlässigt wurde (indem z. B. bei Erbwörtern mit lateinischer Vorform das französische Wort lediglich bis zu dieser zurückgeführt wurde und die Erklärung der Entstehung der lateinischen Etymologie überlassen blieb). Eine befriedigende Etymologie wird aber (in der oben geschilderten Weise) immer beide Aspekte miteinander verbinden müssen. Dabei ist jedoch zu berücksichtigen, daß zur Erklärung der Entstehung (also der Wortbildung) das Quellenmaterial ergiebig genug sein muß. Deshalb hat z. B. Kluge (1911 [1977: 106]) bloßen Wortgleichungen, die die Erschließung der Bildung nicht erlauben (weil keine Basis erkennbar ist), einen “etymologischen Gehalt” abgesprochen, sie also von der Etymologie im engeren Sinn ausgeschieden, selbst wenn die Zahl der an der Gleichung beteiligten Sprachen ein hohes Alter des betreffenden Wortes sichert. Es ist jedoch zweckmäßiger, auch in solchen Fällen von einer Etymologie zu sprechen, dabei aber zu vermerken, daß der Bildungsvorgang nicht erschlossen werden kann. In den neuesten Auflagen des etymologischen Wörterbuchs von Kluge (231995) wird dies mit dem Vermerk Herkunft dunkel getan. Allerdings können die Wörter einer Sprache auch auf andere Weise als durch Wortbildung entstehen. Solche Möglichkeiten, die für eine Behandlung im Rahmen der Morphologie nicht einschlägig sind, sind die Entlehnung, die Urschöpfung (z. B. Lautnachahmung), die Bedeutungs-Übertragung und -Verschiebung, das Erstarren syntaktischer Fügungen und die Wortkürzung. In der Regel wird auch die Entstehung unselbständiger Morpheme (etwa der Suffixe) der Etymologie zugerechnet, doch könnte dieser Bereich zweckmäßiger in einer historischen Wortbildungslehre (und einer historischen Flexionslehre) behandelt werden.
2.
Etymologie und Wortbildungslehre
Damit berührt sich ein Teil der Etymologie (nämlich die Etymologie derjenigen Wörter, die als in der betreffenden Sprache oder einer
1911 Vorstufe zu ihr gebildet erwiesen werden können) sehr eng mit der Wortbildungslehre (der verschiedenen Stadien dieser Sprache). Eine naheliegende, aber nicht ganz richtige Abgrenzung weist dabei die Etymologie der historischen, die Wortbildungslehre der synchronisch beschreibenden Sprachwissenschaft zu. Dieser Standpunkt verkennt, daß die Etymologie einen wichtigen synchronischen Aspekt in sich schließt, nämlich die Frage, wie die Sprecher einer Sprache neue Benennungen finden, und welche Möglichkeiten sie dabei auswählen. Zur Beantwortung dieser Frage bietet die Wortbildungslehre die allgemeinen Muster, die zur Gewinnung eines neuen Wortes herangezogen werden können (und dieser Aspekt ist zugleich auch ein etymologischer); doch sucht die Etymologie darüber hinaus die nicht nur mit dem Bildungstyp, sondern auch mit dem speziellen Wort verbundenen Fragen zu klären: die des Benennungsmotivs (s. 4.), die der Abgrenzung im Wortfeld oder gegenüber sonstigen Konkurrenten, die der Auswahl aus verschiedenen Bildungsmöglichkeiten und ähnlichem. Die Wortbildungslehre hebt das an dem Bildungsvorgang Typische heraus und erfaßt auf der semantischen Seite in der Regel nur die systematische Bedeutung (die durch die Bedeutung der Elemente und die Funktion der Bildungsregeln voll erschließbar ist). Die Etymologie hebt vor allem auf der Bedeutungsseite heraus, wie mit Hilfe dieser systematischen Bildung ein Zugriff auf das mit dem Wort speziell Gemeinte möglich ist. Deshalb ist es durchaus sinnvoll, nach der Etymologie eines ganz neuen Wortes zu fragen, z. B. nach der Etymologie von Freischwinger (womit Stühle bezeichnet werden, bei denen Sitzfläche und Lehne auf einem elastischen Gestell ruhen, weshalb sie schwingen können). In vielen Fällen ist nämlich auch bei neuen Wörtern die Etymologie nicht klar (während die Wortbildung je nach der Stellung, die man der Semantik in ihrem Rahmen einräumt, gegebenenfalls als eindeutig aufgefaßt werden kann). Solche Wörter sind etwa Geisterfahrer oder Radler(maß). So gesehen beschreibt die Etymologie synchronisch den Vorgang der Wortentstehung, zumindest bei den Wörtern, die gebildet werden, um etwas Bestimmtes zu bezeichnen (die Wortprägungen); voll systematisch gebildete und verwendete Wörter, wie etwa Abstrakta zu Verben und Adjektiven, sind an sich kein Gegenstand der Etymologie; ihre Bildung wird durch die Wortbildungslehre voll erklärt. Sie
1912
XXI. Morphologie und Nachbardisziplinen
können allerdings im Gebrauch Besonderheiten annehmen, die zu Lexikalisierungen führen (Entfaltung nach Seebold 1981: 56 f.); für diesen Vorgang ist dann wieder die Etymologie zuständig. Da auch die Prägung von Wörtern nicht notwendigerweise zum normalen Sprachgebrauch gehört und die meisten Sprecher neue Wörter als bereits gebildete vorfinden, sind für sie die Bildungsvorgänge auch ‘Geschichte’, deshalb ist von der Einstellung der Sprecher her gesehen die Einordnung der Etymologie in die historische Sprachwissenschaft gerechtfertigt.
3.
Etymologie und Wortgeschichte
Von der Etymologie zu verlangen ist, wie ausgeführt, so viel von der Wortgeschichte, daß Form und Bedeutung des Wortes zum Zeitpunkt der Bildung mit seiner Form und Bedeutung in der zu untersuchenden Sprache eindeutig in Zusammenhang gebracht werden können. Im Rahmen einer Wortgeschichte kann natürlich noch viel mehr relevant sein, etwa das Auftreten und die Verwendung des Wortes in bestimmten literarischen Bereichen, seine Häufigkeit, die Herausbildung bestimmter syntaktischer Verwendungen usw. Es ist also durchaus möglich, eine von der Etymologie unterscheidbare Wortgeschichte als Bestandteil der historischen Sprachwissenschaft zu betreiben. Allerdings sind von diesen wortgeschichtlichen Feststellungen die meisten in irgendeiner Weise auch für die Etymologie und ganz besonders für die Erschließung der Etymologie relevant, so daß es unzweckmäßig wäre, in zwei verschiedenen Bereichen der historischen Sprachwissenschaft fast das gleiche zu tun. In der Regel wird deshalb zwischen Wortgeschichte und Etymologie kein systematischer Unterschied gemacht. Mit dem Aspekt ‘Wortgeschichte’ gehört die Etymologie selbstverständlich voll in die historische Sprachwissenschaft.
4.
Die Wortprägung
Sucht ein Sprecher einer Sprache eine Benennung für eine neue Sache (oder eine neue Benennung für eine bereits benannte Sache), so hat er bereits zwei Vorgaben für den Bildungsvorgang: Er sucht erstens ein Wort, also eine Benennung für eine Klasse ähnlicher Gegenstände (Vorgänge, Eigenschaften usw.), nicht einen Eigennamen (für etwas Individuelles), nicht eine Typenbezeichnung
(die Austauschbarkeit der bezeichneten Einheiten voraussetzt) und nicht einen Fach-Terminus (der nur strikt nach Definition verwendet werden darf). Zweitens hat der Sprecher den ungefähren Bedeutungsumfang vorgegeben: Er will etwas Bestimmtes benennen. Das heißt nicht, daß die Bedeutung bereits festliegt ⫺ sie wird erst durch den Gebrauch (durch die Abgrenzung vom Gebrauch teilweise konkurrierender Wörter) im einzelnen bestimmt; aber eine ungefähre Vorgabe (sei es in Form von Prototypen oder in Form von Klassen) ist vorauszusetzen. Dieser Bedeutungsumfang wird nun bei der hier zu betrachtenden Art der Wortgewinnung erfaßt durch eine Wortbildung, die als solche zunächst eine systematische Bedeutung hat. Diese systematische Bedeutung kann den geplanten Bedeutungsumfang mehr oder weniger deutlich wiedergeben: In einem Fall wie Holzstuhl fallen beide mehr oder weniger zusammen: Ein Holzstuhl ist ein Stuhl aus Holz, und jeder Stuhl aus Holz kann Holzstuhl genannt werden. Bei Freischwinger dagegen ist lediglich eine charakteristische Eigenschaft genannt: daß er schwingt, wenn man sich hineinsetzt (und es ist durchaus nicht gesagt, daß nach der Absicht des Wortbildners jeder Stuhl, der schwingt, wenn man sich hineinsetzt, Freischwinger genannt werden soll, und schon gar nicht ist alles, was frei schwingt, z. B. ein Pendel oder eine Schaukel, ein Freischwinger; nur ein Stuhl, der frei schwingt, wird so genannt). Dieses zur Benennung herangezogene Merkmal ist das Benennungsmotiv (weniger genau sagt man auch Bezeichnungsmotiv). Natürlich hat auch Holzstuhl ein Benennungsmotiv (‘besteht aus Holz’), doch ist diese Feststellung hier trivial. Das Benennungsmotiv muß durch die systematische Bedeutung erfaßt werden, und dafür gibt es meist verschiedene Möglichkeiten. Bei Säugetier oder Säuger (als Benennung einer Tierklasse) ist das Benennungsmotiv ‘säugt seine Jungen’; das wird im Deutschen durch den Rückgriff auf das Verb säugen erfaßt, im Französischen und Englischen durch die Erwähnung der Zitzen (lat. mammae, deshalb frz. mammife`re, engl. mammal), im Russischen durch die Feststellung der Ernährung der Jungen mit Milch (russ. mlekopitajusˇcˇee) ⫺ das Benennungsmotiv ist gleich, die Benennungsmittel sind verschieden. Bei der Erschließung des Benennungsmotivs sind nun zumindest bei neueren Wörter (wie Geisterfahrer) die Benennungsmittel ohne weiteres zu erkennen (Geister ⫹ Fahrer), während die
1913
174. Etymologie
Präzisierung des Benennungsmotivs häufig Kenntnis der benannten Sache wie auch einen Zugriff durch historische Belege erfordert. Was die Kenntnis der bezeichneten Sache anbelangt, so wird bei der Wortprägung in der Regel nach dem auffälligsten oder am charakteristischsten erscheinenden Merkmal benannt, wobei vor allem die Gebrauchsmöglichkeit eine große Rolle spielt (die Preiselbeere z. B. wird regional als Steinbeere bezeichnet, weil sie als Heilmittel gegen Steinleiden (Gallensteine, Nierensteine) galt; ohne Kenntnis dieses Sachverhalts ist das Benennungsmotiv nicht erschließbar). Das bedeutet gleichzeitig, daß das Erfassen des Benennungsmotivs beim Erwerb des Wortes zugleich etwas über die bezeichnete Sache verrät, und das ist einer der Gründe des starken Interesses auch von Nicht-Sprachwissenschaftlern an der Etymologie. Während sich nun der Gebrauch des Wortes nach den notwendigen Abgrenzungen auf den ursprünglich vorgegebenen Bedeutungsumfang einpendelt (Gebrauchsbedeutung), hat die systematische Bedeutung der Bildung (Bildungsbedeutung) beim Gebrauch an sich keine Funktion mehr; sie kann vernachlässigt werden (obwohl das die Sprecher in der Regel nicht tun) und wird bei zunehmender Lexikalisierung des Wortes mehr oder weniger undurchsichtig (die Bildungsbedeutung und damit das Benennungsmotiv etwa von Junggeselle ist einem nicht historisch geschulten Sprecher des Deutschen nicht mehr zugänglich; s. Art. 150). Solange sie den Sprechern noch durchschaubar ist, kann sie aber den Gebrauch steuern. Im Normalfall ist diese Steuerung (Verdeutlichung nach Seebold 1981: 223⫺230) nicht zu erkennen, weil sie naheliegende Folgen hat; erkennbar wird sie erst, wenn die Steuerung infolge falscher Anschlüsse durch die Sprecher einen Verlauf nimmt, der durch die Herkunft des Wortes nicht gerechtfertigt ist. Dies ist besonders häufig beim falschen Anschluß von Fremdwörtern der Fall (z. B. irritieren, ursprünglich ‘reizen, erregen’, durch falschen Anschluß an irre in Richtung auf ‘unsicher machen, beunruhigen, irre machen’ gesteuert). Das Aktivieren der Bildungsbedeutung bei den Sprechern (das für den normalen Sprachgebrauch nicht notwendig ist) kann eine Reihe von Funktionen haben: Es erleichtert das Verständnis des Wortes, unterstützt die Organisation des Wortschatzes beim einzelnen Sprecher, hält aktive Muster für weitere Wortbildungen bereit, ermöglicht Wortspiele usw. Die Verfüg-
barkeit über die Bildungsbedeutung wird häufig “Motivation/Motivierung/Motiviertheit” (s. Art. 150) oder auch “Durchsichtigkeit” genannt, doch sind diese Begriffe nicht sehr präzise gefaßt: Sie beziehen sich in erster Linie darauf, wie nahe die Bedeutung eines Wortes der vorauszusetzenden systematischen Bedeutung steht, wobei aber nicht unterschieden wird, ob die Zurückdrängung der systematischen Bedeutung auf nachträglicher Lexikalisierung oder auf einem schwer oder gar nicht erkennbaren Benennungsmotiv beruht.
5.
Theoretische Voraussetzungen
Auf die für die synchronische oder diachronische Betrachtung von Sprache allgemein vorauszusetzenden Theorien soll hier nicht eingegangen werden, doch ist eine für die Etymologie speziell wichtige Konzeption besonders zu erwähnen: Die etymologische Untersuchung geht immer von einem Wort einer bestimmten Sprache aus, das etymologisiert werden soll; und im Falle eines nach Wortbildungsregeln gebildeten Wortes ist das Ergebnis der Untersuchung ein Bildungsvorgang, der unter Umständen zeitlich sehr weit zurückliegt, und dessen unmittelbares Ergebnis unter Umständen sowohl lautlich wie auch semantisch von dem Wort der zu untersuchenden Sprache stark verschieden sein kann. Unter diesen Umständen ist es die Frage, ob das Wort der zu untersuchenden Sprache und die gefundene oder erschlossene Bildung der früheren Sprachstufe als das gleiche Wort aufgefaßt werden können (genauer: ob und inwieweit das Wort der späteren Sprache als Nachfolger des Wortes der früheren Sprache aufgefaßt werden kann). Von den Erfordernissen der Etymologie her gesehen können zwei Wörter dann miteinander identifiziert werden, wenn ihre jeweilige Überlieferung auf denselben Bildungsvorgang oder dieselbe Lexikalisierung einer ursprünglich systematischen Bildung zurückführt (Herkunftsgleichheit). Eine gewisse Beeinträchtigung der Gleichheit, die aber durch die Beschreibung ausgeglichen werden kann, besteht bei Bedeutungs-Übertragungen und -Verschiebungen, bei Beeinflussungen durch andere Wörter (Attraktion) und ähnlichem. Keine Gleichheit im etymologischen Sinn können dagegen parallele Bildungen zu verschiedenen Zeiten und/oder an verschiedenen Orten beanspruchen, selbst wenn sie das glei-
1914
XXI. Morphologie und Nachbardisziplinen
che Sprachmaterial auf gleiche Weise verwenden (materielle Gleichheit). Andernfalls würde materielle Gleichheit von Wörtern in verwandten Sprachen eine Rekonstruktion in der gemeinsamen Grundsprache erfordern, was in solchen Fällen zu Ungereimtheiten führt. Dieselbe Schwierigkeit liegt vor, wenn in der Überlieferung dieser Wörter Entlehnungen (s. Art. 152) aufgetreten sind (Verbreitungsgleichheit). So haben etwa die nordgermanischen Sprachen materiell gleiche Wörter für die Briefmarke (dän. frimærke, schwed. frimärke, norw. frimerke, isl. frı´merki), die selbstverständlich auf sekundärer Verbreitung beruhen und nicht die Erschließung eines urnordischen *frija-mark-ja- (n) ‘Briefmarke’ erlauben, denn Briefmarken gibt es auch in den nordischen Ländern erst seit der Mitte des 19. Jahrhunderts. Ein anderer Fall, bei dem die oberflächliche Gleichheit bei der Etymologie aufzugeben ist, ist die Homonymie, besonders wenn eine zusätzliche Attraktion der Bedeutungen eingetreten ist. So ist etwa das Wort Spieß in der Bedeutung ‘Jagdspieß’ etymologisch ein anderes Wort als in der Bedeutung ‘Bratspieß’ (mhd. spiez und spiz). Ähnlichkeit in Form und Bedeutung, die für den Sprachgebrauch durchaus eine Identifizierung als gleiches Wort zulassen (Gebrauchsgleichheit), sind für die Etymologie erst auf ihre historische Haltbarkeit hin zu prüfen. Selbstverständlich ist die Bestimmung als herkunftsgleich oder materiell gleich nicht (oder zumindest nicht immer) unmittelbar aus dem Befund ablesbar, sondern Ergebnis der etymologischen Analyse. (Ausführlicher zu solchen Fragen: Seebold 1981: 26⫺35).
6.
Erschließungsmethoden
Für die Erschließung und Beschreibung der Etymologie sind im Prinzip dieselben Methoden verwendbar wie für die Wortbildungslehre, die (Wort-) Semantik, die Lautgeschichte, die Morphologie-Geschichte usw., sofern den besonderen Erfordernissen der Etymologie (Benennungsmotive usw.) entsprochen wird. Besonderheiten ergeben sich, wenn für das betreffende Wort eine Basis nicht oder nicht unmittelbar erschlossen werden kann. Terminologisch wäre die Möglichkeit der Erschließung einer Basis als Anschluß zu bezeichnen. Liegen verwandte Wörter vor, die auf eine entsprechende Basis zurückführbar sind, die Basis selbst aber nicht, so ist
dies eine (etymologische) Verknüpfung (man spricht auch von Wurzeletymologie, doch ist dieser Ausdruck mehrdeutig und etwas irreführend). Am ungünstigsten ist die Beleglage, wenn nur Wörter der gleichen Struktur verglichen werden können, aber weder eine Basis noch aus der gleichen Basis gebildete andersartige Wörter vorliegen (reine Wortgleichung). In diesem Fall sind in morphologisch einigermaßen durchschaubaren Fällen etymologische Hypothesen möglich, die aber in der Regel nicht ausreichend zu sichern sind (eine Sicherung wäre z. B. durch gewichtige Parallelen möglich). Eine Bemerkung ist noch am Platz zu der Analyse in [Wurzel ⫹ Affixe] gegenüber der Analyse in [Basis gegenüber Bildung]. In den modernen Sprachen ist es angemessen, bei der Bildung eines Wortes seine Basis, sein Grundwort, aufzuweisen und dann den Bildungsvorgang zu beschreiben. In der älteren Etymologie wurde statt dessen vielfach mit dem Konzept der Wurzel operiert, d. h. einem abstrakten Morphem, das allen Bildungen der betreffenden Wortfamilie gemeinsam ist und eine entsprechend abstrakte Bedeutung hat. An diese Wurzel konnten Primärsuffixe treten, um Stämme zu bilden, an diese wiederum Sekundärsuffixe, die SekundärStämme hervorriefen. Als Wort konnten sowohl die Wurzel (Wurzelverb, Wurzelnomen) wie auch ein Stamm verwendet werden. Diese Analyse geht auf die indischen Grammatiker zurück, die auf diese Weise das Sanskrit und das Vedische grammatisch beschrieben (vgl. Art. 5). Nun wird eine solche Analyse durch die Struktur des ältesten Indischen tatsächlich nahegelegt, weil diese Sprache z. B. bei den primären Verben (die bei weitem den Hauptbestand der altindischen Verben ausmachen) kein festes Paradigma kennt, sondern nach Bedarf Verbalstämme, wie Präsens, Perfekt, Aorist usw. und auch nominale Ableitungen bildet, so daß eine Grundform (als die wir im Deutschen etwa das Präsens oder den Infinitiv ansehen) nicht zu existieren schien und dadurch das abstrakte Konzept der Wurzel nahegelegt wurde. Diese AnalyseArt ist aber schon bei dem mit dem Altindischen nahe verwandten Altgriechischen nicht mehr zweckmäßig, weil dort die Paradigmen wesentlich stärker ausgebildet sind, und bei den modernen Sprachen, wie Deutsch, Englisch oder Französisch, sind die Paradigmen so fest, daß keinerlei Anlaß besteht, mit abstrakten Wurzeln statt mit konkreten Wörtern zu operieren. Die Gefahr bei Wurzel-
1915
175. Schriftsysteme
Analysen besteht hauptsächlich darin, daß sie durch ihren abstrakten Charakter leichter an irgendwelche hypothetischen Annahmen anzupassen sind, während konkrete Wörter mit erheblich größerer Präzision (besonders im Bereich der Semantik) bei der Analyse eingesetzt werden können. Die etymologische Forschung hat viel mit frühen, ja, wo dies möglich ist, sogar mit erschlossenen Sprachzuständen zu tun und setzt deshalb eine umfassende Kenntnis der Sprachgeschichte und der Verwandtschaftsund Entlehnungsverhältnisse der betrachteten Sprache voraus. Für die früheren Sprachstadien und ganz besonders für die erschlossenen Sprachstufen ist sie die wichtigste und vielfach einzige Quelle der Erschließung der Wortbildungslehre (und anderer Teile der Sprachbeschreibung). Dies ist nicht nur wichtig für die Beschreibung von unproduktiv gewordenen Affixen und dergleichen, sondern auch für Wortbildungsmittel, die in den lebenden Sprachen der betreffenden Sprachfamilie gar nicht mehr faßbar sind, so etwa für das Deutsche die Vriddhi-Bildung (ein Bildungstyp für Zugehörigkeitsbildungen, der durch Dehnung oder sonstige Verstärkung des Vokals der ersten Silbe charakterisiert ist; Beispiele wären Huhn zu Hahn oder Schwager zu Schwieger), Konsonanten-Geminierung bei emphatischen Bildungen, Wortbildung durch Akzentversetzung und ähnliches. Auf der anderen Seite bietet die präzise Erfassung eines Wortbildungstyps im Rahmen der Wortbildungslehre der Etymologie nicht
nur ein Mittel der Untersuchung und Beschreibung, sondern auch der Bewertung: eine Bildung, deren Merkmale nicht denen ihres Bildungstyps folgen, ist in besonderem Maße erklärungsbedürftig und möglicherweise falsch eingeordnet. Auf diese Weise sind Etymologie und Wortbildungslehre (auch allgemeiner: und Morphologie) gegenseitig voneinander abhängig und können sich gegenseitig erhellen.
7.
Zitierte Literatur
Baldinger, Kurt (1959), “L’e´tymologie hier et aujourdh’hui”. Cahiers de l’Association Internationale des E´tudes Franc¸aises 11, 233⫺264 [Nachdruck in: Schmitt (1977, Hrsg.), 213⫺246] Kluge, Friedrich (1911), “Aufgabe und Methode der etymologischen Forschung”. Neue Jahrbücher für das Klassische Altertum, Geschichte und deutsche Literatur 14, 365⫺376 [Nachdruck in: Schmitt (1977, Hrsg.), 103⫺119] Kluge (231995) ⫽ KLUGE: Etymologisches Wörterbuch der deutschen Sprache. Berlin, New York: de Gruyter [231995, Nachdruck 1999, 23. erw. Aufl., bearbeitet von Elmar Seebold; 242002] Schmitt, Rüdiger (1977, Hrsg.), Etymologie. Darmstadt: Wissenschaftliche Buchgesellschaft Seebold, Elmar (1981), Etymologie: Eine Einführung am Beispiel der deutschen Sprache. München: C. H. Beck
Elmar Seebold, Andechs (Deutschland)
175. Schriftsysteme 1. 2.
4. 5.
Sprachsystem und Schriftsystem Morphologische Aspekte in logographischen Schriftsystemen Morphologische Aspekte in phonographischen Schriftsystemen Zum Parameter der “Tiefe” Zitierte Literatur
1.
Sprachsystem und Schriftsystem
3.
Der überkommenen Auffassung zufolge wäre eine breite Transkription die ideale Schrift ⫺ wenn die Schriftzeichen direkt die kleinsten lautsprachlichen Einheiten abbilden, d. h. grosso modo den Phonemen entsprechen.
Dementsprechend wurde die Geschichte der Schrift als Fortschreiten hin auf das Ideal einer reinen Alphabetschrift gesehen (die klassische Darstellung ist Gelb 1963). Schriften, in denen die Schriftzeichen Wörter, Silben, Morpheme, nur Konsonanten etc. repräsentieren, haben dieser Auffassung zufolge den höchsten Entwicklungsstand noch nicht erreicht. Umgekehrt gelten Alphabetschriften, in denen vom phonographischen Ideal (ein Schriftzeichen ⫺ ein Laut bzw. Phonem) abgewichen wird, als in beklagenswerter Weise defizient (vgl. z. B. Saussure 1916). Diese Sichtweise ist im letzten Viertel des 20. Jahrhunderts zunehmend kritisiert und
1916
XXI. Morphologie und Nachbardisziplinen
revidiert worden. Wiewohl der Schrift ihr phonographischer Charakter nicht abgesprochen werden kann, ist dieser doch nicht auf eine linear-eindimensionale Abbildung der Phonemebene zu reduzieren. Zudem ist die Abbildungsfunktion der Schrift nicht ihre einzige, übernimmt sie doch gleichzeitig auch Modellfunktion für die Analyse (Olson 1993) und Entwicklung der gesprochenen Sprache (Koch & Oesterreicher 1994). Zu unterscheiden ist zwischen einem semantischen Aspekt (Schriftzeichenfolgen repräsentieren Bedeutungen), einem phonographischen Aspekt (Schriftzeichenfolgen repräsentieren die lautliche Seite der Sprache), und einem grammatischen Aspekt (Schriftzeichenfolgen repräsentieren die grammatische Organisation von Äußerungen) (vgl. Maas 1992; Günther 1995). Als Schriftsystem einer Sprache wird die Gesamtheit der die schriftliche Ausdrucksseite der Sprache betreffenden Mittel bezeichnet (Eisenberg 1996 a: 1368 f.), von denen einige in der Lautsprache nicht originär sind (z. B. Abkürzungen, abstrakte Symbole, Groß-/Kleinschreibung, Determinative). Moderne Schriftsysteme weisen stets semantische, phonographische und grammatische Aspekte gleichzeitig auf. So sind die arabischen Ziffern, wo immer sie in Texten verwendet werden, grundsätzlich semantisch bestimmt; in den modernen Alphabetschriften sind die Wortkörper durch die Buchstaben phonographisch, die Interpunktion und der Gebrauch von Großbuchstaben dagegen grammatisch bestimmt. Im folgenden soll anhand exemplarischer Beispiele die Rolle der Morphologie als Brücke zwischen phonographischem und grammatischem Aspekt in Schriftsystemen dargestellt werden.
2.
Morphologische Aspekte in logographischen Schriftsystemen
Im Hinblick auf die Grundbezugsebene lassen sich drei Schrifttypen unterscheiden: In logographischen Schriften beziehen sich die Schriftzeichen in erster Linie auf Bedeutungsträger (z. B. Chinesisch); in Silbenschriften bilden die Schriftzeichen Silben ab (z. B. die japanischen Kana); in Alphabetschriften sind die Schriftzeichen auf Phoneme bezogen (Eisenberg 1996 a: 1372). Der Terminus “logographisch” besagt wörtlich, daß die einzelnen Schriftzeichen dazu dienen, Wörter zu bezeichnen. In den meisten logographischen Schriften wird aber
eher davon die Rede sein, daß Schriftzeichen Morpheme bezeichnen. In isolierenden Sprachen wie dem Chinesischen wird damit im Prinzip jedes Morphem durch ein Schriftzeichen repräsentiert. In flektierenden Sprachen wird dagegen in der Regel nur die Wortbasis durch ein Logogramm repräsentiert. So wird in der sumerischen Schrift anfänglich ein Logogramm für alle möglichen grammatischen Formen eines Lexems verwendet; wenn aber später “syllabisch ausgedrückte grammatische Elemente obligatorisch zu den Logogrammen treten, wird deren Funktion auf den Ausdruck der Wortbasis eingeengt” (Krebernik & Nissen 1994: 279). Das Zitat kennzeichnet eine Veränderung, die überall beobachtet werden kann, wo von einer semantischen hin zu einer phonographischen Schreibweise übergegangen wird: Die ausschließliche Verwendung von Zeichen für Inhaltswörter wird ergänzt durch silbische Elemente; solche Schriftsysteme werden häufig als “morpho-syllabisch” bezeichnet. Denn die Schriftzeichen für Flexionselemente repräsentieren nicht in direkter Weise Flexionsmorpheme (ein Zeichen pro Morphem), sondern indirekt durch Kennzeichnung der Aussprache. Es gibt verschiedene Verfahren, die den Übergang zu einem morphosyllabischen und damit jedenfalls teilweise phonographischen System kennzeichnen. Der bekannteste Fall sind sog. Rebusschreibungen: Ein Wortzeichen wird nach seinem Lautwert, nicht seiner Bedeutung gelesen, besonders bei Homonymen. So wird im Sumerischen das (ursprünglich piktographische) Schriftzeichen für sar ‘Pflanze’ dann auch benutzt für ein homophones Verb mit der Bedeutung ‘schreiben’ (Cooper 1996: 42; mit weiteren Beispielen). In der späteren Zeit werden logographische Zeichen auch für gleich oder ähnlich lautende Flexionselemente verwendet, z. B. dient dann das Logogramm für ga ‘Milch’ auch dazu, als Silbenzeichen den Auslaut der Form duga zu du ‘sagen’ zu schreiben (Krebernik & Nissen 1994: 285). Ein weiterer uneigentlicher Gebrauch logographischer Schriftzeichen sind semantische Determinative (auch Komplemente genannt). Hier wird einem Logogramm für einen homonymen Ausdruck ein weiteres Zeichen aus dem gleichen Bedeutungsfeld beigegeben, das die Lesart indiziert; in einer logographischen Schrift für das Deutsche würde man z. B. dem Wortzeichen für Schloß das Zeichen für Schlüssel beigeben, um die Lesart
1917
175. Schriftsysteme
Türschloß zu spezifizieren. In der ägyptischen Hieroglyphenschrift stehen einmal etablierte Determinative auch bei nicht ambigen Zeichen, um sie inhaltlich näher zu spezifizieren; so kann dem Wort ‘fd.t ‘Kasten’ ein semantisches Determinativ hinzugefügt werden, das ihn als aus Holz, Stein oder Metall gefertigt kennzeichnet (Ritner 1996: 77); es wird so ein umfangreiches semantisches Klassifikationssystem aufgebaut, das in der gesprochenen altägyptischen Sprache keine Entsprechungen hat (Schenkel 1994: 291). Es läßt sich hier durchaus von einer morphologischen Kennzeichnung semantischer Klassen in der Schreibung sprechen, die es im Mündlichen nicht gibt. Eine systematische Aufteilung der logographischen und der phonographischen Ebene bietet das japanische Schriftsystem (vgl. Stalph 1996). Es besteht aus drei funktional und formal differenzierbaren Schriftzeichensets. Das eine, Kanji genannt, ist eine heute auf knapp 2000 beschränkte Menge von Logogrammen, denen durch vielfache Homonymie ca. 4000 Bedeutungen entsprechen (Stalph 1996: 1415 f.). Kanji stellen praktisch immer Basismorpheme dar. Daneben stehen zwei homologe, funktional differenzierte Syllabare. Das eine, Katakana genannt, dient der Schreibung von Namen, Fremdwörtern, Dialektismen etc., also rein phonographischen Zwecken. Das andere, Hiragana genannt, dient der Repräsentation grammatischer Aspekte (Hilfswörter, Verbalsuffixe, Konjunktionen) und repräsentiert mithin (auch) die Flexionsmorphologie des Japanischen. Eine so eindeutige funktionale Aufteilung der Schriftzeichen ist in den Schriften der Welt eher selten.
3.
Morphologische Aspekte in phonographischen Schriftsystemen
3.1. Der Parameter der Tiefe Die Abbildung der Phoneme auf die Buchstaben gilt als Endstufe der phonographischen Entwicklung. In der reinsten Form verkörpert wohl das Altgriechische der vorklassisch-klassischen Periode dieses Prinzip: Hier scheint es keinen Bezug auf die Bedeutungsseite zu geben. Dies wird besonders deutlich durch das Prinzip der scriptio continua: Quasi den mündlichen Sprachstrom direkt abbildend werden im Gegensatz zu semitischen Vorbildern keine Wortgrenzen markiert (vgl. Ludwig in Vorb.). Die Schriftgeschichte zeigt
indes, daß dies nicht der Endpunkt der Entwicklung ist: Wenn sie länger in Gebrauch sind, entwickeln sich in allen alphabetischen Systemen Auszeichnungsmethoden, die mehr als die Darstellung der Lautseite der Sprache zum Ziel haben. Durch typographische Mittel, Einführung der Worttrennung durch das Spatium, Groß-/Kleinschreibung, Interpunktion u. a. m. wird in der schriftlichen Form die “grammatische Artikulation des Textes” repräsentiert (Maas 1992): Absätze organisieren den Textfluß, initiale Majuskel und Schlußpunkt gliedern den Satz als die maximale syntaktische Einheit aus, das Spatium markiert minimale syntaktische Einheiten (Wörter), etc. Auch die Repräsentation morphologischer Aspekte gehört zu diesen Mitteln. Neuerdings wird hier mit dem Konzept der Tiefe eines Schriftsystems gearbeitet. Wohl im Anschluß an die Konzeption einer Tiefenphonologie (Chomsky & Halle 1968) gelten Schriftsysteme als um so flacher, je direkter die Buchstabenfolgen die Oberflächenphonologie einer Sprache reflektieren. So werden im modernen Neuhochdeutschen morphonemische Phänomene wie die Auslautverhärtung in der Schrift nicht ausgedrückt, die noch im Mittelhochdeutschen und Frühneuhochdeutschen auch im Schriftbild erschien, vgl. das Wort Tag: (1) (a) mhd. (b) nhd.
[tak][ta:k]
[tages][ta:ges]
Es wird davon gesprochen, daß das Mittelhochdeutsche eine flachere Graphie als das Neuhochdeutsche aufweist. Die systematischsten neueren Untersuchungen zum Parameter der Tiefe von Schriftsystemen in linguistischer Hinsicht stammen von Trudel Meisenburg. Abb. 175.1 kennzeichnet den Ansatz. Unterschieden werden zunächst zwei durch den Parameter der Tiefe verbundenen Ebenen, eine (oberflächen-)phonologische und eine morphosemantische Ebene. Eine dritte (phonetische) Ebene ist zwar ebenfalls anzusetzen, aber ohne große praktische Relevanz (Meisenburg 1996 a: 25). Außerdem ist zu unterscheiden, ob die Tiefe des jeweiligen Systems in diachroner oder synchroner Hinsicht betrachtet wird. Für die verschiedenen Ebenen werden Beispiele aus dem Lateinischen gegeben (signalisiert im Lateini-
1918
XXI. Morphologie und Nachbardisziplinen
Ebenen der Verschriftung (mit lateinischen Beispielen)
flach
((phonetische ) <ECVS>
Parameter der Tiefe
Ebene: Allophone —> Allophonie)
phonologische Ebene Oberflächenphoneme —> Allomorphie Kontextsensitivität Graphotaktik prosodische Bezüge
morphosemantische Ebene – diachron –: ≈ etymologisch/historisch —> diachrone Morphemkonstanz, graphische Physiognomie
– synchron –: ≈ zugrundeliegende Repräsentationen, ≈ Morphophoneme —> graphische Morphemkonstanz
<EQVVS>
(etymologische oder überlieferte Schreibungen, ermöglichen keine synchronen Bezüge, liefern lexikalische Information, graphische Individuierung von Wortformen, Homonymendifferenzierung)
(zusätzliche morphosemantische Analyse; Überbrückung von Allomorphie durch möglichst konstante Morphemschreibung – auf Kosten regelmäßiger GPK)
tief Abb. 175.1: Der Parameter der Tiefe (aus Meisenburg 1998: 45)
schen sowohl /v/ als auch /u/). Da gibt es zunächst indie Wiedergabe phonetisch bedingter allophonischer Variation durch auch unterschiedliche Schriftzeichen: In den älteren Zeugnissen steht für /k/ vor vorderen Vokalen , vor hinteren , vor Konsonanten⫺ eine Praxis, die aber schon sehr früh aufgegeben wird: verschwindet, und wird nur noch vorverwendet (Meisenburg 1996 a: 39). Auch oberflächenphonologische Variation wird anfänglich berücksichtigt, z. B. urbs ‘Stadt’ Nom. Sg. vs. Gen. Sg. wegen Entstimmung des zugrundeliegenden /b/ vor /s/ in der Nominativform (/urbs/ -[urps]), später nicht mehr; ähnlich die anderen Fälle. Entgegen dem Eindruck der vereinheitlichten Klassikerausgaben gibt es in klassischer und nachklassischer Zeit Schreiberunterschiede bezüglich der “tieferen” Schreibung, z. B. trotz der vorherrschenden Schreibung nach der Aus-
sprache [offero] wegen der morphologischen Struktur ob ⫹ fero, oder, wiewohl selten,trotz der Dissimilation [tantus] wegen des Bestandteils tam (vgl. ausführlich Meisenburg 1996 a: 35⫺47). Die bisher angeführten Beispiele betreffen Schreibungen, die synchron nachvollziehbar sind. Andere Mechanismen beziehen sich u. a. auf die diachrone Ebene (vgl. ausführlich zur lateinischen Graphie Meisenburg 1996 a: 31⫺ 47). Dazu gehört z. B. die Kennzeichnung (ursprünglich) fremder Wörter durch fremde Buchstaben(kombinationen), etwa wenn aus dem Griechischen entlehnte Wörter mit und erscheinen wie in , obgleich wohl nur Gebildete dem auch den entsprechenden, dem Lateinischen fremden Laut [y] zuordnen konnten; die Normalaussprache dürfte [‘fisikus] gewesen sein (Meisenburg 1996 a: 46). Es findet sich auch Homonymendifferenzierung, z. B. [abitus]
1919
175. Schriftsysteme
vs. wegen der Ableitungsstruktur ab-ire. Im folgenden soll der Parameter der Tiefe in Anschluß an Meisenburg (1996 a; 1998) exemplarisch an einigen Schriftsystemen erläutert werden. Die Anordnung soll die zunehmende Tiefe der Systeme spiegeln. 3.2. Spanisch Das spanische Schriftsystem wird von Meisenburg diachron (1996 a) und synchron (1996 b) als nahezu idealtypisch flaches System gekennzeichnet. Die Elemente eines relativ einfachen Phonemsystems werden meist direkt durch lateinische Buchstaben repräsentiert. Allomorphien, ohnehin eher selten und unsystematisch, werden durchgängig schriftlich repräsentiert, ebenso Neutralisierungen oder Assimilationen. Systematisch phonologiebezogen ist auch die graphische Markierung des relativ freien und teilweise distinktiven Wortakzents. Die synchrone Flachheit des spanischen Schriftsystems wird lediglich durchbrochen bei der Akzentmarkierung homophoner Funktionswörter. Eine geringfügige diachrone Tiefe erhält das System durch aus dem Lateinischen stammende Alternationen, die keine morphophonemischen Entsprechungen haben, z. B. , trotz der gleichen Aussprache [to1kar], [to1kes]. 3.3. Koreanisch Das moderne koreanische Schriftsystem Han‘gul, im 15. Jahrhundert entwickelt, ist in vieler Hinsicht einzig unter den alphabetischen Schriftsystemen (vgl. King 1996). Die Basis des koreanischen Schriftsystem ist eine quasi ikonische Abbildung der artikulatorischen Geste bei den Konsonanten; die genaue Abbildung der Lautstruktur ist Ziel dieser Schriftform (vgl. Günther 1988: 54⫺58). Anders als in Lateinschriften werden (ähnlich wie in indischen Schriften) die alphabetischen Symbole zu Silben repräsentierenden Schriftzeichen zusammengefaßt, deren Form und Größe an chinesische Schriftzeichen erinnert. Trotz der Zielsetzung einer möglichst genauen Darstellung der Lautebene gab es von Anfang an auch Bestrebungen, morphologische Aspekte zu berücksichtigen (King 1996: 223). Aufgrund der Zusammenfassung einzelner Phoneme zu Silben ergeben sich bei flektierten Formen nämlich unterschiedliche Schriftzeichen. So lautet der Lokativ des Wortes saram ‘Mensch’ saram-ege ‘Menschlok’. Ersteres ergibt die Silbenfolge sa-ram,
letzteres sa-ra-me-ge. Die zweite Silbe ist also in der Grundform ram, in der flektierten Form ra. Dies aber ergibt unterschiedliche Zeichen: Zwar sind in beiden Silbenzeichen die Elemente für /r/ und /a/ enthalten, doch die Verwendung des Zeichens für /m/ hängt ab von der Syllabifizierung, so daß das komplexe Zeichen unterschiedlich aussehen kann (vgl. Abb. 175.2). In der modernen koreanischen Orthographie gilt die morphemische Schreibweise. (a) saram ‘Mensch’
sa
ra m
(b) saramege ‘Mensch’, Lokativ morphemische Schreibweise
silbische Schreibweise
sa
sa
ra m
’e
ge
ra
me
ge
Abb. 175.2: Schreibweisen des Wortes saramege in Hang’ul
Das koreanische System ist also, betrachtet man die Komponenten der einzelnen Schriftzeichen, wie das Spanische sehr flach; betrachtet man die Komposition der Silbenzeichen, so ergibt sich eine gewisse morphologische Tiefe, die aber die phonologische Flachheit nicht tangiert (s. 4). 3.4. Deutsch Das deutsche Schriftsystem ist durch das Prinzip der Morphemkonstanz bestimmt (Günther 1988: 86⫺94), das in 3.1 schon am Beispiel der deutschen Auslautverhärtung erläutert wurde: Mit wenigen Ausnahmen haben deutsche Wortstämme im Geschriebenen eine konstante Form ungeachtet der Variation in der phonetisch-phonologischen Realisation. Instruktiv ist das Morphem /kø:nig/. In der Standardsprache wird /ig/ im Auslaut [ic¸] ausgesprochen, also lautet die Basisform [kø:nic¸]. Die Ableitung mit Movierungssuffix /kø:nig ⫹ in/ wird [kø:nigin] ausgesprochen, das denominale Adjektiv /kø:nig ⫹ lic¸/ lautet [kø:niklic¸]. Für das Stammorphem gibt es also drei phonetische Realisierungen, aber stets die gleiche Schreibung; synchron repräsentiert die Schreibung direkt die morphologische Ebene. Auch in anderen Fällen signalisiert die geschriebene Form morphologische Struktur,
1920 die an der lautsprachlichen Oberfläche nicht sichtbar ist. So gibt es im gesprochenen Nhd. keine konsonantischen Geminate; wo sie aus morphologischen Gründen auftauchen würden, werden sie im Gesprochenen vereinfacht, in der Schrift jedoch nicht, vgl. Tiefflug [1ti:flu:k], verraten [ve1ra:tn], etc. In der Flexion gibt es zwar auch im Geschriebenen Geminatenvereinfachungen, vgl. <er rät> /re:t ⫹ t/,/feiks ⫹ st/, doch bleibt auch hier die Stammkonstanz erhalten. Ein etwas komplexerer Fall ist der Umlaut. Bei vokalischer Alternation gibt es im Nhd. (anders als im Englischen, s. 3.6) keine graphische Morphemkonstanz, vgl. singe ⫺ sang ⫺ gesungen; ziehen ⫺ Zug, etc. Die Vokale [e], [ø] und [y] und die entsprechenden Langvokale alternieren in einer Reihe von Fällen, z. B. der nominalen Pluralbildung, mit [a], [o] und [u], vgl. /bart/ ⫺ /berte/, / korb/ ⫺ /kørbe/, /hu:t/ ⫺ /hy:te/, teilweise auch in Derivationen, vgl. /alt/ ⫺ /eltlich/, / to:t/ ⫺ /tø:tn/, /vu:t/ ⫺ /vy:teric¸/. Die umgelautete Form wird mit dem Basisvokal plus Trema als Diakritikon geschrieben: , , , <ältlich>, , <wüterich>. Im heutigen Deutsch ist der Umlaut morphologisch, nicht mehr phonologisch geregelt (s. Art. 147; Wurzel 1984). Außerdem müssen <ä>, <ö> und <ü> in der nhd. Orthographie als Grapheme gelten, weil sie auch ohne Umlautbasis vorkommen, vgl. z. B. , wozu es keine * gibt. Natürlich sind diese Schreibungen diachron als Umlaut erklärbar; dies ändert nichts am graphematischen Status der Umlautbuchstaben. In der seit 1998 gültigen neuen Rechtschreibung werden in einigen Fällen diachron erklärbare Fälle synchron umgedeutet, so ist z. B. früher (zu ahd. bentil) nunmehr zu schreiben, weil es synchron zu gestellt wird, d. h. der Bezug der Umlautbuchstaben auf die Basisvokale wird als morphologisches Schreibprinzip angesehen. In die diachrone Dimension gehört die Morphemkonstanz u. a. bei unproduktiven Ableitungen, z. B. das bei (historisch zu drehen), da das Dehnungs-h nur vor auftritt (vgl. Eisenberg 1996 b). Hierher gehört auch die Verteilung von und zur Bezeichnung von /f/ im Anlaut im indigenen Wortschatz: Dem unmarkierten Fall der -Schreibung stehen eine Handvoll -Schreibungen gegenüber, die jedoch großteils sehr häufige Wörter
XXI. Morphologie und Nachbardisziplinen
bezeichnen (vgl. Günther 1988: 94⫺97). Diachrone Tiefe durch Homonymendifferenzierung (Saite ⫺ Seite, Mohr ⫺ Moor, etc.) ist offenbar im 17. Jahrhundert erheblich stärker ausgeprägt gewesen als heute. Ein orthographisches Problem bietet schließlich die Fremdwortschreibung: Fremde Wörter werden zunächst grundsätzlich in der Originalschreibung wiedergegeben, wodurch diachrone Tiefe erreicht wird; die allmähliche (und unsystematische) phonologische und morphologische Angleichung an des deutsche Schriftsystem führt hier zu Inkonsistenzen (z. B. ist die eingedeutschte Form Frisör heute wieder veraltet; das Duden-Rechtschreibwörterbuch (2000) verzeichnet als Hauptlemma Fotografie mit dem Verweis “auch Fotographie”, dagegen als Hauptlemma Lithographie mit dem Verweis “auch Lithografie”; etc.). Daß im heutigen Deutsch in der Tat die konstante Morphemschreibung systematisch dem phonographischen Oberflächenbezug vorgeordnet ist, belegt das Kuriosum der Schreibung von. Phonologisch ist stets von stimmhaftem Auslaut auszugehen, denn es heißt standardsprachlich z. B. im Plural <doofe> [do:ve]. Da jedoch die Grundformen mit geschrieben werden (wohl, weil diese Stämme fast nur so vorkommen), wird die Schreibung auch in den anderen Formen beibehalten, obgleich im Deutschen stets /f/, nie /v/ bezeichnet (vgl. Günther 1988: 96 f.). Insgesamt ist das Deutsche tiefer als z. B. Spanisch oder Koreanisch, aber flacher als die beiden im folgenden besprochenen Schriftsysteme. 3.5. Französisch In noch stärkerem Ausmaß als im Deutschen ist im Französischen graphematische Morphemkonstanz gewahrt. Dies betrifft insbesondere nicht hörbare, aber geschriebene Morphologie, etwa das Plural -s, vgl. <enfant ⫺ enfants> [a˜fa˜] ‘Kind(er)’, [le; le] oder den Genusmarker -e, vgl. [pe1ti ⫺ pe‘tit]. Verkompliziert wird das Ganze durch das Phänomen der Liaison; vor vokalisch anlautenden Folgewörtern wird z. B. das Plural -s hörbar, vgl. les Etats Unis [leze:1tazy:1ni] ‘USA’. Da die Wortgrenze überschritten wird, wenn nur geschriebene Morpheme hörbar werden, spricht Eisenberg (1996 a: 1376) vom Wortformenbezug des Französischen. Wenn unterschiedlich lautende Allomorphe durch graphische Morphemkonstanz aufeinander bezogen bleiben, vgl. z. B. [sˇa˜ ⫺
1921
175. Schriftsysteme
sˇa˜1petr] ‘Feld ⫺ ländlich’ oder[a1za:r - aza:r1de] ‘Zufall ⫺ aufs Spiel setzen’, spricht Catach (1996: 1448 f.) von ‘Morphogrammen’. Bezüglich diachroner Tiefe ist vor allem die morphemidentifizierende Distinktschreibung von Homonymen zu nennen, vgl. z. B. die verschiedenen Schreibungen der [ko˜t] ausgesprochenen Morpheme in ‘Märchen’, ‘Graf’, ‘Rechnung’, wodurch zugleich derivationelle Zusammenhänge verdeutlicht werden, vgl. ‘erzählen’, ‘gräflich’, ‘zählen’. Eine systematische Darstellung der diachronen Entwicklung der französischen Graphie bis heute bietet Meisenburg (1996 a: 68⫺206). 3.6. Englisch Englisch ist für viele Autoren der Prototyp eines “tiefen” Systems; als “deeply morphologized” wird es in Stubbs (1996: 1442) bezeichnet. Die Auffassung, daß es sich um ein notorisch irreguläres System handelt, ist spätestens seit der Behauptung, es handele sich um ein “near to optimal system” (Chomsky & Halle 1968: 49) zumindest fragwürdig geworden. Aufgrund der englischen Sprachgeschichte ist insbesondere die diachrone Tiefe des englischen Schriftsystems wohl einzigartig (zur historischen Entwicklung vgl. Scragg 1974). Hier sind insbesondere zwei Aspekte zu nennen. Der eine ist die Zusammensetzung des Wortschatzes aus einem angelsächsischen und einem romanischen Teil mit unterschiedlichen Wortstrukturen und orthographischen Konventionen, vgl. z. B. die Schreibung des Auslauts [sˇen] in [fresˇen] und [neisˇen]. Der andere ist der Umstand, daß durch den Drucker Caxton die englische Orthographie schon im 15. Jahrhundert weitgehend festgelegt wurde; die nachfolgenden gravierenden Lautveränderungen, insbesondere der “great English vowel shift”, wurden in der Schreibung nicht berücksichtigt. Ähnlich wie das Französische archiviert so das englische Schriftsystem in Paaren wie <sign ⫺ signature> oder morphologische Zusammenhänge, die in der gesprochenen Sprache nicht mehr erkennbar sind, vgl. [sain ⫺ signetsˇer ], [di1vain ⫺ di1viniti]. Ein weiterer morphologischer Aspekt ist die Worttrennung am Zeilenende, wo nicht wie in vielen anderen Alphabetschriften Schreibsilben abgetrennt werden (vgl. Eisenberg 1996 b), sondern Flexionssuffixe, sofern sie
Vokale enthalten, vgl. read-ing, walk-ed, substitut-ion. Obgleich solche Beispiele morphologischen Bezug aufweisen, muß doch festgestellt werden, daß ihre Reichweite und Systematizität ungleich kleiner ist als z. B. im Deutschen oder auch im Russischen (dazu Meisenburg 1998). Es ist deshalb eher davon zu sprechen, daß die Tiefe des englischen Schriftsystems sich auf die Wortebene bezieht; “the invariant unit of English spelling is not the morpheme, but the word” (Stubbs 1996: 1442). In der Tat sind einzelne Autoren so weit gegangen, das englische als ein logographisches Schriftsystem zu bezeichnen ⫺ sicherlich ein Unfug, denn jenseits der Unrichtigkeit von G.B. Shaws Behauptung, das Englische sei so unregelmäßig, daß z. B. auchals [fisˇ] gelesen werden könne ⫺ [f] wie in , [i] wie im Plural <women>, [sˇ] wie in ⫺, gibt es auch im Englischen eine phonographische Grundebene; die Buchstabenfolge kann niemals genauso wie (beides kann ‘Grab’ bedeuten) gelesen werden (Beispiel von Daniels 1996: 654).
4.
Zum Parameter der “Tiefe”
Betrachtet man die vorstehenden Beispiele im Zusammenhang, so wird deutlich, daß die so einfache scheinende Dimension von morphologisch flach bis tief keineswegs homogen ist. Schon in Abb. 175.1 (Meisenburg 1998) werden zusätzlich die synchrone und die diachrone Ebene unterschieden. Letztere ist überdies höchst disparat: Fremdwortschreibung, Homonymendifferenzierung und etymologische Schreibung können nur a posteriori zusammenwirken. Ein einheitliches Gebiet bilden sie nicht; zudem ist morphologischer Bezug für keinen der Bereiche konstitutiv. “Verallgemeinerungen über den Ebenenbezug alphabetischer Schriftsysteme sind [...] noch ein Forschungsdesiderat” stellt dementsprechend Eisenberg (1996 a: 1376) fest, so daß die von ihm tentativ formulierten Tendenzen noch der Bestätigung harren: (a) Sprachen mit wenig Allomorphie neigen eher zu flachen Schriftsystemen (Spanisch). (b) Isolierende Sprachen und solche mit segmental schwer faßbarer Morphologie neigen zu Wortformbezug (Französisch, Englisch, Arabisch).
1922 (c) Sprachen mit segmentaler Morphologie neigen zu Morphembezug (Deutsch, Russisch). Die Notwendigkeit weiterer Forschungen liegt auf der Hand. Naheliegende Umkehrschlüsse z. B. sind offenbar nicht zutreffend. So ist das serbokroatische Schriftsystem ähnlich flach wie das spanische, obgleich es erheblich mehr Allomorphie gibt, z. B. durch die Palatalisierung, die hier im Unterschied zu anderen slawischen Sprachen in der Schrift angezeigt wird (vgl. Comrie 1996), vgl. z. B.‘Arm’; <snaha ⫺ snasi> ‘Schwiegertochter’ (jeweils Nom./Dat. Sg.). Die phonographische Genauigkeit geht so weit, daß sogar Unterschiede zwischen serbischen und kroatischen Dialekten graphisch repräsentiert werden, vgl. serbisch <mleko>, kroatisch <mlijeko> ‘Milch’. Wohlgemerkt gilt dies für die serbokroatische Schrift seit ihrer grundlegenden Reform im 19. Jahrhundert (Feldman & Barac-Cikoja 1996). Eine besondere Stellung nehmen die semitischen Schriftsysteme ein. Auch wenn es Ansätze zur Schreibung von (Lang-)Vokalen gibt, sind sie systematisch als Konsonantenalphabete zu kennzeichnen. Aufgrund der semitischen Morphologie mit (extrem vereinfacht!) drei Konsonanten als Wurzel und vokalischer Variation zur Kennzeichnung von Flexion und Derivation wird die morphologische Ebene in besonderer Weise ausgezeichnet (das Konsonantengerüst signalisiert morphologische Beziehungen); das System scheint morphologisch “tief”. Eine Kennzeichnung der arabischen Schriften als Morphemschriften aber ist unangemessen. So spricht Eisenberg (1996 a: 1376) von einer morphologischen Analyse, die aber erst auf der Ebene der Wortformen strukturbildend wirkt, und ganz explizit wehrt sich Bauer (1996: 1435) gegen die landläufige Vorstellung, “die arabische Schrift diene vor allem zur Schreibung von Wurzeln”. Die vielen Details, worin dies sich zeigt, können hier nicht ausgeführt werden; wesentlich ist das Ergebnis: “Die durch die Defektivschreibung des Arabischen entstehende Tiefe der Schrift ist [...] eher lexikalischer als morphologischer Natur” (Bauer 1996: 1455). Ähnlich war in 3.6 das englische Schriftsystem als weniger morphologisch und vielmehr lexikalisch tief gekennzeichnet worden. Eine solche differenzierte Betrachtung des Parameters der Tiefe ist insbesondere deshalb wesentlich, weil er auch in anderen Wissen-
XXI. Morphologie und Nachbardisziplinen
schaften eine Rolle spielt. In der psychologischen Leseforschung etwa ist er weitgehend als reguläre vs. irreguläre Graphem-PhonemKorrespondenzen interpretiert worden (z. B. Katz & Feldman 1996). Eine entwicklungspsychologische Interpretation besagt dann, daß Schriftsysteme um so leichter zu lernen seien, je flacher sie sind (Wimmer & Goswami 1994 zum Vergleich von Englisch und Deutsch), und eine kognitionspsychologische Deutung vermutet, daß der Erwachsene in flachen Systemen beim Lesen phonologisch rekodiert (Katz & Feldman 1996) oder in tiefen Systemen morphologisch zerlegt (Schreuder et al. 1989; s. Art. 162). Die oben gegebenen Hinweise lassen es als ratsam erscheinen, im jeweiligen Fall genau zu spezifizieren, was unter “flach” bzw. “tief” zu verstehen ist. Schon Bolinger (1946) war die besondere Rolle von “visual morphemes” bewußt. Es scheint, als ob die Analyse von Schriftsystemen in besonderer Weise geeignet ist, die Rolle der Morphologie als Interface von Bedeutung und Form in der menschlichen Sprache und ihre unterschiedliche Ausprägung in den Sprachen zu kennzeichnen. In den letzten Jahren ist die Erforschung der Schriftsysteme weiter vorangetrieben worden, insbesondere die Rolle der Schreibsilbe zwischen segmentaler und morphologischer Ebene steht im Zentrum der Aufmerksamkeit. Systematisch wird dies am Deutschen exemplifiziert in Eisenberg (1998: 286⫺ 311); einen guten Einblick in verschiedene Ansätze vermittelt der Sammelband von Neef et al. (2002, Hrsg.).
5.
Zitierte Literatur
Bauer, Thomas (1996), “Das arabische Schriftsystem”. In: Günther & Ludwig (Hrsg.), 1433⫺ 1436 Bolinger, Dwight L. (1946), “Visual Morphemes”. Language 22, 333⫺340 Catach, Nina (1996), “The French Writing System”. In: Günther & Ludwig (Hrsg.), 1445⫺1450 Chomsky, Noam & Halle, Morris (1968), The Sound Pattern of English. New York: Harper & Row Comrie, Bernard (1996), “Languages of Eastern and Southern Europe”. In: Daniels & Bright (Hrsg.), 663⫺689 Cooper, Jerrold S. (1996), “Sumerian and Akkadian”. In: Daniels & Bright (Hrsg.), 37⫺57
175. Schriftsysteme
1923
Daniels, Peter T. (1996), “English”. In: Daniels & Bright (Hrsg.), 651⫺655
Ludwig, Otto (in Vorb.), Geschichte des Schreibens. Buchmanuskript
Daniels, Peter & Bright, William (1996, Hrsg.), The World’s Writing Systems. Oxford: Blackwell
Maas, Utz (1992), Grundzüge der deutschen Orthographie. Tübingen: Niemeyer
Duden (2000), Die deutsche Rechtschreibung. 22., völlig neu bearbeitete und erweiterte Auflage. Mannheim et al.: Dudenverlag (Duden Band 1)
Meisenburg, Trudel (1998). “Zur Typologie von Alphabetschriftsystemen anhand des Parameters der Tiefe.” Linguistische Berichte 173, 43⫺64
Eisenberg, Peter (1996 a), “Sprachsystem und Schriftsystem”. In: Günther & Ludwig (Hrsg.), 1368⫺1380
Meisenburg, Trudel (1996 a), Romanische Schriftsysteme im Vergleich. Tübingen: Narr
Eisenberg, Peter (1996 b), “Das deutsche Schriftsystem”. In: Günther & Ludwig (Hrsg.), 1451⫺ 1455
Meisenburg, Trudel (1996 b), “Das spanische Schriftsystem”. In: Günther & Ludwig (Hrsg.), 1437⫺1441
Eisenberg, Peter (1998), Grundriss der deutschen Grammatik I: Das Wort. Stuttgart/Weimar: Metzler
Neef, Martin & Neijt, Anneke & Sproat, Richard (2002, Hrsg.), The Relation of Writing to Spoken Language. Tübingen: Niemeyer
Feldman, Laurie B. & Barac-Cikoja, Dragana (1996), “Serbo-Croatian: A Biscriptal Language”. In: Daniels & Bright (Hrsg.), 769⫺776
Olson, David R. (1993), “How Writing Represents Speech”. Language & Communication 13, 1⫺17
Gelb, I[gnaz] (1963), A Study of Writing. Chicago: University Press
Ritner, Robert K. (1996), “Egyptian Writing”. In: Daniels & Bright (Hrsg.), 73⫺84
Günther, Hartmut (1988), Schriftliche Sprache: Strukturen geschriebener Wörter und ihre Verarbeitung beim Lesen. Tübingen: Niemeyer
Saussure, Ferdinand de (1916), Cours de linguistique gene´rale [e´dition critique pre´pare´ par Tullio de Mauro; 1974]. Paris: Payot
Günther, Hartmut (1995), “Die Schrift als Modell der Lautsprache”. Osnabrücker Beiträge zur Sprachtheorie 51, 15⫺32
Schenkel, Wolfgang (1994), “Die ägyptische Hieroglyphenschrift und ihre Weiterentwicklungen”. In: Günther & Ludwig (Hrsg.), 289⫺297
Günther, Hartmut & Ludwig, Otto (1994, Hrsg.), Schrift und Schriftlichkeit: Ein interdisziplinäres Handbuch internationaler Forschung. Berlin: de Gruyter (Handbücher zur Sprach- und Kommunikationsforschung 10.1)
Schreuder, Rob & Jarvella, Robert & Job, Raimo & Sandström, Göran (1989), “Morphologische Strukturen im mentalen Lexikon”. In: Günther, Hartmut (Hrsg.), Experimentelle Untersuchungen zur deutschen Flexionsmorphologie. Hamburg: Buske, 27⫺40
Günther, Hartmut & Ludwig, Otto (1996, Hrsg.), Schrift und Schriftlichkeit: Ein interdisziplinäres Handbuch internationaler Forschung. Berlin: de Gruyter (Handbücher zur Sprach- und Kommunikationsforschung 10.2) Katz, Leonard & Feldman, Laurie B. (1996), “The Influence of an Alphabetic Writing System on the Reading Process”. In: Günther & Ludwig (Hrsg.), 1094⫺1101 King, Ross (1996), “Korean Writing”. In: Daniels & Bright (Hrsg.), 218⫺227 Koch, Peter & Oesterreicher, Wulf (1994), “Schriftlichkeit und Sprache”. In: Günther & Ludwig (Hrsg.), 587⫺604 Krebernik, Manfred & Nissen, Hans J. (1994), “Die sumerisch-akkadische Keilschrift”. In: Günther & Ludwig (Hrsg.), 274⫺288
Scragg, Donald G. (1974), A History of English Spelling. Manchester: Manchester University Press Stalph, Jürgen (1996), “Das japanische Schriftsystem”. In: Günther & Ludwig (Hrsg.), 1413⫺ 1427 Stubbs, Michael (1996), “The English Writing System”. In: Günther & Ludwig (Hrsg.), 1441⫺1445 Wimmer, Hans & Goswami, Usha (1994), “The Influence of Orthographic Consistency on Reading Development: Word Recognition in English and German Children”. Cognition 51, 91⫺103 Wurzel, Wolfgang U. (1984), Flexionsmorphologie und Natürlichkeit. Berlin: Akademie
Hartmut Günther, Köln (Deutschland)
1924
XXI. Morphologie und Nachbardisziplinen
176. Terminology in special languages 1. 2. 3. 4. 5. 6. 7.
Introduction What are ‘special languages’? The lexicon in special languages Term formation: morphological trends Standardisation in term formation Appendix References
1.
Introduction
Special languages are heavily dependent for their development on creative word formation processes. While the specialised vocabulary encountered in special languages has the same range of morphological structures as the general lexicon, it exhibits far greater regularity as a result of the deliberate and often systematic techniques of term creation. This article outlines the distinguishing characteristics of specialised designation (cf. 2) and briefly examines the nature of the lexicon in special languages (cf. 3). The semantic content of these designations and the motivations for their expression forms are explored, and a description is provided of the prevalent morphological variations encountered in English term formation (cf. 4). (For German and French term formation see Drozd & Seibicke 1973; Hoffman 1988; Kocourek 1982.) An appendix (cf. 6) presents guidelines for the creation of terminology in which certain preferences in practical term formation are motivated. The linguistic examples given are drawn from a variety of specialised subject domains since our objective is to make statements which apply to special languages in general.
2.
What are ‘special languages’?
Special languages (or sublanguages) are semiautonomous, complex semiotic systems based on and derived from general language. Their use presupposes special education and is restricted to communication among specialists in the same or closely related fields (cf. Sager et al. 1980). The features of special languages implied in this definition, namely special reference, special speech acts, special message forms and special pragmatic situations belong to quite separate aspects of language. From the point of view of function and structure, special languages highlight the referential and
systematic nature of language, maximally exploit the communicative and classificatory use of language and aim at a clear distinction between their three favoured intentions of speech acts, i.e. the informative, evaluative, and directive. This functional restriction of language use permits deliberate planning of special languages by means of rules for the creation and use of specialised linguistic forms.
3.
The lexicon in special languages
The lexicon of special languages has developed its own particular characteristics which set it apart from the general lexicon. These characteristics reflect and arise from the referential and classificatory needs of special languages. 3.1. Special reference A description of the particular referential needs of special language is fundamental to an account of the lexical trends in specialised designation. (For a detailed exposition of the theory of reference in special languages see Sager et al. 1980: 70⫺86.) The lexicon of a special subject language reflects the organisational characteristics of the conceptual domain by tending to provide as many lexical units as there are distinct concepts conventionally established in the knowledge structure of the domain, and by restricting the reference of each such lexical unit to a well-defined entity within it. It is this enhanced rigour which characterises special reference, and distinguishes it from general reference, where “general knowledge” tends to have a less well-defined, less “disciplined” structure. In addition to a large number of items which have the property of special reference (terms), the lexicon of a special language also contains items of general reference which are not usually specific to any discipline and whose referential properties are vague or generalised (words). The criterion which determines whether a particular lexical item is a term or word is thus the nature of its reference within the discipline concerned. The creation of new lexical items in special languages is a conscious and deliberate process carried out by subject specialists and is
176. Terminology in special languages
thus part of the scientific method. Consequently, the naming of concepts is as subject to procedural regulation and systematisation as scientific investigation and has its own special rules and recommendations (cf. BS 3669 1963 and ISO 704 1986). The standardisation of terminology is discussed further in 5 and 6. 3.2. Linguistic modes of designation Designation in special languages aims to achieve transparency and consistency. The structure of a designation often reflects the major conceptual features it represents, and related terms often have a comparable expression form. The most common mode of term formation is determination, which operates by means of compounding and by derivation. On the basis of the superordinate concept (hyperonym), it creates a more specific term whose designation is motivated by systematic criteria, e.g. according to material (iron chain), function (safety chain), location (door chain), etc. Consecutive stages of determination produce sequences of terms in hierarchy. Determination by means of pre-modification can use any word category as the specifying element (pre-edit, virtual storage, T-iron, four square, feed hole, preshrunk fabric, programming language). When a concept is determined by more than one modifier, it may be necessary to make use of hyphens to clarify the relationship (wide-angle lense). In some cases, the determined element is omitted and the new concept is represented solely by the modifier (floppy disk J floppy). Conjunction is a much less frequent technique of term formation, although it does occur in designations of chemical compounds (nickel steel) and of dual purpose devices (lifting and forcing pump). Disjunction, which joins the extensions of two or more concepts into a new hyperonym by presenting two alternatives as a single concept, is used to express an ‘either-or’ relationship (glide-sail parachute).
4.
Term formation: morphological trends
The three main linguistic methods of designation are: (a) the use of existing resources; (b) the modification of existing resources; (c) the creation of new resources.
1925 The following description of the usage of these various modes of designation in English special languages is not intended to be exhaustive but is rather indicative of the range of possibilities (cf. Sager 1996: 250⫺253 for further references on scientific and technological word formation). 4.1. Use of existing resources The creation of new designations by narrowing or widening the meaning of existing linguistic resources is frequent in pre-scientific enquiry and in popular expositions of science. The two principal methods used are polysemy and redefinition. Simile is a common method of new designation which highlights a distinguishing feature of a concept through comparison (Gclamp, star wheel). Special language also uses many metaphoric expressions based on form, function or position (ear of a hammer). Metaphors are also common in popular naming in biology (bluebell, seahorse). Other transfers of meaning occurring in special reference include the use of proper names to designate objects and measurements (machmeter, watt), the form of an object for another object of that form (square, bulb), and the name of the material for the object made of it (glass, iron). The process of redefinition usually entails reducing the extension of a general language word. For example, speed and velocity are used synonymously in general reference, but in special reference they designate two distinct concepts (cf. Sager et al. 1980: 255 f.). 4.2. Modification of existing resources The principal method of designation in special reference is the modification of existing resources by means of affixation, compounding, conversion or compression. 4.2.1. Affixation Affixation is a particularly important method of designation in special languages because of its ability to contribute to precision of expression and systematic reference. Affixes are far more numerous in special languages than in general language, largely because of the borrowing or adaptation of neoclassical words and methods of word formation, many of which are not required in general language. English freely assimilates word elements and words from other languages, and in science and technology, in particular, makes substantial use of Greek
1926 and Latin prefixes, suffixes, and stems (cf. Nybakken 1959; Flood 1960; Hogben 1969). Prefixes make an important contribution to the systematic structuring of special vocabularies. Contrasting sets of terms can be created easily by the negatives un-, in-, dis-, a(synchronous/asynchronous, organic/inorganic), the privatives un-, de-, dis- (stable/unstable, code/decode), the prefixes of degree or size super-, sur-, sub-, hyper-, ultra-, mini- (subsonic, hypersonic, supersonic), the locatives super-, sub-, inter-, trans- (amination, transamination), the temporal or sequential pre-, post-, re- (post-emphasis, pre-emphasis), and the numerals uni-, mono-, bi-, di- (monovalent, divalent), etc. Such sets maintain terminological unity and appear to be transparent. However, prefixes are not always used systematically and some prefixed words take on special meanings in the process of terminologisation (denature in chemistry means to render ethanol unfit for human consumption). In special languages, certain neoclassical stems are so frequently used that they can be considered prefixes from the functional and structural point of view (auto, equi, ext(e)r, iso, macro, mega, meta, micro, pan, para). The noun-noun transformation suffix -ing is highly productive in the conversion of countable nouns to mass nouns in order to indicate an assembly of objects (wire ⫺ wires ⫺ wiring). Concrete nouns are also derived from verbs by means of the suffix -er and its neoclassical alternative -or (scanner, collator). In this way, terminological agreements can be established between the action and the instrument of the action. The causative Latin-derived pattern -fy, -fier, -fication also contributes to systematicity by providing names for action verbs, agents or instruments and the corresponding processes (amplify ⫺ amplifier ⫺ amplification). Agent nouns can also be formed by -ant and often have matching verbs in -ate (penetrate ⫺ penetrant). Most process nouns end in -ation or -ion (collation). They can also end in -ing but this is less desirable in specialised designation because of the danger of confusion with the identical concrete nouns designating the product (carving, fitting). Properties are designated by adjectives derived from nouns, and by nouns derived from adjectives. Adjectival derivations in -al, -ic, -ive, and -ous are particularly prevalent (trinomial, concentric, coercive, contiguous). Property nouns are derived from adjectives ending mainly in -al, -ent, -ic, -ous (trinomiality, absorbency, concentricity, contiguity).
XXI. Morphologie und Nachbardisziplinen
Terms designating measurable properties frequently end in -ance or -ivity (capacitance, inductivity). Measuring devices are commonly designated by suffix-like expressions which indicate the method of measurement, e.g. -meter (spectrometer), -graph (thermograph). Certain suffixes are restricted to one field in which they are used exclusively or with a special meaning (-ase, -ate, -ol, -ose in chemistry; -asis, -itis, -oma, -ectomy in medicine). However, even within a single field suffixes may have more than one function (-ine and -ic in chemistry). 4.2.2. Backformation The process of backformation, or back-derivation, permits the formation of complex verb forms which correspond very closely in their range of reference to nominal concepts of processes. It thus achieves a great deal of concentration of information while maintaining terminological accuracy. Backformation in special languages occurs most often in complex nouns and in technology rather than science where there is a greater need for verbs to describe processes. It is closely linked to the process of compound verb formation which often terminologises verbs by narrowing their range of reference. The first element may designate the instrument or agent (steam-bend), the purpose (test-drive), the method (colour-code), or the location (factory-test). 4.2.3. Compounding Compounds play a very important role in special language designation where their potential for systematicity and regularity is maximally exploited to create coherent terminological systems. The vast majority of special language compounds are nouns. Noun compounds provide an ideal basis for the building of terminological systems as the nucleus indicates the category to which the concept belongs and the determinant the criterion for the subdivision of the category (cf. Art. 87). Three major types of noun compounds can be identified: those which designate entities, those which designate properties, and those which designate operations and processes. Most noun compounds fall into the first of these groups which exhibits the greatest variety of relationships between determinant and nucleus. Thus, the determinant may compare the nucleus to another object (daisy wheel); specify the material of which the nucleus is
1927
176. Terminology in special languages
made (steel band); express an inherent property of the new concept which is not inherent in the nucleus (pneumatic drill); specify the use to which the nucleus is regularly put (splitting axe); specify the object of an instrument or agent noun (card punch); express the product which is regularly associated with the nucleus (cotton mill); specify the mode of operation of the nucleus (tumble-drier, sliderule); refer to the whole of the nucleus as a part (brake lining); assert identity with the nucleus (fuel oil) (for further exemplification of these and other possible relationships see Sager et al. 1980: 267⫺271). Eponymic noun compounds are also frequent in special languages. While having the advantage of precision through unique differentiation, these compounds lack systematic import (Venn diagram, Eddy’s theorem). Compounds with particles occur in all word categories and form a large number of specialised terms (by -chamber, -effect, -product; on -line, -stream). There are also many phrasal compounds linked by prepositions (on the fly printer). Compounds are often made up of two, three or more elements, but five- or six-element compounds are rare. Multiple-element compounds occur in schedules for classificatory purposes but in other text types they are shortened because contextual reference resolves any possible ambiguity. Compound adjectives are rarely ambiguous and are therefore frequently used in special language where there is a need for economy and transparency of expression. Such compounds are used to determine processes and operations, and to identify new entities (heat-resistant, lead-free, voice-activated). Most compound verbs in special languages are formed by backformation (cf. 4.2.2) or conversion. A third major group are compounds with the particles out, over, under (underwrite, output, overcoal). 4.2.4. Conversion and compression Conversion occurs most frequently from verb to noun and vice versa (a drill, to telex), but there are also cases from adjective to noun (a solvent), from noun to adjective (imitation, model), and from adjective to verb (to dry, to smooth). The productivity of conversion is reduced in scientific English where there is a high proportion of terms derived from Latin and Greek word elements with identifiable noun endings which are unsuitable for conversion.
Lexical compression in special languages is achieved by abbreviating existing expression forms, omitting elements in compounds, creating letter symbols, combining letters and numbers into short designations, and by using pictograms (cf. Sager et al. 1980: 277⫺ 280). 4.3. The use of new resources Absolute invention in special language designation is rare since it is appropriate to reflect in new designations the relationships of new concepts to existing ones. In English, the creation of neologisms is often hard to distinguish from Greek and Latin borrowing. The principal methods of creating special language neologisms are the combination of existing English and anglicised Greek or Latin word elements (videodisc); the use of Latin or Greek stems (apex); blending (transceiver) (cf. Art. 91); the use of eponyms (Mach number); the use of letters, singly or in compounds (beta-ray); the rare use of absolute invention (quark). Borrowing from other languages is relatively infrequent in contemporary English, but is widespread in languages which import new technology, often together with English terminology. Of morphological interest are the methods by which foreign borrowings are adapted to the requirements of inflection or derivation in the target language. This process is relatively smooth in English because of its limited inflectional endings, its weak requirements for identifying word classes and its great openness to conversion. The spelling is anglicised if a word enters into general usage, but even this offers no problems with the dual English tradition of Germanic and Romance vocabulary (cf. Rey 1995: 3⫺85).
5.
Standardisation in term formation
Once the validity of a new concept is generally acknowledged, it becomes established within the specialist community which usually regularises it by agreeing on a suitable linguistic designation. Standardisation is a separate process by which users reach ‘public’ agreement on the adoption of a given term for use in specific circumstances. In addition to fixing meaning, standardisation may also involve a choice from among competing terms using the criteria of economy, precision or appropriateness (cf. Cabre´ 1998: 194 ff.).
1928
XXI. Morphologie und Nachbardisziplinen
The mandatory fixation of form and meaning of a group of terms is exceedingly rare and can be as limited in application as a single legal document (contracts with definition clauses), or cover the entire activities of a professional group (the language used for formal communication between aircraft pilots and air traffic controllers). Contrary to appearances and popular belief, the terminology of so-called “standard vocabularies” is only rarely made mandatory. In most cases it represents a set of terms recommended, for use as established in a sequence of object or procedural standards and collected in the first instance to preserve consistency of reference in an extended series of interrelated standards (cf. Sager 1996: 114⫺128).
6.
Appendix
The “Recommendations for the Selection, Formation and Definition of Technical Terms” (BS 3669 1963) issued by the British Standards Institute are guided by three principles: appropriateness, to the circumstances of communication, precision, of expression as demanded by specialised topics of communication, and economy, of expression as suitable for discourse among specialists. The most important recommendation is that terms should be created systematically with respect to their morphological, syntactic, semantic, and pragmatic characteristics. Terms should conform to the morphology, spelling, and pronunciation conventions of the language for which they are intended, and should not give rise to ambiguities in pronunciation or irrelevant associations. They should be capable of permitting the creation of derivatives so that the expression and content forms can be made consistent wherever possible. Latin or Greek roots are particularly useful for this purpose. Terms should be as short and concise as the requirements of precision and systematicity will permit. As function, mode of operation, production, and even origin can vary, it is recommended that in a complex term the determined constituent should be the genus of the new concept. In running text or spoken language terms may be abbreviated provided that the context is clear. A term should represent only one concept. Homonymy may occur when a general language designation becomes a term or when a term from one area is applied in another. Homonyms should be avoided when two
meanings are very similar, when they may occur in the same context or when a term from one special subject is applied in another with a different definition. A single concept should preferably be expressed by a single term. The creation of synonyms should be avoided because they may create confusion about the identity of a concept and because they burden the memory. Once a term has gained wide acceptance it should not be changed without compelling reasons and a strong certainty that the new term will become accepted as a full substitute.
7.
References
BS 3669 (1963) ⫽ Recommendations for the Selection, Formation, and Definition of Technical Terms. London: British Standards Institution Cabre´, M. Teresa (1998), Terminology: Theory, Methods and Applications, ed. by Juan C. Sager. Amsterdam: Benjamins Drozd, Lubomı´r & Seibicke, Wilfried (1973), Deutsche Fach- und Wissenschaftssprache. Wiesbaden: Brandstetter Flood, Walter Edgar (1960), Scientific Words: Their Structure and Meaning. London: Oldbourne Hoffman, Lothar (1988), Vom Fachwort zum Fachtext. Tübingen: Narr Hogben, Lancelot (1969), The Vocabulary of Science. London: Heinemann ISO 704 (1986) ⫽ Principles and Methods of Terminology. Geneva: International Organization for Standardization Kocourek, Rostislav (1982), La langue franc¸aise de la technique et de la science. Wiesbaden: Brandstetter Nybakken, Oscar E. (1959), Greek and Latin in Scientific Terminology. Ames, Iowa: Iowa State College Press Rey, Alain (1995), Essays on Terminology, transl. & ed. by Juan C. Sager. Amsterdam: Benjamins Sager, Juan C. (1996), A Practical Course in Terminology Processing. Amsterdam, Philadephia: Benjamins [11990] Sager, J[uan] C. & Dungworth, D[avid] & McDonald, P[eter] F. (1980), English Special Languages: Principles and Practice in Science and Technology. Wiesbaden: Brandstetter
Jeanette M. Pugh, Manchester (Great Britain) Juan C. Sager, Manchester (Great Britain)
1929
177. Sprachunterricht
177. Sprachunterricht 1. 2. 3. 4. 5.
1.
Eingrenzung des komplexen Themas Konzepte für das Lehren und Lernen einer Sprache Konzepte pädagogischer Grammatiken für den Fremdsprachenunterricht Grammatik im muttersprachlichen Unterricht Zitierte Literatur
Eingrenzung des komplexen Themas
Sprachunterricht ist in der Praxis wie in der Theorie ein hochkomplexes Geschehen, sowohl was den Unterricht selbst, das Lehren und Lernen, als auch was den Unterrichtsgegenstand, die Sprache, als auch was die Unterrichtsziele, das Beherrschen der Sprache in Wort und Schrift, betrifft. Praxis und Theorie des Sprachunterrichts, des muttersprachlichen wie des fremdsprachlichen, können zudem auf eine lange Geschichte zurückblicken. Sie ist faßbar bereits in den antiken Grammatiken (Art. 6), von denen Donat und Priscian eine überragende Rolle für die Vermittlung des Lateinischen ebenso wie für die allmähliche Herausbildung nationalsprachlicher Grammatiken und Lehrbücher zur Vermittlung der europäischen “Nationalsprachen” gespielt haben, den antiken Rhetoriken, von denen Cicero und Quintilian intensiv rezipiert wurden (s. Gwynn 1926), schließlich auch in den antiken literarischen Werken selbst, insofern an und mit ihnen im Lateinunterricht Sprache gelernt wurde. Die Ziele, die in diesem Unterricht verfolgt wurden, sind vielfältig: Die Lernenden sollten diese Texte verstehen und bei entsprechendem Lernerfolg selbst Texte (prosaische wie Briefe, Reden, Traktate, Verträge und poetische wie Epigramme, Epen, Oden, Dramen) nach den antiken Exempla verfassen können. Außerdem sollten sie, spätestens seit dem Beginn des Humanismus in Italien im 15. Jahrhundert, das Lateinische als lingua franca in Europa in Wissenschaft, Verwaltung, Politik möglichst vollkommen in Wort und Schrift, rezeptiv wie produktiv, zu beherrschen lernen. Während die Komplexität des Unterrichtsgegenstandes “Sprache” keiner weiteren Erläuterung bedarf, ist die Komplexität des Unterrichtsgeschehens selbst wenigstens stichpunktartig anzudeuten: Unterricht zielt durch
Lehrverfahren vielfältigster und unterschiedlichster Art darauf, Lernprozesse anzustoßen, zu ermöglichen, zu steuern, zu sichern, zu evaluieren. Die über der Praxis des Sprachunterrichts operierenden Theorien rezipieren und integrieren in vielfältiger und zum Teil unterschiedlicher Weise Erkenntnisse und Konzepte verschiedener wissenschaftlicher Disziplinen, insbesondere der Pädagogik, der Psychologie, der Linguistik, der Literaturwissenschaft. Insofern es sich um Fremdsprachenunterricht handelt, integrieren sie zudem fremdsprachendidaktische Konzepte sowie Ergebnisse der Zweitsprachenerwerbsforschung (s. Art. 166) und der Fremdsprachenlehr- und -lernforschung. Erst vor dem Hintergrund dieses komplexen Geflechts interdependenter Einflußfaktoren läßt sich das Thema “Sprachunterricht” so eingrenzen, daß seine Behandlung in dem vorgegebenen Rahmen nicht als inadäquate Verkürzung oder Simplifizierung mißverstanden wird. Der Beitrag fokussiert die folgenden Aspekte: ⫺ Konzepte für das Lehren und Lernen einer Sprache (insbesondere im fremdsprachlichen, aber auch im muttersprachlichen Unterricht) in Vergangenheit und Gegenwart in Auswahl und im Vergleich mit dem Schwerpunkt auf der Morphologie (s. 2), ⫺ Konzepte pädagogischer Grammatiken für den Fremdsprachenunterricht (s. 3), ⫺ Grammatik im muttersprachlichen Unterricht (s. 4).
2.
Konzepte für das Lehren und Lernen einer Sprache
2.1. Überblick Überblickt man die Konzepte für das Lehren und Lernen einer Fremdsprache seit der Antike, so lassen sich immer wiederkehrende Grundpositionen feststellen, die einander zum Teil zu widersprechen scheinen: (a) monolinguale Konzepte vs. bilinguale Konzepte (s. 2.2), (b) Konzepte induktiven Lernens, das von Texten ausgeht und Regularitäten durch die Lernenden selbst entdecken läßt, vs. Konzepte deduktiven Lernens, das von den durch die Grammatik und die Leh-
1930 renden zu vermittelnden Regeln und grammatischen Kategorien ausgeht, diese auswendig lernen und danach erst auf Beispielsätze und Texte anwenden läßt (s. 2.3), 2.2. Monolinguale vs. bilinguale Konzepte 2.2.1. Monolinguale Konzepte Monolinguale Konzepte propagieren die möglichst ausschließliche Verwendung der Zielsprache im Unterricht und außerhalb des Unterrichts. Sie versprechen sich davon vor allem zweierlei: (a) das Zurückdrängen interlingualer, potentiell Fehler verursachender Interferenz der Muttersprache auf die zu erlernende Fremdsprache sowie (b) einen dem Erwerb der Muttersprache vergleichbaren direkten Erwerb der Fremdsprache. Prominente Beispiele aus dem 20. Jahrhundert sind (a) die im Kontext der Reformpädagogik der zwanziger Jahre propagierte Direkte Methode, (b) die in den vierziger Jahren in den USA entwickelte “Army Method” für die schnelle Vermittlung von Grundfertigkeiten in ganz unterschiedlichen Sprachen an Angehörige der US-Army, die nach dem Zweiten Weltkrieg zur Audio-lingualen Methode mit dem Fokus auf dem Hören und Sprechen weiterentwickelt wurde, und (c) die in Frankreich am CREDIF ausgearbeitete Audio-visuelle Methode als Variante der Audio-lingualen Methode, die sich im Anfängerunterricht in extensiver Weise visueller Hilfen in Form von Bilderfolgen in Dia-Reihen zur “unmittelbaren” Bedeutungsvermittlung sowie zum Einprägen und Memorieren ganzer Dialoge anhand dieser Bilderserien bedient. Angestoßen wurde die Entwicklung dieser Methoden aber bereits im 19. Jahrhundert, insbesondere durch die unter dem Pseudonym “Quousque tandem” publizierte programmatische Schrift des Anglisten Wilhelm Vietor mit dem Titel “Der Sprachunterricht muß umkehren”, der eine radikale Abkehr von der im 19. Jahrhundert im Unterricht der klassischen Sprachen Latein und Griechisch praktizierten Lehr- und Lernverfahren forderte, die angeblich nur auf die Lektüre antiker Literatur und deren Übersetzung auf der Basis intensiven Grammatikunterrichts zielten. Allerdings greift diese Auseinandersetzung, historisch betrachtet, zu kurz. Denn mit Sicherheit bis weit in das 18. Jahrhundert hinein zielte der Unterricht im Lateinischen auch auf die produktive Beherrschung der la-
XXI. Morphologie und Nachbardisziplinen
teinischen Sprache in Schrift und Wort, war sie doch über Jahrhunderte hinweg die lingua franca Europas, die Sprache nicht nur der Kirche, sondern vor allem auch der Wissenschaften, in der sich die Vertreter aller Fakultäten mündlich wie schriftlich verständigten und ihre Werke publizierten. Das belegen die Unterrichtswerke selbst (beispielsweise schrieb Melanchthon für zwei seiner adligen Privatschüler eine Grammatik des Griechischen für Lernzwecke in lateinischer Sprache), die überlieferten Unterrichtskonzepte (beispielsweise Comenius 1642) und die Schulordnungen für die Lateinschulen jener Zeit, in denen ausdrücklich gefordert wurde, daß sich die Schüler nicht nur in der Schule, sondern auch außerhalb der Schule des Lateinischen befleißigen sollten. In diesen Schulen wurde von Anfang an lateinisch gelesen, geschrieben, gehört und gesprochen; ja selbst die Grundlagen der Grammatik wurden mündlich im katechetischen Verfahren gelehrt und gelernt, bei dem die Schüler ihren Lehrer und der Lehrer seine Schüler nach festgelegten Mustern fragte: “Partes orationis quot sunt? Octo. Quae? Nomen pronomen verbum adverbium participium coniunctio praepositio interiectio. Nomen quid est? Pars orationis cum casu corpus aut rem proprie communiterve significans. Nomini quot accidunt? Sex. Quae? Qualitas conparatio genus numerus figura casus. ... Conparationis gradus quot sunt? Tres. Qui? Positivus, ut doctus, conparativus, ut doctior, superlativus, ut doctissimus ...” (Donati de partibus orationis ars minor, Keil 1864, Hrsg.: 355)
Der Unterschied zwischen diesem einsprachigen Konzept des Lateinlernens und den einsprachigen Konzepten zur Erlernung moderner Fremdsprachen ist freilich gravierend: In diesem Lateinunterricht wurde zunächst und vor allem das grammatische Kategoriensystem der acht Wortarten, meistens durch Beispiele illustriert, gelernt, im Grunde also eine Mischung von Primär- und Metasprache, bevor die Lektüre begann und die sprachlichen Muster durch dauernde mündliche Wiederholung eingeprägt wurden. Daß dieser Unterricht für die noch recht jungen Schüler häufig genug zur Qual wurde, belegen vor allem die heftigen Attacken bedeutender Humanisten (wie Erasmus, Comenius und vieler anderer), die das Nachplappern von nicht Verstandenem als sinn- und erfolglose Mühe verurteilten und statt dessen forderten, daß nur das gelernt werden könnte und sollte, was zuvor in seiner Bedeutung verstanden worden war.
177. Sprachunterricht
Wie wurde und wird in einsprachigen Konzepten die Morphologie der Fremdsprache gelehrt und gelernt? Da auf die Muttersprache als Vergleichsbasis und als Erklärsprache zu verzichten ist, bleiben als Möglichkeiten vor allem (a) die direkte zielsprachliche Präsentation morphologischer Lerninhalte in Form von primärsprachlichen Beispielsätzen und tabellarischen Übersichten (Deklinations- und Konjugationsparadigmata), (b) die metasprachliche Formulierung der jeweiligen Regularitäten in der Zielsprache, (c) das Reproduzieren und Variieren der jeweiligen Formen in zum Teil monotonen Übungen im Stile der “pattern drills” der audio-lingualen Methode zwecks unbewußter Automatisierung dieser Formen, (d) das Hoffen darauf, daß die “eingedrillten” Formen in der “freien” Sprachproduktion automatisch richtig verwendet werden ⫺ eine Hoffnung freilich, die sich nicht erfüllt, wenn nicht vielfältige Übungen zum produktiven Sprachgebrauch hinzutreten. Was demgegenüber bei strikter Handhabung des einsprachigen Prinzips als Möglichkeit ausscheidet, ist die Zuhilfenahme der Muttersprache als Vergleichsbasis und Kommunikationsmittel. 2.2.2. Bilinguale Konzepte Bilinguale Konzepte propagieren den Einsatz der Muttersprache im Fremdsprachenunterricht in vielerlei Hinsicht, (a) zur Vermittlung der Bedeutung von Wörtern und Wortgruppen, (b) zur Formulierung grammatischer Regeln, (c) als Basis für den Vergleich der Zielsprache mit der Muttersprache, (d) zur Herüber- und Hinüberübersetzung sowohl zu Übungszwecken als auch als übergeordnetes Lernziel, (e) als Unterrichtssprache. Sie versprechen sich davon ein schnelleres und genaueres Verstehen der Bedeutung von Lexik und Grammatik, die Vermeidung von Interferenzfehlern durch Herausarbeitung der Unterschiede zwischen Zielsprache und Muttersprache und einen schnelleren Lernfortschritt. Beispiele aus dem 20. Jahrhundert sind: (a) die von Butzkamm (1989) u. a. propagierte bilinguale Methode, bei der in Abkehr von dem zu engen Korsett der audio-lingualen Methode ein maßvoller und jeweils didaktisch genau zu begründender Einsatz der Muttersprache angeraten wird, (b) Suggestopädie und Superlearning, wobei in strenger Dogmatik der Ausgang vom muttersprachlichen Wort, Satzteil und Satz vorgeschrieben wird, denen jeweils die zielsprachlichen Wör-
1931 ter, Formen und Strukturen unmittelbar zugeordnet werden, und (c) der Unterricht in Latein und Griechisch, in dem das Herüberübersetzen authentischer antiker Texte sowie deren Analyse und Interpretation das komplexe, erst allmählich zu erreichende Richtziel ist, auf das hin jedoch alle didaktischen Überlegungen und methodischen Konzepte von Anfang an ausgerichtet sind. Dieses didaktische Konzept des Lateinunterrichts wurde bereits im 19. Jahrhundert entwickelt und als generelles Konzept auf den Unterricht in den modernen Sprachen übertragen ⫺ was dann von Vietor und anderen (wie bereits gezeigt) heftig kritisiert wurde. Gleichwohl wird dieses Konzept der sogenannten Grammatik-Übersetzungsmethode bis in die Gegenwart zumindest in Teilaspekten (z. B. muttersprachliche Erklärungen, Übersetzungsübungen) noch praktiziert. Dies gilt insbesondere für homogene Lernergruppen im Land der jeweiligen Muttersprache, auch wenn die Zielsetzung nicht auf Lesen, Verstehen und Übersetzen reduziert bleibt, sondern daneben und darüber hinaus auch Hören und Sprechen entwickelt werden sollen ⫺ mit häufig unbefriedigendem Erfolg, weil der Unterricht zu sehr durch ausschließlich deduktive Grammatikvermittlung und durch Übersetzungsübungen dominiert wird. Ansätze des bilingualen Konzepts zur Vermittlung der Fremdsprache Latein sind bereits im frühen Mittelalter, wenn auch in sehr bescheidener Form, faßbar und weiten sich erst langsam in dem Maße aus, wie die lingua vernacula sich mit und an dem lateinischen Vorbild lexikalisch und grammatisch entwikkelt; das vollzieht sich im romanisch-sprachigen Raum (Italien, Spanien, Frankreich) wesentlich früher und schneller als beispielsweise im deutschsprachigen Raum, bis auch hier aufgrund großer und zum Teil überspannter Anstrengung das Deutsche aufgrund extensiver Nutzung der Möglichkeiten der Wort(neu)bildung etwa Mitte des 17. Jahrhunderts einen Zustand erreicht hat, der die Sprachwissenschaftler wie Schottel und die Poeten und Poetiker wie Harsdörffer und Zesen nicht nur die Gleichrangigkeit, sondern sogar die Überlegenheit des Deutschen gegenüber dem Lateinischen und aller zeitgenössischen europäischen Sprachen konstatieren läßt. Wie wurde und wird in diesen bilingualen Konzepten die Morphologie der Zielsprache gelehrt und gelernt? In den der antiken Grammatik in Gestalt der Ars Minor des Do-
1932 nat verpflichteten zweisprachigen Grammatiken des Lateinischen seit dem 15. Jahrhundert sind folgende Verfahren zu beobachten: (a) Zunächst und vor allem wird das grammatische Kategoriensystem der octo partes orationis in Frage und Antwort gelehrt, wobei den lateinischen Beispielwörtern für die einzelnen Kategorien Übersetzungen dieser Beispielwörter in die jeweilige Muttersprache folgen. (b) Die metasprachlichen Fragen nach den Kategorien und die Antworten und Definitionen werden in einem weiteren Schritt ebenfalls in die jeweilige Muttersprache übersetzt, wobei die muttersprachliche Formulierung der zielsprachlichen lateinischen nachfolgen oder vorangehen kann. (c) Die Formen des Nomens und des Verbs erscheinen in systematisch geordneter und übersichtlicher tabellarischer Form, manchmal in Form kurzer Sätze in der Zielsprache, wiederum gefolgt von der Übersetzung in die Muttersprache. (d) Das Erlernen des kompletten morphologischen Systems dient (in Verbindung mit dem Erlernen eines Grundwortschatzes anhand von zweisprachigen Vokabellisten oder Wörterbüchern) zunächst dem Übersetzen lateinischer Texte ins Deutsche, sodann in zunehmendem Maße auch dem Verfassen lateinischer Texte; dabei wird implizit und kontinuierlich das fremdsprachliche morphologische System geübt. (e) Der Einbezug der Muttersprache in den Sprachunterricht dient vorrangig dem Erwerb der Zielsprache Latein, darüber hinaus aber auch der Entwicklung und Ausbildung der muttersprachlichen Kompetenz. (f) Gelegentlich wird die Einführung in die Grammatik der Muttersprache auch dem Lateinunterricht vorgeschaltet und verstanden als Propädeutikum des Lateinlernens. Im gegenwärtigen Lateinunterricht finden sich viele dieser Merkmale wieder; ein wesentlicher Unterschied besteht freilich darin, daß von Anfang an mit lateinischen Texten gearbeitet wird, die es zu verstehen und zu übersetzen gilt; anhand dieser Texte erfolgt die Herausarbeitung morphologischer und syntaktischer Regularitäten, die in Form von Tabellen und Regeln sowie anhand von Beispielsätzen in grammatischen Übersichten in den Lehrbüchern und in eigenen Grammatiken des Lateinischen, in denen die Meta- und Erklärsprache die Muttersprache ist, zusammengestellt sind. Analoge Übersichten über die Morphologie und Syntax finden sich in Lehrbüchern und Grammatiken der modernen Fremdspra-
XXI. Morphologie und Nachbardisziplinen
chen; diese Übersichten kommen entweder ganz ohne Regelformulierungen aus oder bieten solche Formulierungen in möglichst einfacher Form in der Zielsprache oder in der Muttersprache. Sofern sie zweisprachig sind, kontrastieren sie darüber hinaus die morphologischen Systeme der Muttersprache und der Zielsprache, um das Verstehen zu erleichtern und um mögliche Interferenzfehler zu verhindern. 2.3.
Konzepte induktiven vs. deduktiven Lernens 2.3.1. Konzepte induktiven Lernens Konzepte induktiven Lernens im Sprachunterricht zielen auf einen Lernprozeß, in dessen Verlauf die Lernenden selbst sprachliche Regularitäten entdecken und internalisieren und dabei zugleich die kognitiven Fähigkeiten des Vergleichens, Abstrahierens und Generalisierens entwickeln. Diese Konzepte werden des weiteren mit dem Hinweis darauf begründet, daß die selbst entdeckten Regularitäten besser verstanden und behalten werden als vorformulierte Regeln, die auswendig zu lernen sind. Erkenntnisse der Gehirnforschung, nach denen Behaltensleistungen nach Differenziertheit und Dauer von der Intensität und Tiefe der kognitiven Verarbeitung von Informationen abhängen, tragen zur Fundierung dieser Konzepte bei: Je mehr Gehirnregionen gleichzeitig und parallel intensiv arbeiten, desto wahrscheinlicher ist es, daß die jeweilig bearbeiteten Sachverhalte aufgrund stabiler molekularer Veränderungen im sog. Langzeitgedächtnis gespeichert werden können. Dabei ist jedoch zu beachten, daß viele Regularitäten einer Sprache aufgrund ihrer Kompliziertheit, Komplexität und Anzahl der “Ausnahmen” auf diesem Wege von den Lernenden selbst nicht oder nicht in einem überschaubaren und vertretbaren Zeitraum zu entdecken sind. Beispiele solcher komplexen und komplizierten Regularitäten sind: Gebrauch des definiten, des indefiniten und des Null-Artikels im Deutschen, der von der Linguistik mittels 55 Regeln ⫺ die Ausnahmen nicht mitgerechnet ⫺ beschrieben worden ist; Deklination der Nominal-Gruppe im Deutschen. Daneben gibt es aber eine Vielzahl von Regularitäten, die aufgrund ihrer Reichweite (keine oder nur ganz wenige Ausnahmen) und internen, vergleichsweise einfachen Struktur von den Lernenden in einem methodischen Drei-Schritt selbst entdeckt
177. Sprachunterricht
werden können: (a) Sammeln/Ordnen (b) Vergleichen/Systematisieren und (c) Beschreiben. Die Propagierung und Realisierung dieser Konzepte induktiven Lernens stehen in einem unlösbaren Zusammenhang mit grundlegenden pädagogischen Konzepten, welche die Lernenden, ihre Lernprozesse und Lernstrategien konsequent fokussieren, denen erst in einem zweiten Schritt geeignete Lehrprozesse, welche die Lernprozesse ermöglichen, fördern, und sichern sollen, zuzuordnen sind. Beispiele für solche Konzepte finden sich überwiegend erst im 20. Jahrhundert, sind aber ansatzweise und indirekt bereits in früheren Jahrhunderten nachweisbar, zum Beispiel in der von Comenius propagierten und praktizierten Trias exempla ⫺ praecepta ⫺ exercitia, bei der die Beispiele den Regeln immer vorangehen sollen: “Per Exempla, Praecepta, Exercitia, docentur et discuntur omnia ... Exemplum semper praecedat, Praeceptum semper sequatur, Imitatio semper urgeatur ... Per Exempla facilius discitur, quam per Praecepta. Facilius tamen per utrumque junctim. Sed ut Exempla praecedant ...” (Comenius 1642: 102 ff.)
Der Prozeß der Ermittlung der Regularitäten (“praecepta”) auf der Basis konkreten sprachlichen Materials (“exempla”) wird in Konzepten induktiven Lernens also von jedem Lernenden selbst vollzogen. Diese Konzepte haben zusammen mit anderen Einflußfaktoren etwa seit den siebziger Jahren des 20. Jahrhunderts zur Entwicklung von Fremdsprachen-Lehrwerken und von pädagogischen Grammatiken geführt, in denen selbständiges und selbsttätiges Lernen im Mittelpunkt aller didaktischen Überlegungen und methodischen Konkretisierungen steht. Wie wurde und wird in Konzepten induktiven Lernens die Morphologie der Zielsprache gelernt? Die Lernenden selbst analysieren, spontan oder durch entsprechende Aufgabenstellungen dazu aufgefordert, das präsentierte sprachliche Material mit dem Ziel, Regularitäten der Formenbildung und des Gebrauchs dieser Formen zu entdecken und in Paradigmata und Regeln (zunächst in der Muttersprache, dann zunehmend auch in der Zielsprache) abzubilden. Diese Analyse läßt sich in die drei bereits erwähnten Analyseschritte gliedern: (a) Sammeln/Ordnen (b) Vergleichen/Sytematisieren und (c) Beschreiben. Sie operiert auf und vor dem Hintergrund sowohl des muttersprachlichen Kön-
1933 nens und des sich allmählich parallel herauskristallisierenden Wissens, als auch des jeweils erworbenen zielsprachlichen Wissens und Könnens. Der Analyse müssen sodann vielfältige Übungen im Erkennen und im Gebrauch der jeweiligen Formen folgen. Zu beachten ist, daß die einzelnen morphologischen Subsysteme (z. B. das Tempussystem) weder in einem einzigen Zugriff noch in toto behandelt und internalisiert werden können, sondern in mehreren, aufeinander aufbauenden Zugriffen und in einem Grad von “Vollständigkeit”, der didaktisch (auf der Ebene der Lernzielermittlung und -beschreibung) und nicht primär linguistisch (im Sinne des Vollständigkeitspostulats) zu definieren ist. Hochkomplexe Strukturen mit einem hohen Maß an Irregularitäten jedoch sind mit Konzepten induktiven Lernens allein nicht zu bewältigen. 2.3.2. Konzepte deduktiven Lernens Konzepte deduktiven Lernens im Sprachunterricht zielen auf einen Lernprozeß, in dessen Verlauf die Lernenden die in z. T. komplexen und zudem kompliziert formulierten Regeln generalisierten Regularitäten einer Sprache lesen und auf ausgewählte, meist kontextlose Beispielwörter und -sätze anwenden. Die Lernenden werden dabei von Anfang an mit abstrakter linguistischer Metasprache in komprimierter Form konfrontiert, welche das Verstehen erschwert, wenn nicht gar verhindert, weil sie vor dem konkreten Gegenstand, der Sprache, die abstrakte und komprimierte Beschreibung dieses Gegenstands kennenlernen und verstehen sollen. Ein solches Verfahren läßt sich pädagogisch und lernpsychologisch schwerlich begründen. Gleichwohl scheint es sich immer noch einer gewissen Beliebtheit zu erfreuen. Über die Gründe dafür läßt sich nur spekulieren: Sicherheit seitens der Lehrperson aufgrund ihres überlegenen Wissens, scheinbare Zeitersparnis, auch in anderen Fächern praktizierte Vermittlung von Fachwissen, Verankerung in der Tradition der Vermittlung des Lateinischen und Griechischen. Wie wurde und wird in Konzepten deduktiven Lernens die Morphologie der Zielsprache gelehrt und gelernt? Lehrbücher und Lehrpersonen präsentieren die jeweiligen Formen der zu erlernenden Sprache in enger Anlehnung an die linguistische Beschreibung dieser Formen in linguistischen Grammatiken in Form von Paradigmata, Regeln und
1934
XXI. Morphologie und Nachbardisziplinen
Ausnahmen von den Regeln in der Muttersprache der Lernenden; bei Einbeziehung des kontrastiven Aspekts erfolgt darüber hinaus ein Vergleich mit den jeweiligen Formen der Muttersprache, insbesondere dann, wenn systematische Unterschiede negativen Transfer möglich oder wahrscheinlich erscheinen lassen. Als Beleg für die jeweiligen Regeln und Ausnahmen folgen sodann Beispielwörter und Beispielsätze. Die Lernenden müssen als erstes die schwierigen metasprachlichen Regelformulierungen verstehen, bevor sie sie auf primärsprachliche Wörter, Sätze und Texte anwenden. Während dieses Verfahren bei der Rezeption von Texten in schriftlicher Form sinnvoll sein kann, kann es bei der schriftlichen und allemal bei der mündlichen Produktion dazu (ver)führen, Sätze durch parallele Anwendung einer Vielzahl von Einzelregeln konstruieren zu wollen, was die Konzentration auf den Inhalt der Aussage erschwert und die Flüssigkeit der Äußerung erheblich beeinträchtigen kann. Deshalb kommen auch deduktive Konzepte ohne viele und vielfältige Übungen, die auf “Automatisierung” der zu erlernenden Formen zielen, nicht aus.
3.
Konzepte pädagogischer Grammatiken für den Fremdsprachenunterricht
Konzepte pädagogischer Grammatiken für den Fremdsprachenunterricht basieren nicht alleine auf linguistischen Beschreibungen der jeweiligen Zielsprache, sondern darüber hinaus auf lernpsychologischen Erkenntnissen allgemein sowie Erkenntnissen der Kognitionswissenschaft (Cognitive Sciences) und
der Zweitsprachenerwerbsforschung (s. Art. 166) sowie zunehmend auch der Fremdsprachenlehr- und -lernforschung. Da dem Fremdsprachenunterricht aufgegeben ist, das Verstehen und den Gebrauch der Fremdsprache bei den Lernenden zu entwickeln, führt das zu Funktionsbestimmungen und Realisierungsformen von Grammatik, die sich von Funktionsbestimmungen und Realisierungsformen rein linguistischer Grammatiken zum Teil deutlich unterscheiden. Autoren pädagogischer Grammatiken für den Fremdsprachenunterricht bedienen sich deshalb auch morphologischer Theorien und der auf diesen Theorien basierenden Darstellungen der Morphologie in eklektischer und souveräner Weise: Sie sichten den Formenbestand einer Sprache und die den Formen, ihren Funktionen und ihrem Gebrauch innewohnenden Regularitäten. Sie analysieren diese Regularitäten im Hinblick auf ihre Reichweite (den Grad der ihnen innewohnenden Irregularitäten) und ihren Komplexitätsgrad. Die Ergebnisse dieser Analysen sind eine wichtige Basis für die Auswahl, Progression und Präsentation unter lernpsychologischen Gesichtspunkten. Der Wert einzelner morphologischer Theorien, wie beispielsweise des Strukturalismus, bestimmt sich danach, ob und gegebenenfalls bis zu welchem Grad die auf ihnen basierenden Beschreibungen den Anforderungen des Lehr- und Lernprozesses entsprechen. Abstrakte Darstellungen sind von weit geringerem Wert als Darstellungen, die anhand von kontextualisierten Beispielen die jeweiligen Regularitäten besser verstehbar und behaltbar machen. Metasprachlich möglichst einfache Regelformulierungen mit Fokussierung
Linguistische Grammatik
Lerner-Grammatik
Totalität (Ausnahmen von der “Regel” besonders wichtig)
Auswahl
Abstraktheit (der Beschreibung/Darstellung)
Konkretheit/Anschaulichkeit (der Abbildung/Darstellung)
Kürze (der Darstellung)
Ausführlichkeit (der Darstellung der als wichtig erkannten Elemente)
Keine lernpsychologischen Vorgaben/ Rücksichten
Lernpsychologische Kategorien (Verstehbarkeit, Behaltbarkeit, Anwendbarkeit)
Tab. 177.1: Wesentliche Unterschiede zwischen einer linguistischen und einer Lerner-Grammatik (Schmidt 1992: 163)
177. Sprachunterricht
des Wesentlichen sind von weit größerem Wert als in zum Teil eigenwilliger Metasprache gefaßte Regelformulierungen mit Fokussierung aller nur denkbaren Ausnahmen und Ausnahmen von den Ausnahmen. Pädagogische Grammatiken für den Fremdsprachenunterricht lassen sich deshalb auch nicht einfach aus linguistischen Grammatiken durch Kürzung und Vereinfachung deduzieren, sondern sind Sprachbeschreibungen sui generis (s. Art. 11). Zentrale Funktionen solcher Grammatiken sind: (a) Sie sollen die Einsicht der Lernenden in die Regularitäten grammatischer Formen, Strukturen und Funktionen fördern. (b) Sie sollen das Einprägen, das Behalten und die Abrufbarkeit dieser Regularitäten unterstützen. (c) Sie sollen die möglichst fehlerlose Beherrschung der Fremdsprache entwickeln helfen. (d) Sie sind damit “nur” Instrument, nicht Selbstzweck.
4.
Grammatik im muttersprachlichen Unterricht
Grammatik im muttersprachlichen Unterricht unterscheidet sich prinzipiell ebenso von rein linguistischen Grammatiken wie Grammatik im Fremdsprachenunterricht, insofern auch bei dieser Grammatik lernpsychologische Erkenntnisse sowie pädagogische und didaktische Konzepte eine entscheidende Rolle spielen. Grammatik im muttersprachlichen Unterricht setzt ⫺ im Unterschied zu Grammatik im Fremdsprachenunterricht ⫺ die je nach Altersstufe und individuellen Lernbiographien unterschiedlich weit fortgeschrittene Beherrschung der Muttersprache voraus. Vor diesem Hintergrund dient sie vor allem der kognitiven Auseinandersetzung mit den Formen und Strukturen der eigenen Sprache sowie deren Funktionen und Leistungsfähigkeit in Wort und Schrift. Damit lassen sich Zielsetzungen auf unterschiedlichen Ebenen verbinden, die zudem noch nach Altersstufen variieren können, z. B.: Unterstützung beim Erlernen der Orthographie (auf der Primarstufe) mittels spielerischem und entdeckendem Umgang mit der eigenen Sprache; Einsicht in grundlegende Regularitäten und Funktionsweisen der eigenen Sprache zwecks
1935 Verbesserung der mündlichen und Entwicklung der schriftlichen Kompetenz sowie zwecks Förderung der Fähigkeit zur Textanalyse und -interpretation (auf der Sekundarstufe I); Entwicklung der Fähigkeit zu Analyse, Generalisierung und Abstraktion (“formale” Bildung); Kennenlernen und Verstehen linguistischer Kategorien und Terminologie(n) (und damit in gewisser Weise Vorbereitung und Entlastung des nachfolgenden oder parallel einsetzenden Unterrichts in einer ersten Fremdsprache). Welchen Stellenwert hat Morphologie in diesen Zielhorizonten? Die Behandlung morphologischer Themen im muttersprachlichen Unterricht dient vor allem dazu, die Morphologie der eigenen Sprache zu entdecken und die metasprachliche Beschreibung morphologischer Sachverhalte zu verstehen. Das impliziert zunächst und vor allem das (Kennen-)Lernen grundlegender Kategorien, insbesondere der Wortarten und ihrer Akzidentien (z. B. Verb; die Tempora wie Präsens, Perfekt usw., die Genera wie Aktiv und Passiv, die Modi wie Indikativ, Konjunktiv usw.); unterschiedliche linguistische Verfahren, Modelle und Terminologien stellen dabei ein großes Problem dar, insofern sie zu erheblichen Konfusionen in den Köpfen der Lernenden führen können, wenn sie beispielsweise auf der Primarstufe mit anderen Begriffen konfrontiert werden als auf der Sekundarstufe I, oder wenn zeitlich parallel im muttersprachlichen Unterricht andere Begriffe für dieselbe Sache verwendet werden als im fremdsprachlichen Unterricht. Des weiteren war und ist Grammatik im muttersprachlichen Unterricht immer beeinflußt von Entwicklungen in der Sprachwissenschaft: Während bis weit ins 20. Jahrhundert hinein Verfahren, Modelle und Kategorien der lateinischen Grammatik die Grundlage grammatischer Darstellungen in Sprachbüchern gewesen sind, haben etwa seit der Mitte des 20. Jahrhunderts Entdeckungsprozeduren, Modelle, Kategorien und Begrifflichkeiten des Strukturalismus verstärkt in Sprachbüchern und damit auch im Sprachunterricht Platz gegriffen. Exemplarisch verwiesen sei auf den “Sprachspiegel” von Hans Glinz (1969, Hrsg.), in dem Entdeckungsund Analyseprozeduren des Strukturalismus zu von den Schülern handhabbaren operationalen Werkzeugen für den entdeckenden Umgang mit der eigenen Sprache gemacht werden: Der Permutationstest wird zur Umstellprobe, mit deren Hilfe Satzglieder ent-
1936 deckt und abgegrenzt werden sollen; der Deletionstest wird zur Weglaßprobe, mit deren Hilfe Kernsätze entdeckt werden sollen; die Infinitivierung wird dazu verwendet, das Subjekt (oder die Nominalphrase des Knotens Satz im strukturalistischen Syntax-Modell) von dem Rest des Satzes (oder der Verbalphrase des strukturalistischen Modells) zu trennen, und so fort. Zugleich werden allerdings auch neue Begrifflichkeiten eingeführt (z. B. Grundgröße für Subjekt, Gleichgröße für Prädikatsnomen, Anteilgröße für Genitivobjekt, Zuwendgröße für Dativobjekt usw.), die sich weder im muttersprachlichen Deutschunterricht generell noch gar im Fremdsprachenunterricht durchsetzen konnten. Neuere Entwicklungen innerhalb der Linguistik, insbesondere die Erweiterung des Analysegegenstands über Morphem, Syntagma und Satz hinaus auf Text, Kommunikation, sprachliches Handeln, Spracherwerb und die damit einhergehende Untergliederung der Linguistik in weitere Teildisziplinen (Textlinguistik, Pragmalinguistik, Psycholinguistik u. a.) haben sich ebenfalls auf den muttersprachlichen Unterricht und den Teilbereich “Reflexion über Sprache”, innerhalb dessen Grammatikunterricht wiederum einen Teilbereich darstellt, ausgewirkt. In ihrem Buch “Der andere Grammatikunterricht” fragen Boettcher & Sitta “… nach Begründungen, Prinzipien und Realisierungsmöglichkeiten von Grammatikunterricht im Rahmen des Lernfelds ‘Reflexion über (sprachliche) Kommunikation’ im Kontext eines insgesamt schüler- und situationsbezogenen Deutschunterrichts” (Boettcher & Sitta 1978: 5). Welche Bedeutung kommt in einem solchen Konzept noch der Vermittlung der Einsicht in die Morphem-, Phrasen- und Satzstrukturen der Muttersprache zu? Generell läßt sich feststellen, daß die Reflexion über diese Strukturen instrumentalisiert wird für die Lösung von sprachlichen und kommuni-
XXI. Morphologie und Nachbardisziplinen
kativen Problemen in für die Schüler relevanten Situationen außerhalb der Schule, aber auch innerhalb der Schule und innerhalb einzelner Fächer. Es bleibt aber zu fragen, ob und in welcher Weise in diesem anderen Grammatikunterricht eine auch situationsunabhängige systematische Vermittlung der o. g. Strukturen erfolgen kann oder darf, welche die Schüler kennengelernt haben müssen, wenn sie mit ihnen arbeiten sollen.
5.
Zitierte Literatur
Boettcher, Wolfgang & Sitta, Horst (1978), Der andere Grammatikunterricht. München usw.: Urban & Schwarzenberg Butzkamm, Wolfgang (1989), Psycholinguistik des Fremdsprachenunterrichts. Natürliche Künstlichkeit: Von der Muttersprache zur Fremdsprache. Tübingen: Francke Comenius, Johann Amos (1642), Didacticorum Operum Pars II, (Text II: Novissima Linguarum Methodus, Fundamentis Didacticis solide superstructa &c.) [Konstanz: Archiv für Fremdsprachenvermittlung 1978] Glinz, Hans (1969, Hrsg.), Deutscher Sprachspiegel: Sprachgestaltung und Sprachbetrachtung. Düsseldorf: Schwann Gwynn, Aubrey (1926), Roman Education from Cicero to Quintilian. Oxford: Oxford University Press [New York: Teachers College Press o. J. (Classics in Education 29)] Keil, Heinrich (1864, Hrsg.), Probi Donati Servii Qui feruntur De arte grammatica libri. Leipzig: o. V. [Hildesheim: Olms 1961] Schmidt, Reiner (1992), “Das Konzept einer Lern(er)-Grammatik für DaF und seine linguistischen und lernpsychologischen Grundlagen”. In: Leirbukt, Oddleif & Lindemann, Beate (Hrsg.), Psycholinguistische und didaktische Aspekte des Fremdsprachenlernens. Tübingen: Narr, 159⫺177
Reiner Schmidt, Bielefeld (Deutschland)
Verzeichnis der Abkürzungen / List of Abbreviations
1. Deutsch ⵰ 1 2 3 abl abs adjr advr ag akk akt all aor appl art asrt aug aux dat def dem dir distr do du dyn erg exist f fakt fok fut gen ger habit imp impf inakt
Null 1. Person 2. Person 3. Person Ablativ Absolutiv (Kasus/Funktion des ergativischen Systems) Adjektiv(alis)ator Adverbialisator Agens Akkusativ Aktiv Allativ Aorist Applikativ Artikel Assertiv Augmentativ Auxiliar Dativ definit Demonstrativ Direktional Distributiv direktes Objekt Dual dynamisch Ergativ Existential femininum Faktitiv Fokus Futur Genitiv Gerundium Habitualis Imperativ Imperfekt inaktiv
inch ind indef inf instr int ints io ipfv irr kaus kln kmpr komit komp kond konj konn konsek kop lig lok m med medpass n neg nom nprt nr obj obl oblg opt part pass pat pf pfv pl plup poss
Inchoativ Indikativ indefinit Infinitiv Instrumental Interrogativ Intensiv indirektes Objekt Imperfektiv Irrealis Kausativ Nominalklasse n Komparativ Komitativ Kompletiv Konditional Konjunktiv Konnektor Konsekutiv Kopula Ligatur Lokativ maskulinum medial (nicht-finaler Satz) Mediopassiv neutrum negativ Nominativ Nicht-Präteritum Nominalisator Objekt Oblik Obligativ Optativ Partizip Passiv Patiens Perfekt Perfektiv Plural Plusquamperfekt Possessiv (pronominales Element)
1938 pot pr‰s pr‰t priv prog prolat proz rdp refl rell rez s sg stat sup top ven vok vr zirk
Verzeichnisse und Sprachenkarten
Potentialis Präsens Präteritum Privativ Progressiv Prolativ Prozessiv Reduplikation reflexiv Relational(isator) reziprok intransitives Subjekt Singular Stativ Superlativ Topik Venitiv Vokativ Verbalisator Zirkumstantiell
2. English ⵰ 1 2 3 abl abs absl abstr acaus acc acnnr acr act adess adjr adm advr ag agnr agr agt all ana and
zero 1st person 2nd person 3rd person ablative (separative) absolutive (case/function of ergative system) absolute (free form of noun) abstract anticausative (deagentive) accusative action nominalizer actor active adessive adjectiv(al)izer admonitive adverbializer agentive agent nominalizer agreement agent allative anaphoric andative
anim aor apass appl art asrt ass at aug aux avers ben caus cl cln cmpr coll comit compl cond conj conn consec constr cont cop d1 d2 d3 dat de decl def dem des det detr dim dir dist distr don dr ds du dub dur elat emph eqt erg evid
animate aorist antipassive applicative article assertive (as)sociative actor topic augmentative auxiliary aversive (abessive) benefactive causative classifier noun class n comparative collective comitative completive conditional conjunctive connector consecutive construct continuous copula deictic of 1st person deictic of 2nd person deictic of 3rd person dative dual exclusive declarative definite demonstrative desiderative determiner detransitivizer diminutive directional distal (remote) distributive donative direct different subject dual dubitative durative elative emphasizer/emphatic equative ergative evidential
1939
Verzeichnis der Abkürzungen
exist exper f fact fin foc fut gen ger gnr habit hest hodpst hon hort hum ill imm imp impf impr inact inan inch incompl ind indef indep iness inf infr ingr instnr instr int ints inv invis io ipfv ips irr iter juss loc log lt m mannr med mid
exist(ential) experiential feminine factitive finite focus future genitive gerund (conjunctive participle) generic habitual (customary) hesternal hodiernal past honorific (ad-/co-)hortative human illative immediate imperative imperfect impersonal inactive inanimate inchoative incompletive indicative indefinite independent inessive infinitive inferential ingressive (inceptive) instrument nominalizer instrumental interrogative intensive (reinforcement marker) inversive invisible indirect object imperfective impersonal passive irrealis iterative jussive locative logophoric locative topic masculine manner nominalizer medial (non-final clause) middle (medio-passive)
n narr neg nfin nfut nhum nom npst nr obj obl oblg obv opt part pass past pat patnr pe pf pfv pi pl plup pnct poss pot pres pret priv proc prog proh propr prox prtv prx pt purp quot rec recpst rdp refl rel rell rem rempst rep res rls
neuter narrative negative non-finite non-future non-human nominative non-past nominalizer object oblique obligative obviative optative participle passive past patient patient nominalizer plural exclusive perfect perfective plural inclusive plural pluperfect punctual possessive (pronominal element) potential present preterite privative processive progressive prohibitive proprietive proximal (local deixis) partitive proximate (pronominal) patient topic purposive quotative reciprocal recent past (immediate past) reduplication reflexive relative relational(izer) remote remote past repetitive resultative realis
1940 rprt s sbj sens seq sg spec sr ss stat subess subj
Verzeichnisse und Sprachenkarten
reportative intransitive subject subject sensory evidential sequential singular (restricted) specific subordinator same subject stative subessive subjunctive
sup supess top trnsl trl trr ugr ven voc vol vr vis
superlative superessive topic translative trial transitivizer undergoer venitive vocative volitive verbalizer visible
Sprachenkarten/Language maps 1. Afrika/Africa
119. 121. 139. 140. 141.
English (Indo-European: Germanic) Franc¸ais (Indo-europe´en: Roman) Turkana (Nilotic) Twi (Kwa) Kinyarwanda (Bantu)
119. English (Indo-European: Germanic)
121. Franc¸ais (Indo-europe´en: Roman)
137. Diyari (Pama-Nyungan)
1942 Verzeichnisse und Sprachenkarten
2. Australien/Australia
119. English (Indo-European: Germanic) 120. Deutsch (Indogermanisch: Germanisch) 121. Franc¸ais (Indo-europe´en: Roman)
122. Russisch (Indogermanisch: Slawisch) 123. Altgriechisch (Indogermanisch) 124. Finnish (Finno-Ugric)
125. Hebrew (Semitic) 126. Türkisch (Turk) 127. Hunzib (North-East Caucasian)
Sprachenkarten
1943
3. Europa und Nahost/Europe and Near East
122. Russisch (Indogermanisch: Slawisch)
128. Ketisch (Jenisseisch)
1944 Verzeichnisse und Sprachenkarten
4. Nordasien/North Asia
1945
Sprachenkarten
5. Nordamerika/North America
119. English (Indo-European: Germanic) 121. Franc¸ais (Indo-europe´en: Roman) 129. West Greenlandic (Eskimo)
130. Koyukon (Athapaskan) 131. Montagnais/Innu-aimun (Algonquian)
1946
Verzeichnisse und Sprachenkarten
6. Südamerika/South America
119. English (Indo-European: Germanic) 121. Franc¸ais (Indo-europe´en: Roman) 132. Guaranı´ (Tupi-Guaranı´)
134. Quechua (Quechua) 135. Yagua (Peba-Yaguan)
136. Tagalog (Austronesian)
138. Wambon (Awyu)
142. Vietnamesisch (Viet-Muong)
Sprachenkarten
1947
7. Südostasien/South East Asia
Namenregister / Index of names A Aaseb, Ivar 146 Abbi, Anvita 546, 561 Abbott, Miriam 1097 Abercrombie, David 97 Abhyankar, Kahinatha Vasudev 42 Abney, Steven 1896 Abondolo, Daniel 1080 Abraham, Werner 1274 Achma´nova, Ol’ga S. 133 Ackema, Peter 409 Ackrill, John L. 211 Acson, Veneeta 585 Adams, Catherine 1820 Adams, G. Brendan 88 Adams, Karen 1016, 1475 Adams, Valerie 960, 155, 1255 Adelaar, Willem F. H. 164, 1206, 1453 f., 1458, 1462, 1654 Adelung, Johann Christoph 91, 113, 668, 698, 1267, 1682, 1690 Admoni, Vladimir G. 611, 665, 668, 677, 681 f., 1270, 1906 Adodurov, Vasilij 1300 Adontz, Nicolaus 59 Adrados, Francisco Rodriguez 122 Aelfric, Abbot of Eynsham 86, 696 Aerts, Willem Joan 656 Aeschylus 1316 Agha, Asif 993 Agricola, Mikael 1328 Ahlquist, Anders 88 Aichinger, Carl Friedrich 698 Aikhenvald, Alexandra Y. 1024, 1031 f., 1034⫺1042, 1070, 1197 Aissen, Judith L. 324, 1135 Aitchison, Jean 206, 386, 1771⫺ 1774, 1781, 1783⫺1785 Ajnewtegin, M. V. 315 Akamatsu, Tsutomu 493 Akmajian, Adrian 818, 1005 Aksakov, Konstantin S. 608 Aksu-Koc¸, Ayhan 1364, 1800 Al, Bernard P. F. 163 Al-Astarabadhi, Radi 67, 71 f. Aldama y Guevara, Joseph 1433 Alerten, D. J. 6
Alexander der Große 55 Al-Farra?, Abu Zakariyya 67, 72 Algeo, John 954, 957 f. Ali, Abdul Sahib Mehdi 1750 Allan, Keith 1016, 1018, 1060, 1067 Allen, Barbara J. 924 f. Allen, James 1893 Allen, John 1896 Allen, Margaret Reece 189, 199, 351⫺353, 356⫺358, 411, 541, 907, 948 Allen, Robert E. 1883 Allen, Shanley E. M. 1821 Allen, William Sidney 93, 218, 437 Allerton, D. J. 696 Allie`res, Jacques 488 Allport, D. A. 1790 Allwood, Jens 1195 Almazan, M. 1822 Almeida, Matthew 767 Al-Mubarrad, Ibn Yazid 68, 70, 703 Alonso, Da´maso 1582 Alpher, Barry 719, 1034 Alsdorf-Bolle´e, Annegret 1708 Alsina, Alex 325 Althaus, Hans Peter 435 Altmann, Gabriel 1239⫺1243 Altner, Günter 16 Ameka, Felix 719, 1092 Amith, Jonathan D. 711 Ammann, Hermann 667 Amritavalli, A. 859 Anceaux, Johannes C. 164, 166, 803 Andersen, Elaine 750 Andersen, Henning 273, 383 Andersen, Karl Gustaf 1620⫺ 1623 Andersen, Paul Kent 1086, 1149, 1159, 1212, 1216 Andersen, Roger 1654, 1806⫺ 1808 Anderson, Arthur J. O. 1452 Anderson, John M. 507, 812 f., 1596 Anderson, Stephen R. 6, 20, 27, 32 f., 153, 197, 199⫺202, 205, 207, 233, 235, 261, 268, 316, 320 f., 325 f., 328 f., 337, 339, 341, 346 f., 350 f., 353,
358, 361, 365 f., 373, 395 f., 406, 410, 412, 429, 437, 439, 442, 453, 456, 460, 476, 489, 499 f., 502, 508, 526, 542, 546, 549, 571, 580, 584 f., 587, 589, 625, 627 f., 716, 722, 735, 739, 742 f., 753, 784, 814, 842, 851, 854, 877, 882, 976, 983, 986, 988, 991⫺993, 1005, 1047, 1049, 1252, 1655, 1732, 1772, 1790, 1796 f., 1801 Andrade, Manuel J. 560 f. Andre´, Jacques 1218 Andreev, Nikolaj D. 133 Andrews, Avery D. 877 Andrews, Edna 279 Andrews, J. Richard 621 f., 1440, 1447 Anshen, Frank 206, 346, 348, 852 Antinucci, Francesco 1807 Anttila, Raimo 100, 121, 281, 856, 1086, 1662, 1667 Antworth, Evan L. 1833, 1896, 1898 Anward, Jan 710 Aoki, Haruo 998 Aoun, Joseph 1265 Aoyama, Takashi 988 Aphek, Edna 1346 Apollonius Dyscolus 55, 64 f. Appel, Elsbeth 1271 Appleyard, D. L. 809 Apresjan, Jurij D. 514, 1546 Arce-Arenales, Manuel 1162 Archaimbault, Sylvie 697 Archangeli, Diana 346 Arens, Hans 7, 91⫺93, 674 Arensen, Jonathan E. 1100 Ariel, Mira 983 Ariel, Shlomo 1345, 1347 f. Aristarchus 55 Aristophanes 53, 1316 Aristoteles 2, 53 f., 57, 77, 83, 211 f., 264, 459, 600, 695, 732, 734, 1168, 1172, 1182, 1326 Arlotto, Anthony 1664 Armagost, James 1005 Armon-Lotem, Sharon 1796 ´ rnason, Kristja´n 1607 A Arnauld, Antoine 698 Arndt, Walter W. 1269
1950 Arnim, Ioannes von 54 f. Arnott, Donald Whitehorn 762, 1134 Aronoff, Mark 31, 165, 189, 196⫺198, 200⫺203, 205 f., 230, 249, 252 f., 290, 308 f., 341, 346, 348, 350, 364, 408, 410, 412, 414, 419, 423, 430, 433, 445, 477, 479, 541, 546, 583, 596, 633, 657, 834, 838, 841 f., 848, 852, 854, 856, 868, 870, 873, 878⫺880, 882, 887, 899, 932, 1255 f., 1259⫺ 1262, 1637, 1639 f., 1642 f., 1780, 1790 f. Aronson, Howard I. 597 Arrive´, Michel 1291 Arutjunowa, N. D. 680 f. Asdahl Holmberg, Märta 890, 1277 Asher, R. E. 738 Ashton, Ethel O. 762, 1120, 1123 Askedal, John Ole 1268, 1273 Asporius (Asperius, Asper) 78 Atkins, Beryl T. 1885 Atkins, John 178 Attla, Catherine 1409 Augst, Gerhard 420, 905, 1267 f., 1272, 1622 f., 1686, 1689, 1692 Augustinus 76 Auroux, Sylvain 39, 693 f., 697 f., 701 Austerlitz, Robert 1217 Austin, Peter 987, 1078, 1088, 1158 f., 1202⫺1205, 1491⫺ 1495, 1497 f. Averroes 77 Avicenna 77 Avram, Andrei 493 Axelrod, Melissa 917⫺919, 1402 Axelsson, Monica 1413 Ayto, John 1674 Azzopardi-Alexander, Marie 760
B Baarle, Peter van 823 Baayen, Rolf Harald 164 f., 206, 299, 346, 1639, 1773, 1779, 1898 Babby, Leonard H. 1157 Bach, Adolf 1905 Bach, Emmon 201, 727, 851 Bacher, Wilhelm 74, 703 Back, Otto 1563 Bacon, Francis 1816 Bacon, Roger 92 Baddeley, Alan D. 1822 Badecker, William 364, 1784, 1817, 1819
Verzeichnisse und Sprachenkarten Bader, Franc¸oise 1325 Bader, Thomas 1481 Bahner, Werner 104 Bai, Juyi 1737 Bailey, Charles-James N. 289 Bailey, Nathalie 1806 Bakema, Peter 1048 Baker, Mark C. 199, 203 f., 315, 321⫺326, 329⫺332, 368, 407, 542, 916, 922, 924 f., 1135, 1137, 1140 f. Baker, William 1823 Bakker, Peter 1646 f., 1659 Balbi, Adriano 91, 93 Baldi, Philip 1214, 1640 Baldinger, Kurt 1706 f., 1911 Ballweg, Joachim 1273 f. Bally, Charles 211, 216, 248, 514, 520, 945 Bammesberger, Alfred 823 Banerjee, Satya Ranjan 42 Bani, Ephraim 1034 Barac-Cikoja, Dragana 1922 Bar-Adon, Aaron 1343 Baratin, Marc 694 Barbaud, Philipp 837 Barbe, Katharina 452 Barber, Charles C. 759 Baresˇ, Karel 967 Barkali, Sha’ul 1349 Barnes, Janet 1022 Barnetova´, Vilma 1305 f., 1308 f. Barnhart, Clarence L. 836 f. Barnhart, Robert K. 930, 1674 Barre´, Louis 16 Barret, Anthony 1507 Barron, Roger 1019, 1024 Bartke, Susanne 1269, 1820 f. Barton, John 88, 696 Bartsch, Werner 1273 Barz, Irmhild 356, 887⫺889, 891⫺894, 898 f., 902, 931 f., 950, 958, 1267, 1276 Basbøll, Hans 341, 384 f., 433, 584 Bassarak, Armin 1359 f., 1363 Basset, Louis 694 Bastiaanse, Roelien 1818 Bat-El, Outi 556 Bates, Dawn 560 Bates, Elizabeth 289, 795, 1800, 1807 Ba´tori, Istva´n 984 Battistella, Edwin 273, 276, 278 f. Baudouin de Courtenay, Jan 19 f., 91, 98, 125 f., 134, 205, 344 f., 405, 417, 420, 436 f., 451 f., 454, 460, 490, 492, 508, 613, 1601 f., 1616, 1620 Bauer, Heinrich 109 Bauer, Laurie 5, 155, 220, 235, 237, 248, 251 f., 319, 350,
354, 356, 405, 463, 474, 526, 532 f., 542 f., 546, 556, 582, 588, 617, 622, 638, 832⫺835, 837 f., 874, 886, 888⫺895, 899, 901, 931, 933, 938, 945, 955, 957, 968, 979, 993, 1047, 1248, 1255, 1259 f., 1264, 1626⫺1629, 1631 f., 1636⫺ 1638, 1640 f., 1643, 1781, 1888, 1890, 1922 Bauer, Winifred 766 Baum, Richard 689 Bausch, Karl-Heinz 1274 Bautista, Maria L. S. 1481 Bayer, Josef 364, 829, 1817 f. Bazell, Charles E. 1225 f. Beard, Robert 153, 200, 205, 247, 250, 266, 320, 332, 353, 367, 417, 454, 511, 622, 627, 657, 842, 849, 866 Beauze´e, Nicolas de 698 Be´bian, Roch-Ambroise Auguste 1555 Bech, Gunnar 1267 f. Be´chade, Herve´ D. 1287, 1291, 1294 Becker, Alton 999 Becker, Claudia 1560 Becker, Karl Ferdinand 19, 104 f., 107, 109 f., 113 f., 157, 162, 669 f., 672 Becker, Thomas 161, 231, 298, 456, 866, 883, 1275, 1278 Becker, Ulrich 1563 Be´clard, Pierre Augustin 17 Beekman, John 1831 Behaghel, Otto 112, 159, 753, 1233, 1680, 1690, 1692 Behnstedt, Peter 662, 1742 f. Behrens, Leila 1067 f. Beifuss, Karin 984 Beito, Olav Toreson 147 Bellmann, Günther 894 Bellugi, Ursula 1822 f. Belvalkar, Shripad Krishna 41 Ben Amotz, Dahn 1356 Ben Asher 74 Ben Asher, M. 1345 Ben Labrat, Dunash 703 Bendix, Edward 178, 1091 Bendor-Samuel, David 1194 Bendor-Samuel, John T. 587 Benedict, Paul K. 1733 f., 1737 Benesˇ, Brigit 97 Bennis, Hans 829 Benson, Morton 1890 Benveniste, E´mile 122, 142, 277, 286, 614, 751, 785, 813, 984, 998, 1214, 1218, 1295, 1592, 1595 Berendsen, Egon 389 Berezin, Fedor M. 135, 1308 Berg, Helma E. van den 1367, 1370⫺1373
1951
Namenregister Berg, Thomas 1773, 1791 Bergenholtz, Henning 405, 408, 439, 445, 452 f., 489, 526, 542, 596, 600, 676 f., 683, 699 f., 867, 891, 893 f., 933, 1275, 1615, 1888, 1890 f. Bergman, M. W. 1786 Bergmann, Rolf 122, 904, 1684 Bergsland, Knut 146 f., 1389 Bergström, Gustav A. 1678 Berko, Jean 300, 1615, 1619, 1623, 1766, 1769, 1795, 1797 f. Berman, Ruth Aronson 1343, 1345, 1347⫺1349, 1795⫺ 1799, 1802 f. Bernard, John 832 Berndt, Rolf 1673 Bernhardi, A. F. 108, 698 Bernstein Ratner, Nan 1798 Berthelsen, Christian 1389 Bertinetto, Pier Marco 364, 660 f., 1176 Bertocchi, Alessandra 1213 Besch, Werner 1267, 1681 f., 1684 Besnier, Niko 577 f., 580 f. Best, Günter 1507 Best, Karl-Heinz 121 Bettelhäuser, Hans-Jörg 1269 Beurden, Lisan van 165 Bever, Thomas G. 1799 f. Bhartrøhari 42 Bhat, Darbhe N. Shankara 720, 725 f., 759, 761, 764, 1112, 1209 f., 1214, 1216 Bhattacharya, Sudhibhusan 1004 Bhayani, Harivallabh Chunnilal 1722 Bickel, Balthasar 1178, 1835 f., 1853 Bickerton, Derek 1183, 1764 Biedermann, Johann 125, 135, 697 Bierbach, Mechtild 1704 Biermann, Anna 1054, 1060, 1064, 1097 Bierwisch, Manfred 160, 268⫺ 270, 667, 908, 910, 1099, 1788 Biesold, Horst 1555 Biggs, Bruce G. 578 Bills, Garland 1157 Binnick, Robert I. 1168, 1172, 1175, 1180 Bird, Charles S. 569, 589 Bisang, Walter 1657 Bishop, Brian R. 1563 Bishop, Dorothy V. M. 1820 f. Bittner, Andreas 294, 297, 299, 519, 636, 1272 Bittner, Dagmar 1269
Black, James R. 389 Black, Jeremy 35 Blainville, Henri Marie Ducrotay de 16 f. Blair, Hugh E. 1563, 1567 Blake, Barry J. 640, 642 f., 809 f., 815, 825, 1083, 1085⫺ 1088, 1135 f., 1482, 1491, 1493 f. Blake, Frank R. 1474 f., 1479 f., 1486 Blanc, Haim 1065, 1742, 1749 Blanche-Benveniste, Claire 1285 Blank, Andreas 1616 Blanke, Detlev 1563⫺1565, 1571 Blanken, Gerhard 1817 Blatz, Friedrich 106 Blau, Joshua 1741 Bleek, W. H. I. 96 Bleser, Ria de 364, 1817, 1819 Bloch, Bernard 178 f., 180, 185, 218 f., 241, 379, 437, 440 f., 453 Bloch, Jules 1712, 1724 Blois, Kornelis F. de 589 Bloom, Lois 1807 Bloomfield, Leonard 20, 26, 55, 71, 100, 109, 121, 165, 170⫺ 189, 191, 196⫺198, 202 f., 205, 218 f., 236 f., 252, 304, 319, 336 f., 345, 350⫺352, 356⫺359, 379 f., 404⫺406, 418, 420, 427, 437 f., 440, 444 f., 453, 463, 490, 494, 508, 542, 561, 582, 599, 677, 698, 721, 735, 872 f., 1005, 1012, 1233, 1416, 1474, 1476⫺1479, 1481, 1483, 1486, 1488, 1579, 1789, 1836 Blust, Robert A. 423 f. Boas, Franz 173, 176⫺178, 186, 547, 760, 824 Boase-Beier, Jan 908 Bochner, Harry 407 Bock, Kay 1780 Boeder, Winfried 1091, 1136 Boersma, Paul 1613 Boettcher, Wolfgang 1936 Bogaards, Paul 1891 Bogdanov, Sergej I. 420, 422 Bogorodickij, Vasilij A. 126 Bohas, Georges 70, 702 f. Bohnhoff, Lee E. 985, 1006 Böhtlingk, Otto 437 Bois, John W. Du 983 Boisson, Claude 693 Bokamba, Eyamba G. 811 Bokarev, Evgenij A. 1367 Bol, Gerard 1821 Bolinger, Dwight L. 247, 249, 252, 262, 418, 423, 766, 953, 1136, 1233 f., 1591, 1619 f., 1623, 1884, 1889, 1922
Bolozky, Shmuel 1345 Bonda´rko, Aleksandr V. 132, 609, 611, 613⫺615, 886 f. Bonifatius 76, 79, 81 f. Bonvillain, Nancy 920, 993 Boogaart, Ronny 1178 Booij, Geert E. 32, 163, 165 f., 195, 198, 203, 230 f., 318, 327, 336 f., 339⫺342, 351 f., 361, 363, 365⫺367, 380, 384 f., 395, 414, 433, 476, 541, 845, 854, 861 f., 864, 879, 883, 889, 891 f., 902, 926, 931, 933, 976, 979, 1214, 1288, 1806, 1809 Boost, Karl 668 Bopp, Franz 6⫺8, 19, 59, 93, 95⫺98, 100, 537, 1590, 1592 Borer, Hagit 314, 331 Boretzky, Norbert 290, 1644 f., 1647, 1649, 1651 Borg, Albert 760 Borg, Alexander 1748 f. Borgman, Donald M. 1216 Borgstrøm, Carl Hjalmar 147, 439 Börjars, Kersti 656 f., 662 Bork, Hans Dieter 890, 1698, 1704 Bornemann, Eduard 1320 Borsche, Tilman 696 Bortolini, Umberta 1820⫺1822 Bosa´k, Ja´n 424 Bosch, A. van den 1898 f. Botha, Rudolf P. 165 f., 195, 318, 351 f., 357, 845, 878 f., 1636, 1816 Bouchard, Denis 1265 Bouma, Lowell 1809 Bouquiaux, Luc 1832 Bowe, Heather 715 Bowerman, Melissa 1798 Bowern, Claire 1491 Boyes Braem, Penny 1556 Bradley, Dianne C. 1818 Brady, Ronald H. 16 Braine, Jean Critchfield 803 Branch, Michael 1079 f. Brandt Corstius, Hugo 772, 775 Braune, Wilhelm 100, 1683, 1686 f., 1689 Brauner, Siegmund 715 Bre´al, Michel 19 Brecht, Richard 1157 Breedveld, Anneke 1510 Breeze, Mary 1007, 1011 Brekle, Herbert E. 19, 54, 91 f., 161, 834 f., 837, 906 f. Brend, Ruth M. 192⫺194 Brenninkmeyer, Leo 1760 Bresnan, Joan 199, 346, 784, 1135 Brewster, E. Thomas 1832
1952 Brewster, Elizabeth S. 1832 Bright, William 553, 803 Brinker, Klaus 114, 1273 Brinkmann, Hennig 116, 1693 Brinton, Laurel 815 Brisard, Frank 983 Brittain, Julie 1419 Broadwell, George Aaron 584 f. Brockelmann, Carl 1096 Brockhaus, Wiebke 431 Broeder, Peter J. F. J. 1813 Bromley, Myron H. 1142, 1203, 1205 Bronckart, Jean-Paul 1807 Brøndal, Viggo 146, 273 f., 679 f., 699 f. Broschart, Jürgen 683, 717, 721, 725, 938 Broselow, Ellen 206, 548, 555, 560, 564 Brouwer, De´de´ 1050 Brown, Dunstan 155 Brown, Roger 540, 1001, 1011, 1795, 1797 Browne, Wayles 571 Browning, Robert 809 Bruce, Les 1036, 1041 Bruche-Schulz, Gisela 135 Brugman, Claudia 1596 Brugmann, Karl 10, 100 f., 112, 158, 782, 1626 Brunner, Karl 1670 Bryan, Margaret A. 1077, 1079 Buba, Malami 990 f., 993 Bubenik, Vit 1712 Buchholz, Oda 1213, 1218 Buck, Carl Darling 1126 Buckingham, Henry W. Jr. 1789 Buckingham, Hugh W. 1819 Bugge, Aage 1389 Bühler, Karl 110, 163, 249, 667, 682, 685, 983 Bühler, W. 58 Bulacho´v, Michail G. 135 Bulygina, Tat’jana V. 26 f., 133, 526 Bumann, Waltraud 98 Bunt, Harry C. 1067 f. Burani, Cristina 1772, 1778 f. Burdach, Karl Friedrich 16⫺18 Burgschmidt, Ernst 834 Burling, Robbins 1196 Burrow, Thomas 1150, 1159 Bursill-Hall, Geoffrey L. 696 f. Burt, Marina K. 1806 Burzio, Luigi 813 Buscha, Joachim 677, 683 Busla´ev, Fe¨dor I. 125 Bußmann, Hadumod 436, 463 Butterworth, Brian 1768, 1771 f., 1778 f., 1783, 1790, 1818 Butzkamm, Wolfgang 1931 Buyssens, Eric 211, 214, 216
Verzeichnisse und Sprachenkarten Buza´ssyova, Kla´ra 424 Bwenge, C. 1886, 1888 Bybee, Joan L. 27, 32 f., 122, 203 f., 220, 231, 261, 266 f., 286 f., 294, 346, 348, 361, 366, 371 f., 374⫺376, 407, 442⫺444, 536⫺543, 548, 550, 600, 604, 606, 626, 633, 656 f., 659⫺662, 708, 712, 753 f., 799, 804, 806, 814, 869, 871 f., 929 f., 938, 976, 1137, 1140, 1167, 1173, 1175, 1177 f., 1180, 1184 f., 1191, 1194⫺1196, 1199, 1234, 1247, 1249⫺1252, 1407, 1423, 1591, 1595, 1603, 1773, 1778, 1781, 1795⫺1797 Bynon, Theodora 97, 121, 1223, 1225, 1228 f. Byrd, Roy J. 206
C Cabre´, M. Teresa 1927 Cadiot, Pierre 699 Caesar 1077 Caferog˘lu, Ahmet 1579 Calder, George 228 Caldwell, Robert 96, 813 Caluwe, Johan de 166 Camaj, Martin 978 Campbell, Lyle 442, 760, 979, 1666 Canello, Ugo Angelo 452 Canger, Una 1450 Cann, R. 681 Cannon, Garland 835, 952 f., 955 f. Cantwell, Dennis P. 1823 Capell, Arthur 373, 1479 Capidan, Theodor 1651 Caplan, David 1818 Caramazza, Alfonso 347, 364, 1769, 1772, 1778, 1780 f., 1783, 1817⫺1819 Cardona, George 41, 43⫺46, 48, 50 f., 1712 Carlson, Greg 862, 1261 Carlson, Robert 757, 761, 764, 1018, 1022 f., 1025 Carochi, Horacio 1433 Carpenter, Kathie 1018, 1028 f. Carrier-Duncan, Jill 559, 561, 1477 f. Carroll, Janet F. 1096 Carroll, John M. 837 Carroll, Lewis 953, 955 Carroll, Mary 1810 Carroll, Susanne 1812 Carstairs, Andrew 27, 154, 200, 231, 338, 346, 389, 460, 463, 469, 478 f., 514, 516 f., 597,
599, 603, 617, 626 f., 630, 632, 634⫺636, 638 f., 645 f., 1269, 1514, 1766, 1772 Carstairs-McCarthy, Andrew 155, 289, 453, 618, 623 f., 635, 637, 639, 1255 Carter, Michael 74 Carvalho, Paulo de 1076 Casacchia, Giorgio 703 f. Casad, Eugene H. 1121, 1129 Cassirer, Ernst 212, 1091 Castre´n, M. Alexander 96 Castren, Mathias A. 1376 Catach, Nina 1921 Caubet, Dominique 1742 ´ avar, Damir 398 C Cavigneaux, Antoine 37 Caxton, William 1668 Cerro´n-Palomino, Rodolfo 1453, 1455 Chafe, Wallace L. 1098, 1197 Chalilov, Majid Sh. 1367 Chanet, Anne-Marie 1319, 1323 Channon, Rachel 1157 Chantraine, Pierre 121, 1325 Chao, Yuen Ren 316, 1730, 1732, 1736⫺1738 Chapman, Shirley 987, 993, 1040 Chappell, Hillary 1093 Charax 58 Charette, Monik 506 Charitonova, Irina I 677, 685, 699 Charles, Elena 922 f. Charles, George 922 Charmasson, Henri 966 Charney, Jean O. 917 Charpin, Franc¸ois 696 Chaucer, Geoffrey 1642 Chaumeil, Jean-Pierre 1464 Chen, Matthew Y. 1732 Cheng, Chin-chuan 1732 Cherpillod, Andre´ 1563 Cherubim, Dieter 1858 Chervel, Andre´ 698 Chevalier, Jean-Claude 698 Chevillard, Jean-Luc 703 Chi, Tedee 950 Chialant, Dorian 1780 f., 1783 Chierchia, Gennaro 1070, 1213 Childs, G. Tucker 719 Chiu, Bonnie Hui-Chun 549 Chmura-Klekotowa, Maria 1802 Cholewa, Jürgen 1817, 1819 Cholodo´vicˇ, Aleksandr A. 132, 1137, 1140, 1146 f. Chomsky, Noam A. 26, 28, 31, 121, 143, 155, 160, 165, 170, 176, 179, 188 f., 194⫺200, 202, 204 f., 252, 267, 276, 321, 329, 331, 341, 344⫺347, 378, 385, 405, 408, 471, 478,
1953
Namenregister 497, 500, 507, 541, 555, 667, 671, 818, 835, 841, 948, 1131, 1250, 1262, 1576, 1602, 1612 f., 1640, 1778, 1789 f., 1822, 1894, 1917, 1921 Chou, Fa-kao 1731 f., 1735 Chouinard, Michelle M. 1798 Chrakovskij, Viktor S. 132, 1134 Christaller, Johann G. 1518 Christie, Agatha 1641 Christie, William M. 422 Christmann, Hans Helmut 109 Christoph, Ernst-Michael 1901 Chrysoloras, Manuel 59 Chung, Sandra 814, 1134, 1197 Churchward, Clerk Maxwell 577, 1136 Cicero 76, 1929 Cinque, Guglielmo 899, 904 Civil, Miguel 35 Clahsen, Harald 1767, 1799, 1813 f., 1820⫺1822 Clairis, Christos 241, 488 Clajus 697 Clark, Douglas 836 Clark, Eve V. 445, 835, 853, 947⫺950, 1093, 1796, 1798, 1800⫺1803 Clark, Herbert H. 445, 835, 853, 947⫺950, 1115, 1796, 1801 Clarke, Sandra 1412 Claudi, Ulrike 1094, 1099 Clements, George N. 506, 563, 1005 Cloarec-Heiss, France 529 Cloquet, Hippolyte 17 Coates, Jennifer 1194 f. Cobb, Howard 1783, 1819 Coenen, Jose´e 1810 Cohen, Antonie 164 Cohen, David 486 f., 1741 Cohen, Marcel 142 Cohen, Sophia R. 1802 Cohn, Abigail C. 341, 414 Cole, Jennifer 554 Cole, Peter 714, 1091 f., 1106, 1137⫺1139, 1194, 1199, 1454, 1459, 1863 Cole, Ronald A. 540 Coleman, Robert 639, 641, 1087 Colombat, Bernard 693, 697, 699 Coltheart, Max 1770, 1773, 1817 Combrink, Johan G. H. 166, 858 Comenius, Johann Amos 1930, 1933 Comrie, Bernard 14, 245, 252, 282, 374, 401, 407, 481, 538, 543, 601, 603, 609⫺611, 613, 640, 712, 715, 734, 753, 797, 809, 812⫺815, 817, 846, 848, 939, 1076 f., 1080 f., 1083,
1085, 1096, 1108 f., 1120 f., 1137 f., 1152, 1165, 1174, 1180 f., 1186, 1223 f., 1247, 1432, 1556, 1780, 1863, 1868, 1871 f., 1922 Condillac, Etienne 698 Conklin, Harold 178 Conklin, Nancy F. 1016 Connell, Phil J. 1821 Conrad, Rudi 436, 536 Conrady, August 1730 Consentius 85 Constantino, Ernesto 1473 f. Contini-Morava, Ellen 1189 Contreras, Heles 1236 Cooke, Joseph 716, 1013 Cooper, Jerrold S. 1916 Cooper, William E. 540 Corbett, Greville G. 29, 155, 282 f., 318, 362, 770, 787, 1033 f., 1042, 1053 f., 1057⫺ 1060, 1062⫺1065, 1067⫺ 1069, 1071, 1080 Corbin, Danielle 139, 143 f., 583, 834, 837 f., 841, 852, 869, 871, 873 f., 1285⫺1292, 1294⫺1296 Corbin, Pierre 143, 1286, 1290⫺ 1292, 1294 Corblin, Francis 983 Cornish, Francis 983 Correa, Julio 1430 Corriente, Federico 1741, 1748 Coseriu, Eugenio 7, 115, 222, 682 f., 686, 688, 834, 1221 f. Coulmas, Florian 988, 992 Couper-Kuhlen, Elisabeth 1136 Coupez, Andre´ 1527 Court de Ge´belin, Antoine 698 Courte´s, Joseph 210 Coute´, Bernard 1285 Couturat, Louis 1563, 1565 Cowan, William 588, 1415 Coward, Harold 42 Cowell, Mark W. 547 f., 765 Cowgill, Warren 1237, 1240 Cowie, Anthony P. 1884, 1888⫺ 1890 Cowper, Elizabeth 569 f., 572 Crago, Martha B. 1820 f. Craig, Colette G. (s. Grinevald, Colette) Crates 60 Creider, Chet A. 1135, 1507 Creider, Jane Tabsubei 1135, 1507 Creissels, Denis 1092 Crocco Gale`as, Grazia 458, 460, 585, 875, 1706 Croft, William 258, 261 f., 275, 277 f., 709, 714, 728, 758, 785, 788, 796 f., 986, 1094, 1112⫺1115, 1139, 1141, 1169,
1171, 1248, 1250, 1252, 1422 f., 1426 Cromer, Richard 1822 Crookston, Ian 1212 Crössmann, Helga 693 Crowhurst, Megan J. 1474 Crowley, Ellen T. 953 Crowley, Terry 1078 Crowther, Jonathan 1884 Cruse, David A. 596, 871 Crystal, David 437, 677, 693 Cubar, Ernesto H. 1473 ˇ urganova, Valerija G. 419, 424 C Curtiss, Susan 1822 Curtius 1167 Cusihuama´n Gutie´rrez, Antonio 1455 Cutler, Anne 432, 539 f., 550, 867, 941, 1080, 1251, 1779 f., 1783, 1785, 1788, 1790 f. Cuzzolin, Pierluigi 1214 Cyr, Danielle 1412⫺1414, 1416 Czapkiewicz, Andrzej 1746 Czech, Franz Herrmann 1555 C ¸ abej, Ekrem 1647
D Daalder, Saskia 984 Dabir-Moghaddam, Mohammad 1161 Dacia, Martin von 83⫺85 Daelemans, Walter 1896⫺1899 Dahl, Östen 32, 374⫺376, 539, 609⫺611, 613, 656, 661, 797, 806, 809, 1175 f., 1180, 1184 f., 1208, 1527, 1595 Dahlstrom, Amy 1151, 1413, 1419 Ía´i Xuaˆn Ninh 1550 f. Dakin, Karen 1447 Dal, Georgette 143, 1292 Dal, Ingerid 1268 Dalalakis, Jenny 1821 Daly, Margarita H. 193 Daniels, Peter T. 1921 Danon-Boileau, Laurent 983 Danso, Mary 1519 Darmesteter, Arse`ne 138⫺143, 1287 Darski, Jo´zef 1271 f. Darwin, Charles 8, 17, 98 Daum, Edmund 1309 Dauzat, Albert 140, 142 Davane, G. V. 1722 Davidson, Donald 727 Davies, John 713, 767, 800, 1202 f. Davis, Carol 1781 Davis, John F. 1203 f. Davis, Phillip W. 923 Davis, Stuart 430, 560
1954 Day, A. Colin 1833 Day, Christopher 979 Dayley, Jon P. 764, 766, 1131, 1215 Debrunner, Albert 121, 452, 1316, 1320, 1325 DeCamp, David 1653 De´chaine, Rose-Marie 1413, 1419 Declerck, Renaat 1166 Dede, Müs¸erref 1360 Dedenbach, Beate 1271 Dederding, Hans-Martin 1278 Deegener, Günter 300 Deely, John 211 Dees, Anthonij 1701 DeGraff, Michel 1654 Deigthon, Len 836 Delacroix, Euge`ne 163 DeLancey, Scott 1010, 1027, 1082, 1592, 1595 f. Delbridge, Arthur 832 Delbrück, Berthold 100 f., 104 f., 112, 642 Delfitto, Denis 1176 Delgaty, Alfa 1098 Dell, Franc¸ois 143, 338, 1288 f. Deloria, Ella 547 Delsing, Lars-Olof 1069 DeMatteo, Asa 1560 Demers, Richard 721, 723 Demuth, Katherine A. 1019, 1043, 1799 Dench, Alan 1009, 1079, 1086 f., 1204, 1491, 1493 Denes, Gianfranco 1818 f. Denny, J. Peter 922 f., 993, 1016, 1018, 1121, 1417 Denz, Adolf 1746 f. Depraetere, Ilse 1166 Derbyshire, Desmond C. 742 f., 919 f., 979 f., 987, 993, 1024, 1040, 1078, 1098, 1202 f., 1208 Derwing, Bruce L. 420, 854, 866, 1766 Desˇerieva, Tamara I. 604 Desloges, Pierre 1554 Deuchar, Margaret 1556 Deutschbein, Max 814 DeWolf, Charles M. 1475, 1479, 1481 Dez, Jacques 702 Dı´az Cahuachi, Pedro 1471 Dı´az, Manungo 1471 Dibble, Charles 1452 Diderichsen, Paul 146 Didymus 62 Diem, Werner 703, 1096, 1741 f., 1748, 1750 Diessel, Holger 983 Diesterweg, Adolph 109 Dietrich, Rainer 1808
Verzeichnisse und Sprachenkarten Diewald, Gabriele 1273 Diez, Friedrich 139 f., 158 Diffloth, Ge´rard 1128 Dijkhoff, Marta 1658 Dik, Simon C. 165 f., 196, 973 f., 1105, 1109, 1130 f. Dillon, Myles 569, 571 Dimmendaal, Gerrit J. 637, 751, 758, 766, 822, 1032 f., 1507, 1509⫺1511, 1514 f. Dineen, Francis P. 54 Diogenes Laertius 54 Dionysios Thrax 55⫺58, 77, 87, 674, 694, 708, 732, 745, 1149 Dirven, Rene´ 1049 Dittmann, Jürgen 1817 Dixon, Robert M. W. 341 f., 373, 400, 432, 443 f., 621, 637, 685 f., 714⫺716, 719, 728, 742, 757, 767, 796 f., 817, 985, 1004, 1018, 1021, 1026, 1032, 1035, 1038, 1040⫺1042, 1080⫺1083, 1108, 1111⫺1116, 1135, 1137, 1158, 1197, 1209, 1491, 1493⫺1495 Dobson, William Arthur Charles Harvey 1053 Doerfer, Gerhard 1579, 1649, 1651 Dogil, Grzegorz 1819 Doke, Clement Martyn 719, 854, 1538 Dokulil, Milosˇ 446, 888 Doleschal, Ursula 1275 Dolphyne, Florence 1518 Dominicy, Marc 92, 96, 211 Domokos, Pe´ter 330 Don, Jan 165 f., 446, 950 Donaldson, Bruce 658 Donaldson, Tamsin 822, 1004, 1196, 1198 f. Donatus 63⫺65, 77 f., 82⫺84, 674, 696, 1929, 1931 f. Donegan, Patricia Jane 401, 1594 Donhauser, Karin 1274 Donnellan, Keith S. 983, 985 Donner, Kai 1376 Donnet, Daniel 59, 65 Dordian, Monique 1818 Dorian, Nancy 1043, 1655 f. Doron, Edit 327, 328 Downer, Gordon B. 588, 1731 f. Downing, Pamela 352, 837, 852, 907, 909, 1258, 1760 Dowty, David 201, 907, 909, 1168 Drabbe, Peter 1501 Drach, Erich 108 Dragunov, A. A. 677 Drapeau, Lynn 1412, 1414, 1418 Dressler, Wolfgang Ullrich 26, 31, 33, 122, 161, 219 f., 270,
289 f., 292⫺294, 296, 336, 338 f., 361, 424, 431, 433, 455⫺457, 460, 485, 490, 492, 507, 514 f., 518, 530⫺533, 550, 555 f., 579, 583⫺585, 737, 753, 755, 781, 838, 868, 874, 879, 891, 934, 936, 940, 979, 1045 f., 1049 f., 1276, 1602 f., 1623, 1789, 1817⫺ 1820, 1822, 1902 Driem, George van 767, 1214 Dringel-Techt, Eva 290 Drivaud, Marie-He´le`ne 1883 Droixhe, Daniel 93 Dromi, Esther 1795, 1821 Drossard, Werner 1475, 1479, 1483, 1486 Drozd, Lubomı´r 1924 Drude, Sebastian 1836 f., 1851 Dryer, Matthew 539, 1053, 1062 Dubnova, Yelena Zinov’evna 1216 Dubois, Claude 1886 f., 1889 f. Dubois, Jean 139, 142 f., 536, 1285, 1291, 1885⫺1887, 1889 f. Duchan, Judith F. 983 Dudas, Karen 568 Duhoux, Yves 122 Dul’zon, Andrei P. 1376, 1382, 1384, 1386 Dulay, Heidi 1806 Dulicˇenko, Aleksandr D. 1563 Dunnigan, Timothy 785, 788 Durand, Jacques 500 Durie, Mark 329, 521, 713, 717, 781, 978, 1058, 1082 Durkheim, Emile 4, 142 Durnovo´, N. N. 127 Durovic, Lubomir 697, 1309 Durrell, Martin 1685 Dyk, Janet 362
E Eades, Diana 992, 1095, 1097 Ebbinghaus, Horst 1560 Ebeling, Carl 164, 1081, 1084 Ebert, Karen H. 985 Ebert, Robert Peter 1685, 1688 f., 1693 Eckert, Gabriele 1707 Eckman, Fred 276 Eco, Umberto 211 f., 882 Efremova, Tatjana F. 131, 889 Egede, Paul 1389 Egerod, Søren 546, 548 Ehrlich, Susan 1174 Eichinger, Ludwig M. 931, 1278, 1683 Eichler, Ernst 1905 Eijk, Jan P. van 701 f.
1955
Namenregister Einhauser, Eveline 100 Eisele, John C. 1746 Eisenberg, Peter 108, 1269, 1916, 1920⫺1922 Ekdahl, Muriel 1210 Elbert, Samuel 717 Elffers-van-Ketel, Els 105 Elimelech, Baruch 571, 573 Elson, Benjamin F. 442, 489, 541, 1831 Emenanjo, E. Nolue 934, 1128 Emmorey, Karen 541, 1560 Emonds, Joseph E. 681, 811, 1257 Encreve´, Pierre 1700 Endzelin, Jan 1650 Engberg-Pedersen, Elisabeth 1559 Engelen, Bernhard 671, 688 Engelhardt, Wolf von 16 Engelien, August 113 England, Nora C. 1012, 1033, 1100 Englebretson, Robert 1116 Engler, Rudolf 216 Eoin, Gearo´id Mac 627 Epstein, Richard 985 Erasmus 1930 Erbaugh, Mary S. 1027 Erben, Johannes 160, 677, 891⫺ 895, 898, 902, 1273, 1278, 1621, 1686, 1694 f. Erdman, Paul 836 Erdödi, Jo´zsef 823 Erickson, Barbara 193 Erlebach, Peter 1677 Erlenkamp, Sonja 1560 Erlinger, Hans Dieter 104 f., 110, 115 Ernout, Alfred 120 Eroms, Hans-Werner 1278 Ersen-Rasch, Margarete I. 1106 Ertel, Suitbert 1811 f. Erwin, Wallace 1742, 1747 Escalante, Fernando 1092 Eschenlohr, Stefanie 1275 Espinosa Pe´rez, L. 1464 Ettinger, Stefan 1045, 1047, 1051 Euler, Wolfram 122 Eutyches 77, 80 Evans, Nicholas D. 709, 711, 743 f., 753, 918, 926, 1033 f., 1042, 1057, 1079, 1086 f., 1108, 1491, 1854, 1880 Evans, Roger 1896, 1899 Everett, Daniel L. 1214, 1216, 1465 Evett, Lindsay J. 1771 Ewen, Colin J. 507 Eynde, Karel van den 222, 1285
F Fabb, Nigel 412, 882, 1255 f. Fabricius, Anne H. 1056, 1061 Fabricius-Hansen, Cathrine 904, 1272⫺1274 Fagan, Sarah 1149 Faiß, Klaus 1627, 1668 f., 1671, 1673⫺1675 Fanselow, Gisbert 161, 904⫺ 906, 908⫺910, 1278 Faraclas, Nicholas 759 Farrar, Frederic W. 16, 19 Faust, Manfred 157 Fay, Donald 1784, 1790 f. Fearn, John 92 Fedson, Vijayarani Jotimuttu 1138 Fejos, Paul 1464 Feldman, Laurie B. 206, 1783, 1922 Fe´lice, The´odore de 1285 Fellbaum, Christiane 1821 Fellman, Jack 1343 Fenk-Oczlon, Gertraud 289 Ferguson, Charles A. 1741 Ferguson, Charles J. 179, 1654 Fettes, Mark 1564 Feuillet, Jack 983, 993, 1274 Fife, James 482 Filin, Fedot P. 135 Fill, Alwin 1619 f., 1623, 1633 f. Fillmore, Charles J. 260, 667, 983, 1099 Finck, Franz Nikolaus 1835, 1837 Firestone, Homer L. 711 Firth, John Rupert 149, 150 f., 153, 455, 494, 554, 589, 616 Fischer, Olga 1590 Fischer, Renate 1555 Fischer, Susan 1555, 1559 f. Fischer, Wolfdietrich 821, 1746⫺ 1751 Fischer-Jørgensen, Eli 489, 506, 508 Fisher, Herbert A. L. 52 Flacourt, Etienne de 702 Flämig, Walter 678, 699 Fleisch, Henri 70 Fleischer, Wolfgang 160, 250, 355 f., 744, 886 f., 889, 891⫺ 894, 898 f., 902, 931 f., 950, 958, 970, 1267, 1276, 1635, 1902 Fleischman, Susanne 1184, 1708 Fliedl, Günther 453 Flood, Walter Edgar 1926 Fludernik, Monika 983, 985, 988 Flydal, Leif 146 f. Fodor, Jerry A. 286, 1822 Fogel, Michael 92 Foley, James 1285
Foley, William A. 760, 795, 859, 1012, 1032, 1036, 1038, 1041, 1110, 1120, 1126⫺1128, 1142, 1192 f., 1211, 1377, 1475, 1479, 1501 Fo´nagy, Ivan 214 Forchheimer, Paul 716, 786, 1000, 1003 f., 1008 Ford, Alan 871, 1415 Ford, Kevin 1041 Forrest, R. A. D. 1731 ˚ ke 104 f., 109, Forsgren, Kjell-A 113, 115, 665 f., 668⫺670, 672 Förstemann, Ernst 1615, 1620⫺ 1623 Forster, Keith I. 1790 Forster, Kenneth 1766, 1768 f., 1771 f., 1779, 1781, 1784 Fortescue, Michael 254 f., 768, 922, 979, 981, 1005, 1009, 1107, 1216, 1389 f., 1394, 1400, 1407, 1840 Fortunatov, Filipp F. 126⫺130, 1308 Foster, George 1004 Foster, Mary 1004 Fought, John 173, 175 f. Fourquet, Jean 1267 f., 1270, 1272 f. Fowler, Anne E. 1822 Fowler, C. A. 206 Fowler, Carol F. 1783 Fowler, George 958, 961 Fox, Anthony 250, 1667 Fox, Barbara 1099 Fradin, Bernard 143 Fradkin, Robert A. 1060 Fraiberg, Selma 999 Frajzyngier, Zygmunt 546, 548, 1042, 1058 Francis, W. Nelson 253 Frank, Paul 1135 Franks, Steven 789 Frantz, Donald G. 924 f. Fraser, Norman 155 Frauenfelder, Uli 347, 1772⫺ 1774, 1799 Frede, Michael 54 Fredkin, E. 1894 Freeland, L. S. 804 Frei, Henri 386, 435, 984, 1099 French, Koleen Matsuda 547 f. Freud, Sigmund 1782 Frey, Eberhard 1270 Fried, Mirjam 397 Friederici, Angela D. 1818 f. Friedrich, Paul 1596 Friend, Jewell A. 106 Fries, Charles C. 668, 672 Fritz, Helmut 438 Fromkin, Victoria A. 507, 541, 558, 589, 1788⫺1791
1956 Fruyt, Miche`le 1049 Fuhrhop, Nanna 458, 905, 1276, 1908 Fukuda, Shinji 1821 Fukuda, Suzy 1821 Fulmer, S. Lee 543 Funaioli, Hyginus 56 Funk, Wolf-Peter 1319 Funke, Otto 693 Funnell, E. 1790
G Gabelentz, Georg von der 9, 105, 1590, 1858, 1870 f. Gabelentz, Hans Georg Conon von der 95, 97, 99 f., 1835 Gallmann, Peter 1270 Galloway, Brent D. 1217 Gamillscheg, Ernst 1616 Gandour, Jack 1029 Ganschow, Gerhard 122 Garde, Paul 383, 708 Gardiner, Alan H. 216 Gardiner, Donna B. 924 f. Garey, Howard B. 1166 Garlandia, Johannes de 83 Garnham, Alan 983 Garrett, Merill F. 1788 f., 1791 f. Gasparov, Boris M. 127 Gathercole, Susan E. 1822 Gawełko, Marek 1708 Gaynor, Frank 536 Gazdar, Gerald 331, 1896, 1899 Geckeler, Horst 1702 Geeraerts, Dirk 1050 Geerts, Guido 166 Geiger, Richard A. 983 Geiger, Wilhelm 1712 Gelb, Ignaz 1915 Genetti, Carol 1205 f. Geniusˇiene˙, Emma S. 1131 f., 1146 f., 1157, 1159 Gentner, Dedre 374, 755, 795 f. George, Kenneth E. M. 958 Georgius Choeroboscus 58 Gerd, Aleksandr S. 420 Gerhardt, Julie 1001 Gersbach, Bernhard 933, 1267 Gesenius, William 1345, 1347⫺ 1349 Gessinger, Joachim 97 Gessmann, Albert M. 453 Ghomeshi, Jila 1061 Giacalone Ramat, Anna 1590, 1808 Giannakis, Giorgios K. 1320, 1323 Gibbon, Dafydd 1899 Gibbon, Edward 52 Gibson, Eleanor 1769, 1783
Verzeichnisse und Sprachenkarten Giegerich, Heinz J. 1780 Gil, David 721, 737, 1059, 1061, 1067, 1070⫺1072, 1475 Gildea, Spike 987 Gillie´ron, Jules 1615 Gilligan, Gary 538⫺540, 750, 752, 867, 941 Gilman, Albert 1001, 1011 Ginneken, Jacobus van 813 Gipper, Helmut 95 Girard, Abbe´ Gabriel 92, 672, 698 Girault, Louis 1461 Giurescu, Anca 1677 Givo´n, Talmy 286, 371, 401, 407, 537⫺540, 727, 733, 735, 784, 786, 795, 799, 810, 814, 823, 941, 983, 986, 1081, 1086, 1091 f., 1115, 1148, 1161 f., 1202, 1204 f., 1210, 1229, 1232, 1248, 1423, 1590⫺1593, 1595 f., 1871 Gla´dkij, Aleksandr V. 132 Glazov, Jurij Ja 1237 Gleason, Henry A. 219, 345 Glebova, Jvetta 1550 Gledhill, Christopher 1564 Gleitman, Henry 837 Gleitman, Lila R. 837 Glinert, Lewis 1096, 1343, 1345, 1347⫺1349 Glinz, Hans 104, 107⫺110, 113⫺116, 668 f., 672, 677, 1273, 1935 Glück, Manfred 64 Goad, Heather 1821 Goddard, Cliff 266, 481, 603, 1076 Goddard, Pliny E. 764 Goddard, R. H. Ives 1004, 1413, 1416, 1418 Gode, Alexander 1563 f., 1567 Godel, Robert 217, 946 Goethe, Johann Wolfgang von 4, 6⫺10, 16⫺20 Goffman, Erving 984 Goldsmith, John A. 150, 189, 259, 339, 496, 500, 506, 880, 1898 Golla, Victor 1012, 1399 Golston, Chris 433, 472, 476, 1791 Go´mez Lopez, Paula 243 Gonc¸ales, Cristina Helena R. C. 919 Gonzales, Andrew B. 1475 Goodenough, Ward H. 178, 248 Goodglass, H. 1769 Goossens, Jan 1582 f. Gopnik, Myrna 1820 f. Gopsill, Peter 1564 Goral, Donald R. 1021 Gorbacˇevskij, Anton A. 520
Gordon, Lynn 1009 f. Görlach, Manfred 1620 Goswami, Usha 1922 Gottsched, Christoph 1682, 1690 Götz, Dieter 1619, 1627, 1885 Gove, Philip B. 1883, 1888 Goyvaerts, Didier L. 1829 Graf, Rainer 933, 1267, 1274 Gragg, Gene 1001 Graham, A. C. 1734 f. Grammont, Maurice 577, 1552 Granberry, Julian 702 Grasserie, Raoul de la 548 f. Graßhoff, Ludwig 1555 Gray, Louis H. 249, 254 Green, John N. 788, 813 Greenbaum, Sidney 316, 790, 947 Greenberg, Joseph H. 121, 202, 273, 275, 351, 366, 373, 405, 418, 443, 537, 539, 540, 542, 548, 550, 715, 738, 741, 745 f., 752 f., 772⫺776, 781, 784 f., 791 f., 805, 854, 940 f., 1000, 1003, 1008, 1028, 1041 f., 1053, 1055, 1057 f., 1062, 1223, 1225, 1229, 1235⫺1244, 1247⫺1249, 1252, 1400, 1598, 1609, 1796 Greene, Elton 917⫺919, 925 Greenough, James Bradstreet 1619, 1621 f. Gregor, Bernd 1269 Gregores, Emma 1107, 1423, 1427⫺1429 Greimas, Algirdas J. 210 Grenoble, Lenore A. 985 Gre´sillon, Almuth 901, 1286 Greule, Albrecht 1903 Grevisse, Maurice 571, 932 Grewendorf, Günther 672 Grice, H. Paul 985 Grimes, Joseph E. 541, 1202, 1210, 1831 Grimm, Jacob 6⫺8, 16, 19, 96, 109, 157⫺159, 162, 898, 905 f., 1050, 1267, 1269, 1683, 1693 Grimm, Ursula 1625, 1628 Grimm, Wilhelm 16, 162, 1050 Grimshaw, Jane 203, 327, 329, 331, 407, 864, 1130 Grinevald, Colette (Craig, Colette G.) 1018, 1020 f., 1024 f., 1028 f., 1119 f., 1123, 1135 f., 1206 Grodzinsky, Yosef 1769, 1818 Grønvik, Ottar 1273 Groot, A. Willem de 163 f. Groot, Casper de 166, 823 f., 1109 f. Gross, Gaston 699, 1287 Grosseteste, Robert 79, 83
1957
Namenregister Grotjahn, Rüdiger 1240 Grotzfeld, Heinz 1742 Grube, Henner 447 Gruber, Jeffrey S. 165, 1637 Guasch, Antonio 711 Gudschinsky, Sarah C. 253, 1832 Guilbert, Louis 142, 1287, 1883 Guillaume, Jean-Patrick 70, 142, 702 f. Guillaume, Paul 1798 Guinet, Lynn 1769, 1783 Guiraud, Pierre 142, 1288 Gulya, Ja´nos 92 Gulyga, Elena 666 Gundel, Jeannette 791 f. Günther, Hartmut 122, 907, 909, 1278, 1768 f., 1773 f., 1919 Gussmann, Edmund 503, 505⫺ 507, 880 Guzman, Videa P. de 1477, 1479 f. Gvozdanovic´, Jadranka 273, 638, 640, 772 Gvo´zdev, Aleksandr N. 126, 1799 f. Gwynn, Aubrey 1929 Gwynn, J. P. L. 1033 Gyarmathi, Samuel 92 f.
H Haaften, Ton van 862 Haarmann, Harald 91, 1241 Haarmann, Henk J. 1818 Haas, Mary R. 538, 814, 1592 Haas, Walter 582, 585 Haas, William 150⫺152, 218, 440 f., 534, 626, 946 Haas, Wim de 166 Habermann, Mechthild 1684 Häckel, Susanne 1693 Hacken, Pius ten 166, 350 f., 353 f., 357⫺359 Häcki Buhofer, Annelies 656 Hackl, Martin 1818 Hadj-Sadok, Mahammed 1750 Haeckel, Ernst 17⫺19 Haegeman, Liliane 321, 829 Haenisch, Erich 554, 568 Haensch, Günther 890 f. Haeringen, Coenraad Bernardus van 163, 1585 Haftka, Brigitta 668 Hage`ge, Claude 916, 923 f., 1005 f., 1009 Hahn, Michael 1732 Haider, Hubert 681, 1822 Haiman, John 259, 282, 286, 375, 397, 548 f., 568, 573, 712, 1012, 1091, 1097, 1202⫺ 1205, 1232 f., 1252
Hajdu, Pe´ter 330 Haji-Abdolhosseini, Mohammad 1059, 1061 Häkkinen, Kaisa 1328 Hakulinen, Auli 1330 Hale, Kenneth 328, 521, 582, 584, 821, 1009, 1119 f., 1123, 1135 f. Hale, William Gardner 1126 Hall, Christopher J. 453, 537⫺ 540, 550, 1251 Hall, Robert A. 237, 381, 925 Hall, Robert A. Jr. 249, 251, 1657 f., 1756 Hall, T. Alan 386 Halle, Morris 26, 31, 33, 189, 197 f., 200, 202⫺205, 252, 267 f., 341, 346 f., 378, 385, 408, 427 f., 471⫺479, 497, 500, 507, 541, 841, 948, 1256, 1262, 1272, 1602, 1612 f., 1640, 1732, 1766, 1771, 1778, 1790 f., 1894, 1917, 1921 Halliday, Michal A. K. 222, 977, 1193, 1195, 1198, 1202 Halpern, Aaron L. 389, 395⫺ 397, 400 Halpern, Abraham M. 711 Hamanaka, Toshihiko 1817 Hammarström, Göran 457 Hammerich, Louis L. 380, 701 Hammond, Michael 20, 32, 220, 548, 564, 627 Hamp, Eric Pratt 453, 1235 Hamze´, Hassan 702 Hancock, Ian 1647 Hankamer, John A. 1785 Hankamer, Jorge 347, 364 Hanks, Patrick 1883 Hanks, William F. 983⫺986, 988, 993, 1121 Hansen, Björn 132 Hansen, Detlef 1821 Hansen, Klaus 893 Hansson, Kristina 1821 Happ, Daniela 1560 Harada, Sin I. 993 Hardman, Martha 1003, 1009 Hardy, Heather K. 547, 582, 584 Hare, Mary 1767 Harell, Richards 765 Harford, Carolyn 802 Hargreaves, David 809, 813 Hargus, Sharon 433, 546 Harley, Heidi 1060 Harman, John 809 Harning, Kerstin E. 1748 Harnisch, Rüdiger 289, 1272, 1908 Harre´, Rom 1006 Harrell, Richard S. 1742 Harries, Lyndon 804 Harris, Alice C. 1082, 1085, 1137
Harris, James W. 200 Harris, John 430 Harris, Maria De Boe 193 Harris, Martin 599 Harris, Zellig S. 6, 121, 150, 164, 174, 176⫺182, 184⫺187, 189, 194 f., 200, 205, 345, 379, 405, 418, 436, 440 f., 453, 478, 490⫺492, 494, 496, 554, 573, 582, 584, 589, 620, 622, 666, 672 Harrison, Sean P. 559, 1135 Hartman, Lawton 1732 Hartman, Reinhard R. K. 568, 597 Hasan, Ruqaiya 1202 Haselbach, Gerhard 669 Hashimoto, Mantaro J. 1147 Hashimoto, O. K. Y. 781 Haspelmath, Martin 361 f., 549, 658, 660 f., 663, 709, 719, 799, 1041, 1132, 1137, 1140, 1142, 1158, 1213, 1218, 1580, 1590, 1871, 1877 Hasselrot, Bengt 1045, 1048 f., 1051 Hatcher, Anna Granville 1705 Haudricourt, Andre´ 1545, 1731, 1734 Haudry, Jean 122 Haupenthal, Reinhard 1563 Hauri, Hans Walter 1324 Häusler, Frank 508 Hausmann, Franz Josef 694 Haverkate, Henk 1192 Haverkort, Marco 1819 Havers, Wilhelm 253 Haviland, John B. 993, 1092, 1158 Hawkins, John A. 538⫺540, 550, 750, 752, 867, 941, 983⫺985, 1251, 1267 f., 1780 Hay, Jennifer 1786 Hayes, Bruce 431, 1262, 1613 Hays, David G. 667 Head, Brian 788, 1001 Healey, Alan 1829 Healey, Phyllis M. 558, 561 Heath, Jeffrey 1010, 1023, 1041, 1493, 1495, 1745, 1750 Hecht, Barbara F. 1798, 1802 Hedberg, Nancy 791 f. Hedin, Eva 1187 Heeschen, Claus 1818 Heeschen, Volker 983, 993 Hegel, Georg Wilhelm Friedrich 2 Heger, Klaus 211, 222, 685 Heidolph, K. E. 678, 684 Heine, Bernd 32, 743, 1019, 1032, 1035⫺1037, 1039 f., 1042 f., 1092, 1094, 1161, 1232, 1507, 1590⫺1593, 1595 f., 1654, 1752
1958 Heinicke, Samuel 1554 Heinimann, S. 87 Heinz, Sieglinde 1700 Helbig, Gerhard 109, 114, 665, 667, 671, 676⫺679, 683, 690, 693 Helias, Petrus 84 Heller, Louis G. 960 Helmbrecht, Johannes 1053, 1060, 1249, 1253 Helwig, Christoph 697 f. Hempel, H. 677 Hempen, Ute 1272, 1688 Henderson, Leslie 1769, 1771, 1773, 1775, 1779, 1783⫺1785 Hendrick, Randall 784 Hengeveld, Kees 728 f., 938, 974 f., 1105, 1107, 1192, 1199 Henry, David 1399 Henry, Patrick Leo 1649 Henzen, Walter 19, 160, 891, 898, 901, 1275, 1684 Herbermann, Clemens-Peter 250, 598, 1278 Herbers, Birgit 1694 Herbert, Robert K. 1043 Hercus, Luise 1491 Herder, Johann Gottfried von 7 Herling, Simon Heinrich Adolf 105, 669 Hermann, Eduard 104, 106, 677 Hermon, Shula 1822 Herms, Irmtraud 1045, 1048 Hertog, Cornelis H. den 162 Herva´s y Panduro, Lorenzo 91 Hervey, Sa´ndor G. 213 f. Hess, Thom 701 f. Heßmann, Jens 1560 Hetzron, Robert 750, 1077 Hewitt, Brian George 800, 802, 824, 980 f., 1004, 1038, 1096, 1203, 1205, 1215 Hewitt, Helen-Jo Jakusz 122 Heynatz, Johann Friedrich 108 Heyse, Karl W.L. 104 f. Higginbotham, Jim 862 Hilgard, Alfred 54, 56⫺59 Hill, Jane H. 926 Hill, Kenneth C. 926 Hillis, Argye E. 1818 Himmelmann, Nikolaus P. 717, 721 f., 983, 1475 f., 1479 f., 1482, 1486 Hinch, H. E. 373 Hinds, John 1094, 1215 Hinnebusch, Thomas 791, 1099 Hinrichs, Erhard 1170 Hinüber, Oscar von 1712 Hirshfeld, H. 74 Hirt, Hermann 1686, 1691 Hjelmslev, Louis 3, 24, 59, 95, 146 f., 210 f., 220 f., 235, 273, 275, 378, 380, 405 f., 452,
Verzeichnisse und Sprachenkarten 608, 688, 834, 1076, 1084, 1086, 1127 Ho oˆ` Leˆ 1545, 1550 Hocart, A. M. 577 Hock, Hans Henrich 121, 402, 475, 576 f., 1667 Hockett, Charles Francis 26, 53, 150⫺152, 155, 178⫺181, 183⫺189, 204, 211, 214, 219, 235⫺237, 241, 248 f., 350, 379, 383, 405⫺407, 409 f., 418, 427, 439⫺441, 447, 452 f., 458, 463 f., 475, 546, 573, 582, 588, 622, 627, 656 f., 660, 694, 743, 866, 1225, 1732 Hoddinott, William G. 1080 Hoek, Karen van 1823 Hoeksema, Jacob 153, 165 f., 201, 206, 851 f., 855 f., 861, 863, 900, 1255 Hoekstra, Teun 166, 346, 864, 1250 Hoenigswald, Henry M. 100, 118, 345, 1887 Hoff, Berend J. 164, 166 Hoffman, Lothar 1924 Hoffmann, J. 720, 725 f. Hoffmann, Walter 1688 Höfler, Manfred 1698, 1705 Hofmann, Johann Baptist 1217 Hofstede, Ben T. M. 1818 Hogben, Lancelot 1926 Hohenberger, Annette 1560 Höhle, Tilman 161, 351⫺354, 892 Holdcroft, David 214 Holes, Clive 750, 1741 f. Holm, John 1654, 1657 Holmberg, Märta A. 905 Holmquist, Jonathan Carl 1036 Holton, David 1217 Holtus, Günter 1703 Holtz, Louis 63, 696 Homer 55, 58, 122, 1064, 1149, 1582, 1611 Honda, Isao 1208, 1210 f. Honey, Patric J. 1552 Hooper, Joan Bybee 189, 336, 348, 496 f., 1582, 1602 Hoppe, Gabriele 1275, 1277 Hopper, Paul J. 259, 287, 401, 727 f., 747, 764, 795, 799 f., 809, 1084, 1170, 1174, 1232, 1590⫺1592, 1595 f. Horn, Laurence 285 Hornby, A. S. 1884 Horne Tooke, John 1228 Hotzenköcherle, Rudolf 1684 Houghton, Donald E. 1643 Houis, Maurice 1202 Householder, Fred W. 406 Housen, Alex 1808
Hovdhaugen, Even 242 f., 582, 1107 Howard, David 1783, 1818 Howe, Stephen 1270 Hristea, Theodor 584 Hualde, Jose´ Ignacio 1069 Huber, Walter 811 Hübschmann, J. H. 100 Huck, Geoffrey J. 549 Huddleston, Rodney D. 248, 602, 656, 810, 816 Hudson, Grover 337, 463 f., 467, 474, 479, 1216 Hudson, Richard A. 313, 555, 1891 Huges, Richard 1619 Hugutio von Pisa 83 Hulst, Harry van der 165, 842 Humboldt, Alexander von 7⫺ 10, 12 Humboldt, Wilhelm von 19, 32, 92, 94 f., 98 f., 108, 116, 127, 157, 742, 840, 998, 1221⫺ 1224, 1228, 1247, 1474, 1575 f., 1590, 1835, 1873 Humphreys, Glyn 1771 Hunnius, Klaus 1285, 1701 f. Hurch, Bernhard 1651 Hurford, James R. 771 f., 774, 1064, 1791 Hutchison, John P. 374, 1135, 1205 Hyman, Larry M. 330, 383, 1099, 1232 Hymes, Dell 173, 175 f.
I Iacobini, Claudio 1706 Iatridou, Sabine 818 Ibn øUsfur (al-?Ashibili) 68 f., 73 f. Ibn al-?Anbari, Abu l-Barakat 69, 73 Ibn al-Sarraj, Abu Bakr 68 f., 71⫺73, 703 Ibn Faris, Ahmad 71 Ibn Jinni, Abu l-Fath 67⫺69, 70 f., 73 Ibn Qureish 74 Ibn Sina, Abdallah 69 Ibn Yaøish, Ibn øAli 71, 73 Ickelsamer, Valentinus 675 Ifrah, Georges 437, 773 Igla, Birgit 1644, 1650 Ilson, Robert 1889 f. Ineichen, Gustav 1235 Ingham, Bruce 1742, 1746, 1748 Ingo, Rune 647 f. Ingram, David 716, 784, 1003 Inhoff, Albrecht 1770 Inkelas, S. 150
1959
Namenregister Innes, Gordon 1135 Irigoin, Jean 1322 Irmer, Roland 446 Irvine, Judith T. 984 Isaac, Luc 1285 Isacenko, Alexander V. 129, 134, 609, 1091, 1099, 1303, 1305 f., 1309 Isakov, Isak A. 1367 Ishihara, Shinichiro 1060 Isserlin, M. 1818 Iturrioz-Leza, Jose´ Luis 238 f., 242, 1054, 1056, 1058⫺1061, 1064 Ivanova, Irina P. 420 Iverson, Gregory K. 878 Ivo, Hubert 676, 696
J Jaberg, Karl 521 Jachnow, Helmut 1309 Jachontov, Sergej E. 1237, 1239 f., 1552 Jackendoff, Ray 161, 188, 197 f., 346, 667, 671, 681, 841 f., 845, 847, 849, 900, 948, 1067⫺1069, 1250, 1790 Jackson, Howard 1884, 1886⫺ 1889 Jacobsen, Thorkild 37 Jacobsen, William H. 721⫺723 Jacobsen, William H. Jr. 985, 1012, 1088, 1202⫺1204 Jacobson, Steven 922 Jaeger, Jeri 1778 Jaeggli, Osvaldo A. 398 Jaggar, Philip J. 990 f., 993 Jahr, Ernst Ha˚kon 147 Jain, Dhanesh K. 1712 Jakobson, Roman 26, 33, 124, 127, 129, 134, 164, 204, 217 f., 238 f., 267⫺270, 272⫺ 277, 279 f., 287 f., 305, 336, 338, 340 f., 380, 436, 438, 442, 444, 493, 536, 608⫺612, 871, 1076 f., 1192, 1197, 1349, 1377, 1732, 1817 f., 1862 James, Deborah 1414 Janda, Richard D. 153, 401 f., 437, 580, 585, 877, 936, 1666 Jänicke, Otto 1698 Janis, Wynne 1559 Jankowsky, Kurt R. 1612 Janssen, Theo A. J. M. 985⫺ 988, 993, 1274 Janssen, Theo M. V. 851, 856 Janssens, Gerard 1742 Jansson, Tore 396 Janton, Pierre 1564 Janus 65
Jarceva, Viktoria N. 122, 135 Jarema, Gonia 1820 Jarvella, Robert J. 983, 1770, 1783 f. Jastrow, Otto 1742, 1746⫺1751 Javed, Naseem 966 Jaxontov, Sergej Je. 1177 Jedlicˇka, Alois 135 Jeffers, Robert J. 401 Jelinek, Eloise 721, 723 f., 1092 Jellinek, Max Hermann 113, 669, 672, 675, 697 f. Jendraschek, Gerd 1056 Jensen, Hans 1212, 1214 Jensen, John T. 20, 217, 346, 405, 573, 622, 628, 641, 835 Jerome 76 Jespersen, Otto 18, 110 f., 145 f., 201, 344, 352, 381 f., 446, 680, 688, 699, 758 f., 945, 948 f., 958, 1068, 1112, 1114, 1584, 1637, 1858, 1871 Jette´, Jules 1399⫺1401 Jisa, Harriert 1803 Joannes Balbus (John of Genoa) 84 f. Joeres, Rolf 1684 Johanson, Lars 753 Johnson, Janet H. 40 Johnson, Kyle 323, 329 Johnson, Mark 286 Johnson, Robert E. 423 Johnson, Samuel 1638 Johnson, Trevor 1557, 1560 Johnson-Laird, Philip N. 1091 Johnston, Raymond Leslie 560 Jolley, Catherine 1151 Jolly, Julius 106 Joly, Andre´ 694, 697, 699 Jonasson, Kerstin 986 Jones, Daniel 149, 251, 383 Jones, Eliza 1399⫺1401 Jones, Peter 985 Jones, Sir William 18, 93 f. Jong, Arie de 1563 Jönsjö, Jan 1677 Joos, Martin 179 Jordens, Peter 1810⫺1812 Jørgensen, Mogens Wied 1269 Joseph, Brian D. 330, 397 f., 401 f., 815, 1611, 1666 Josephs, Lewis 990 Joshi, Shivaram D. 508 Jouison, Paul 1559 Joyce, James 1621 Juilland, Alphonse 1637 Jung, Walter 677 Junger, Judith 1345, 1347⫺1349 Jungraithmayr, Herrmann 555 Jurafsky, Daniel 1049, 1893 Jurjani, øAbd al-Qahir 67, 73 Justinian 63
K Kaburaki, E. 734 Kageyama, Taro 882 Kahane, Henry 1889 Kahane, Rene´e 1889 Kahn, Daniel 336 Kainz, Friedrich 1621 Kaisse, Ellen M. 393, 411, 1264 f. Kalocsay, Ka´lma´n 1565, 1567 Kaltz, Barbara 693 f., 696⫺700 Kamenskich, Uljana K. 1388 Kamhi, Alan G. 1820 Kamp, Hans 911, 1170, 1179 Kamptz, Hans von 122 Kanakin, Igor A. 1376 Kandler, Günther 418 f. Kanngießer, Siegfried 887 Kant, H. 888 Kapesh, An Antane 1419 Kaplan, Lawrence 922 Kaplan, Ron 1894 Karau´lov, Jurij N. 135 Karcevski, Serge 719 Karger, Nikolaj K. 1376 Kari, James 412, 1400 Karius, Ilse 446 Karlgren, Bernhard 1730⫺1732, 1735 Karlsson, Fred 648, 652, 980, 1114, 1127, 1330 f., 1338 Karlsson, Göran 651 Karlsson, Keith E. 1709 Karmiloff-Smith, Annette 983, 1800 Ka´roly, Sa´ndor 1132 Karpf, Annemarie 290, 294, 1615, 1619, 1623 Karstien, Hans 547⫺550 Karttunen, Frances 400 Karttunen, L. 1896 Kasevicˇ, Vadim B. 1237, 1239 f. Kastenholz, Raimund 1092 Kaster, Robert A. 58 Kastovsky, Dieter 19, 195, 306, 441, 446, 465, 834, 847, 882, 906, 1275, 1626, 1628, 1631 f., 1674⫺1676 Katamba, Francis 155, 405, 436, 541, 622, 641 Katharina II 91 Katre, Sumitra Mangesh 45 Katwijk, Albert van 165 Ka¯tya¯yana 42 f. Katz, Leonard 1922 Kaufman, Terence 1645 f. Kay, Martin 1894 Kay, Paul 1093 Kaye, Alan S. 546, 1080 Kaye, Jonathan D. 430, 582 Kayne, Richard S. 323 Kazakov, Dimitar 1898
1960 Kaznelson, Solomon D. 677, 679 f., 682 f., 687⫺689 Kean, Marie-Louise 341, 1769, 1818 Keen, Sandra L. 1079 Keenan, Edward L. 245, 716, 853, 983, 986, 988, 991⫺993, 1153, 1161 Keil, Heinrich 53, 60, 62⫺65, 1930 Keizer, M. Eveline 1105 Kelemen, Jo´zef 1235 Keller, Jörg 1560 Keller, Rudi 1250 Kemmer, Suzanne 285, 1134, 1150, 1158 Kemp, Alan 694, 1149 Kempen, Willem 166 Kempgen, Sebastian 697, 1303, 1308 f. Kendall, Martha B. 569 Kennedy, Benjamin Hall 16 Kennedy, George A. 1731, 1736 Kenstowicz, Michael 429, 431 f., 469, 499 f., 502, 508, 1613 Kent, Roland G. 120 Kerleroux, Franc¸oise 143, 1286 Kern, Barbara 1214, 1216 Kern, Franz 105, 107, 110 f., 162 Kern, Peter Chr. 1683, 1692 Kessler-Mesguich, Sophie 703 Kettemann, Bernhard 458 Key, Mary Ritchie 488 Keyser, Samuel J. 506, 570 Khaidakov, S. M. 1038 Khrakovsky, Viktor 1146 f. Kibbee, Douglas 696 Kibrik, Aleksandr E. 482, 715, 1033, 1070 f., 1367 f., 1829 Kibrik, Andrej A. 1136 Kiefer, Ferenc 195, 1049 Kienle, Richard von 1683 Kihm, Alain 1658 f. Kilani-Schoch, Marianne 220, 289, 290, 533, 555 f., 579, 585, 962 Kilbury, James 337, 489, 494 Kiliaan, H. N. 951 Killingley, K. S.-Y. 781 Kim, Kong-On 568 f. Kimball, Geoffrey 1056, 1060 Kimenyi, Alexandre 1135, 1139, 1527⫺1531, 1533, 1535, 1538 f., 1541 Kimhi, Rabbi David 1345, 1348 King, Gareth 482, 660, 792 King, Robert D. 1612 King, Ross 1919 Kinkade, M. Dale 702, 720, 723 Kiparsky, Paul 26, 31, 121, 165, 189, 203, 331, 339, 346, 367, 412, 433, 471, 505, 508, 570, 653, 841, 854, 873, 879, 880,
Verzeichnisse und Sprachenkarten 946, 948, 1256, 1263, 1612 f., 1790 Kirsner, Robert S. 983, 993, 1099 Kirwan, Brian Edmond Renshaw 1187 Kisseberth, Charles 500, 502 Kitagawa, Yoshihisa 560 Kittredge, George Lyman 1619, 1621 f. Kiyomi, Setsuko 1061 Klaiman, Miriam 1078, 1085, 1149 f. Klamer, Marian 362 Klaus, Georg 3 Klausenburger, Jürgen 1602 f. Klavans, Judith L. 327, 393 f., 397 f., 400 Kleiber, Georges 990, 1290, 1296 Klein, Hamet 1024 Klein, Harriet E. Manelis 1421 Klein, Henny 1217 Klein, Thomas 1691 Klein, Wolfgang 983, 1180, 1273, 1810 Kleinhenz, Ursula 386 Kleinschmidt, Samuel 1389 Klimaszewska, Zofia 1045 f., 1049 f. Klimov, Georgij A. 1082 Klingebiel, Kathryn 1705 Klingenschmitt, Gert 122 Kloeke, Gesinus G. 163 Kloeke, Wus van Lessen 419, 1268 Klokeid, Terry J. 400 Kluge, Friedrich 121, 1616, 1684, 1911 Knobloch, Clemens 104 f., 669, 672, 680, 685, 687 Knobloch, Johann 1622 Knox, Robert 16 f. Kobler-Trill, Dorothea 1275 Koch, Harold 1491, 1579 Koch, Max 1905 Koch, Peter 1702, 1916 Koch, Walter A. 211, 214 Kocourek, Rostislav 1292, 1924 Kodzasov, Sandro V. 1367 Koefoed, Geert A. T. 233, 1574, 1580 Koehn, Edward 1214 Koehn, Sally 1214 Koenraads, Willy Henri August 1685 Kofod, Frances M. 1080 Kohrt, Manfred 1680 Koivulehto, Jorma 1328 Kolk, Herman 1818 Kollien, Simon 1560 Kølln, Herman 1312 Kolmer, Agnes 1067
Kölver, Ulrike 737, 1042 König, Ekkehard 362, 1205 Konrad, Reiner 1560 Kooij, Jan G. 1250 Köpcke, Klaus Michael 289, 479 f., 1035, 1036, 1268 f., 1272, 1808 f. Koptjevskaja-Tamm, Maria 658, 939, 976, 980 f., 1072, 1093 Korhonen, Mikko 701 Kornfilt, Jaklin 1079 Koschmieder, Erwin 613, 685 f., 688 Koskenniemi, Kimmo 207, 1833, 1894 Köstling, Pamela 1323 Koul, Omkar N. 1062 Kouloughli, Djamel Eddine 703 Koutsoudas, Andreas 628 Kouwenberg, Silvia 236, 1658 f. Kozinceva, Natalija A. 1134 Kozinsky, Ivan S. 1162 Koziol, Herbert 945, 1615 f., 1619, 1621, 1694 Krahe, Hans 1674 Kra´msky, Jirˇ´i 249, 252⫺254, 717 Krashen, Stephen D. 1806 Kraus, Alfred 1817 Krause, Scott Russell 560 Krauss, Michael E. 1022 f., 1028, 1399 Krebernik, Manfred 1916 Kreidler, Charles W. 957, 960 Krejnovicˇ, Eruchim A. 1376 f., 1380 Krifka, Manfred 1070 Krishnamurti, Bhadriraju 1032 f., 1043 Kristoffersen, Lars 1393 Kröber, Günter 2 Kroeber, Alfred 916, 1237 Kroeger, Paul R. 1475, 1481 Krohn, Robert 1640 Kronecker, Leopold 782 Kruisinga, Etsko 163, 427, 944 f. Krupa, Viktor 582, 1238⫺1244 Kruszewski, Nikolaj 126, 1602 f., 1606 Kryk-Kastovsky, Barbara 983 Kubrja´kova, Elena S. 133, 406, 419, 421 f., 424, 886, 888, 892⫺894 Kucˇera, Antonin 1888 Kucˇera, Henry 253 Kuczaj, Stan A. 1798, 1801 Kudo, Takayaki 1819 Kuen, Heinrich 1703 Kuhn, Dorothea 16 Kuhn, Wilfried 1904 Kühner, Raphael 655 Kühnhold, Ingeburg 355 f., 890, 933 Kuiken, Folkert 1821
1961
Namenregister Kuipers, Aert H. 1020, 1024 Kujore, Obafemi 1312 Kukich, Karen 1897 Kukuczka, Elena 1097 Kummer, Werner 811 Kuno, Susumu 734 Kupfer, Peter 676, 1736 f. Kuroda, Sige-Yuki 988 Kürschner, Wilfried 160, 901, 907, 1278 Kuryłowicz, Jerzy 32, 121 f., 365, 376, 456, 610, 612, 680, 682, 758, 813, 931, 946, 984, 1080, 1590, 1593, 1613 Kusters, Wouter 1655 Kutscher, P. 1760 Kuzneco´v, Petr S. 128 Kuznecova, Ariadna I. 131, 889
L La Brosse, Jean-Baptiste 1412 Labov, William 241, 243 f., 1665, 1761, 1764 Laca, Brenda 873 Ladd, Robert 899, 1232 Lafford, Barbara A. 272 Lafitte, Pierre 1199 Lafon, Rene´ 1002 Lagarde, Jean-Pierre 693, 698⫺ 701 Laine, Matti 1341, 1819 Lakoff, George 283, 286, 297 f., 611, 1018 Lakoff, Robin 993 Lalleman, Josine A. 1807 Lallemand-Rietkötter, Annette 890 f. Lallot, Jean 694 Lambert, Fre´de´ric 694, 696 Lamberterie, Charles de 1325 Lambrecht, Knud 984 Lancelot, Claude 697 f. Landaburu, Jon 740 Landau, Sidney I. 1887⫺1889 Landeweerd, Rita 1174 Landsberger, Benno 35, 37 Langacker, Ronald W. 257, 259⫺262, 289, 297, 524, 727, 926, 983, 1068 f., 1097, 1121, 1129, 1203, 1205, 1436 f., 1597 Langdon, Margaret 569 f., 1023, 1063 f., 1088, 1204 Langendoen, Donald Terence 150 Lapenna, Ivo 1564 Lapointe, Steven G. 197, 199, 321, 327 f., 346, 395, 900 Lardiere, Donna 1813 f. Large, Andrew 1563 Larochette, Joe 221 f.
Larrimore, Bonni 711 Larsen-Freeman, Diane 1806 Laskowski, Roman 289, 433, 596 Lass, Roger 430, 465, 507, 637 Laudanna, Alessandro 1772, 1778 f. Launey, Michel 1433, 1437, 1439, 1445, 1447⫺1449, 1451 Laurendeau, Paul 699 Laury, Ritva 983, 988, 993 Lausberg, Heinrich 655, 1703 Law, Vivien A. 60, 63, 86, 696, 703 Laycock, Donald C. 1756 Lazard, Gilbert 990 Leˆ Quang Trinh 1552 Leˆ Vaˇn Ly 1549, 1552 Leˆ Xuaˆn Thi 1551 Le, Tam Duy 814 Leau, Le´opold 1563 Leben, William R. 588 Lecle`re, Christian 699 Lecours, Andre´ Roch 1817 Lee, Jennifer R. 926 Leech, Geoffrey 596, 607, 609 f., 747, 837, 1627 f., 1632 Leer, Jeff 1404 Lees, Robert B. 188, 195 f., 204, 352, 357, 906, 907, 1255 Lefebvre, Claire 828 Lehfeldt, Werner 597, 633, 1239⫺1245, 1308 f. Lehmann, Christian 12 f., 32, 242, 244, 262, 660 f., 683, 685, 714, 737, 754, 809, 859, 864, 1065, 1086, 1202, 1229, 1248, 1250, 1481, 1590, 1593 f., 1611, 1836, 1844, 1860, 1870 f. Lehmann, Thomas 761 Lehmann, Winfred P. 121 f., 1594 Lehnert, Martin 1637, 1639 Leibniz, Gottfried Wilhelm 1555 Leinonen, Marja 614, 1183 Leirbukt, Oddleif 1273 Leisi, Ernst 1619 Leiss, Elisabeth 641, 1274 Lejeune, Michel 585 Lely, Heather K. J van der 1820 f. Lemare´chal, Alain 689, 702, 722, 1475 Lemmens, Marcel 1891 Lemoine, Georges 1412 Lene´, Gustav 1709 Leo´n, Maria Lourdes de 1021, 1025 f., 1028 Leonard, Laurence 1820⫺1822 Leontiev, V. V. 315 Leopold, Werner F. 1615, 1621, 1623
Lepsius, Richard 96 Leser, Martin 891 Lessen, Jacoba H. van 163 Leumann, Manu 120 f. Leuninger, Helen 1769 Levelt, Willem 1769, 1780 Leven, Regina 1560 Levi, Judith 352 f., 358, 909, 1255, 1262 Levin, Aryeh 71 Levin, Beth 845, 847 Levin, Juliette 559 Levin, Saul 441, 444 Levinsohn, Stephen H. 1197, 1454 Levinson, Stephen C. 983 f. Levy, Yonata 1797, 1799, 1822 Le´vy-Bruhl, Lucien 1091 Lewandowski, Theodor 436 Lewis, Geoffrey L. 232, 314, 351, 825, 969, 1073, 1097, 1106, 1191, 1193 f., 1197 f., 1215, 1583 Lewy, Ernst 1229, 1311 Leys, Odo 1269, 1905 Li, Charles N. 240, 749, 810, 940, 950, 1020, 1113, 1191 f. Li, Fang-kuei 1730, 1733 Li, Jinxi 704 Liang, Franklin Mark 1897 Libera, Alain De 696 Liberman, Mark 899, 1232 Lichtenberk, Frantisek 999, 1097 Liddell, Scott 423, 1557, 1560 Lieb, Hans-Heinrich 1836 f., 1851 Lieber, Rochelle 31, 199 f., 202, 318⫺321, 337, 350, 354, 357, 408 f., 411, 414, 433, 445 f., 479, 568 f., 572⫺574, 606, 860 f., 871, 873, 880, 900, 904, 908, 946, 948, 1255 Lightfoot, David W. 1265 Lightner, Theodore M. 421, 515 Lima, Susan D. 1770 Limburg, Machiel J. 823 Lindemann, Margarete 1707 Lindenfeld, Jacqueline 1078 Lindner, Thomas 1704 Linebarger, Marcia C. 1818 Linell, Per 338, 507, 1792 Lipka, Leonhard 122, 446, 464 f., 834 f., 837, 887, 894, 901, 906, 1626 f., 1632, 1634 Littell, Robert 836 Littre´, E´mile 16, 18 Liu, Jian 703 Ljungerud, Ivar 1270 Llamzon, Teodoro A. 1479 Lombardi, Linda 548 Lomono´sov, Michail V. 125, 1300 Lomtadze, E`lizbar A. 1367
1962 Long, Antony A. 54 Longacre, Robert E. 179, 192 f., 1202, 1501, 1831 f. Lönnrot, Elias 1338 f. Loos, Eugene 1141 Lopatin, Vladimir V. 132, 419, 424, 892, 1309 Lopez, Cecilio 1474, 1486 Loporcaro, Michele 452, 1702 Lorentz, James P. 1271 Lorentz, Ove 147 Lorimer, David L. R. 766 Lotz, John 436 Lötzsch, Ronald 1569 Lounsbury, Floyd 176, 178 f., 181⫺183, 186 f. Loveland, Katherine 999 Lü, Shuxiang 1736 Lüdtke, Helmut 1250, 1685 Ludwig, Otto 1917 Lukatela, Georgije 206, 1783 Luraghi, Silvia 639, 641⫺645 Luria, Aleksander R. 1816 f. Luschützky, Hans-Christian 289, 454, 458 Luther, Martin 1692 Lyall, Gavin 836 Lynch, John 1204 Lynge, Kristoffer 1397 Lyon, Otto 105 Lyons, Christopher 327 f., 983 f. Lyons, John 28, 32, 149, 212, 228, 248⫺250, 252⫺254, 277, 496, 596, 671, 686⫺688, 762, 809, 811, 815, 834, 837, 844, 973 f., 983, 986, 1071, 1104, 1111, 1115, 1192, 1194 f., 1213, 1625, 1627, 1630, 1632 f., 1735, 1887, 1889⫺ 1891
M Ma, Jianzhong 703 Maas, Utz 1917 Macaulay, Monica 546, 766, 1062 MacDonald, John D. 836 MacDonald, Lorna 1191, 1195 MacKay, Donald G. 576, 1791 Mackenzie, J. Lachlan 974 f., 979 f. MacKenzie, Marguerite E. 1412 Mackinnon, Roderick 571 Mackridge, Peter 814 f. Macris, James 960 Macrobius 64 MacWhinney, Brian 206, 289, 345 f., 364, 795, 879, 1798 f. Madhavan, P. 546 Maes, Hubert 701
Verzeichnisse und Sprachenkarten Mager, Karl Wilhelm Eduard 698 Mahapatra, Bijaya P. 1041 Mahmud, Ushari Ahmad 1752 Maiden, Martin 302, 456, 1069 Mailhot, Jose´ 1412 Majewicz, Alfred F. 526 Malchukov, Andrej L. 1151 Malicka-Kleparska, Anna 433, 834, 878, 1255, 1260 Malkiel, Yakov 536, 542 f., 890⫺ 892, 1708 f. Mallinson, Graham 825, 1083, 1086 f. Malmberg, Bertil 147 f. Malotki, Ekkehart 987, 993 f. Manandhar, Suresh 1898 Manandise, Esme´ralda 437, 817 Manaster-Ramer, Alexis 1475 Manchester, Martin 1221 Man´czak, Witold 121, 1252, 1613 Manelis, Leon 1785 Mann, Thomas 1621 Mannheim, Bruce 1004 Manning, Christopher D. 1893, 1898 Manzelli, Gianguido 739 Marantz, Alec 204 f., 321 f., 411, 413, 467, 561 f., 565 f., 935, 1131 f., 1141, 1477 Maratsos, Michael P. 1799 Marchand, Hans 160, 195, 198, 350, 355 f., 359, 444⫺446, 834 f., 867, 870⫺872, 874, 887, 889⫺894, 898 f., 901 f., 904⫺908, 931 f., 935, 944⫺ 949, 958, 965, 971 f., 1255 f., 1259 f., 1263 f., 1276, 1631⫺ 1633, 1674 f., 1678, 1789 Marchese, Lynell 984, 987, 994 Marcos, Habte M. 569, 573 Marcos, Imelda 284 Marcus, Gary F. 479 f., 1798 Mardirussian, Galust 916, 924 Markey, Thomas L. 511, 519 Marle, Jaap van 32, 164 f., 225 f., 230⫺233, 265, 305, 307 f., 310, 317, 342, 347, 365, 836, 841, 848, 877 f., 880, 933, 1578, 1580 f., 1585 f., 1641, 1643 Marouzeau, Jules 140, 142 Marr, N. Ja. 127, 135 Marre, Aristide 1474, 1479 Marshall, J. H. 696 Marslen-Wilson, William D. 412, 539 f., 1770, 1773 f., 1778, 1780 f. Martin, Jack 583⫺585 Martin, James H. 1893 Martin, Maisa 1341 Martin, Pierre 488
Martin, Samuel E. 1114, 1732 Martinet, Andre´ 24, 142, 214, 217, 220, 247⫺249, 251 f., 338, 406, 454, 485⫺488, 490, 492⫺494, 693, 1285, 1571, 1645, 1870 Martı´nez de Sousa, Jose´ 958 Martı´nez Moreno, Annette 1699 Martini, Francois 1552 Marx, Karl 2, 142 Mascaro´, Joan 340 Masica, Colin 1092, 1712 Ma´slov, Jurij S. 129, 612, 648, 650 Maspero, Henri 1545 Mater, Erich 1272, 1637 Mates, Benson 54 Mathesius, Vile´m 668 Matisoff, James A. 1593 Matonis, Ann 88 Matsuda French, Koleen 1474, 1476⫺1478 Matsumoto, Yo 1018 f. Mattens, Willy H. M. 165 Matthaei, Rupprecht 16 f. Mattheier, Klaus 1682 Matthew 284 Matthews, George Hubert 1197 f. Matthews, John 1820 Matthews, Peter Hugo 5, 27, 52, 72, 122, 152⫺155, 172⫺177, 179, 187⫺189, 195, 200 f., 205, 217, 248 f., 254, 266⫺ 268, 320, 346, 350, 356, 363, 373, 405, 409, 437, 453, 460, 463, 474, 491, 526, 529, 546, 548, 588, 595 f., 605, 616, 618 f., 622⫺624, 626 f., 638, 641, 647, 655, 657, 816, 837, 841, 1074, 1226, 1475 Matthiessen, Christian 1202 Mauro, Tullio de 699, 1638 Maximus Planudes 54, 59, 1076 Maxwell, Mike 1899 Mayer, Erwin 1615, 1617, 1620⫺1622 Mayer, Franz Joseph Carl 17 Mayer, Karl 1623, 1768 Mayerthaler, Willi 122, 161, 220, 270, 274, 278 f., 288⫺292, 294, 443 f., 453, 1650 McArthur, Harry S. 981 McArthur, Lucille E. 981 McBain, Ed 836 McCarthy, John J. 151, 189, 205 f., 341 f., 385, 410, 431, 458, 467, 501, 546⫺548, 553⫺556, 560⫺566, 569 f., 572 f., 970, 1045, 1255, 1348, 1613 McCaughey, Terence P. 1095 McCawley, James D. 385 f., 508, 965
1963
Namenregister McClelland, James L. 1766, 1780, 1807 McCloskey, James 328 McCoard, Robert W. 1186 McConnell-Ginet, Sally 1213 McElhanon, Karl 1501 McFarland, Curtis D. 1473 f., 1479, 1481, 1485 f. McGregor, Karla 1821 McGregor, William B. 1093, 1203, 1863 f., 1871 Mchombo, Sam 330, 477, 784 McIntosh, Mary 624 f. McIntyre, Linda Lee 563 Mcire, Ephrem 87 McKay, Graham 1008 McMillan, James B. 546 McNeill, David 300, 540 McQuown, Norman A. 1869 Meder, Gregor 447, 650 Meeussen, Achiel Emiel 166, 253, 589, 1527, 1532, 1540 Mei, Tsu-lin 1731 Meid, Wolfgang 1674 Meier, Georg Friedrich 4, 91 f., 218, 435, 437, 683, 685, 687, 693, 698 f., 701 Meier, Michael 1325 Meier, Richard P. 1823 Meigret, Louis 697 Meillet, Andre´ 19, 122 Meillet, Antoine 140, 142, 214, 245, 249, 537, 813, 1042, 1590 f., 1611, 1645, 1664⫺ 1666 Meiner, Johann Werner 698 Meinhof, Carl 1094 Meisel, Jürgen 1799 Meisenburg, Trudel 1917⫺1919, 1921 Mejiers, Guust 1783⫺1785 Mejlach, Michail 1237, 1241 Mel’cˇuk, Igor A. 23, 26, 28, 32 f., 129 f., 135, 235, 237⫺ 239, 266 f., 419 f., 424, 436, 438, 440, 442, 444, 453, 456 f., 481 f., 511 f., 514 f., 517, 523, 525 f., 528, 530, 532, 536, 542, 552, 582 f., 585, 588, 601, 604, 656, 660, 709, 712, 841, 867, 935 f., 945, 1079 f., 1084, 1130, 1146 f., 1160, 1309 Melanchthon, Philipp 697, 1930 Melchert, Craig 641 Me´lis-Puchulu, Agne`s 1292 Me´ndez Dosuna, Julia´n 291 Meneses, Porfirio 1462 Menn, Lise 879, 1796, 1798, 1818 f. Menninger, Karl 774 Menovsˇcˇikov, Georgij Alekseevicˇ 1651
Menyuk, Paula 1820 Meringer, Rudolf 1623, 1768 Merlan, Francesca 237, 744, 751, 835, 916, 924, 926, 978, 1034, 1082 Merlini Barbaresi, Lavinia 220, 290, 457, 460, 583 f., 934, 1045 f., 1050, 1817 Merrifield, William R. 405, 622 Mervis, Carolyn B. 264 Merx, Adalbert 59 Metcalf, G. J. 92 f. Mey, Jacob 1091 Meyer, Gustav 1649 Meyer-Klabunde, Ralf 910 f. Meyer-Lübke, Wilhelm 1698, 1705 Meys, Willem Johannes 165 f., 763, 834 Miceli, Gabriele 1818 f. Michael, Ian 697 Michels, Victor 1683 Micla˘u, Paul 214, 220 f. Miettinen, Erkki 1615, 1622 f. Mifsud, Manwel 1056, 1745, 1750 Mignot, Xavier 1325 Mikes, Melanie 1797 Milewski, Tadeusz 134 f. Miller, Carol 1821 Miller, George A. 1091 Miller, Philip Harold 392, 395, 402 Miller, Robert L. 249 Miller, Roy Andrew 51, 703 Miller, Ruth 1807 Miller, Wick R. 1211 Mills, Anne 1800, 1820 Milner, George B. 578 Milner, Jean-Claude 835 Miloslavskij, Igor’ G. 422 Mimouni, Zohra 1820 Miner, Kenneth L. 920, 924 f. Minnen, Guido 1899 Mitchell, Terence F. 555, 620 Mitchell, Tom M. 1898 Mithun, Marianne 204, 315, 322, 442, 521, 711, 736 f., 785, 801, 805, 889, 916, 919 f., 924⫺926, 975, 1022 f., 1026, 1028, 1058 f., 1065, 1135, 1202, 1800 Miyaoka, Osahito 1147, 1162 f. Moder, Carol Lynn 633, 1778 Moens, Marc 1168 f., 1179 Mogford-Bevan, Kay 1823 Mohanan, Karuvannur P. 203, 479, 507, 1790 Mohanan, Tara 713 Mok, Q. I. M. 1285 Molina, Alonso de 1433 Molinier, Guilhem 85, 88 Monnot, M. 960
Monod-Becquelin, Aurore 1855 Monsell, Stephen 1770, 1781 Montague, Richard 201, 851, 909, 1182 Montler, Timothy R. 547⫺549, 582, 584 Moortgat, Michael 165 f., 201, 842, 860, 863 Moravcsik, Edith A. 205, 280, 287, 546⫺549, 558, 561, 710, 759, 935, 941, 1085, 1645 Morel, Mary-Annick 983 Moreno, Martino Mario 577 Morev, Lev Nikolaevicˇ 1216 Morgan, William 316 f., 1098, 1136 Morgenstern, Christian 456 Morin, Yves-Charles 363, 582 f., 877, 1603 Moritz, Karl Philipp 108 Morpurgo-Davies, Anna 91⫺93, 95, 98 Morris, Charles W. 211, 213, 680 Morris, Henry Francis 1187 Morrow, Daniel G. 258, 261 Morton, John 1790 Morvan, Danie`le 1291 Mosel, Ulrike 242 f., 582, 993, 1070, 1107, 1764, 1859, 1869 f., 1872 Moser, Hugo 1271 Moser, Virgil 1680, 1683 Moshi, Lioba 1135 Moskal’skaja, Olga I. 678 f., 1273 Moskalev, Aleksej Alekseevicˇ 1216 Motapanyane, Virginia 389 Motley, Michael T. 576 Motsch, Wolfgang 160 f., 832 f., 886, 906, 1278, 1773 Mous, Maarten 1646 Mueller-Vollmer, Kurt 94 Mufwene, Salikoko S. 811, 1070 Mugdan, Joachim 20, 236 f., 251, 405, 408, 417, 439 f., 445, 447, 451⫺453, 460, 474, 477, 485, 489, 508, 526, 536, 542, 596, 600, 631, 650, 867, 891⫺894, 933, 1268⫺1270, 1275, 1601 f., 1616, 1620, 1638, 1773, 1882, 1888⫺1891 Mühlhäusler, Peter 1006, 1756, 1758 f., 1761 f., 1764 f. Mulder, Jan W. F. 213 f. Mulisch, Herbert 1309 Müller, Beat Louis 665 Müller, Friedrich 1474, 1479 Müller, Max 91, 98, 101, 162 Müller, Peter O. 1684 Müller, Walter M. 1741 Müller, Wolfgang 1890
1964 Müller-Schotte, H. 901 Mundt, Christoph 1817 Munro, Pamela 814, 1109, 1142, 1203 f. Muratori, Carlo 772 Muravyova, Irina A. 477, 538 Murethach 79 f. Murjasov, Rachim Z. 1546 Murray, Ben 1491 Murru, Furio 59 Musaev, Kenesbaj M. 439 Mussies, Gerard 164 Muysken, Pieter C. 165, 412, 541, 1140, 1454, 1459, 1461, 1647, 1656 f. Myhill, John 921
N Naes, O. 679 Na¯ges´abhatøtøa 41 Napps Shirely E. 1783 Naro, Anthony J. 1654 Nash, David 391, 396, 400, 558 Nattiez, Jean-Jacques 211 f. Naumann, Bernd 92, 115, 698, 886, 892, 894, 898, 932, 1275 f. Nauta, A. H. 228 Naville, Adrien 215 Naylor, Paz B. 1475 Nedjalkov, Vladimir P. 315, 916, 918, 1132 f., 1142, 1177 Nedyalkov, Igor 1151, 1160 Neef, Martin 468 f., 472, 479 f., 1269, 1275, 1922 Neeleman, Ad 165 f. Nekra´sov, Nikolaj P. 125 Nemcˇenko, Vasilij N. 894 f. Nespor, Marina 341, 385, 413 Nespoulous, Jean-Luc 1818 Neumann, Werner 104 Neuß, Elmar 908 Nevis, Joel Ashmore 389, 397, 400 f., 546 Newfield, Madeleine 279 Newman, Paul 553, 786, 788 Newman, Stanley 1119 Newmeyer, Frederick 321 Newport, Elissa 1557, 1823 Ngai, G. 1898 Nguyeeˆ˜ n, Huy Caaœˆ n 1551 Nguyeeˆ˜ n, Kim Thaœ n 1547 Nguyeeˆ˜ n, Phu´ Phong 984, 988, 992- 994, 1550 Nguyeˆn, Íı`nh-Hoa` 720, 1216 Nichols, Johanna 458, 553, 975, 1045, 1081, 1093, 1097, 1119, 1122, 1197, 1203, 1227, 1229 f., 1248, 1400, 1491, 1515 Nicodemus, Lawrence 720
Verzeichnisse und Sprachenkarten Nida, Eugene A. 6, 20, 179⫺ 182, 184 f., 193, 217, 219, 236 f., 336, 350, 405, 407, 418, 438, 440⫺442, 452 f., 463⫺465, 467, 474, 526, 538, 541⫺543, 573, 582, 588, 880, 1249, 1789, 1831 Niemi, Jussi 1341 Niemikorpi, Antero 1329 Nikolaeva, Irina 1215 Nikola´eva, Tat’jana M. 133 Ningomba, M. S. 759, 1209 Nissen, Hans J. 1916 Nissim, Gabriel M. 590 Nitti-Dolci, Luigia 42 Noailly, Miche`le 1296 Nöldeke, Theodor 1741 Noll, Craig 628 Nolte, Gabriele 613 Noonan, Michael 20, 32, 220, 799, 1200, 1203, 1859, 1863 Nooteboom, Sieb G. 540, 1791 Nordlinger, Rachel 718 Noreen, Adolf 452 Noreen, Alfred 110, 147 f. Norman, Jerry 724 f., 937, 1734 Noske, Manuela 1512 f. Nöth, Winfried 210 f. Noyer, Robert Rolf 603 Nübling, Damaris 1271 f., 1903 Nuessel, Frank 1570 Nunberg, Geoffrey 983 Nybakken, Oscar E. 1926 Nyrop, Krystoffer 139, 1285, 1288
Oliverius, Zdene˘k F. 132 Olmos, Andrei de 1433 Olmsted, David 439 Olschansky, Heike 1635, 1905 Olsen, Gloria 1025 Olsen, Susan 161, 444⫺446, 892, 894, 898, 902⫺905, 933, 1278, 1767 Olson, David R. 1916 Olson, Michael 1110 Omar, Margaret 1797 Onions, Charles Talbut 1616 Onischi, Masayuki 1032, 1043 Orlandini, Anna 1213 Ornan, Ozi 1345 Ortmann, Albert 466 Ortmann, Wolf Dieter 1272 Ortner, Hanspeter 355, 898, 1278 Ortner, Lorelies 355, 898, 1278 Osada, Toshiki 720, 726, 1209 Osbern von Gloucester 83 Osborne, Charles R. 373, 801 Oshima-Takane, Yuriko 999 Ossner, Jakob 699 f. Osthoff, Hermann 10, 100, 510, 514, 519 Osumi, Midori 764 Oswalt, Robert 1202 f. Otanes, Fe T. 390, 399, 543, 569, 721 f., 743, 749, 760, 786, 1203, 1474⫺1478, 1481⫺ 1483, 1486 Ouhalla, Jamal 330, 1818 Owens, Jonathan 67, 69 f., 72, 702 f., 1742, 1752
O P o` Cro`inin, Donncha 569, 571 ´ Cuı´v, Brian 88 O ´ Siadhail, Mı´chea´l 503 O O’Dochartaigh, Cathair 1215 O’Leary, Lacy de 1096 Ober, Dana 1041 Oberlies, Thomas 1712 Obler, Loraine 1818 f. Odden, David 559, 565, 587 Odden, Mary 559, 565 Oehrle, Richard 818 Oesterreicher, Wulf 1916 Oetting, Janna 1821 Oettinger, Norbert 122 Oflazer, Kemal 1899 Ogulnick, Karen A. 446 Ohnheiser, Ingeborg 888 f. Oinas, Felix J. 401 Ojeda, Almerindo E. 549 Oka, I. Gusti Ngurah 999 Okrand, Marc 577⫺579 Oksaar, Els 1645 Olbertz, Hella G. 1195 Ölinger, Albert 697
Paccia-Cooper, Jeanne 540 Padden, Carol 1557 Padley, G. Arthur 697 f. Pagliuca, William 443, 537 f., 550, 814 Pajunen, Anneli 1333 Pallas, Peter Simon 91 Palmer, Abraham Smythe 1621 Palmer, Frank R. 151, 667, 809, 811, 817, 1197 Palomäki, Ulla 1333 Pa¯nø ini 41⫺51, 93, 179, 186, 437, 508, 703, 762, 854, 877⫺879, 1712 f., 1715 f., 1718 f. Panagl, Oswald 122, 289, 1213 Panfilov, Valerij 716, 1552 Panhuis, Dirk 1324 Pano´v, M. V. 132 Panzeri, Maria 1819 Papias 83 Papp, Ferenc 135 Paradis, Michael 1820 Paris, Marie-Claude 992
1965
Namenregister Park, Jeong-Woon 1120 Parker, Elizabeth 988 Parker, Gary J. 1453, 1455, 1458 Parks, Douglas R. 804, 1196 Pasch, Helma 1040, 1094 Passerieu, Jean-Claude 703 Patan˜jali 42 f. Paul, Hermann 10, 19, 100, 110, 112, 157⫺160, 162, 670, 675, 878, 898, 901, 945 f., 1579, 1590, 1594, 1597, 1601, 1612 f., 1623, 1626, 1683 f., 1686⫺1689, 1691, 1766 Paul, Jean 109 Paulissen, Det 165 Paunonen, Heikki 1341 Pawley, Andrew 713, 1142 Payne, David L. 987 Payne, Doris 978 f., 919 f., 1018, 1022⫺1025, 1141, 1421, 1428, 1465, 1467⫺1469, 1472 Payne, John R. 1208 Payne, Thomas E. 978 f., 1154, 1464 f., 1467, 1861 Peano, Giuseppe 771 Pearsall, Judy 1883 Pecci, Antonio 1430 Pedersen, Annegrethe 1559 Pedersen, Holger 145, 177 Pe´e, Willem 163 Pei, Mario 536 Peirce, Charles Sanders 212 f., 220, 282, 288, 290, 300 f. Pellegrini, Ines Angela 1047 Pelletier, Francis J. 1067 Peltzer-Karpf, Annemarie 1823 Pelzter, Louise 990 Pen˜a Cahuachi, Hilario 1471 Pennanen, Esko V. 444⫺446, 872 Pensado, Carmen 291 Penzl, Herbert 381, 1681 Percival, W. K. 97 Perel’muter, Ilja Aronovicˇ 1318 Pe´rennec, Marcel 699 Perkins, Revere D. 443, 550 Perlmutter, David M. 321, 361, 366, 399 f., 1131, 1140, 1148, 1448 Perner, Josef 983 Perpillou, Jean-Louis 1325 Perrot, Jean 142 Pesetsky, David 203, 320, 414, 863, 1261, 1790 Pesˇko´vskij, Aleksandr M. 127, 612 Peters, Ann M. 1796 Peters, Elizabeth 836 Peters, Toine 1237, 1239 Petersen, Walter 905 Peterso´n, M. N. 127 Petraeus 701 Peuser, Günther 1818
Peytard, Jean 142 Pfeffer, J. Alan 1271 Pfeifer, Wolfgang 16 Philippe de Thaun 1620 Philippe, Benoıˆt 1564 Phillips, Betty S. 1809 Phocas 63, 77 Picabia, Le´lia 994 Picard, H. B. 18 Pichon, Edouard 139, 142, 1288 Pickett, Velma B. 440, 442, 489, 541, 1831 Piel, J. M. 1698 Pierce, Joe E. 1237 f. Pierer, Heinrich August 15 Pike, Evelyn G. 192⫺194, 1831 Pike, Kenneth Lee 179, 191⫺ 194, 245, 337, 341, 418, 431, 452, 484, 587 f., 1270, 1831 Pillon, Agnesa 1785 Pinault, Georges-Jean 703, 1214 Pinchon, Jacqueline 141, 1285 Pinker, Steven 880, 882, 1778, 1807 Pinkster, Harm 813 f. Piron, Claude 1570 Pischel, Richard 1712, 1719 Pitman, Isaac 149 Pittman, Richard 1486 Plag, Ingo 1255, 1259, 1780 Plank, Frans 94, 255, 289, 298, 361, 420, 465, 478 f., 550, 585, 618, 620, 624, 626, 639, 645, 647 f., 653, 684 f., 687, 710, 746, 754, 833, 866, 878, 887 f., 890, 893, 895, 930, 999, 1057 f., 1061, 1079, 1081, 1085, 1491, 1615, 1620, 1623, 1868 Plante, Elena 1820 Platon 53 f., 56 f., 264, 417, 694 f., 953 Platt, Carol B. 1798 Platz, Bärbel 698 Ple´h, Csaba 1797 Pleines, Jochen 667 Ple´nat, Marc 143 f., 583, 1285 f., 1289 Plotkin, Vul’f Ja. 1376 Plungjan, Vladimir A. 442, 445 Pogonowski, Jerzy 526 Poitou, Jacques 684 f. Polenz, Peter von 1275, 1278 Polinskaja, Maria S. 916, 918 Pollak, Hans W. 452 Pollok, Jean-Yves 330, 332, 818 Pomorska, Krystyna 272, 275, 277, 280 Poppe, Nicholas N. 442 Porı´zka, Vincenc 1094 Porotova, Tel’mina 1382 Port, Robert F. 1135 Porter, M. F. 1896, 1899
Porzig, Walter 116, 1325, 1623 Pospelov, Nikolaj S. 127, 677 Postal, Paul M. 199, 428, 1131, 1140, 1148, 1265 Posti, Lauri 1337, 1340 Potebnja´, Aleksandr A. 125, 610, 612 Pott, A. F. 9 Pottier, Bernard 141 Poulisse, Nanda 1812 Powlison, Esther 1465, 1467 Powlison, Paul 1464 f., 1467 Pratchett, Terry 249 Prcic, Tvrtko 1889 Prell, Heinz-Peter 1684, 1694 Premper, Waldfried 737 Press, Ian 760, 823, 979, 1063, 1065 Press, Margaret L. 767 Preuss, Fritz 446 Pride, Leslie 588 Priestly, Tom M. S. 1043 Prince, Alan S. 206, 341 f., 385, 431, 467, 471, 555 f., 562⫺ 566, 569 f., 1045, 1613, 1807 Prior, Arthur 1180, 1182 Priscian 53, 58⫺60, 62⫺65, 77⫺ 79, 82⫺86, 674, 695 f., 732, 1929 Probus 63 Prochazka, Theodore 1742, 1744 Proctor, Paul 1885, 1889 f. Progovac, Ljiljana 789 Prokosch, Erich 1752 Protagoras 1031 Przyluski, Jean 1545 Puhvel, Jaan 1214, 1216 Pulleyblank, Douglas 541, 558, 561, 563 Pulleyblank, Edwin 1730⫺1735, 1737 Pullum, Geoffrey K. 297, 389 f., 617, 818, 965, 967, 969, 972, 1078, 1255, 1265, 1666 Pustet, Regina 768, 1100 Putschke, Wolfgang 1680 Putten, Frans van der 166, 864 Putzer, Oskar 933 Pye, Clifton 1795 Pylyshyn, Zeon Walter 1766 Pym, Noreen 711
Q Qafishesh, Hamdi A. 980 Quine, Willard V. O. 1067 f. Quintilian 64, 1076, 1929 Quirk, Randolph 316, 648, 760, 765, 790, 947, 1105, 1215, 1255, 1625 f.
1966 R Rabain-Jamin, Jacqueline 1001 Rabin, Chaim 1741, 1743 Radtke, Petra 1273 f. Raecke, Jochen 1309 Rainer, Franz 308, 347 f., 367, 584 f., 873, 878, 912, 1049, 1291, 1704 f., 1710 Raja, K. Kunjunni 42 Rajaona, Simeon 702 Ramat, Paolo 300, 817, 1222 Ramers, Akrl Heinz 1276 Ramos, Teresita V. 1479, 1481 Ramus, Petrus 697 f. Randall, Janet 864 Ranjeva, Henri 702 Rappaport, Malka 845, 847 Rask, Rasmus Kristian 7, 95, 145, 1127 Raster, Peter 1244 Ratcliffe, Robert R. 1055, 1057, 1062 Rathay, Wilfried 446 Rauh, Gisa 983 Ravid, Dorit 1801 Rayner, Keith 1770 Raynouard, Franc¸ois-JusteMarie 139 Reaney, P. H. 1677 Redder, Angelika 1274 Reed, Carrol E. 1271 Reed, Irene 1082 Reed, Judy 987 Reesink, Ger P. 1003, 1011 f., 1056, 1062 f., 1202⫺1205 Reeves, Jennifer Elaine 435 f. Reforma´tskij, Aleksandr A. 128, 130, 508, 513, 526 Reh, Mechthild 802, 805, 1091, 1095 f., 1107, 1142, 1590, 1592, 1595 Rehg, Kenneth 1022 Reich, Carl Gottlob 1555 Reichenbach 1180, 1182 f., 1186 f. Reichling, Anton 163 f., 852 Reichmann, Oskar 1681 Reid, Lawrence A. 1474 Reid, Nicholas 1042 Reifferscheid, Augustus 58 Reiner, Erica 35, 40 Reinhart, Tanya 789 Reis, Marga 1273 Remigius von Auxerre 79, 82 f., 85 Remmius Palaemon 62, 696 Rennison, John R. 759 f., 766, 768 Renou, Louis 437, 703 Restrepo, Maria Adelaide 1821 Retsö, Jan 1746 Rettig, Wolfgang 1268 f.
Verzeichnisse und Sprachenkarten Reuland, Eric J. 789, 818 Reuther, J. G. 1491 Re´vzin, Isaak I. 132 Rey, Alain 1883, 1927 Reyle, Uwe 911, 1170, 1179 Rhodes, Richard A. 1049 Rhys Jones, Thomas J. 572 Rice, Keren D. 407, 412, 569 f., 572 Rice, Mabel L. 1820 f. Richards, Man˜a 1020 Richter, Heide 983 Richter, Jean Paul 157 Riegel, Martin 1286 Ries, John 106, 108, 110, 112 f., 665, 672 Riis, Hans N. 1518 Rijkhoff, Jan 1068 Rijksbaron, Albert 1318 f. Rinco´n, Antonio del 1433 Risch, Ernst 121 f., 1320, 1325 Rischel, Jørgen 1389 Rissanen, Mati 1590 Ritchie, Graeme D. 155, 1896 Ritner, Robert K. 1917 Ritter, Elizabeth 1060 Rivet, Paul 1464 Riwle, Ancrene 1642 Rix, Helmut 122, 1311, 1319 Rizzi, Luigi 328 Roberts, Ian 323, 329 Roberts, John R. 1202⫺1204 Roberts, Lawrence D. 985 Robin, Christian 1741 Robins, Robert Henry 19, 56, 59, 91 f., 100, 151 f., 264, 344, 552, 627, 674, 694, 708, 770, 1074, 1226, 1235 Robinson, Orrin W. 1603 Roche, Emmanuel 1896 Roche´, Michel 1707 Rodrigues, Aryon D. 742 Roeper, Thomas 356 f., 862, 1259, 1261, 1263 Rogers, David E. 175, 437 Rohrer, Christian 906 Rolecke, Thorsten 1680 Romaine, Suzanne 1638⫺1640, 1756, 1761 Römer, Paul G. 1659 Romero-Figueroa, Andre´s 1216 Rondal, Jean A. 1822 Ronneberger-Sibold, Elke 293, 455, 520, 1268, 1270 f., 1275, 1684 Rooij, Vincent de 245 Rosch, Eleanor 264 Rose, James H. 834 f. Rosen, Carol 925 Rose´n, Haiim B. 655 f., 1343, 1345 Rosen, Sarah Thomas 204, 322, 919, 924
Rosenhouse, Judith 1742 Rosetti, A. 249, 252 f. Rosier, Ire`ne 84, 696 f. Ross, John Robert 264, 699, 976 Ross, Malcom D. 1135 Ross, Sir David 264 Rosten, Leo C. 971 Roth, Juliana 1650 Rothweiler, Monika 1767, 1820 f. Rousseau, Andre´ 983 Rowley, Anthony 1271 Royen, Gerlach 163 Rozental, Ditmar E. 536 Rozwadowska, Bozena 863 Rubach, Jerzy 203, 339, 341 f., 384, 395, 414, 541 Rubin, Gary 1774 Rubino, Carl R. G. 1476 Rubio, Lisardo 1076 Rude, Noel 1082, 1087 Rudes, Blair A. 514, 606 Ruegg, David Seyfort 437, 703 Rugemalira, Josephat Muhozi 1124 Ruhlen, Merritt 577 Ruipe´rez, Martı´n Sa´nchez 1319 Ruiz de Montoya, Antonio 1421 Rukeyser, Alison 1056 f., 1060 Rumelhart, David E. 1766, 1807 Rumsey, Alan 1001 Runciman, Sir Steven 63 Rundgren, Frithiof 74 Rupp, Heinz 1273 Russ, Charles 1683, 1687 Russell, I. Willis 958 Russell, Kevin 471 Rutherford, William 276
S Saadiah Gaon 74, 703 Sabeau-Jouannet, Emilie 1001 Sˇa´chmatov, A. A. 127 Sadler, Louisa 327 f. Sadock, Jerrold M. 204, 327, 331 f., 394, 922⫺924, 1191, 1395 Saeed, John Ibrahim 1135 Sager, Juan C. 1924 f., 1927 f. Sagey, Elizabeth 574 Saint-Hilaire, Auguste Prouvensal de 16 f. Sakaguchi, Alicja 1563 S´a¯kalya 43 Sala, Marius 1648, 1651 Sallust 76 Salminen, Tapani 1328 Salmon, P. B. 91 Saloni, Zygmunt 1889, 1891 Saltarelli, Mario 569, 759, 761, 813, 817, 980, 1078, 1085
1967
Namenregister Saltveit, Laurits 1274 Salzmann, Zdenek 773 Samarin, William J. 1759, 1829 Sanctius 697 Sandberg, Bengt 690, 895, 1275 Sanders, Gerald 363, 442 f., 446, 628, 750, 871, 946 f., 950 Sanders, Lawrence 836 Sandfeld, Kristian 1649 Sandmann, Manfred 677, 680, 683, 690 Sandoval, Marı´a 877 Sandra, Dominiek 421, 1768 f., 1770, 1772, 1774 f., 1779 Sands, A. Kristina 1034, 1037, 1042 f. Sankoff, David 243, 245 Sankoff, Gillian 1755 f., 1761 Sˇanskij, Nikolaj M. 132, 1309 Santandrea, Stefano 1094 Santen, Ariane J. van 164⫺166 Sanzeev, G. D. 1218 Sapir, Edward 163, 170, 172⫺ 178, 186, 202, 218, 248, 253, 373, 376, 383, 432, 507 f., 513, 524, 526, 538 f., 561, 588, 716, 757, 764, 916, 923⫺925, 932, 935, 940, 1213, 1221, 1223⫺1225, 1228, 1232 f., 1235 f., 1248, 1311, 1400, 1647, 1655 Saporta, Sol 440 Sardzˇveladze, Z. A. 87 Sarmiento, Ramo´n 697 Sasse, Hans-Jürgen 728, 890, 919, 975, 1068, 1646 Sassen, Albert 165, 231, 233 Sauer, Hans 1627, 1630, 1632⫺ 1634, 1676 Sˇaumjan, Sebast’an K. 1309 Saunders, Ross 923 Saussure, Ferdinand de 91, 100, 126 f., 139, 141 f., 154, 171, 210 f., 214⫺217, 219, 221, 225 f., 239, 244, 304, 344, 404, 437 f., 441, 609, 698 f., 851, 1349, 1565, 1576, 1612, 1615, 1622, 1633, 1915 Saussure, Rene´ de 1565 f. Sauvageot, Serge 987, 994 Savis¸ır, I˙skender 1364 Scalise, Sergio 198 f., 201, 220, 320 f., 351 f., 354 f., 361, 407 f., 540, 542, 583, 859, 869, 875, 879, 882, 1047, 1255, 1288 Scarborough, D. L. 1770 Sˇcˇeglov, Jurij K. 1216 Sˇcˇerba, Lev V. 126, 129 Schaaik, Gerjan van 1195 Schabes, Yves 1896 Schachter, Paul 390, 399 f., 543, 558, 569, 589, 710, 714,
718 f., 721 f., 743, 746, 749, 760, 784, 786, 790 f., 815, 817, 1113, 1124, 1203, 1474⫺ 1478, 1481⫺1483, 1486 Schaeder, Burkhard 676 f., 683, 699 f. Schaff, Adam 212 Schane, Sanford A. 1285 Schaner-Wolles, Chris 1822 Schaub, Willi 978, 1006, 1137, 1192, 1199 Schebben-Schmidt, Marietheres 1684, 1694 Scheer, Tobias 1272 Scheerer, Eckart 1766 Scheerer, Otto 1479 Schegloff, Emmanuel A. 984 Scheidweiler, Gaston 1050 Scheler, Manfred 1619, 1674 Schellinger, Wolfgang 1058 Schenk, Werner 1309 Schenkel, Wolfgang 1917 Scherer, Wilhelm 1680 Schick, Brenda 1559 Schifko, Peter 435 f. Schiller, Friedrich von 16, 1621 Schindler, Jochem 1611 Schipper, Joleen 165 Schirmunski, Victor M. 1688, 1691 Schlachter, Wolfgang 122 Schlegel, August Wilhelm von 7, 94, 1221, 1223, 1247, 1698 Schlegel, Carl Wilhelm Friedrich von 7, 18, 94, 1221⫺1223, 1247 Schleicher, August 8⫺10, 18 f., 91 f., 97⫺99, 217, 1221, 1223, 1227 f., 1247 Schleicher, Charles 925 Schleiermacher, Friedrich 108 Schleyer, Johann Martin 1563 Schlieben-Lange, Brigitte 676 Schmid, Günter 16⫺18 Schmid, Stephan 290 Schmid, Wolfgang P. 680, 699⫺ 701 Schmidt, Annette 1029, 1043, 1655 f. Schmidt, Gernot 122 Schmidt, Günther Dietrich 351, 355 f., 892, 931 Schmidt, Johann 1563, 1567 Schmidt, Karl Horst 1687 Schmidt, Reiner 1934 Schmidt, Wilhelm 675, 677, 693, 699 Schmitt, Christian 1704, 1708 Schmitter, Peter 91, 95 Schmitthenner, F. 157, 698 Schneider, Iris 1050 Schneider, Klaus P. 1046, 1049 f. Schooneveld, Cornelis H. van 125, 127 f., 277, 279
Schøsler, Lene 1699 Schott, Wilhelm 96 Schottelius, Justus Georg 19, 157, 1682, 1931 Schreiner, Markus 1616 Schretter, Manfred K. 35 Schreuder, Robert 347, 1922, 1770, 1772⫺1774, 1779, 1799, Schreyer, Rüdiger 92 Schubert, Klaus 1563, 1565, 1570 Schubert, Lanhart K. 1067 Schuchardt, Hugo 1653, 1659 Schuh, Russel G. 1094, 1216 Schultink, Hendrik 163⫺166, 195, 201, 305, 342 Schultz-Lorentzen, Christian Wilhelm 1389 Schuster-Sˇewc, Heinz 1648 Schütz, Albert 1136, 1139 Schütze, Carson T. 400 Schütze, Hinrich 1893, 1898 Schwartz, Catherine 837 Schwartz, Linda 275, 785, 788 f. Schwartz, Myrna F. 1817 f. Schwarz, Catherine 1883 Schwarz, Hans 1684 Schwarze, Christoph 514, 889 f., 892, 895 Schwarzwald, Ora 1345, 1349 Schwegler, Armin 1698 Schweickard, Wolfgang 1708 Schwyzer, Eduard 1312 f., 1316, 1320 Sciullo, Anna-Maria Di 31, 201, 249, 315 f., 319, 322, 325 f., 350, 354, 358, 418, 596, 860⫺862, 900, 904, 911, 1255, 1287, 1295 Scott, Graham 1078, 1081, 1203 f. Scragg, Donald G. 1921 Sebeok, Thomas A. 91, 135, 210, 213 Sedulius Scottus 79 f., 83, 86 Seebold, Elmar 1049, 1912⫺ 1914 Segui, Juan 1780 Seibicke, Wilfried 1903, 1924 Seidenberg, Mark 1780 Seiler, Hansjakob 237, 242, 247, 251, 253, 735, 740, 1019, 1091, 1097, 1101, 1224 Seiler, Walter 1022 f., 1026, 1028 Selkirk, Elisabeth O. 31, 33, 189, 200, 202 f., 319, 338, 352, 354, 357, 408, 540 f., 550, 838, 860⫺862, 880, 899 f., 904 f., 907 f., 911, 1255, 1264, 1288 f., 1296 Senft, Gunter 1012 Serbat, Guy 638, 641, 645, 1076
1968 Serebrennikov, Boris A. 129 Se´re´gal, Philippe 1272 Serzisko, Fritz 1018 f., 1024, 1040, 1098 f., 1904 Seuren, Pieter 815 Sextus Empiricus 56 Sgall, Petr 1224 Shafer, Robert 1733 Shakespeare, William 1002, 1674, 1678 Sharma, Rama Nath 45 Shaumyan, Sebastian 1152 Shaw, George Bernard 1921 Shaw, J. Howard 901 Shaw, Patricia A. 411 Sheckley, Robert 836 Shepherd, Susan 814 Sheppard, Helen E. 953 Sheridan, Richard Brinsley 1621 Sheridan, Thomas 1638 Sherwood, David 407 Sherzer, Joel 1077, 1080 Shibatani, Masayoshi 1013 f., 1123, 1131 f., 1134, 1137, 1140, 1148, 1153⫺1156, 1160⫺1162, 1225, 1479 Shkarban, Lisa I. 1475 Shopen, Timothy 1861 Showalter, Catherine 987, 994 Sibawayhi, Ibn øUthman 67⫺ 73, 702 Sibomana, Leonidas 1538 Siegel, Dorothy L. 189, 198 f., 202 f., 320, 341, 350 f., 354, 411, 433, 541, 879, 948, 1255 f., 1790 Siegel, Muffy 356 f. Siemund, Peter 1070 Sievers, Eduard 100 Siewierska, Anna 809, 813, 1140, 1160 f. Sifianou, Maria 1050 Sil’nickij, Georgij G. 1132, 1235, 1237⫺1243 Silverstein, Michael 788, 1011, 1082 f., 1153, 1493 Sˇimecˇkova´, Alena 892 Simmel, Georg 3 Simmler, Franz 1685, 1687 Simon, H. F. 677 Simon, Walter 725 Simons, Gary F. 1836 Simons, Linda 1097 Simpson, Jane 330, 470 Simpson, John A. 1882 Sims-Williams, Nicholas 602 Sinclair, Hermine 1796, 1807 Sinclair, John 1884, 1888 Singh, Rajendra 871 Sitnikova, Antonina 1550 Sitta, Horst 1936 Sittig, Ernst 1074 Skalicˇka, Vladimir 133 f., 293, 630, 913, 1224, 1235
Verzeichnisse und Sprachenkarten Skorik, Pjotr Ja. 314 f. Skousen, Royal 571, 854, 866, 1338, 1602 f., 1766 Slavicˇkova´, Eleonora 1235 Slobin, Dan I. 290, 300, 348, 796, 1769, 1778, 1795⫺1797, 1799 f. Slotkin, Alan R. 1641 Sluiter, Ivoke 60 Smaragdus 80, 83 f. Smedts, Willy A. J. 165 Smeets, Ineke 1194 f., 1197, 1205 Smirnickij, Aleksandr I. 23, 129, 656, 660 Smirniotopoulos, Jane C. 330 Smirno´v, S. 135 Smith, Adam 92, 1221 f. Smith, Carlota S. 1166, 1168 f., 1172 f., 1179 Smith, Geoff P. 1756 Smith, John Ch. 984 Smith, Neil 1822 Smith, Norval 555, 1121, 1656⫺ 1658, 1863, 1871 f. Smith, Philipp T. 1771 Smith, Wayne H. 1557 Smith-Lock, Karen M. 1820 Smith-Stark, T. Cedric 1059 f. Smith-Stark, Thomas C. 711, 736 Smits, Caroline 1587 Smoczynska, Magdalena 1799 f. Smolensky, Paul 471, 566, 1613, 1766 Smyth, Herbert 1150 Sneddon, James Neil 241 Snodgrass, John G. 1770 Snyman, J. W. 797 Soberano, Rosa 1474 Soboleva, Polina A. 648, 650, 1309 Sodmann, Timothy 1681 Soeteman, Carl 1272 Sohn, Ho-min 577 Sokrates 53 Solms, Hans-Joachim 1684, 1686, 1688⫺1691, 1693 f. Sommer, Bruce A. 751 Sommer, Ferdinand 120 Sommerfeldt, Karl-Ernst 693 Sommerfelt, Alf 147 Son, Jos van 1237, 1239 Sonderegger, Stefan 1681 f., 1684 f., 1687 f., 1691⫺1693 Song, Jae Jung 1137, 1139, 1142 Sophronius 58 Sørensen, Finn 882 Sorenson, Arthur P. 374, 803 Sorin, Carmen Dobrovie 400 Sosenskaja, Tat’jana B. 1215 Spang-Hanssen, Henning 214 Speas, Margaret 330 Spence, N. C. W. 1707
Spencer, Andrew 5, 23, 27, 32, 155, 188 f., 315 f., 319, 321 f., 325⫺327, 329⫺331, 340, 391, 405, 409, 413 f., 451, 463, 466, 471, 474, 481, 489, 583, 627, 639, 647, 841, 854 f., 1256, 1259, 1261, 1263 f., 1780 Spencer, Herbert 1258 Spitzer, Leo 1290, 1645 Spring, Cari Louise 563 Sproat, Richard W. 199, 321, 326, 332, 413, 899, 904, 1894 Spruit, Arie 1109 Sridhar, Shikaripur N. 203, 252, 414, 541, 880, 993, 1137, 1256 Staal, J. Frits 165, 703 Städtler, Thomas 87, 89, 451, 696 Stalin, Joseph 127 f. Stalph, Jürgen 1917 Stampe, David 288, 401, 770, 772, 1594 Standop, E. 814 Stanley, Richard 385 f., 429 Stanners, Robert F. 1770, 1773, 1783, 1785 Sta˚rck, John 1688 Stark, Jacqueline A. 1820 Stark, John A. 1789 Starosta, Stanley 1086 Stassen, Leon 710, 759, 766, 1107, 1212, 1216 Steedman, Mark 1168 f., 1179 Steele, Susan 396, 400 f., 806, 817 f. Steever, Sanford B. 920 f., 1080 Stegmann, Carl 655 Stehr, A. 92 Steinberg, Danny 1640 Steinhauer, Anja 1275 Steinitz, Renate 1277 Steinthal, Heymann 9, 92, 97⫺ 99, 105 Sˇtejnfel’dt, E˙vi A. 652 Stemberger, Joseph Paul, 206, 345 f., 364, 392, 1766, 1788, 1817 Stepanov, Arthur V. 1238 Stepanov, Jurij S. 417 Stepanova, Marija D. 665, 667, 690, 892 f., 931, 1565 Steriade, Donca 430, 436, 490, 562⫺564 Stern, Theodore 1735 Stevens, Alan M. 400, 560 Stewart, John 1517 f. Stiebels, Barbara 380, 902, 905 Stocking, George 173 Stockwell, Robert P. 1791 Stokes, Bronwyn 1078 Stokoe, William C. 1555, 1557
1969
Namenregister Stolz, Thomas 693 Stong-Jensen, Margaret 346, 628 Stonham, John T. 411, 578, 581, 584 f. Stopp, Hugo 169 Stork, Francis C. 568, 597 Strang, Barbara M. H. 835 Strauß, Gerhard 694 Strauss, Steven L. 203, 880, 1256 Strauss, Susan 993 Strawson, Peter Frederick 974 Strong, David R. 815 Strunk, Klaus 1318 Stubbs, Michael 1921 Stump, Gregory T. 153, 321, 367, 442, 541, 784, 877, 1047 Sturtevant, Edgar H. 121, 866 Sua´rez, Jorge A. 1077, 1107, 1216, 1423, 1427⫺1429 Sucho´tin, A. M. 132 Sugamoto, N. 755 Summers, Della 1884 Sun˜er, Margarita 327 f. Sunik, Orest P. 129 Supalla, Ted 1022, 1557⫺1560 Sütterlin, Ludwig 104, 106⫺113, 115, 670, 672, 675 f., 684, 699 Suyuti, Jalal al-Din 73 Svartvik, Jan 747 Sˇvedova, Natal’ja J. 1303, 1305 f., 1308 f. Swadesh, Morris 178, 186, 251, 452, 490 f., 508, 721, 781 Sweet, Henry 149, 344, 346, 944 Sweetser, Eve 1591 Swiggers, Pierre 93, 211, 216, 218⫺220, 222, 245, 249, 697 f., 1285 Swisher, Linda 1821 Syncellus 59 Szantyr, Anton 1217 Szczepaniak, Renata 1904 Szemere´nyi, Oswald 121, 772 Szerdahelyi, Istva´n 1564 Szober, St. 134 Szpyra, Jolanta 252 Szymanek, Bogdan 266, 267, 353, 433, 511, 526, 529, 532, 877, 881, 1255
T Tabakowska, Elz´bieta 1049 Taeldeman, Johan 165 Taft, Marcus 1766, 1768⫺1772, 1774, 1778 f., 1784 Tagliavini, Carlo 458 Talen, Jan G. 162 Tallal, Paula 1820 Talmy, Leonard 258 f., 261, 374, 796, 1125 f., 1402
Tamba, Ire`ne 988, 992, 1296 Tanenhaus, Michael K. 837, 1771 Tanz, Christine 983 Tatwine 78 f., 84 Taylor, Ann 398 Taylor, Charles 253, 768, 980 f., 1210 Taylor, Gerald 1455 Taylor, John R. 392, 424, 1049, 1091 Tedeschi, Philip J. 809 Tegey, Habibullah 399, 546 Tekinay, Alev 1363 Teleman, Ulf 148 Telenkova, Margarita A. 536 Temple, Martine 143, 1292 f. Terence 76 Terrell, Peter 1886 Tesak, Jürgen 1817 f. Tesar, Bruce 1613 Tesnie`re, Lucien 107, 109, 667, 680, 689, 699 f., 1117⫺1119, 1130 Testen, David 556 Teuber, Oliver 1271 Tharp, David A. 1785 Theobald, Elke 1272 Theodorides, Chr. 58 Theodorus von Gaza 59 Theodosius 57⫺59, 63 Thiel, Gisela 836, 906, 1638 Thiele, Johannes 888 f., 895, 1287, 1291 Thieroff, Rolf 1273 f. Thomas von Aquin 54 Thomas von Erfurt 7, 86, 697 Thomas, Jacqueline M. C. 1832 Thomason, Sarah Grey 1645⫺ 1647, 1651 Thompson, Chad 1400 Thompson, E. David 1100 Thompson, Laurence 254, 577 f., 580, 936, 1546, 1734 Thompson, M. Terry 577 f., 580, 936 Thompson, Sandra A. 240, 287, 374, 712, 727 f., 747, 749, 758, 764, 795, 799 f., 809 f., 846, 848, 882, 939 f., 950, 1020, 1084, 1112 f., 1116, 1191 f., 1202, 1637 Thomsen, Vilhelm 145 Thomson, David 82, 87 Thorndike, Edward L. 1235 Thornton, Anna M. 294 Thrane, Torben 818 Thümmel, Wolf 655 Thun, Nils 1679 Thurman, Robert 1204 Thurneysen, Rudolf 1215, 1219 Thwaites, Reuben Gold 1411
Tichonov, Aleksandr N. V. 677, 1307, 1309 Tiefenbach, Heinrich 894 Tiersma, Pieter Meijes 166, 278, 337 Timberlake, Alan 814, 1197 Ting, Pang-hsin 1732 Tobin, Yishai 1343, 1345⫺1347, 1350, 1352 f., 1356 Todaeva, B. X. 1218 Todd, Loreto 1759 Togeby, Knud 379⫺381, 1704 Toivainen, Jorma 1341 Tolskaya, Maria 1215 Toman, Jindrich 161, 900, 904, 908, 1278 Tomblin, J. Bruce 1820 Tommola, Hannu 1186 Tonelli, Livia 289 Tonkin, Humphrey 1564 Tooke, John Horne 92 Toorn, Maarten C. van den 165 Topping, Donald M. 560, 568 Torero, Alfredo 1453 Törnqvist, Nils 905 Totanes, Sebastian de 1474 f. Touratier, Christian 1076 Tournier, Jean 1903 Tovar, Antonio 91 Tracy, Rosemarie 1799 Trager, George L. 184, 219, 379, 453 Trask, R. Larry 655, 813, 970, 1061, 1667 Traugott, Elizabeth Closs 259, 263, 401, 814, 1199, 1205, 1232, 1590, 1592, 1595 f. Treffry, Diana 1883 Trefry, David 1491 Trier, Jost 116 Trnka, Bohumil 134 Troll, Wilhelm 16 f. Trommelen, Mieke 165 f., 201, 860 Trost, Harald 1896 Troupeau, Ge´rard 703 Tru’o’ng Íoˆng San 1551 Truax, Catherine Elaine 584 f. Trubetzkoy, Nikolaj S. 132, 235 f., 241, 244, 267, 272, 275, 335 f., 383, 427, 431⫺ 433, 442, 490, 492⫺494, 772 Tschenke´li, Kita 439, 442 Tschirch, Fritz 1694 Tsimpli, Ianthi-Maria 1822 Tsunoda, Tasaku 1158 f. Tucker, Archibald N. 1077, 1079, 1095 Tuite, Kenin 1215 Twaddell, W. Freeman 178 Tyler, Lorraine K. 540, 1775, 1778, 1783 f., 1786, 1819
1970 U Uc Faidit 696 Uehara, Satoshi 1114 Uhlenbeck, C. C. 163 Uhlenbeck, Eugenius M. 164, 166, 867, 951 Uhlig, Gustav 56, 59, 63 f. Uhrström, Wilhelm 1678 Ułazyn, Henryk 134, 490 Uldall, Hans Jørgen 254, 452 Ullman, Michael 1821 Ullmann, Stephen 251 f., 1615, 1621 f., 1633 Ultan, Russell 547⫺550, 576 f., 790, 937, 1093, 1595 Uluchanov, Igor S. 888, 1309 Ulvestadt, Bjarne 1272 Underhill, Robert 711, 981 Ungeheuer, Gerold 115 Ungerer, Friedrich 1615, 1619 Upa¯dhya¯ya, Baladeva 50 Urbaniak, Gertrud 902, 1276 Usˇako´v, D. N. 127
V Vachek, Josef 133 Vago, Robert M. 621 Vaidya, P. L. 1722 Vairel, He´le`ne 999 Valin, Robert van 859, 1192 f., 1377, 1475 Vanvolsem, Serge 697 Varro, Marcus Terentius 56, 59⫺62, 82, 696 f., 732 Vasu, S´rı¯s´a Chandra 437 Vater, Heinz 681, 698, 1270, 1274 Vater, J. S. 91 Veith, Werner 1272 Vela´zquez-Castillo, Maura 1233, 1421⫺1428, 1430 Vendler, Zeno 1165, 1168⫺1173, 1176 Vendrye`s, Joseph 693, 813, 1218, 1611 Veneziano, Edy 1796 Vennemann, Theo 290, 337, 468, 498, 577, 840, 1620, 1905 Vergnaud, Jean-Roger 203, 347 Verkuyl, Henk J. 1167, 1170 f. Verner, Genrich K. 1377, 1382 Verner, Karl 100, 145, 1673 Verrac, Monique 693 Versaw, Larry 1836 Versteegh, Cornelis H. M. 54, 57, 60, 702 Versteegh, Kees 70, 74, 1741, 1752 Vesper, Wilhelm 104, 115, 669
Verzeichnisse und Sprachenkarten Vet, Co 1174 Vietor, Wilhelm 1930 f. Vihman, Marily M. 1799 Vilkuna, Maria 1183 Villa Dei, Alexander von 79 Villiers, Jill de 1823 Villiers, Reinette De 352 Vincent, Nigel 442, 648, 656, 661 f., 808, 814 Vinogradov, Viktor V. 127 f., 1307⫺1309 Vinoku´r, G. O. 127 Virgilius 76 Vjalkina, L. V. 1218 Voegelin, Charles F. 178, 186, 508, 1235 Voeltz, Erhard F. K. 1099 Vögeding, Joachim 902 Vogel, Irene 341, 385, 413 Vogel, Petra Maria 932, 936, 1275 Vogel, Thomas 1808 Vogt, Hans 146 f. Voigt, Wolfram 4 Volodin, Aleksandr P. 443 Voorhoeve, Clemens L. 164, 166 Voorhoeve, Jan 589, 950, 1660 Vooys, Cornelis G. N. de 163 Vorlat, Emma 697 Vries, Jan W. de 164 Vries, Lourens J. de 1205, 1501 f., 1506 Vries, Wobbe de 163
W Waag, Albert 110 Wackernagel, Jacob 121, 396 f., 401, 644 Waddell, Helen 65 Waddy, Julie 1033 Wagner, Robert-Le´on 142, 1291 Wahrig, Gerhard 1884 Walch, Maria 1693 Waldron, Ronald A. 1620 f. Wali, Kashi 1062 Wall, Margarita N. 1376, 1388 Wallace, Anthony 178 Wallis, Ethel 569, 572 Wallis, John 19, 149, 660 Walter, Henriette 485⫺488 Walter, Heribert 687, 709 Wang, Li 1731, 1735⫺1738 Wanner, Dieter 1701 Ward, Dennis 958 Wares, Alan C. 193 Waringhien, Gaston 1565, 1567 Warnke, Camilla 2 Warren, Beatrice 1630 Wasow, Thomas 810, 818 Waterhouse, Viola G. 192 Waterson, Natalie 150
Watkins, Calvert 122, 286, 998, 1663, 1665 Watkins, Laurel J. 917 Watkins, Ruth 1820 Watson, Janet C. E. 1742 Waugh, Linda R. 272, 275, 279, 493 Webb, Charlotte 569 Weber, David 1011, 1454, 1833 Weekley, Ernest 1618, 1620⫺ 1622 Weerman, Fred 166 Wegener, Heide 1268 f., 1808 f., 1814 Wegera, Klaus-Peter 1684⫺ 1686, 1691⫺1694 Weggelaar, Cas 926 Weiers, Michael 1830 Weigand, Georg 669 Weigand, Hans 976 Weiner, Edmund S. C. 1882 Weinhold, Karl 1688 Weinreich, Uriel 178, 243, 519, 1647, 1650 f. Weinrich, Harald 1274 Weir, E. M. Helen 921, 1136 Weisgerber, Leo 109, 115, 160, 1278, 1620⫺1622 Weiß, Joseph 74 Weißenborn, Jürgen 983 Weist, Richard M. 1807 Wekker, Herman 1891 Welke, Klaus 668 Wellens, Ineke 1752 Wellmann, Hans 890, 902, 933 Wells, Christopher J. 1681 Wells, John C. 484 f., 507 Wells, Rulon S. 173, 184, 186⫺ 189, 226, 381, 406, 452, 464, 475, 672 Welmers, William E. 588 f., 1038, 1112, 1538 Wentworth, Harold 954 Werner, Otmar 91, 289, 301, 519, 738, 1267⫺1270, 1272, 1274, 1680, 1683, 1902 Westney, Paul 662 Westrum, Peter N. 987, 1100 Wetzer, Harrie 766 f., 1113 Wexler, Ken 1821 Wheatley, Julian K. 747 Wheeler, Deirdre W. 878 Wheeler, Max W. 294 Whitaker, Katherine P. K. 589 Whitney, William Dwight 98, 358 f., 568 Whorf, Benjamin Lee 883, 1450, 1563 Wicentowski, R. 1898 Wiederspiel, Brigitte 986 Wiegand, Herbert Ernst 1888 Wiersma, Robinia 1501 Wierzbicka, Anna 32, 259, 262, 286, 457, 708, 719, 758, 1046,
1971
Namenregister 1048⫺1050, 1068 f., 1077, 1083, 1113, 1423 f. Wiese, Bernd 1270 Wiese, Richard 476, 558, 900, 904, 1275 Wiesemann, Ursula 785, 983, 987, 1013, 1100, 1832 Wiesinger, Peter 1267 Wiik, Kalevi 1338 Wijk, Nicolaas van 162 f. Wilbur, Ronnie Bring 566 Wilder, Chris 398 Wildgen, Wolfgang 4, 7 Wilkins, David 1203 Willett, Thomas 1197 f. Williams, Edna R. 86 Williams, Edwin S. 31, 200⫺203, 249, 268 f., 271, 315 f., 319 f., 322, 325 f., 350, 352, 354, 358, 366, 408, 418, 540, 596, 603, 855, 860⫺862, 900, 904, 911, 1141, 1255, 1287, 1295 Williams, Jennifer 1161 Williams, Stephen J. 823 Williamson, Kay 1128 Willis, Penny 478, 482 Wills, Dorothy D. 987, 991, 994 Wilmanns, Wilhelm 19, 104, 106, 158 f., 898, 901, 1683 f. Wils, J. 254 Wilske, Ludwig 887⫺889, 892 f., 950 Wilson, William H. 1094 f. Wimbish, John S. 1833 Wimmer, Hans 1922 Windfuhr, Gernot L. 746, 1086 Winfield, W. W. 766 Winford, Donald 1665 Winkler, Christian 453 Winkler, Gertraud 1684 Winter, Werner 584 f., 1202, 1204 Winters, Margaret E. 289 Wirth, Jessica 280 Wisch, Fritz-Helmut 1823 Wise, Mary Ruth 1135 f., 1138 f. Wissing, Daniel P. 165 Withgott, Margaret 330 Wittgenstein, Ludwig 1233 Wittwer, Michael 1315 Woidich, Manfred 1742 Woisetschlaeger, Eric 259 Wölcken, Fritz 958 Wolf, Heinz Jürgen 1708 Wolf, Lothar 16 f. Wolf, Norbert Richard 1681, 1684, 1687, 1691 f.
Wolfart, H. Christoph 1096, 1233, 1413, 1419 Wolfenden, Elmer 1478 f. Wolff, John U. 1474, 1476, 1478 Wolff, Susanne 875 Wolfram, Walt 241 Wonderly, William L. 546 Woodard, Roger D. 122 Woodbury, Anthony 1152 Woodbury, Hanni 919, 924, 1844 Woodworth, Nancy L. 374, 376, 800 f. Worth, Dean S. 87 Wouters, Alphonse 56 Wrede, Ferdinand 163 Wright, Elizabeth Mary 1668, 1670 Wright, Fiona 1761 Wright, Joseph 16, 1668, 1670 Wright, William 555 Wuest, Jakob 1702 Wüllner, Franz 59, 608 Wunderlich, Dieter 380, 902, 905 Wundt, Wilhelm 105, 110, 122, 162 Wurm, Stephen A. 999, 1010, 1501, 1756 Wurst, Raimund Jakob 405 Wurzel, Wolfgang Ullrich 6, 160 f., 195, 200, 231, 266 f., 279, 288⫺294, 338 f., 433, 479, 514, 516, 519 f., 556, 570, 573, 597, 604 f., 622, 632⫺637, 648 f., 653, 878, 894, 1267⫺1272, 1277, 1603, 1607
X Xenophon 1326 Xerberman, K. R. 1212, 1214 Xhuvani, Aleksande¨r 1647
Y Yallop, Colin 1079 Yamada, Jeni E. 1822 Yang, L. 1162 Yang, Paul Fu-mien 1734 Yarowsky, D. 1898 Yaska 703 Yip, Moira 587, 1045 Young, Philip D. 810
Young, Robert 316 f., 1098, 1136 Youssi, Abderrahim 487 Yu, Alan C. 1058, 1062 Z Zacharski, Ron 791 f. Zaenen, Annie 809 Zager, David 456 Zajjaji, Abu l-Quasim 67 f., 71, 73 Zajjaji, Ibn ?Ishaq 73 Zaliznjak, Andrej A. 129⫺132, 1308 f., 1868, 1876 Zamaxshari, Abu l-Qasim 71, 73 Zamenhof, Ludwig L. 1564 f. Zandvoort, Reinhard W. 835 Zavadovskij, Jurij 1746 Zavala, Roberto M. 1020 f., 1024, 1026 Zawadowski, Leon 655 Zeegers Vander Vorst, Nicole 1323 Zemb, Jean Marie 700 Zemskaja, Elena A. 132, 419, 421, 424, 1309 Zgusta, Ladislav 250, 1887⫺ 1890 Zhang, Zheng-Sheng 559 Zhen, Yuan 1737 Zhou, Xiaolin 1780 Zide, Norman 1128 Zimmer, Karl 195 f., 907, 909, 1093 Zipf, George Kingsley 286 f., 289, 1233, 1252 Zˇirmu´nskij, Viktor M. 129 Zˇivova, Galina T. 1377 Zonneveld, Ronaldus M. van 165 Zonneveld, Wim 165, 201, 860, 970 Zribi-Hertz, Anne 990 Zubin, David A. 613, 1035 f., 1269, 1811 Zubizaretta, Maria-Luisa 1780 Zˇuravlev, Aleksandr P. 423 Zutt, Herta 1683, 1692, 1694 Zwanenburg, Wiecher 266, 847, 1287 f. Zwettler, Michael 1741 Zwicky, Arnold M. 200, 266 f., 297, 319, 346, 389 f., 392 f., 395, 401, 451, 464, 472, 482, 546, 617, 815, 861, 877, 899, 965, 967, 969, 984, 1008, 1191, 1255, 1262, 1264 f., 1666
Sprachenregister / Index of languages A Abchasisch/Abkhaz 739⫺741, 798, 800, 802, 823 f., 980 f., 1004, 1008, 1038, 1096, 1109, 1227 ⫺ Abaza 1039 Abipon 541 Achagua 1040 Acholi (Acoli) 539, 757 Acoma 1211 Adnyamathanha 1491 Adygeisch/Adigei (Adyge) 330, 1096 Afar 543 Afrikaans 13, 166, 232, 307, 352, 365, 658 f., 845, 847, 858, 879, 1186, 1578 f., 1586 afrikanische Sprachen/African languages 91, 166, 548, 588 f., 867, 1001, 1005⫺1007, 1013, 1016, 1019, 1031 f., 1040, 1043, 1112, 1128, 1208, 1512, 1515, 1647 ⫺ west-/West 592, 1128, 1183, 1659 afroasiatische Sprachen/Afroasiatic languages 91, 540 f., 554, 981, 1013, 1032⫺1034, 1037, 1040, 1096, 1216 Aghem 590 Agta 558 f., 561 f. Aguakatekisch/Aguacatec 980 Ainu 927, 1123, 1131, 1159 f., 1162 Akan 558, 560 f., 563, 589, 717, 790, 800, 815, 1517 f. ⫺ Fante 539 Akasele 541 Akhvakh 1035 Akkadisch/Akkadian 35⫺40, 780 ⫺ Alt-/Old- 35 ⫺ Altbabylonisch/Old Babylonian 35 ⫺ Standardbabylonisch/Standard Babylonian 35 Alabama 547, 549, 583 f. Alamblak 1033 f., 1036, 1038, 1041, 1128 Alawa 540, 1010 Albanisch/Albanian 516, 710, 717, 978, 1310, 1647, 1649, 1651, 1662 f., 1665
Aleutisch/Aleut 926, 1651 algonkianische Sprachen/Algonquian languages 476, 541, 926, 940, 1004 f., 1010, 1012, 1032, 1151, 1411⫺1416, 1418 Altägyptisch/Ancient Egyptian 99, 1917 altaische Sprachen/Altaic languages 741 f., 1077 f., 1083⫺ 1085 Alur 1142 Alutor 532 Alyawarra 539, 1079 amazonische Sprachen/Amazonian languages 1078, 1214, 1216, 1428 Amele 1203 f. Amerikanische Gebärdensprache/American Sign Language 1022, 1027, 1556 f. amerikanische (amerindische) Sprachen/American (Native American, Amerindian) languages 91, 94, 99, 121, 176, 180, 548, 553, 588, 739, 742, 1006, 1010, 1016, 1019, 1021 f., 1024, 1028, 1032, 1045, 1065, 1077, 1082, 1087, 1202⫺1204, 1216, 1247, 1425, 1433, 1453 ⫺ meso-/Meso- 1018, 1021 ⫺ nord-/North 94, 736, 1071, 1125, 1135 ⫺ süd-/South 94, 1019, 1038, 1041, 1208, 1432, 1469 Amharisch/Amharic 540, 552, 555, 981, 1216, 1741 anatolische Sprachen/Anatolian languages 585, 645, andische Sprachen/Andean languages 1037, 1206 Andoke 740 Angloromani 1647 Anindilyakwa 1033, 1038 Annamitisch/Annamite (s. Vietnamesisch) Apabhram ˙ s´a 1712, 1719⫺1722, 1725 Apalai 1214 Apurin˜a 1032 Arabana 748, 1491, 1494 Arabisch/Arabic 60, 67⫺72, 74, 151, 206, 238, 431, 486 f., 528, 533 f., 547⫺549, 554,
⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺
⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺
556, 558, 587, 620, 702 f., 715, 717, 732, 737, 740 f., 750, 758, 777, 779, 781, 798, 800 f., 821⫺823, 827, 830, 841, 934, 940, 1007, 1045, 1056 f., 1060, 1062, 1069, 1071 f., 1080, 1096, 1100, 1188, 1215, 1348, 1441, 1740⫺1742, 1745 f., 1748, 1750⫺1753, 1797, 1835, 1840, 1853, 1921 f. ägyptisches/Egyptian 620, 1742⫺1744, 1746⫺1749, 1751 f. algerisches/Algerian 1742 f., 1750 f. anatolisches (Daragözü)/Anatolian 1741, 1747 f., 1750 f. andalusisches/Andalusian 1748, 1751 Baghdad 1746, 1748 beduinisches/Bedouin 555, 1742, 1744⫺1746, 1748 ˘ iblah 1751 G Golf/Gulf 750, 980 Hassa¯niyya 1742, 1746 irakisches/‘Iraqi 1744, 1747 jemenitisches/Yemeni 1742, 1748⫺1750 Juba/Juba 1752 f. Kairener/Cairene 151, 555, 1751 Khartoum 1752 f. klassisches/Classical 72, 540, 553, 555, 579, 935, 953, 1055, 1058, 1096, 1213, 1740, 1742⫺1751 libanesisches/Lebanese 1742, 1751 libysches/Libyan 1742 maghrebinisches/Maghrebine 1746, 1750 marokkanisches/Moroccan 765, 953, 1742⫺1751, 1813 mesopotamisches/Mesopotamian 1742 Nag˘dı¯ 1746, 1748 Neu-/New 1096, 1741 f., 1744, 1746, 1748⫺1750 Neusüd-/Modern South 1750 Ost-/Eastern 1742 palästinensisches/Palestinian 1751 saudiarabisches/Saudi-Arabian 1744, 1746, 1751
1973
Sprachenregister ⫺ sudanesisches/Sudanese 1742, 1748, 1751 ⫺ syrisches/Syrian 547, 765, 1057, 1742, 1746⫺1749, 1751 ⫺ tschadisches/Chad 1742, 1748, 1751 ⫺ tunesisches/Tunisian 579, 1217, 1742, 1751 ⫺ usbekisches/Uzbekistan 1741, 1747⫺1750 ⫺ zypriotisches/Cypriot 1748 f., 1751 f. Arapesh 633 Arawa´ 1032 Arawak 773, 798, 821, 823, 827 f., 830, 927, 1032, 1035, 1039 f., 1039 f., 1135 f., 1139 ⫺ Ignaciano 1039 ⫺ Nord-/North 1039, 1043 Archi 482, 715, 1034, 1038 Arikara 1082 Arinisch/Arin 1376 Armenisch/Armenian 59, 94, 1041, 1064, 1134, 1185, 1310 ⫺ Alt-/Old 122 Asamiya 1712 Asheninca 987 asiatische Sprachen/Asian languages 1016, 1018, 1021, 1028, 1032, 1113 ⫺ ost-/East 993, 1004 , 1187 ⫺ südost-/Southeast 1053, 1216 As´oka (As´okan) 1239, 1242, 1712, 1719 f., 1722 Assamese 1042 Assanisch/Assan 1376 Atayal 546⫺549 athapaskische Sprachen/Athabaskan languages 763, 926, 1012, 1022, 1028, 1031, 1136, 1399⫺1401, 1403, 1407, 1409, 1467 Athapaskisch-Eyak-Sprachen/Athapaskan-Eyak languages 1399 (s. auch Eyak) Äthiopisch/Ethiopian 1013 Atjeh/Acehnese 713, 934, 1082 atlantische Sprachen/Atlantic languages ⫺ ost-/East 1061 ⫺ west-/West 1038 Atsugewi 796, 934, 1125, 1126 Auetö 773 Australische Gebärdensprache/ Australian Sign Language (Auslan) 1556 australische Sprachen/Australian (Aboriginal) languages 432, 540, 548, 603, 643 f., 744, 748, 751, 999, 1003, 1007, 1009⫺1012, 1016, 1021 f., 1026, 1032⫺1034, 1037 f., 1041⫺1043, 1056 f., 1060 f.,
1076⫺1080, 1082⫺1088, 1092, 1095, 1097, 1112, 1135 f., 1158, 1202⫺1205, 1208, 1491, 1493⫺1495 austroasiatische Sprachen/ Austroasiatic languages 549, 561, 1128, 1731, 1545 austronesische Sprachen/Austronesian languages 362, 540, 548 f., 560 f., 577, 760, 1012, 1017, 1032, 1053, 1056, 1082, 1094, 1124, 1214, 1473, 1476, 1501 Avarisch/Avar 715, 1081, 1083, 1085, 1142 Awyu 1501 f., 1506 Aymara 1003, 1009 aymarische Sprachen/Aymaran languages 1453, 1461 Azerbaijani (Azeri) 1578 f.
B Babungo 978, 1006, 1137, 1192, 1199 Bagvalal 1035 Bahwana 1043 Balante 785 Balbi 93 Balochi 541 baltische Sprachen/Baltic languages 548, 1069, 1214, 1310 baltofinnische Sprachen/BaltoFinnic (Baltic-Finnic) languages 407, 1042, 1078, 1080, 1083, 1328, 1650 baltoslawische Sprachen/BaltoSlavic languages 401, 642 Bambara 589, 592, 715, 717, 1092 Bamileke ⫺ Dschang 590 ⫺ Ghoma´la´? 590 Banda-Linda 529 Bangla (s. Bengali) Baniwa 1039 Bantusprachen/Bantu languages 134, 182, 323 f., 329, 336, 530, 538, 750, 777, 791, 854, 1013, 1019, 1033, 1037⫺ 1040, 1042, 1070, 1077, 1120, 1123, 1186, 1189, 1092, 1208, 1210, 1216, 1527⫺1533, 1538, 1542 Barai 815, 1110 Bare 1043 Baskisch/Basque 94, 100, 569 f., 741, 759, 761, 806, 813, 816 f., 970, 980, 1000, 1002, 1007, 1011, 1013, 1045, 1054, 1056, 1061, 1064, 1067, 1078, 1085, 1199, 1214, 1651
⫺ West-/West 1061 ⫺ Zentral-/Central 1061 Bayso 1054 f., 1057 f. Bedauye (Beja) 400, 555, 777 Belhare 1178 Bella Coola 1101 Belorussian (s. Weißrussisch) Beludzhi 1043 Bemba 757, 1092, 1210 Bembe 1114 Bengali (Bangla) 737, 779, 1043, 1048, 1078, 1085, 1092, 1239, 1242, 1712, 1723⫺1729 Benue-Congo-Sprachen/BenueCongo languages 774, 778, 1043 Berber 330 Berbice 236, 1658 f. Berik 987, 1100 Be´te´ 738 Bini 569 Blackfoot 1048 Bodisch/Bodic 1205 Bodo-Garo 1733 Bolanci 971 Botocudo 773 Bowili 1036 Bretonisch/Breton 328 f., 367, 442, 516, 760, 823, 979, 1063, 1065 Britische Gebärdensprache/British Sign Language 1556 Budukh 1071 Bulgarisch/Bulgarian 29, 400, 492, 520, 814, 1064, 1176, 1245, 1649 f. ⫺ Alt-/Old 1649 Burjat 1218 Burmesisch/Burmese 745, 1013, 1021, 1029, 1732 f., 1736, 1887 Buruschaski/Burushaski 766, 1038 Behza¯nı¯ 1751
C Cabecar 1020 f. Caddoan 926 Cahuapana 927 Callahuaya 1461 Cambodian (s. Kambodschanisch) Capanahua 1141 Cariban languages (s. karibische Sprachen) Caspian languages (s. kaspische Sprachen) Catalan (s. Katalanisch) Caucasian languages (s. kaukasische Sprachen) Cayuga 728, 926, 1022
1974 Cayuvava 927 Cebuano 1153 f., 1156 Celtic languages (s. keltische Sprachen) Cemuhi 589 Ch’ol 935 Chacobo 539 Chadic languages (s. tschadische Sprachen) Chaha 569 f., 572 Chama 488 Chamalal 1038 Chamorro 545, 560 f., 568 Chatino 588 Chechen (s. Tschetschenisch) Chemakua 1032 Chemehuevi 569, 767, 1077 Cherokee 1023 Cheyenne 541, 805 Chibcha-Sprachen/Chibchan languages 1021 Chi-Bemba 1039 Chichewa 322⫺324, 477, 536, 784 Chickasaw 584, 1110, 1142 Chimariko 538 Chi-Mwi:ni 324 chinantekische Sprachen/Chinantec languages 539 Chinesisch/Chinese 92, 94, 96⫺ 99, 134, 164, 227, 239, 241, 251, 258, 316, 454, 525, 536, 559, 561, 591 f., 677, 682, 703 f., 714, 720, 724⫺726, 728, 758, 807, 913, 934, 936 f., 940, 947, 950, 960, 992, 1004, 1017, 1020, 1027⫺ 1029, 1045, 1051, 1070, 1151, 1218, 1222, 1232, 1545, 1730 f., 1733⫺1738, 1808, 1813 f., 1819, 1843, 1875, 1897, 1903, 1916, 1919 ⫺ Alt-/Old 937, 1147 f., 1053, 1730⫺1735, 1737 ⫺ Amoy 1734 ⫺ archaisches/Archaic 1730 ⫺ Changsha 591, 1735 ⫺ Chungan 1734 ⫺ Fuzhou 1734 ⫺ kantonesisches/Cantonese 589, 781, 1736 ⫺ Kienyang 1734 ⫺ klassisches/Classical 588, 940, 1735, 1737, 1871 ⫺ Mandarin 239, 376, 727, 736, 748 f., 773⫺775, 781, 800, 810, 937, 1113, 1191, 1192, 1730, 1732, 1735⫺1738 ⫺ Min 1734 ⫺ Mittel-/Middle 1730⫺1735, 1737 ⫺ Neu-/Modern 937, 1730 ⫺ Neumandarin/Modern Mandarin 1731 f.
Verzeichnisse und Sprachenkarten ⫺ Proto- 1731 ⫺ Shanghai 592 ⫺ Spätmittel-/Late Middle 1730 f. Chinook 1007, 1048 ChiNyanja 588 Chipewyan 546 Chirag 1070 Chishona 802 Chitimacha 781, 785 Choctaw 538, 584, 625, 1000 Chontal 547 Chrau 934 Chuave 1204 Chuckchee (s. Tschuktschisch) Chukotko-Kamtchatkan languages (s. Tschukotko-Kamtschatka-Sprachen) Chuvash (s. Tschuvaschisch) Clallam 578, 936 Cocama 798 Cœur d’Ale`ne 553 f., 720 Comanche 917, 1005, 1077, 1433 Comox 923, 926 Coos 537 Cora 1129 Cree 1096, 1233, 1411, 1419, 1647 Cross River 1043 Cushitic languages (s. kuschitische Sprachen) Czech (s. Tschechisch)
D Dagbani 593 daghestanische Sprachen/Daghestanian languages 526, 1070 f. Dakota 442, 517, 547⫺549, 739, 806, 1082, 1099, 1100, 1835 Dangbon 1042 Dani 1142, 1203, 1205 Dänisch/Danish 95, 145⫺147, 249, 253, 378, 380, 382, 384, 386, 584, 741, 784, 832, 960, 1058, 1070, 1160, 1328, 1389, 1905, 1914 ⫺ Alt-/Old 146 Dargwa (Dargi) 526, 543, 1203 Deutsch/German 1, 3, 11, 13⫺ 15, 26, 61, 91⫺93, 96, 104 f., 108⫺111, 114, 118, 122, 135, 157, 159 f., 163, 174, 216, 236⫺239, 241, 249 f., 255, 266⫺269, 273 f., 291, 298 f., 301, 338, 340 f., 351⫺353, 355 f., 362, 366, 378⫺380, 383, 412, 418, 420, 423, 440⫺442, 445, 454⫺459, 468, 479, 482, 485, 491 f., 495, 497 f., 517 f., 520, 528, 535 f.,
⫺
⫺ ⫺ ⫺ ⫺ ⫺
⫺
⫺ ⫺ ⫺
541 f., 546, 568⫺570, 573 f., 576, 582⫺584, 595⫺597, 599 f., 602, 604 f., 607, 612 f., 621 f., 626, 628, 630⫺637, 646, 648⫺650, 655 f., 659, 661, 670, 674⫺677, 682, 684 f., 693, 697⫺699, 703, 732, 734⫺739, 744 f., 749, 751 f., 754 f., 774, 777, 779⫺ 782, 787⫺793, 798, 800, 803, 811, 829, 833 f., 836, 858, 881, 887⫺895, 899 f., 902⫺ 905, 907⫺909, 911, 929⫺937, 939⫺941, 943⫺945, 950, 956, 958, 960⫺962, 970, 975, 1007, 1013, 1035⫺1037, 1046⫺1051, 1054⫺1056, 1061, 1063 f., 1067 f., 1072, 1076, 1078 f., 1087, 1091, 1093, 1099 f., 1114, 1135 f., 1139 f., 1142, 1147, 1149, 1153, 1160 f., 1167, 1178, 1181, 1212, 1217 f., 1224, 1228 f., 1255, 1267⫺1278, 1309, 1325, 1328, 1491, 1546, 1564, 1582, 1600, 1602 f., 1608 f., 1612, 1615⫺1621, 1625⫺1634, 1636⫺1638, 1643, 1648, 1650, 1671⫺ 1674, 1681⫺1686, 1690 f., 1694, 1705, 1756⫺1758, 1767 f., 1770, 1773, 1775, 1783, 1797, 1799 f., 1803, 1808⫺1814, 1817, 1819⫺ 1821, 1835 f., 1838, 1840⫺ 1843, 1852⫺1854, 1860, 1862⫺1864, 1873, 1878, 1880, 1884⫺1888, 1890, 1894, 1897 f., 1901, 1903⫺ 1906, 1912⫺1916, 1919⫺ 1922, 1924, 1931 f., 1936 Althoch-/Old High 13, 122, 159, 636, 930, 936, 956, 1604⫺1610, 1618, 1637, 1680⫺1694, 1910 Bayrisch/Bavarian 1067 Fränkisch/Franconian 584 Hessisch/Hessian 475, 582, 584 Mittel-/Middle 1069, 1682, 1691 Mittelhoch-/Middle High 649, 930, 1217, 1604⫺1606, 1609 f., 1616 f., 1621, 1637, 1680⫺1694, 1914, 1917 Neuhoch-/Modern High 675, 1087, 1228, 1602⫺1604, 1626, 1631, 1680⫺1686, 1688⫺1694, 1910, 1917, 1920 Nieder-/Lower 1681 Nord-/Northern 1625 Ober-/Upper 1682, 1685, 1688, 1690, 1692 f.
1975
Sprachenregister ⫺ Sächsisch/Saxonian 1681, 1692 ⫺ Süd-/Southern 1178, 1186, 1625 ⫺ Westmittel-/Western Middle 1689 ⫺ Westober-/Western Upper 1689 Deutsche Gebärdensprache/German Sign Language 1554⫺ 1561 Dholuo 802, 1142 Dhuwal 1003 Dieguen˜o 569 f., 1023, 1088 Digul-Wambo 1501 Dinka 571, 851 Diyari 987, 993, 1035, 1078, 1158, 1159, 1490⫺1500 Djabugay 258, 262 dravidische Sprachen/Dravidian languages 561, 813, 1013, 1032, 1037, 1039, 1043, 1077 f., 1080, 1085 Dutch (s. Niederländisch) Dyirbal 481, 643, 714, 716, 719, 742, 758, 817, 1004, 1011, 1029, 1034⫺1038, 1043, 1080, 1112, 1116, 1655 f. ⫺ Giramay 1011 Dyola 541 ’Dongo-ko 1040
E Efik 778 Englisch/English 13 f., 16, 19, 38, 52 f., 58, 61 f., 64, 82, 86⫺88, 95, 98, 118, 120, 133, 135, 149, 152, 154 f., 163, 172, 175, 178, 180 f., 185, 195 f., 198⫺203, 205, 216 f., 232, 238, 245, 248⫺253, 255, 258, 263⫺267, 272⫺279, 282, 284⫺286, 288, 296⫺300, 308, 312⫺314, 316, 318⫺320, 322⫺324, 329, 331, 335 f., 340, 342, 346⫺348, 351 f., 354 f., 358 f., 362⫺365, 367, 370⫺373, 379⫺381, 383, 393⫺395, 402, 408, 410, 412⫺414, 427 f., 430, 433, 440⫺442, 453, 457, 464⫺466, 471⫺474, 481, 484 f., 490⫺ 494, 497, 500⫺502, 506 f., 512, 514 f., 517, 525, 529⫺ 531, 533 f., 536 f., 541, 545⫺ 547, 568⫺571, 576⫺578, 583, 587, 589, 596, 599, 601 f., 607, 610, 617 f., 620 f., 625 f., 628, 633, 648 f., 651, 653, 655⫺658, 660⫺663, 668, 670, 675, 682, 684, 708, 713⫺718,
⫺
⫺ ⫺ ⫺ ⫺
727 f., 732, 741, 743⫺747, 749, 751, 755, 757, 759 f., 762⫺765, 770⫺773, 775, 777⫺782, 787⫺793, 795⫺ 801, 803, 806⫺808, 810⫺818, 824, 827 f., 830, 832⫺834, 840⫺843, 845 f., 849, 853, 860, 864, 870, 874, 881 f., 886, 889⫺891, 893⫺895, 899⫺905, 908 f., 911⫺913, 921 f., 925 f., 930 f., 935⫺937, 939⫺941, 943⫺951, 953 f., 957⫺962, 964⫺981, 988 f., 998⫺1004, 1006⫺1009, 1011, 1013, 1020, 1033 f., 1036, 1043, 1053, 1060 f., 1063, 1067⫺1069, 1072, 1077, 1086 f., 1094, 1105, 1107 f., 1112⫺1116, 1118 f., 1121 f., 1127, 1130, 1134⫺1136, 1138, 1147, 1149, 1153 f., 1156 f., 1161 f., 1165⫺1178, 1180 f., 1183⫺1189, 1193, 1197⫺ 1200, 1202, 1205, 1208⫺ 1210, 1213⫺1215, 1217, 1222⫺1224, 1226, 1228 f., 1232⫺1234, 1236, 1255⫺ 1266, 1276, 1278, 1329, 1334, 1356, 1400, 1411 f., 1415, 1417, 1422, 1424, 1433 f., 1437, 1439, 1443, 1466, 1473, 1480, 1512, 1529, 1533, 1541, 1546, 1564, 1567, 1584, 1591, 1593 f., 1596, 1612, 1615⫺ 1622, 1630⫺1633, 1636⫺ 1640, 1642 f., 1653, 1656 f., 1664⫺1666, 1668, 1671⫺ 1679, 1694, 1704⫺1706, 1724, 1752, 1755⫺1758, 1762 f., 1766, 1769 f., 1773⫺ 1775, 1778, 1780, 1782⫺ 1785, 1789, 1796⫺1798, 1800⫺1803, 1806⫺1809, 1812⫺1814, 1819⫺1822, 1831, 1836, 1839 f., 1852, 1862, 1882⫺1890, 1894, 1896⫺1899, 1912, 1914, 1920⫺1922, 1924⫺1927 Alt-/Old 86 f., 118, 372, 573, 576, 599, 696, 1043, 1080, 1087, 1228, 1236, 1239, 1242, 1256, 1343, 1616⫺1618, 1620, 1627⫺1630, 1632, 1634, 1636, 1641, 1666, 1668⫺1677, 1679 amerikanisches/American 180, 772, 970, 1264 f., 953, 1473, 1616, 1620, 1625, 1633 angelsächsisches/AngloSaxon 95, 1236, 1621, 1921 anglo-irisches/Hibernian English 1649 britisches/British 1625, 1628, 1633, 1071, 1264 f., 1883
⫺ chinesisches Pidgin/Chinese Pidgin 1657 ⫺ Frühneu-/Early Modern 1668, 1671⫺1675, 1677⫺ 1679 ⫺ Hindi-Englisch/Hindi English 232 ⫺ Inglis (Scottis) 1668⫺1670 ⫺ kentisches/Kentish 1668, 1670 ⫺ Mercian 1668, 1669 ⫺ Midland 1668, 1670, 1672 f. ⫺ Mittel-/Middle 88, 649, 1579, 1617, 1626, 1630, 1633, 1642, 1668, 1671⫺1679 ⫺ Neu-/Modern (Present-day) 118, 1239, 1242, 1256, 1343, 1626⫺1631, 1633 f., 1642 f., 1663, 1666, 1668, 1671, 1674 f., 1678 f. ⫺ Neufundland-/Newfoundland 1070 ⫺ Neuseeland-/New Zealand 1071 ⫺ Nord-/Northern 1668⫺1670, 1672⫺1674 ⫺ Spätneu-/Late Modern 1668, 1677 ⫺ Süd-/Southern 1668, 1672 f. ⫺ tasmanisches/Tasmanian 1070 ⫺ West Saxon 1668 f., 1671 ⫺ West Somerset 1070 Ennemor 569 f., 573 Epun 1822 eritreische Sprachen/Eritrean languages 1808 Erzya Mordva/Erza Mordvin 821, 823, 825⫺830 Ese’ejja 488 Eskimo 204, 254, 380, 539, 739, 922 f., 926, 1005, 1009, 1152, 1162, 1223, 1236 f., 1239, 1241 f., 1400, 1407, 1651 Eskimo-aleutische Sprachen/ Eskimo-Aleut languages 1032, 1077 Esperanto 453, 933, 1563⫺1571 Estnisch/Estonian 546, 774, 778, 814, 1042, 1076, 1085, 1218, 1328 Etsako 571, 573 europäische Sprachen/European languages 91, 367, 381, 620, 1001, 1016, 1019, 1067, 1071, 1147, 1157, 1183⫺1186, 1208, 1218, 1228 f., 1319, 1564, 1570 f., 1647⫺1649, 1682, 1657, 1704, 1706, 1708, 1886 f., 1931 ⫺ nord-/North 1189 ⫺ west-/West 381, 1189, 1647 Even 1151
1976 Evenki 1151, 1160 Ewe 30, 719, 751, 1005 f., 1036, 1092, 1094 Eyak 1023, 1028, 1399
F Fang 469 Färöisch/Faroese 1268 Farsi 779 Fidschi/Fijian (Fiji) 1108, 1208, 1135 f., 1139 ⫺ Boumaa 1108 Filipino (s. Philippinisch) Finnisch/Finnish 252, 330, 388, 389⫺391, 396, 400, 569 f., 622, 647, 649⫺652, 701, 753, 758, 809, 811, 817, 825, 941, 980, 988, 1060, 1072, 1079 f., 1084⫺1086, 1114, 1121, 1127, 1160, 1185, 1186, 1328⫺ 1342, 1602 f., 1650, 1797, 1894, 1897 finnopermische Sprachen/FinnoPermic languages 401 finnougrische Sprachen/FinnoUgric languages 541, 1056, 1083, 1218, 1328 Flämisch/Flemish 829, 1038 Fore 1078, 1081, 1127, 1203 f. Französisch/French 16 f., 86⫺89, 93, 109, 118, 135, 138⫺142, 144, 159, 175 f., 183 f., 216, 232, 235⫺237, 245, 249, 253, 274 f., 279, 291, 297, 316, 318, 338, 352, 358, 361, 363, 378⫺383, 385, 393, 395, 398 f., 402, 454, 457 f., 476, 484⫺488, 492, 495, 506, 512, 514⫺517, 521, 525, 535 f., 571, 582⫺585, 599, 618 f., 622, 626, 655, 659 f., 662, 675, 680, 696⫺698, 702 f., 708, 732, 745, 750, 772, 774, 776 f., 781, 800, 812, 832, 833 f., 841, 843 f., 846⫺849, 851 f., 859, 869, 872⫺874, 881, 886, 888⫺890, 892, 894, 908, 911⫺913, 929, 935, 940, 950, 957⫺960, 962, 964, 966, 979, 1006 f., 1011, 1013, 1016, 1031, 1037, 1039, 1042, 1051, 1063, 1067⫺1069, 1112, 1160, 1166 f., 1173⫺ 1175, 1177⫺1179, 1181, 1186, 1217, 1228 f., 1258 f., 1264, 1285⫺1297, 1411, 1417, 1564, 1581, 1583, 1590, 1592, 1594⫺1596, 1603, 1615⫺ 1618, 1620⫺1622, 1628, 1630, 1634, 1637 f., 1640⫺ 1642, 1647 f., 1653, 1666,
Verzeichnisse und Sprachenkarten 1671, 1675⫺1677, 1679, 1698⫺1710, 1745, 1750, 1785, 1799⫺1801, 1803, 1808, 1811 f., 1819, 1821, 1836, 1843, 1852, 1859, 1861, 1863, 1883⫺1885, 1887⫺ 1889, 1891, 1897, 1910⫺ 1912, 1914, 1920 f., 1924, 1931 ⫺ Alt-/Old 86 f., 278, 645, 1086, 1579, 1616 f., 1621, 1698⫺ 1709 ⫺ Altnord-/Old Northern 1618 ⫺ elsässisches/Alsatian 1617 ⫺ kanadisches/Canadian 488 ⫺ Mittel-/Middle 1698⫺1701, 1704 f., 1707 ⫺ Neu-/Modern 1699⫺1705, 1707⫺1709 ⫺ normannisches/Norman 1257 ⫺ West-/West 444 Französische Gebärdensprache/ French Sign Language 1560 Friesisch/Frisian 95, 278, 1070, 1255 ⫺ West-/West 444, 754 Fula 568⫺571, 573 f., 762, 790, 1040, 1047, 1049, 1134 f., 1150 Fur 1042
G Gaagadju 1034 Gälisch/Gaelic 228, 1656 ⫺ schottisches/Scottish 1655 ⫺ hebridisches/Hebridean 623 Gallisch/Gallic 1218 ˙ a¯mid 1744 G Garo 537, 1196 Georgisch/Georgian 87, 325, 439, 442⫺444, 521, 541 f., 597, 814, 934, 1079, 1082, 1085, 1136⫺1138, 1186, 1215 ⫺ Alt-/Old 754, 1079 ⫺ Neu-/Modern 1215 germanische Sprachen/Germanic languages 94⫺96, 121, 157 f., 300, 314, 335, 338, 341, 361, 363, 382, 384, 386, 444, 520, 531, 541, 548, 642, 653, 849, 857 f., 861, 880, 911, 937, 950, 1045, 1070, 1078, 1080, 1167, 1172, 1186, 1228, 1257 f., 1262, 1264, 1268, 1310, 1563 f., 1596, 1609, 1643, 1670, 1674, 1675, 1681, 1685⫺1687, 1689, 1691⫺ 1693, 1704, 1707, 1709, 1801, 1903, 1927 ⫺ nord-/Northern 1914 ⫺ nordost-/North East 380
⫺ süd-/Southern 1681 ⫺ Urgermanisch/Proto-Germanic 520, 1268 ⫺ west-/West 118, 1255, 1267 Giha 1527 Gimi 549 Gimira 1007, 1011 Godie´ 984, 987, 994 Godoberi 1035, 1037 Gokana 588, 590, 592 Gola 741 Golin 1053 Gombe Fulani 538, 814 Gooniyandi 1078, 1203, 1871 Gotisch/Gothic 93, 96, 104, 159, 936, 1056, 1132, 1139, 1141, 1239, 1242, 1621, 1662, 1666, 1685, 1691 Grebo 1135 Greenlandic (s. Inuktitut) Griechisch/Greek 18, 40, 52⫺65, 74, 77⫺79, 86 f., 92⫺96, 139, 152, 174, 242, 245, 251, 253, 304, 367, 388, 406, 422, 435, 451 f., 457, 539, 561, 563 f., 642, 644 f., 654, 694, 696, 701, 709, 748, 777 f., 780, 809, 812, 814 f., 818, 869, 874, 901, 945, 957, 968, 971, 1031, 1074, 1076, 1139 f., 1145, 1149 f., 1156, 1181, 1222, 1224, 1239, 1242, 1247, 1261, 1264, 1288, 1310, 1313⫺1321, 1323, 1325, 1343, 1617, 1619 f., 1648, 1662⫺1666, 1671, 1704 f., 1709, 1750, 1835, 1862, 1878, 1918, 1925, 1927 f., 1930⫺ 1932 ⫺ aiolisches/Aeolic 1310 ⫺ Alt-/Ancient 119, 121 f., 239, 291, 354, 516, 624, 732, 1056 f., 1074, 1076⫺1078, 1080, 1086 f., 1214, 1218, 1224, 1228, 1239, 1310⫺ 1326, 1616, 1621, 1853, 1914, 1917 ⫺ anatolisches/Anatolian 585 ⫺ arkadisch-kyprisches/ArcadoCypriot 1310 ⫺ attisches/Attic 956, 1311, 1313, 1316, 1319, 1321 ⫺ dorisches/Doric 1310 ⫺ homerisches/Homeric 1064, 1133, 1242, 1310, 1315, 1325 ⫺ ionisch-attisches/Ionic-Attic 1310, 1325 ⫺ klassisches/Classical 342, 658, 988, 1149 f., 1155, 1157, 1159, 1233 ⫺ Koine´ 1311 ⫺ makedonisches/Macedonian 585
1977
Sprachenregister ⫺ mykenisches/Mycenaean 1310 ⫺ Neu-/Modern 58, 235, 359, 662 f., 1055, 1134, 1187, 1217, 1239, 1242 ⫺ Nordwest-/Northwest 1310 ⫺ Nord-/Northern 585 ⫺ West-/West 1310 Grönländisch (s. Inuktitut) Gta? 553 f., 570 Guahibo 1032 Guajajara 1194 Guajiro 1034 Guana 773 Guaranı´ 711, 927, 1107, 1233, 1421⫺1432 Guinea-Bissau Portuguese Creole (s. Portugiesische Kreolsprache von Guinea-Bissau) Gujarati 1712, 1723⫺1729 Gumbaynggir 992, 1095, 1097 Guoyu 1730 Gur 936 Gurindji 796, 1003 Guugu Yimidhirr 797, 993, 1092, 1099 Gu-Yalanji 539 Gwari 592 Gyarong 1082
H Haitisch/Haitian 237, 1653, 1657 f. Harakmbet 927 Harari 540 Haruai 817 Hatam 1056, 1062 f. Hausa 29, 467, 518, 590, 593, 714 f., 718, 741, 752, 758, 777, 786⫺788, 793, 817, 990 f., 993, 1035, 1216 Havyaka 1208⫺1210 Hawaiisch/Hawaiian 1095, 1862 Haya 590 f. Hebräisch/Hebrew 79, 331, 364, 458, 474 f., 538, 702 f., 711, 737, 781, 799, 957, 1061, 1092, 1094, 1096, 1139, 1228, 1343⫺1356, 1650, 1796 f., 1799⫺1803, 1819, 1821 ⫺ Alt-/Old 362 ⫺ klassisches/Classical (Biblical) 362, 1343, 1345, 1352 f. ⫺ mittelalterliches/Medieval 1343 ⫺ Neu-/Modern (Israeli Hebrew, Contemporary Hebrew, Ivrit) 1039, 1343, 1345, 1352 f., 1355 ⫺ rabbinisches/Rabbinical (Mishnaic) 1343
Herero 1094 Hethitisch/Hittite 35, 122, 641⫺ 643, 735 f., 738 f., 1042, 1239, 1242, 1310, 1662 f. Hidatsa 584, 1197, 1198 Hiligaynon 1061 Hindi 232, 245, 546, 713, 954, 1039, 1094, 1138, 1142, 1161, 1712, 1723⫺1729 Hixkaryana 739, 741⫺743, 823, 979 f., 1098, 1203, 1208 Hoka/Hokan 541, 773, 1125, 1126 Hopi 746, 987, 994, 1064, 1433 Hottentot (s. Nama) Houailou 1094 Hua 547, 549, 568⫺570, 573, 712, 1012, 1204 f., 1232 f. Huahtla 926 Hualapai 1063 f. Huichol 242 f., 536, 1054, 1056, 1058⫺1063, 1433 Hungarian (s. Ungarisch) Hunzib 1062, 1367⫺1375 Hurrisch/Hurrian 1079
I Iatmul 1036, 1127 Igbo 589, 767, 934, 1112, 1128 Ijo 741, 1658 f. Ik 1092, 1098 f. Ika 1135 Ila 777 Illyrisch/Illyrian 1310 Ilokano/Ilocano 971, 1061 Imonda 1023, 1026 f. Indisch/Indic 1042, 1085, 1914, 1919 ⫺ Alt-/Old 96, 100, 548, 642, 644 f., 1056, 1316, 1879, 1914 ⫺ Mittelindische Umgangssprache/Middle Indic vernacular 42, 50 indoarische Sprachen/IndoAryan languages 737, 1013, 1078, 1209, 1712⫺1728 ⫺ alt-/Old 1712⫺1723, 1728 ⫺ mittel-/Middle 1712, 1719⫺ 1722, 1725⫺1729 ⫺ neu-/New 1712, 1720, 1722⫺ 1729 indogermanische (indoeuropäische) Sprachen/Indo-European languages 18, 94⫺96, 100, 119, 121 f., 126, 145, 147, 149, 157 f., 361, 376, 396, 536, 539, 541, 548 f., 561, 568, 602, 631, 633, 639, 642, 644 f., 700, 710, 725, 735, 737⫺739, 746, 748, 772, 809, 911, 940, 945, 971, 1013,
1031⫺1035, 1040⫺1043, 1056, 1068, 1076⫺1078, 1080, 1083⫺1087, 1112, 1141, 1147, 1149 f., 1156, 1182, 1203, 1205, 1215⫺ 1218, 1227 f., 1237, 1301, 1307, 1310, 1312 f., 1317, 1328, 1334, 1336, 1343, 1412, 1439, 1560, 1649, 1655, 1663, 1686 f., 1689, 1712, 1717 f., 1724, 1730, 1752, 1853 ⫺ Urindogermanisch/ProtoIndo-European 100, 158, 514, 612, 642, 1087, 1142, 1159, 1214, 1216, 1328, 1662⫺1667 indoiranische Sprachen/Indo-Iranian languages 122, 1085 f., 1310 indonesische Sprachen/Indonesian languages 94 Indonesisch/Indonesian 241, 414, 540 f., 549, 676, 715, 721, 739, 758, 849, 921, 934, 951, 1011, 1017, 1134, 1218, 1501, 1503, 1853 Inga 1197 Ingrisch/Ingrian 1328 Ingusch/Ingush 1039 Interlingua 1563⫺1571 Inuktitut (Grönländisch/Greenlandic) 254 f., 538, 701, 768, 922, 924, 1121, 1152, 1218, 1400, 1800, 1821 ⫺ Ostgrönländisch/East Greenlandic 1389 ⫺ Polareskimo/Polar Eskimo 1389 ⫺ Westgrönländisch/West Greenlandic 979, 981, 1107, 1216, 1389⫺1398, 1840 Irakw 1035 iranische Sprachen/Iranian languages 314, 541, 602, 642, 1042, 1808, 1651 Irisch/Irish 328, 500, 503⫺505, 517, 570 f., 574, 627 f., 806, 823, 826 f., 1095, 1218, 1233, 1622, 1649 ⫺ Alt-/Old 87, 1215, 1219, 1664 f. ⫺ Neu-/Modern 1215 f., 1218 f. irokesische Sprachen/Iroquoian languages 926, 975, 1022, 1263 ⫺ süd-/Southern 1023 Isländisch/Icelandic 444, 500, 502, 584, 620, 1610, 1618, 1085, 1268, 1914 ⫺ Alt-/Old 13, 86, 1604, 1606, 1607 ⫺ Neu-/Modern 13, 1601, 1607, 1609
1978 Isoko 593 Italienisch/Italian 14, 59, 95, 198, 199, 201, 249 f., 255, 330, 352, 358, 361, 364, 366, 378, 384, 398, 444, 447, 452, 456⫺458, 469, 471, 485 f., 542, 585, 587 f., 661, 697, 718, 743, 750, 775 f., 812 f., 832, 841, 859, 869, 874 f., 878, 882, 890, 892 f., 934, 936, 988, 1031, 1046 f., 1049, 1051, 1068, 1216⫺1218, 1264, 1288, 1563 f., 1592, 1617 f., 1671, 1709, 1741, 1745, 1750, 1784, 1797, 1808, 1810, 1813 f., 1819, 1821, 1931 Italisch/Italic 1310 Itelmenisch/Itelmen 443 Itza´ 257 Iwaidja 711 f.
J Jakaltekisch/Jacaltec 516, 979, 1017, 1021, 1026⫺1028, 1094 Jakutisch 1236, 1239, 1242 Japanisch/Japanese 14, 88, 236, 238, 244, 375, 536, 712, 717, 745, 755, 805, 807, 832, 882, 961, 969, 987⫺989, 992 f., 998, 1000, 1003 f., 1006, 1009, 1012, 1017, 1019, 1028, 1045, 1051, 1060, 1063 f., 1077, 1094, 1114, 1128, 1129, 1146, 1150 f., 1153, 1156, 1161, 1185, 1187, 1208, 1215, 1755, 1819, 1821, 1844, 1862, 1897, 1916, 1917 ⫺ Alt-/Old 970 ⫺ Neu-/Modern 969 Javanisch/Javanese 95, 568⫺570, 883, 951, 1012 ⫺ Alt-/Old 999 Jarawara 1034, 1038 Jawony 1010, 1034, 1042 Jeˆ 1032 Jenissei-Sprachen/Yenissei languages 1376 Jiddisch/Yiddish 243, 482, 971, 1186, 1650 Jivaro 539, 1654 Jukagirisch/Yukaghir 543
K K’iche 1800 Kabardian 1071 Kabyl 539 Kachin 1733
Verzeichnisse und Sprachenkarten Kaingang 1035, 1832 Kalam 713, 795 f. Kaliana 773 Kalkatungu 1084⫺1087, 1136 Kalmuk 1218 Kambata 467, 474 Kambera 362 Kambodschanisch/Cambodian 1225 Kannada 758, 761, 765, 1137, 1139, 1208, 1211 Kanukuru 798 Kanuri 374, 807, 1135, 1205 Karaboro 1183 Karaimisch/Karaim 439, 441 f. Karelisch/Karelian 1086, 1328 karibische Sprachen/Cariban languages 1208 ⫺ Ge-Pano 541 ⫺ Kreolsprachen/Creoles 1655 Karok 541, 546, 553, 781, 803 kartvelische Sprachen/Kartvelian languages (s. kaukasische Sprachen) Kaschmiri 1062 Kasem 576 Kashaya 1203 kaspische Sprachen/Caspian languages 541 Kassitisch/Kassite 35⫺37 Katalanisch/Catalan 619 f., 622 Katcha 1035 kaukasische Sprachen/Caucasian languages 482, 541, 548, 1032, 1035, 1041, 1077, 1084 f., 1096, 1203, 1205 ⫺ kartvelische Sprachen/Kartvelian languages 1082, 1215, 1218 ⫺ nordost-/Northeast 1034, 1038, 1083, 1085, 1142, 1203 ⫺ nordwest-/Northwest 549, 980 f., 1004, 1008, 1038, 1096, 1109, 1136, 1227 ⫺ süd-/South 1032 Kawi 94 f. Kayardild 708 f., 711, 717 f., 744, 753 f., 1057, 1087, 1108, 1880 Kazakh 529 Kegboid 1043 Kekchi 536, 541, 543 keltische Sprachen/Celtic languages 93 f., 327, 478, 503, 569⫺571, 574, 1071 f., 1218, 1310, 1620, 1665 Kerasan 1211 Ketisch/Ket 536, 1035 f., 1376⫺ 1388 Kewa 1011, 1128 Khinalug 1033 Khmer-Sprachen/Khmer languages 539, 720, 805 Khmu 934
Khoisan-Sprachen/Khoisan languages 541 ⫺ Nord-/Northern 1032 ⫺ Zentral-/Central 1037 Kihehe 559, 561, 565 Kikuyu 1527 Kilivila 1012 Kimberley 1038 Kimbundu 1160 Kinande (s. Nandi) Kinubi (s. Nubi) Kinyarwanda (Rwanda) 1058, 1062, 1135, 1139, 1527⫺ 1544, 1867, 1904 Kiowa 917 Kiowa-Tanoa-Sprachen/KiowaTanoan languages 926, 1042 Kiranti 1082 Kiriwinisch/Kiriwinian 757 Kirundi 530 f., 1038, 1527, 1536 Kiswahili (s. Swahili) Kitimaat 723 Kituba 1654 Kiwai 536, 541 Koasati 584 Kobon 713, 736, 767, 1202 f. Kolami 1035, 1039 Kombai 1501 Komi 330, 1183 Komy-Permyak 1085 Konkani 767, 1712, 1724 kordofanische Sprachen/Kordofanian languages 1037, 1208 Koreanisch/Korean 275, 568⫺ 570, 708, 714, 716, 1012, 1064, 1077 f., 1120, 1128, 1151, 1153, 1160 f., 1208, 1663, 1842, 1897, 1919 Koromfe 759, 760, 766⫺768 Koryak 528 Kottisch/Kott 786, 1376 Koyo 741 Koyra Chiini 1006 Koyukon 917⫺919, 1399⫺1410 Kresh 1094 Krongo 1091, 1095, 1098, 1107 Kru 1043 Kui 539, 766 Kunjen 751 Kunwinyku 1010 Kurdisch/Kurdish 1186, 1741 Kuria 1037 kuschitische Sprachen/Cushitic languages 555, 577, 777, 919, 927, 1007, 1013, 1054 f., 1057 f., 1077, 1135 ⫺ Ost-/East 474 ⫺ Süd-/Southern 1647 Kuyani 1491 Kwa 569 Kwakw’ala 739, 742 f., 934 Kwara’ae 577, 580 Kxoe 743, 1036, 1042
1979
Sprachenregister L Lahu 587, 589 Lak 516, 715, 1038 Lakhota 758, 846 Lango 802, 1142, 1203 Lappisch/Lappish 95, 330, 786 Larike 1057 Latein/Latin 7, 11, 13 f., 18, 25, 28, 52⫺65, 76⫺79, 81⫺83, 86⫺89, 93⫺96, 118⫺120, 139⫺141, 153 f., 159, 164, 174, 187, 227 f., 235, 237, 250, 268 f., 271, 277 f., 291, 293 f., 304, 362 f., 367, 376, 393, 396, 406, 418, 422, 536⫺538, 548, 561, 563 f., 576, 584 f., 588, 597⫺600, 602⫺605, 618, 620, 631, 634, 637⫺639, 642⫺645, 647 f., 651, 654⫺657, 662, 674 f., 694, 696 f., 701, 704, 732 f., 735⫺738, 743 f., 746⫺748, 750⫺752, 754, 761, 770, 772, 774, 777 f., 780⫺782, 799, 808, 812⫺814, 852, 857, 869, 874, 881, 901, 911, 935, 940, 945 f., 950, 956 f., 966 f., 988, 1000, 1006, 1031, 1039, 1051, 1064, 1074⫺1077, 1079, 1081, 1084⫺1087, 1126, 1135 f., 1139, 1140, 1145, 1147, 1149, 1173, 1181, 1186 f., 1202, 1213⫺1215, 1217 f., 1222, 1224⫺1226, 1228 f., 1247, 1264, 1273, 1288, 1295, 1310, 1314, 1316, 1325, 1343, 1491, 1545, 1563 f., 1613, 1615⫺1619, 1621, 1628⫺1630, 1633 f., 1642, 1648 f., 1651, 1662⫺ 1666, 1671, 1675 f., 1681, 1684, 1687, 1698⫺1710, 1834⫺1836, 1838⫺1840, 1842, 1852⫺1854, 1861 f., 1877⫺1879, 1911, 1917⫺ 1919, 1926⫺1933, 1935 ⫺ Alt-/Old 1706 ⫺ Früh-/Early 1611 f. ⫺ klassisches/Classical 278, 660, 1612, 1698⫺1701, 1704 ⫺ Spät-/Late 538, 1617, 1649, 1703, 1706, 1708 ⫺ Vulgär-/Vulgar 1592, 1698⫺ 1703, 1705, 1708 Laz 1082 Lendu 588 Lesgisch/Lezgi (Lezgian) 517, 604 f., 658, 660, 663, 1041, 1080, 1137, 1871 Lettisch/Latvian 92, 693, 1069 f., 1650 Limbu 1214
Lingala 810 Lingua Franca 1653 Litauisch/Lithuanian 96, 401, 619, 1056, 1069, 1132, 1140, 1142 Livisch/Livonian 1070, 1328, 1650 Logo 1095 Lokono 1034 Lotuxo 772 Lude 1328 Luganda 593, 757, 1039, 1070, 1120, 1123 f., 1527 Luisen˜o 396, 400, 818, 1204 Lule 1835 Lumbi 767 Lunda 1042 Luo 1188 Luoravetlan 927 Lwo 802 Lye´le´ 987, 994
M Ma 1037 Ma’a 1647 Maasai 516, 530, 536, 588 Maba 1079 Macushi 1097 Madagassisch/Malagasy 702, 991 f. Madura/Madurese 400 Maidu 539, 1203 Maipura-Sprachen/Maipuran languages 927, 1467 Maithili 1161 Makedonisch/Macedonian 29, 327, 391, 396, 398, 400, 585, 1176, 1245 Makro-Geˆ Sprachen/Macro-Ge languages 773 Maku´-Puinave 921 Malabang 1762 Malaiisch/Malay 528, 547, 549, 751, 784, 1045, 1048, 1758 ⫺ Bazaar 1501 Malak-Malak 1039 Malayalam 537, 546 malaiopolynesische Sprachen/ Malayopolynesian languages 1061, 1473 Maltesisch/Maltese 658, 760, 1056, 1071, 1741, 1745, 1748, 1750 Malto 1035, 1040 f. Mam 258, 262, 541, 1012, 1022, 1026, 1033, 1100 mamische Sprachen/Mamean languages 1028 Manam 999, 1022 Manambu 1033, 1036⫺1038 Manchu (s. Mandschu)
Mandan/Mandan 1024 Mandara 546, 548 Mandarin (s. Chinesisch) Mande-Sprachen/Mande languages 588, 1092 Mandinka 1659 Mandobo 1501 Mandschu/Manchu 495, 554, 568⫺570, 1738 Mandschu-tungusische Sprachen/Manchu-Tungus languages 1032 Mangarayi 744, 750 f., 835, 978, 1082 Mangbetu 539 Manipuri 758 f., 1209 Manjaku 1659 Maori 29, 766, 935 Mapuche 1194 f., 1197, 1205 Mara 998 Marathi 1712, 1724, 1729 Margi 540 Mari 330 Maricopa 1009, 1203 Marind 1038 Marrithiyel 1042 Marshallese 935 Mayasprachen/Mayan languages 532, 539, 541, 774, 927, 1012, 1028, 1098, 1100, 1152, 1835, 1842, 1863 Mayali 529, 711 f., 918, 926, 1042, 1060, 1854 Mba 1040, 1094 Mbangi 1099 Mbugu 1647 Media Lengua 1647 Mednyj Aleut 1651 Meglenorumänisch/Meglenoromanian 1651 Melanesisch/Melanesian 1045, 1755 f. Mende 569 f., 572, 574, 588 Menomini 735 Miao 1734 Mikasuki 584 Mikir 1733 mikronesische Sprachen/Micronesian languages 1021 f., 1024 Miskito 548, 750, 1216 Miwok 555 f., 578, 803 f. Mixtekisch/Mixtec 431, 591, 927, 1021, 1027 ⫺ Caotzoquitengo 1021 ⫺ Chalcatongo 766 Mixtepekisch/Mixtepec 541 Miya 1216 Mohawk 711 f., 924, 993, 1027, 1029, 1800 Mokil/Mokilese 559 f., 563 Mongolsprachen/Mongol languages 1217
1980 Mongolisch/Mongolian 1217 f. Mon-Khmer-Sprachen/Mon Khmer languages 548, 1128, 1734 Montagnais (Innu-aimun) 588, 1411⫺1419 Mordvinisch/Mordvinian 330 Motu 935 Motuna 1032, 1039, 1041 Mparntwe Arrernte 1203 Mpi 590 Mubi 554 Munda-Sprachen/Mundan languages 548, 553, 927, 1004, 1013, 1128 Mundani 988, 1209 Mundari 720, 725 f., 728 Munduruku´ 919, 1022, 1024 Mupun 1042 Mura 927 Murle 1100 Muskogee-Sprachen/Muskogean languages 584, 814, 926, 1088, 1110, 1142 Mutsun 578 f. Muyu 1501 Mwera 804, 1173
N Nade¨b (Nadeb) 921, 1136 Nahuatl 442, 541, 701, 711 f., 739, 797, 1433⫺1452 ⫺ Huasteca 1056, 1060 ⫺ Isthmus 1436 ⫺ klassisches/Classical 538, 621 f., 741, 926, 939, 1433, 1436, 1442, 1835 Naiki 1039 Nakanai 548, 560 f., 798 Nakhsprachen/Nakh languages 1032, 1039 Nama (Hottentot) 539, 763, 1045 Nanai 1085 Nandi (Kinande) 591, 593, 1135, 1507 Nasioi 1032 Naskapi 1411, 1419 Nass 524 Naukan 778 Navajo 316, 330, 539, 739, 796, 1005, 1011, 1098, 1100, 1136 Ndu 1032 Nembe 593 Nenets 1101 Neo-Melanesian (s. Tok Pisin) Nepali 1043 Nera 1100 Neuaramäisch/Neo-Aramaic 1741 f.
Verzeichnisse und Sprachenkarten Neuguinea-Pidgin/New Guinea Pidgin (s. Tok Pisin) Newari 1205 f. Nez Perce 1082, 1087 Ngäbe´re 810 Ngalakan 1034 Ngamini 1491 Ngandi 1010, 1023 Ngan’gityemerri 1042 Ngiyambaa 394, 396 f., 822, 1004, 1195⫺1199 Nicobarese 803 Niederländisch/Dutch 13, 162⫺ 166, 198, 229⫺232, 250, 300, 304, 306⫺310, 318, 336⫺342, 347, 351 f., 357⫺359, 361⫺ 366, 367, 412, 467 f., 476, 482, 516 f., 541, 715, 728, 800, 803, 821, 824, 829 f., 833, 841, 848, 856⫺862, 864, 883, 902, 904, 911, 926, 933, 943 f., 949 f., 977, 979, 1046, 1048⫺1051, 1063, 1070⫺ 1172, 1177 f., 1198, 1214, 1218, 1255, 1272, 1288, 1328, 1578⫺1581, 1583⫺1586, 1617, 1641, 1643, 1653, 1659, 1783 f., 1807⫺1813, 1819, 1821, 1891, 1894, 1897 f. ⫺ amerikanisches/American 232, 1579, 1587 ⫺ Ost-/East 1037 Niger-kordofanische Sprachen/ Niger-Kordofanian languages 937 Niger-Kongo-Sprachen/NigerCongo languages 541, 561, 978, 1006, 1032, 1042 f., 1094, 1112, 1135, 1137, 1187, 1192, 1199, 1208, 1517, 1538 nilo-saharanische Sprachen/NiloSaharan languages 772, 1077, 1135, 1208, 1507⫺ 1509, 1513 nilotische Sprachen/Nilotic languages 772, 1032, 1507, 1513 Nimboran 539, 803 Niue 1059, 1061 Nivkh 475, 716 Nkore-Kiga (Runyankore, Runyankole) 253, 768, 980 f., 1187, 1189, 1210, 1527 Nomatsiguenga 1136, 1138 Noni 590 Nootka 539, 721⫺723, 935, 940 ⫺ Nitinaht 251, 400, 722 Nordische Sprachen/Nordic languages 1604 ⫺ Alt-/Old 95, 1621, 1914 ⫺ Urnordisch/Proto-Nordic 1601, 1604⫺1606, 1608⫺ 1610, 1914
Norwegisch/Norwegian 146 f., 384, 784, 960, 1100, 1617, 1914 nostratische Sprachen/Nostratic languages 121 Nubi (Kinubi) 1654, 1752 Nubische Sprachen/Nubian languages 1752 Nuer 569 Nungali 1980 Nunggubuyu 1003, 1008, 1041 Nyigina 1078
O O’odham 806 Obolo 759 Ojibwe (Ojibwa) 785, 788, 1049 Ok 1501 Okrand 579 Okzitanisch/Occitan 85 f., 88, 1186 Ollari 1035 Omotische Sprachen/Omotic languages 1007, 1013 Onondaga 1844 Oriya 774, 778, 1043 Ormuri 541 Osage 1835 Ossetisch/Ossete 1085 Ostyak 774 Otomangue-Sprachen/Otomanguean languages 541, 1021, 1032 Otomi 569 f., 572 ozeanische Sprachen/Oceanic (Oceanian) languages 1016, 1021, 1025, 1094, 1097, 1135 f., 1142
P Paiute 716 Palantla 539 Palauan 990 f. Pa¯li 1712, 1719⫺1722 Palikur 1039 Pama-Nyunga-Sprachen/Pama-Nyungan languages 481, 987, 993, 1035, 1077⫺1079, 1085, 1158 f., 1491, 1493 f. Pamir-Sprachen/Pamir languages 541 Panara 1022 Panjabi 1723 Panoan 927, 1467 Panyjima 1009, 1079 Papago 584, 801, 1204 Papiamentu 1654, 1656⫺1659 Papuasprachen/Papuan languages 987, 1012, 1022,
1981
Sprachenregister 1032, 1037, 1041, 1078, 1100, 1110, 1120, 1126 f., 1142, 1202⫺1205, 1501, 1506 Parji 1035, 1039 Pashto 399 f., 541, 546 Paumarı´ 987, 993, 1041, 1043, 1071 Pawnee 543, 804, 1196 Peba 1464 Peba-Yagua-Sprachen/Peba-Yaguan languages 1464 Penuti-Sprachen/Penutian languages 541, 770, 1077 Persisch/Persian 93⫺95, 458, 520 f., 739, 741, 746, 1043, 1061, 1086, 1093, 1161, 1185, 1186, 1236, 1649, 1651 ⫺ Alt-/Old 520, 1239, 1242 ⫺ Avestisch/Avestan 520 ⫺ Frühneu-/Early Modern 541 ⫺ Neu-/Modern 1239, 1242, 1649 Philippinisch/Filipino 1473 Philippinische Sprachen/Philippine languages 1152⫺1156, 1473, 1479 Phrygisch/Phrygian 1310 Pima 801, 1204 Pipil 760, 979 Piraha˜ 1053 Pitjantjatjara 715 Pitta-Pitta 1491 Plains Cree 1151 Polnisch/Polish 134, 290 f., 318, 335, 339, 365, 391, 395 f., 414, 427 f., 432 f., 437, 441 f., 469, 490, 508, 520, 529, 548, 583⫺585, 661, 863, 881, 936 f., 1048⫺1050, 1068 f., 1085, 1245, 1801, 1809, 1819, 1889, 1904 polynesische Sprachen/Polynesian languages 94, 99, 585, 780, 1045, 1225, 1263 Pomo 1203, 1208 ⫺ Zentral-/central 1058 Ponape 467, 1022, 1024 Popoloc 593 Portugiesisch/Portuguese 271, 380 f., 697, 737, 799, 806, 1032 f., 1039, 1047, 1051, 1217, 1564, 1653, 1656 f., 1659, 1710, 1813 f. ⫺ brasilianisches/Brazilian 1042 Portugiesische Kreolsprache von Guinea-Bissau/Guinea-Bissau Portuguese Creole 1658 Potawatomi 476 f., 742 f. Potou-Tano 1517 Prakrit 41, 50, 1712, 1719⫺1722 Puglian 1069 Puluwat 717
Pumpokolisch/Pumpokol 1376 Puquina 1461 Putonghua 1730
Q Qafar 1035 Q’anjob’al (Kanjobal) 1020 f., 1024⫺1028 Qawasqar 488 Quechua 412, 701, 714, 739 f., 746, 828, 1004, 1091, 1101, 1106, 1137, 1140, 1194, 1199, 1453⫺1463, 1647, 1656 ⫺ Amazon 1654 f. ⫺ Ancash 1219, 1455, 1458, 1460 f. ⫺ Ayacucho 1455⫺1462 ⫺ bolivianisches/Bolivian 1139, 1460 ⫺ Cajamarca 1457 ⫺ Cochabamba 1157 ⫺ Cuzco 1457⫺1461, 1654 f. ⫺ ekuadorianisches/Ecuador 1458, 1460 f. ⫺ Huallaga 1011 ⫺ Huancayo 1455, 1459⫺1461 ⫺ Imbabura 1092, 1137 f., 1457, 1459, 1461 ⫺ Pacaraos 1454, 1456⫺1460 ⫺ Puno 1458 ⫺ San Martı´n 1457 ⫺ San Pedro de Cajas 1458 ⫺ südperuanisches/Southern Peruvian 1004 ⫺ Tarma 1454⫺1462 Quiche´ (s. K’iche) Quileute 547, 549, 560 f., 564, 565 f.
R Rama 1119, 1136, 1206 Rembarrnga 1042 Romani 1645, 1648⫺1651 ⫺ lettisches/Latvian 1070 romanische Sprachen/Romance languages 88, 141, 159, 323, 327, 358, 361, 367, 380⫺382, 384, 398, 422, 444, 521, 538, 645, 658, 660, 814 f., 849, 857, 911 f., 950, 961, 1045, 1051, 1070, 1077, 1157, 1167, 1175 f., 1184, 1186, 1217, 1228, 1233, 1261, 1264, 1310, 1440, 1563⫺1565, 1567, 1570 f., 1596, 1603, 1616, 1643, 1651, 1674, 1698, 1700⫺1704, 1706 f., 1710, 1798, 1801, 1810, 1904, 1921, 1927, 1931
⫺ balkan-/Balkan- 1649 ⫺ ibero-/Ibero- 1654, 1656 ⫺ räto-/Rheto- 444, 521, 1651, 1701 ⫺ west-/West 1087 Rotuman 475, 529, 577, 580, 936 Ruc 1734 Rukai 541 Rukiga 1527 Rumänisch/Romanian 26, 444, 529, 533 f., 584, 625, 645, 657, 741, 1047, 1086 f., 1217, 1648⫺1650 ⫺ Istro- 1648 Runyambo 1123 f., 1129 Runyankole (s. Nkore-Kiga) Runyoro 1527 Russisch/Russian 11, 14, 125, 127⫺133, 135, 155, 237, 239, 251, 254, 261, 269 f., 273, 277, 282, 285, 314, 338, 414, 418⫺421, 427, 431⫺433, 444, 466, 477, 481, 489 f., 492 f., 498, 513, 515 f., 518, 520 f., 525, 528, 531, 534, 535 f., 583, 601 f., 605, 607⫺613, 626, 632 f., 650⫺652, 655 f., 663, 697, 700, 708⫺711, 714 f., 717, 719, 735⫺737, 743, 746, 770⫺772, 776 f., 779, 782, 806, 814, 832, 886, 889, 892⫺894, 929, 930, 934⫺937, 939 f., 950, 957 f., 961, 975, 1033⫺1035, 1037, 1039, 1067⫺1072, 1077, 1080, 1085, 1132 f., 1139 f., 1142, 1157, 1159, 1161, 1167, 1172 f., 1175 f., 1181, 1183 f., 1186, 1188, 1210, 1234, 1245, 1256, 1277, 1300⫺1309, 1328, 1376, 1441, 1546, 1551 f., 1564, 1601, 1651, 1799, 1809, 1867 f., 1876, 1879, 1912, 1921 f. ⫺ Alt-/Old 937, 1070, 1218 ⫺ Mittelgroß-/Middle Great 1300 ⫺ Neu-/Modern 1218 ⫺ Nordgroß-/North Great 1300 ⫺ Südgroß-/South Great 1300 Rwanda (s. Kinyarwanda) S Saami (Saame, Sa´mi) 330 ⫺ Enontekio 401 f. ⫺ South 147 Saek 1216 Sait 1501 Sakao 539 Salish 411, 702, 721⫺724, 726, 729, 923, 926, 1032, 1119, 1208
1982 ⫺ Lummi 400, 578 ⫺ Saanich 411, 548 f., 578 Samoanisch/Samoan 242 f., 541, 549, 560 f., 564 f., 582, 758, 786, 1071, 1107 f., 1214, 1217 samoanisches Plantagen-Pidgin/ Samoan Plantation Pidgin 1755, 1758 samojedische Sprachen/Samoyedic (Samoyed) languages 330, 1328 Sango 757 Sanskrit 18, 41⫺43, 45, 49⫺51, 93⫺96, 98 f., 121, 174, 253, 358 f., 383, 454, 568 f., 623, 703, 737, 749, 758, 762 f., 908, 911, 956, 971, 1057 f., 1061, 1085 f., 1142, 1149 f., 1155⫺1157, 1159, 1214, 1222 f., 1228, 1236, 1239, 1242, 1247, 1592, 1662⫺ 1666, 1712, 1717⫺1722, 1726, 1728, 1837, 1862, 1914 Sanuma 1216 Saramaccan 1657⫺1659 Sayula Popoluca 622 Schwedisch/Swedish 14, 147 f., 380, 384, 648, 650, 662, 741, 902, 960, 1038, 1051, 1068⫺ 1072, 1189, 1328 f., 1648, 1793, 1821, 1903 f., 1914 Sekani 546 semitische Sprachen/Semitic languages 35, 83, 99, 134, 205, 314, 336, 339, 364, 410, 548 f., 552, 554, 558, 562, 741, 750, 869, 1056, 1077 f., 1080, 1085, 1095 f., 1215, 1228, 1343, 1655, 1741, 1750, 1835, 1917 ⫺ alt-/Old 1096 Seneca 1034, 1040, 1082, 1098 Sepik 1211, 1032, 1037, 1042 Serbo-Kroatisch/Serbo-Croatian 206, 327, 400, 520, 542, 603, 803, 1080, 1245, 1783, 1797, 1922 ⫺ Kroatisch/Croatian 1648, 1922 ⫺ Serbisch/Serbian 1648, 1922 Sesotho 1019, 1799 Seychellen-Kreolsprache/Seychelles Creole 1218 Shilluk 1062, 1232 Shona 1038 Siane 549 sibirische Sprachen/Siberian languages 475 ⫺ ost-/East 475 ⫺ paläo-/Paleo- 477 Sidamo 577 Sierra Popoluca 1004 Si-Luyana 846 f., 939
Verzeichnisse und Sprachenkarten Sinhala/Sinhalese 1086 sinotibetische Sprachen/SinoTibetan languages 549, 805, 1731, 1737 Sioux-Sprachen/Siouan languages 926, 1032, 1037 Siriono 711 Siuslaw 1080 Sizang (Siyin) Chin 1735 skandinavische Sprachen/Scandinavian languages 145, 911, 1621, 1819 ⫺ festland-/continental 13 Skandinavische Gebärdensprache/Skandinavian Sign Language 1556 Slave 407, 412, 798 slawische Sprachen/Slavic (Slavonic) languages 314, 318, 340, 375, 422, 444, 520, 548, 588, 648, 650, 772, 1034, 1042, 1045 f., 1049, 1068, 1071, 1142, 1157, 1167, 1172, 1176, 1228, 1301, 1307, 1310, 1647 f., 1651, 1800, 1904, 1922 ⫺ (Alt-)kirchenslawisch/Old Church Slavonic (Old Slavonic) 87, 517, 520, 548, 1040, 1071, 1239, 1242, 1300, 1308 ⫺ Ost-/East 1245 Slowakisch/Slovak 133, 477, 1245 Slowenisch/Slovenian 291, 1056, 1245 Solomoninsel-Pidgin/Solomon Islands Pijin 1755 Somali 778, 1093, 1096, 1099, 1135 Songhay 1006 Soninke 569 f. Sonrai 1160 Sora 547, 549, 1086 Sorbisch/Sorbian 362, 1056, 1648 ⫺ Ober-/Upper 318, 1245 Sotho 1186 Spanisch/Spanish 29, 95, 183 f., 193, 236 f., 243 f., 272 f., 275 f., 327, 371, 373, 390 f., 398 f., 485, 493, 497, 513, 517 f., 530, 542, 583 f., 621 f., 626, 658, 660, 697, 786⫺793, 796, 800, 806, 869, 873, 880, 883, 911⫺913, 926, 950, 954, 957 f., 960 f., 970, 988, 1009, 1019, 1034, 1036, 1039, 1041, 1049, 1069, 1071, 1082 f., 1086, 1064, 1091, 1106, 1186, 1194, 1217, 1229, 1264, 1421 f., 1426, 1433, 1437, 1461, 1464, 1467, 1471, 1473 f., 1564, 1582, 1592,
1603, 1617, 1646 f., 1653, 1656 f., 1666, 1671, 1709, 1807 f., 1813 f., 1919⫺1922, 1931 ⫺ Alt-/Old 576, 881 ⫺ argentinisches/Argentinian 327 ⫺ kantabrisches/Cantabrian 1036, 1042 ⫺ mexikanisches/Mexican 953 ⫺ Lena 1069 Squamish 539, 1020, 1024, 1027, 1080 Sranan (Sranan Tongo) 950, 1657, 1660 Sumerisch/Sumerian 35- 40, 751, 1217, 1916 Sundanesisch/Sundanese 151, 545, 548 f., 846, 935, 939, 951 Supyire 757, 761, 764 Svan 1215 Swahili (Kiswahili) 546, 711, 743, 762, 786, 789, 791, 793, 814, 934 f., 939, 1031⫺1033, 1037 f., 1042, 1045, 1047 f., 1062, 1094, 1135, 1141, 1187, 1189, 1236, 1239⫺1242, 1508, 1527, 1529, 1533, 1542, 1655, 1842 f., 1880, 1888 Syrisch/Syriac 59, 74
T Tabassaranisch/Tabassaran 516, 642, 1083 Tadzhik 458 Tagalog 269, 390, 396, 399 f., 409, 411, 528, 530, 541, 543, 547 f., 559, 560 f., 569, 702, 718, 721 f., 724, 726, 729, 743, 749, 760, 786, 841, 1070, 1072, 1123 f., 1203, 1218, 1473⫺1488, 1836 ⫺ Taglisch/Taglish 1473 Tahitisch/Tahitian 780, 798, 990 Tai 1027 Taino 1617 Taiwanische Gebärdensprache/ Taiwan Sign Language 1557 Tajik 520 Takelma 757, 764, 814, 926, 1101 Talysch/Talysh 541 Tamil 557, 558⫺561, 563, 703, 738, 761, 920 f., 1035, 1037, 1039, 1060, 1080, 1085, 1097, 1138, 1142, 1189 Tangkisch/Tangkic 1087 Tarafit 823 Tarahumara 1433 Tarascan 1027, 1077, 1085 Tariana 1039
1983
Sprachenregister Tauya 1191 f., 1195 Tazerwalt Shilh 1041 Telugu 1033, 1138 Temiar 539 Temne 465 Tera 553 Terena 587, 1210 Teso-Turkana-Sprachen/Teso-Turkana languages 1507, 1509 Thaisprachen/Thai languages 548, 1545 Thai 92, 251, 561, 716, 745, 758, 941, 1012 f., 1017 f., 1020 f., 1028 f., 1186 Thargari 747 Thirrari 1491 Thrakisch 1310 Tibetisch/Tibetan 96, 1085, 1593, 1731⫺1734, 1736 f. ⫺ klassisches/Classical 1732 f. ⫺ Lhasa- 1595, 1597 ⫺ zentrales/Central 1592 tibetobirmanische Sprachen/ Tibeto-Burman languages 591, 1010, 1013, 1032, 1209, 1733⫺1735, 1737 Tigre 806 Timucua 701 Tinrin 764 Tiwa 924 f. Tiwi 539, 801, 926, 1034, 1038 Tlapanec-Subtiaba 927 To’abaita 1097 Toba 1024 Tocharisch/Tocharian 1310 Tojolabal 805 Tok Pisin (Neo-Melanesian, New Guinea Pidgin) 1755⫺ 1765 Tolai 993, 1070, 1758, 1762⫺ 1764 Tondano 541 Tongisch/Tongan 717, 721, 724⫺ 726, 729, 934, 938, 1136, 1173 Tonkawa 1085 Totonakisch/Totonac 442, 541, 927, 1869 trans-neuguineische Sprachen/ Trans-New Guinean languages 1208 Trique 715 Trukese 549, 806, 938, 1160 Trumai 1855 Tsachurisch/Tsaxur 1215 Tsakwambo 1501 Tsangla 1733 tschadische Sprachen/Chadic languages 553 f., 1013, 1094, 1216 Tschechisch/Czech 133, 253, 440 f., 471, 491, 584, 956, 1245, 1643, 1905 f.
Tschetschenisch/Chechen 1039, 1058, 1062, 1203, 1227 Tschukotka-Kamtschatka-Sprachen/Chukotko-Kamtchatkan languages 541, 1032, 1263 Tschuktschisch/Chuckchee (Chuckchi) 252, 314 f., 477 f., 528, 538, 541, 560, 916⫺918, 980, 1010 f., 1162 Tschuvaschisch/Chuvash 753, 799, 1072 Tsimshian 524, 760, 926 Tsova-Tush 1034, 1039 Tuareg 540 Tübatulabal 508 Tuburi 1006 Tucano 374, 803, 1039, 1040 Tümpiso (Tümpisa) 766, 1215 tungusische Sprachen/Tungusic languages 1151, 1160, 1208 Tupı´ 701, 927 Tupi-Guaranı´-Sprachen/Tupi-Guaranı´ languages 1421 Tupinamba 742 Türksprachen/Turkic languages 314, 531, 1013, 1032 Turkana 588, 637, 751, 758, 766, 822, 1032, 1507⫺1516 Türkisch/Turkish 11, 28 f., 98, 150 f., 227⫺229, 232, 275, 285, 292 f., 314, 317, 330, 347, 351, 364, 370, 429, 465, 492, 495 f., 528, 539, 604, 619, 623⫺625, 630, 711, 715, 717, 733, 736⫺742, 753, 778, 786, 805, 825, 846, 939⫺941, 969 f., 981, 1000, 1056, 1061, 1064, 1073 f., 1077, 1079⫺ 1081, 1083- 1085, 1096, 1106 f., 1133 f., 1138, 1191, 1193⫺1195, 1197⫺1199, 1215, 1222⫺1224, 1226, 1237, 1239, 1242, 1358⫺ 1366, 1578 f., 1583⫺1585, 1649⫺1651, 1741, 1785, 1797, 1800, 1809 f., 1813, 1835, 1841 f., 1873, 1877, 1879, 1894, 1897 Tuscarora 916⫺919, 922, 925, 975 Tututni 1012 Tuyuca 1022 Twi 561, 1062, 1517⫺1526 ⫺ Akuapem 1517⫺1519, 1522 ⫺ Asante 1517 f. Tzeltal 1020, 1028, 1831 Tzotzil 1017, 1020, 1025 f., 1098, 1134 Tzutuhil (Tzutujil) 532, 760, 764, 766
U Udihe 1215 Udmurtisch/Udmurt 330, 541, 799, 1132 Ugaritisch/Ugarit 35 Uigurisch/Uigur 539 Ukrainisch/Ukrainian 520, 633, 1244 f. Ungarinjin 1034, 1039 Ungarisch/Hungarian 13, 166, 252, 285, 298 f., 330, 338, 340, 469, 477, 495, 516 f., 583, 598 f., 618 f., 621, 632, 644, 657 f., 717, 737, 746, 753 f., 771, 778, 821, 823⫺ 826, 829, 846, 939 f., 1049, 1064, 1068, 1080, 1085, 1097, 1109 f., 1132 f., 1214, 1218, 1227, 1233, 1263, 1312, 1797, 1801, 1820 Uradhi 1078 Ural-altaische Sprachen/UralAltaic languages 99, 134 uralische Sprachen/Uralic languages 330, 1032, 1077, 1083, 1085, 1121, 1208, 1328 Urdu 1138, 1142 Urhobo 589 Usan 1003, 1011 f., 1203⫺1205 Usbekisch (Özbekisch)/Uzbek 1651, 1741 uto-aztekische Sprachen/Uto-Aztecan languages 541, 917, 926, 1005, 1077, 1088, 1097 f., 1129, 1203, 1205, 1433
V Vanuatu Bislama 1755 Vedisch/Vedic 41⫺43, 1214, 1712, 1716⫺1719, 1914 ø gvedic) 1239, ⫺ Rigvedisch (R 1242, 1724 Venetian 1217 Veps 401, 1328 Vietnamesisch (Annamitisch)/ Vietnamese 30, 201, 254, 636, 708, 720, 724 f., 745, 813 f., 937, 988, 992 f., 1013, 1021, 1070, 1216, 1223⫺ 1226, 1236, 1239, 1241⫺ 1243, 1247, 1545⫺1552, 1731, 1734 f., 1897 Vogul 778 Volapük 1563⫺1571 voltaische Sprachen/Voltaic languages 541, 1042 Vot/Vote 401, 1328
1984 W Wagaya 1034 Wakaschisch/Wakashan 1032 Walisisch/Welsh 328, 478, 481 f., 532, 572, 574, 619, 660, 717, 774, 792, 823, 1071, 1218 f. Walmatjari 796 Walpiri (s. Warlpiri) Wambaya 718 Wambon 1501⫺1506 Wangkangurru 1491, 1494 f. Wangkumara 1034, 1491 Wappo 939 Warekna 1043 Wargamay 985, 1209 Wari’ 1214, 1216 Warlpiri 391, 396, 400, 469, 558 f., 561, 643, 718 f., 796, 821, 1003, 1056, 1800 Warndarang 1003 Warray 1034, 1041 Warrungu 1158 Warumungu 470 Weißrussisch/Belorussian 1245 Wemba-Wemba 643 Western-Desert-Sprachen/Western Desert languages 603, 621 Western Torres Strait 1034 f., 1041 Wik-Munkan 993 Winnebago 1123 Woarani 927 Wolof 779, 939, 987, 991⫺994, 1108
Verzeichnisse und Sprachenkarten Worora 540, 1038 Wunambal 809, 810, 1038 X Xhosa 1217 Y Yaø g Dii 985, 1006 Yagua 978 f., 1017, 1020, 1464⫺ 1472 ⫺ Cahocuma 978 Yameo 1464 Yami 561 Yana 539, 926 Yandrruwandha 1491 Yankunytjatjara 603 Yanyuwa 1038 Yao 1734 Yap 1062 Yaqui 1077 f., 1086, 1092 Yawarrawarrga 1491 Yessan-Mayo 1041 Yiddish (s. Jiddisch) Yidiny 341 f., 400, 467, 644, 796, 1021, 1135, 1158 Yimas 796, 1032, 1036, 1038, 1127, 1211 Yokuts 622, 1004 Yonggom-Wambo 1501 Yoruba 151, 539, 558, 560 f., 563, 745, 800, 934, 1094, 1182, 1185 f.
Yukaghir (s. Jukagirisch) Yukatekisch/Yucatec 247, 736 f., 739 f., 749, 809, 984⫺987, 989⫺991, 993, 1022, 1065, 1838, 1841⫺1843, 1852⫺ 1854, 1861, 1863, 1871, 1878 Yukulta 1079, 1087 Yuma-Sprachen/Yuman languages 569 f., 710, 739, 741, 773, 1022 f., 1077, 1088, 1202, 1208 Yupik (Yup’ik) 922 f., 926, 1081, 1389 ⫺ zentrales/Central 1147, 1162 ⫺ sibirisches/Siberian 537 Yurok 552 f., 770
Z Zande 1037, 1042 Zapotekisch/Zapotec 541 f. Zend 96 Zhuang 1216 Zoque 546, 548 f., 569 f., 927, 934, 1077, 1085 Zulu 539, 636, 854, 939 f., 1048, 1186, 1904 Zuni 925 f.
! !Kung 541, 797, 804 !Xu 1032
Sachregister / Index of subjects A abbreviation (s. Abkürzung) Abbreviaturbildung 1308 abessive 1085 ability, acquired 1194 ability, intrinsic 1194 Abkürzung/abbreviation 958, 1875 Ablaut/ablaut 174, 568, 620, 971, 1271 f., 1275, 1671, 1683, 1685 ⫺ Ablautausgleich 1680 ⫺ Ablautklasse 1272 ⫺ Ablauttyp 1272 ⫺ guna ablaut 568 ⫺ vrøddi ablaut 568 Ableitung (s. Derivation) absentative 1413 absolute 748, 1434, 1441 absolutive 739 f., 1081, 1514, 1727 abstraction 747 abstractness of underlying representations 500 Abundanzplural 1057 accent (s. Akzent) accidence 344 accommodation 1793 accountability principle 1665 accountability, total 183 accusative 1493 f. accusative language 1081 accusative system 1081 acquisition process 1741 acronym 956, 958, 1256, 1619, 1678 actant 1146 activa tantum 1150 active language 1082 active nominalizer 1458 actuation 308 ad hoc formation (s. Einzelprägung) adaptation 1581, 1587 ⫺ morphologische Adaptation/ morphological adaptation 1581, 1645 adessive 1080 adjacency condition 199, 411 Adjektiv/adjective 1111, 1114, 1301, 1306, 1369, 1466, 1510, 1537 ⫺ adjectivization 940
⫺ ⫺ ⫺ ⫺ ⫺ ⫺
deadjectival adjective 1709 denominal adjective 1708 deverbal adjective 1708 possessive adjective 743 relational adjective 1708 semantic type for adjectives 1112 adjustment rule 198 admonitive 1191, 1440 adstratal language 1741 Adverb/adverb 675, 1301, 1315, 1466 ⫺ adverbial suffix 1709 ⫺ mimetic adverb 1128 affective 1085 affirmation 1207, 1210 affixation 527, 750 ⫺ affix substitution 165 ⫺ Affix/affix 257, 406, 524, 1218, 1255, 1853 ⫺ Affixoid/affixoid 892, 931, 1566 ⫺ affix-stripping 1779 ⫺ derivational affix 1801 ⫺ Halbaffix 892 ⫺ phrasal affix 395 ⫺ relatives Affix 892 agent 979, 1081, 1146 Agglutination/agglutination 1271, 1615, 1655 agglutinierende Sprache/agglutinative language 254, 619, 630, 1235, 1222, 1247, 1300, 1358, 1415, 1465 Agrammatismus/agrammatism 364, 1769, 1817 agreement (s. Kongruenz) agreement class 1033 Ähnlichkeit 236 Aktionsart 1165, 1277, 1509 ⫺ dynamic Aktionsart 1168 ⫺ phasal Aktionsart 1167, 1172 ⫺ stative Aktionsart 1168 Aktor 1100 Akzent/accent 251 1319 ⫺ Akzentuierungsmuster 1244 ⫺ Akzentverschiebung 1063 alienability 1025 ⫺ alienable noun 740 ⫺ alienable possession 1748 ⫺ inalienable noun 740 ⫺ inalienable possession 1748 alignment 341 allativ/allative 1494, 1521
alliterative concord 1038 allocutive 1002 Allomorphie/Allomorphik/allomorphy 848, 874, 1268, 1289, 1311, 1655, 1876 ⫺ Allomorph, Auswahl 1270 ⫺ Allomorph/allomorph 336, 463, 473 ⫺ stem allomorphy 363 Alphabet 1564 ⫺ Alphabetschrift 1916 Alternation/alternation 336, 485, 489, 527, 1685 ⫺ automatische Alternation/automatic alternation 178, 464 ⫺ morphologisches Alternationsverhalten 1244 ⫺ non-neutralizing phonetic alternation 497 ⫺ phonetic alternation 497 ⫺ revised alternation condition 339 ⫺ weak beat alternation 1539 Ambiguität 1270 ambiguity, structural 837 analog-holistisch 888 Analogie/analogy 363, 1574, 1611, 1637, 1680, 1766 analogization 1615 analytisch/analytic/analytical 1217, 1306, 1671, 1698 ⫺ analytic structure 1223 ⫺ analytical form 524 ⫺ analytisches Ausdrucksverfahren 1268 ⫺ Analytismus/analyticity 1238, 1742 anaphoric island 199 animacy 1016, 1412 ⫺ animacy hierarchy/empathy hierarchy 734, 1068, 1762 ⫺ animate 747, 1465 ⫺ inanimate 747 Anschluß 1914 anterior 1177 anticausative 1132 antipassive 1148, 1152 antonym, gradable 1115 antonymy 1115 aorist 1176, 1715 Aphasie 1817 ⫺ amnestische Aphasie 1819 ⫺ aphasic 1782 ⫺ aphasic speech 1789
1986 Apophonie/apophony 527, 529, 568, 620, 739, 1687 ⫺ accentual apophony 530 ⫺ segmental apophony 529 ⫺ tonal apophony 530 ⫺ zero apophony 531 Appellativ 1901 applicative 1134 ⫺ benefactive applicative 1134 ⫺ comitative applicative 1135 ⫺ comprehensive applicative 1135 ⫺ directive applicative/goal applicative 1135 ⫺ instrumental applicative 1135 ⫺ locative applicative 1135 appositional 1002 appropriateness principle 1928 Äquivalenzdual 1057 Arbitraritätsprinzip 1767 archiphoneme 336, 492 argument 1104, 1130, 1434 ⫺ argument position 1434 ⫺ argument structure 845 argumentation 1863 Aristotelian classification 264 articulation, double 1232 Artikel/article 674, 717, 783, 1752 ⫺ Artikelflexion 1692 ⫺ bestimmter Artikel/definite article 791, 1315, 1700 ⫺ demonstrative article 792 ⫺ indefinite article 792, 1700, 1748 ⫺ interrogative article 793 ⫺ partitive article 1069 ⫺ possessive article 792 Aspekt/aspect 1165, 1302, 1509, 1702, 1746 ⫺ Aristotelian aspect 1172 ⫺ aspect marker 1536, 1743 ⫺ aspectual derivation 1402 ⫺ aspectual system 1414 ⫺ Aspektform 1274 ⫺ Aspektstamm 1319 ⫺ Aspektualität/aspectuality 1165, 1569 ⫺ grammatical aspect 1166 ⫺ lexical aspect 1166 ⫺ phasal aspect 1172 ⫺ situation aspect 1166 ⫺ successive aspect 1737 assimilation 1390 assistive 1138 atelic 1168 Attraktion 1913 f. attribution 1655 ⫺ attribute 1112 ⫺ attributiv/attributive 1113, 1301 Aufbauprozeß 1645 Augenblicksbildung (s. Einzelprägung)
Verzeichnisse und Sprachenkarten augment 1529 augmentative 743, 1045 f., 1049 Ausgleichsprozeß 1688 austauschbar 235 autolexical syntax 204 autonomous action/state/event 1346 autosegmental analysis 189, 554 Autosemantika 1565 auxiliary 1002, 1199, 1595 auxiliate 809 aversive 1086
B back formation/derivation (s. Rückbildung) background 1170, 1174 bahuvrihi compound 1631, 1675 Basisform/base form/basic form 336, 464, 489, 499, 604, 865, 932 ⫺ base form inflection 604 ⫺ morphological structure of the base 1046 ⫺ postbase 1468 beat addition 1539 beat deletion 1539 beat movement 1539 Bedeutung/meaning 404, 1278 ⫺ basic meaning 607 ⫺ Bildungsbedeutung 1913 ⫺ derivational meaning 524 ⫺ Gebrauchsbedeutung 1913 ⫺ general meaning 607 ⫺ Gesamtbedeutung 259, 1076, 1841 ⫺ grammatical meaning 257, 524 ⫺ Grundbedeutung 259 ⫺ lexical meaning 257 ⫺ Sachbedeutung 677 ⫺ systematische Bedeutung 1911 ⫺ verallgemeinerte abstrahierte Bedeutung 677 bedingt segmentierbar 419 Bedingung 235 Beibehaltung der Abteilungsrichtung 1608 beneficiary 1122 Benennungsmittel 1912 Benennungsmotiv 1912 Bestimmungswort 1275 Bewegung/motion/movement 736, 1469, 1557 Bewegungsverb 1557 Bezeichnungsmotiv 1912 Bildungsbedeutung 1913 bilingual dictionary 1885 bilinguales Konzept 1931 Bilinguismus/bilingualism 1647
bioprogrammatic 1764 bleaching 813, 1593 bleeding 505 blending 957, 1256, 1678 blocking 308, 653, 848, 933, 1642 blocking, morphological 347 body-part tally system 1504 borrowing (s. Entlehnung) bottom-up progression 1872 boundary 377 f. boundedness/bounding 1068, 1166, 1168 ⫺ bound 1864 ⫺ obligatorily bound 406 ⫺ bound pronoun 783 ⫺ unbounded 1469 bracket erasure convention 411 bracketing paradox 413, 841 broad phonetic transcription 1875 Broca-Aphasie 1817
C calque (s. Lehnübersetzung) case (s. Kasus) category 261, 264, 1865 category, derivational 265 causal 1086 causation 1427 causative 1418, 1449, 1485, 1718, 1728, 1733, 1742, 1762 f. ⫺ causative suffix 1370 ⫺ direct causative 1138 ⫺ immediate causative 1138 ⫺ indirect causative 1138 cause 1122 change of state 1168 change, analogical 1611 change, morphological 1574 change, sound 1611 child-etymology 1621 circumfix (s. Zirkumfix) circumfixation (s. Zirkumfigierung) circumstantial 1117 citation form (s. Zitierform) class (s. Klasse) classification of lexemes 1864 classifier (s. Klassifikator) clause linkage 1514 clause ⫺ clause-chaining 1501, 1592, 1595 ⫺ implicated clause 1497 ⫺ lest clause 1498 ⫺ sequential clause 1498 cleft-sentence 1872 climbing 390 clipping 956, 1256, 1678
1987
Sachregister clitic 253, 719, 1002, 1264, 1465 ⫺ clitic boundaries 1852 ⫺ clitic doubling 391 ⫺ cliticization 1594 ⫺ simple clitic 393 ⫺ special clitic 393 coercion 1179 cognitive domain 1870 cohesion 389 ⫺ cohesion, internal 252 coinage 835, 966, 1801 ⫺ analogical coining 165 ⫺ creative coining 1576 ⫺ expressive coinage 963 ⫺ unusual coinage 963 collection 1071 collective 737, 747, 1071, 1706, 1736, 1749, 1752 ⫺ collective count noun 1068 ⫺ collective mass noun 1068 ⫺ collective number 1071 ⫺ collective numeral 1071 combining form 895 comitative 1085, 1122 comment 1434 commercial name 966 communicative domain 1870 comparative (s. Komparation) comparative method 1661 comparison (s. Komparation) complementary 1115 completed (form of the verb) 1440 completive 1172 complexity 1796 complexity, level of 1872 component of a word form 523 component, lexical 345 component, long 189 composition (s. Komposition) compositionality 519, 851, 1286, 1288 compositionality principle 840 compound (s. Komposition) compression 300, 1927 conceptualisation/construal 259 concession 1122 concomitant 1085 concord, grammatical 1799 conditional (s. konditional) configuration (of entities) 1017 confixation 528 conflation, inclusive 1010 conjugation 631 connectionist view 1807 conservative mutation 1595 consistency (of entities) 1017 consonant (s. Konsonant) constituent, immediate 407 constituent, prosodic 413 construct state 362 construction ⫺ coverb construction 1393
⫺ experiential construction 1187 ⫺ mensurative construction 1072 ⫺ periphrastic construction 1199, 1595 ⫺ pseudopartitive construction 1072 contamination 953 content (s. Significatum) contentive 249 continuative 1725, 1734 continuous 1725 contraintes 1289 ⫺ contraintes de disponibilite´ 1289 ⫺ contraintes historiques 1290 ⫺ contraintes phonologiques 1289 ⫺ contraintes se´lectionnelles 1289 ⫺ contraintes se´mantiques 1289 ⫺ contraintes structurelles 1289 conventionalization 1577 converb 362 convergence (diachronic process) 1741 conversion (s. Konversion) coordination 1202 co-ordinator 1492 copula 718, 811, 1737 coreference 1000, 1465 corepresentable linguistic units 511 corpus 1860, 1869 cranberry morph 893 creativity, morphological 164 creativity, rule-changing 1576 creativity, rule-governed 1576 creole 1653, 1761 creolisation 1752, 1755, 1759⫺ 1761 culture-specific 1050 cumulation 812 cumulative reference condition 1068 cyclic rules 340 cyclicity 340
D data 1860 data source 1461 Dativ/dative 1092, 1494, 1862 dative, ethical 1002 deaccusative 1132 Deadjektiva 1566 declarative 1191 declension (s. Deklination) decompositional hypothesis 1778 de-compression 301 defective paradigm 1067
defectivity 602, 745 defektiv segmentierbar 419 definite 1434 definiteness 986, 1413, 1747 definition 1863 deglutination 1615 degree 1115 deixis 983, 1120, 1724 ⫺ deictic base 993 ⫺ deictic center 999 ⫺ deictic infix 993 ⫺ deictic prefix 993 ⫺ deictic reduplication 994 ⫺ deictic suffix 993 ⫺ far deixis 1751 ⫺ middle deixis 1751 ⫺ near deixis 1751 ⫺ vertical deixis 1011 Deklination/declension 631, 733, 1668 ⫺ Hauptdeklinationsklasse 1305 ⫺ Deklinationsklasse/declension class 735 f., 1316 ⫺ Indeklinabilium 1305 Dekomposition 1771 Dekompositionshypothese 1768 demonstrative 1011, 1391, 1433 f., 1701 demotivation 1615 denominative 1718 deobjective 1131 dependence, conceptual 1422 dependency 754 dependency stemmata 1879 dependent marking 1122, 1491 depersonalization 1010 deponent 1139 deprecative 971 depronominalization 1010 Derivation/Ableitung/derivation 344, 499, 595, 887, 1199, 1275, 1325, 1416, 1467, 1567, 1693, 1867, 1889 ⫺ abgeleitetes Wort 1546 ⫺ Ableitungsmorphem 1236 ⫺ adjectival derivate 1708 ⫺ affixal derivation 1258 ⫺ aspectual derivation 1402 ⫺ derivateme 524 ⫺ derivational affix 1801 ⫺ derivational category 265 ⫺ derivational meaning 524 ⫺ derivational morphology 370, 833, 1248, 1426, 1801, 1867 ⫺ derivational suppletion 515 ⫺ derivational type 266 ⫺ Derivationsformant 1546 ⫺ Derivationstiefe 1307 ⫺ derivative 932 ⫺ derived form 604 ⫺ direction of derivation 866
1988 ⫺ explizite Derivation/explicit derivation 890, 932, 1566 ⫺ implizite Ableitung/implizite Derivation/implicit derivation 893, 932, 1275 ⫺ kombinatorische Derivation 890 ⫺ Konfixderivat 895 ⫺ L-derivation 266 ⫺ lexical derivation 266 ⫺ multiple derivation 1762 ⫺ Nullableitung/zero derivation 444, 932, 1275, 1531, 1660, 1801 ⫺ prefix derivation 1531 ⫺ primary derivation 1417 ⫺ proper derivation 932 ⫺ Pseudoderivat 894 ⫺ secondary derivation 1417 ⫺ semantic derivative 931 ⫺ syntactic derivative 931 ⫺ taddhita derivate 1717 ⫺ verbal derivational suffix 1459 description 1863, 1869 ⫺ descriptive concept 1862 ⫺ descriptive device 1875 ⫺ descriptive grammar 1860 desiderative 1718 Determination ⫺ Determinans-DeterminatumStruktur 887 ⫺ determinant 931 ⫺ Determinativkompositum 1275 ⫺ determinatum 931 diachronic analysis 1591 diagrammaticity 867 dialektale Gliederung 1267 Diathese/diathesis 1140, 1146, 1318 dictionary 1874 ⫺ dictionary entry 1886 ⫺ learners’ dictionary 1884 ⫺ monolingual dictionary 1882 Diffusion 1647 diglossia 1741 diminutive 743, 1045, 1048, 1173, 1413, 1586, 1737 diminutive tantum 1046 direct discourse 1001 directional marker 1438 directionality, principle of 866 discontinuity, morphological 546 discourse role 988 disguise 1010 disharmonic 1009 Distanz (des Sprechers vom Sachverhalt) 1274 Distanzkompositum 1277 Distanzstellung 889 distinctive unit 493 Distribution 676
Verzeichnisse und Sprachenkarten distribution, non-contrastive 185 distributive reference condition 1068 ditransitive 1434 documentation of the language 1869 domain, abstract 260 domain, concrete 260 Dual/dual 1056, 1465, 1743, 1749 ⫺ Äquivalenzdual 1057 ⫺ arbiträrer Dual 1057 ⫺ Dualia Tantum 1061 ⫺ Oppositionsdual 1057 ⫺ Paritätsdual 1056 ⫺ pseudo-dual 1749 durative 1168 f., 1171, 1734 durativity 1746 Durchsichtigkeit 1774 dyadic kin term 1493 Dyslexie 1769
E eclecticism 1858 economy (s. Sprachökonomie) Eigenname 1901 Einzelprägung/Augenblicksbildung/ad hoc formation/ hapax legomenon 1627 elative 1213 ellipsis 962 empathy hierarchy (s. animacy hierarchy) endoclitic 392, 546 Endungsreihe 1244 Enklise 1271 ⫺ enclitic 392, 1002, 1396, 1456, 1465 ⫺ enclitic presyllable 1734 ⫺ locative enclitic 1536 Entfaltung 1912 entity 733, 973 ⫺ first-order entity 974 ⫺ fourth-order entity 974 ⫺ second-order entity 974 ⫺ third-order entity 974 Entlehnung/borrowing 1586, 1644, 1646, 1656 entrapment 549, 937 Entwicklungsdysphasie 1820 environment 614 equative form 1213 equativity 1218 ergative 1081 ergative language 1081, 1493 f. ergative morphology 1391 ergative system 813, 1081 Erleichterungsrückbildung 894 error 1789, 1796 etabliert 1091 Etymologie/etymology
⫺ e´tymologie-histoire du mot 1911 ⫺ e´tymologie-origine 1911 ⫺ folk etymology 1587, 1615 ⫺ popular etymology 1615 evaluation, domain of 1193 evaluation, target of 1192 evaluative 1048 evaluative morphology 361 event 1165, 1168 event point 1182 evitative 1086 example 1880 Expansion/expansion 301, 887, 931 expletive 970 explicative 1048 exponence ⫺ cumulative exponence 619, 1074 ⫺ exponent 616 ⫺ extended exponence 622 ⫺ fused exponence 621 ⫺ main exponent 623 ⫺ multiple exponence 590, 622, 812 ⫺ separate exponence 618 expression type 370 expressive process 524 ⫺ grammatical expressive process 524 ⫺ morphological expressive process 524 ⫺ non-grammatical expressive process 524 ⫺ non-morphological expressive process 524 extramorphological characteristic 633 F feeding rules relationship 504 feminin/feminine 1305, 1748 field structure 611 finiteness 1334 flectional language (s. fusionierende Sprache) Flexion 676 ⫺ contextual inflection 365, 1806, 1809 ⫺ edge inflection 395 ⫺ first inflection 1796 ⫺ Flexionsabbau 1270 ⫺ Flexionsklasse/inflection class 200, 363, 597, 630, 1269, 1877 ⫺ Flexionsmorphem/inflectional morpheme 370, 1423, 1236 ⫺ Flexionsmorphologie/inflectional morphology 1248, 1600, 1819
1989
Sachregister ⫺ Flexionsprinzip 1685 ⫺ Flexionstyp/inflectional type 647, 1269 ⫺ general inflectional rules 363 ⫺ inflection 174, 344, 595, 1198, 1467, 1867 ⫺ inflection class (s. Flexionsklasse) ⫺ inflectional generality 656 ⫺ inflectional meaning 524 ⫺ inherent inflection 365, 1806 ⫺ productive inflectional rule 363 ⫺ pronominale Flexion 1095, 1270 ⫺ schwache Flexionsklasse 1269 f. ⫺ starke Flexionsklasse 1269 f. ⫺ stem inflection 604 ⫺ Trennbarkeit von Kasusund Numerusflexion 1268 ⫺ verbal inflection 1671 Flexivhomonymie 1305 flexivische Profilierung 1684 fluctuation 486 focal point 653 focus 1434 foreground 1170, 1174 form (s. Significans) formal segmentierbar 419 formation ⫺ ad hoc formation (s. Einzelprägung) ⫺ dynamic formation 1483 ⫺ established formation 1627 ⫺ nonce formation 835 ⫺ playful formation 838 formative element 172 Formenlehre 1683 frame (s. semantic domain) frame-external 1117 frame-internal 1117 free form, minimum 252, 379 free, potentially 406 frei 240 Fremdwort/foreign loan 1268, 1750 frequency 298, 1252, 1636, 1807 Fugenelement 447, 891, 1276, 1683 full-listing hypothesis 1778 function (s. Funktion) function word 248, 258, 1474, 1538 function, metasemantic 1049 functional constraint 1613 functional domain 1873 functor 249 Funktion/function 3, 1017, 1122, 1467 funktionaler head 681 Fusion 25 fusionierende Sprache/fusional (inflecting, flectional) lan-
guage 254, 619, 630, 1222, 1235, 1247 Futur/future 1184, 1274, 1303, 1521, 1703, 1746 ⫺ future perfect 1187 ⫺ future stem 1725 ⫺ futurum exactum 1703 ⫺ imperfektives Futur 1303 ⫺ perfektives Futur 1303 G gap, accidental 834 gating 1782 Gebärde 1556 Gebärdenraum 1556 Gebrauchsbedeutung 1913 Gebrauchsgleichheit 1914 gemischte Flexionsklasse 1269 gender (s. Genus) general drift 1741 generality 259 generalization 1577 genitive 744, 1514, 1748 ⫺ genitive classifier 1021 ⫺ genitively marked object 1069 ⫺ objective genitive 1078 ⫺ subjective genitive 1078 genus verbi 1304 ⫺ aktivisch 1304 ⫺ (s. voice) Genus/gender 735, 1007, 1031, 1269, 1304, 1316, 1406, 1412, 1509, 1568, 1698, 1713, 1743 ⫺ controller gender 1033 ⫺ double gender 1033 ⫺ gender assignment 1070 ⫺ gender system 1019 ⫺ Genusprofilierung 1692 ⫺ multiple gender 1033 ⫺ semantic gender resolution 1039 ⫺ syntactic gender resolution 1039 ⫺ target gender 1033 gerund 1373, 1482 Gesamtbedeutung 259, 1076, 1841 glossary 1875 goal 1120, 1121, 1135 government/rection 365 gradation 1212, 1424 grammar (s. Grammatik) grammatical category label 1842 grammatical content 373 grammatical feature 1877 grammatical function-dependent 346 grammatical item 1867 grammatical meaning 257, 524 grammatical process 523
grammatical relation 1131 grammatical rule 297 grammaticalization (s. Grammatikalisierung) grammatically corepresentable 512 grammaticography 1857 Grammatik/grammar 1857, 1863 ⫺ categorial grammar 201 ⫺ comparative grammar 1859 ⫺ contrastive grammar 1859 ⫺ general comparative grammar 1859 ⫺ generative Grammatik 1267 ⫺ grammar and dictionary 1870, 1874 ⫺ grammar checker 1897 ⫺ grammar-lexicon interface 1865 ⫺ historical grammar 1859 ⫺ historical-comparative grammar 1859 ⫺ lexical grammar 199 ⫺ lexical-functional grammar 199 ⫺ onomasiological grammar 1873 ⫺ prescriptive grammar 1860 ⫺ reference grammar (s. Sprachlehre) ⫺ semasiological grammar 1872 ⫺ structure of grammar 1869 ⫺ subdivision of grammar 1873 ⫺ theory of grammar 1858 Grammatikalisierung/grammaticalization 12, 257, 262, 300, 930, 1232, 1250, 1590, 1736 ⫺ grammaticalization of zero 1178 ⫺ Grammatikalisierungsprodukt 1273 grammatische Wechsel 1685 grammeme 524, 616 Graphem 1564 grounding 261 Grundbedeutung 259 Grundform 1268 Grundmorphem 1565 Grundwort 1275 Gruppenflexion 1692
H habitual past 1458 habitualis/habitual 1520, 1725 habituality 1746 Halbaffix 892 Handlungsnumerus 1058 hapax legomenon (s. Einzelprägung) Haplologie/haplology 583, 893 ⫺ haplologic word 953
1990 ⫺ morphological haplology 392 harmonic 1009 Hauptdeklinationsklasse 1305 Hauptmarker 1609 head 201 ⫺ head class 1033 ⫺ head movement 368, 398 ⫺ head operation 201, 861 ⫺ righthand head rule 201 ⫺ morphological head 165 ⫺ head-marking 1122 Herkunftsgleichheit 1913 hesternal past 1187 heteroclisis 738 hierarchical asymmetry 1010 hierarchy of grammatical levels 1872 Homonymie/homonymy 250, 618, 944, 1841, 1914 honorific 993, 1724 honorific formative 993 hortative 1191 host 388 human 747 Humboldtian principle 1575 hybrid 1033 Hybridisierung 887 hypercharacterization 1581, 1587 hyphenation, automatic 1897 hypostatization 835, 974, 1628, 1640 hypotaxis 1202
I iambic reversal 1539 iconicity (s. Ikonizität) Identitätsoperation 437 ideophone 570, 719, 964, 1128, 1538 idiom 249 idiom suppletion 515 idiomatization 837, 930, 1626 Ikonismus, konstruktioneller 443 Ikonizität/iconicity 618, 867, 1559, 1757 illocution 1190 ⫺ basic illocution 1190 ⫺ illocutionary force 1050 ⫺ illocutionary modification 1192 Imperativ/imperative 1191, 1209, 1274, 1372, 1521, 1569, 1703, 1716, 1726 imperfect 1176, 1715, 1746 ⫺ imperfect indicative 1702 ⫺ imperfect subjunctive 1703 imperfektiv/imperfective 1166, 1302, 1481, 1725 ⫺ derived imperfective 1176 ⫺ imperfektives Futur 1303
Verzeichnisse und Sprachenkarten imprecative 1191 inactive 1425 inanimacy 1412 inanimate 747 inceptive 1172 inchoative 1172, 1371, 1737 incorporation (s. Inkorporation) indefinite phrase 1434 indefiniteness 1742, 1747 Indeklinabilium 1305 Index 1237 index of grammatical items 1875 indexical 998 indikativ/indicative 1191, 1274, 1303, 1569, 1702 ⫺ imperfect indicative 1702 ⫺ pluperfect indicative 1702 ⫺ present indicative 1702 indirect discourse 1001 individual (entity type) 1104 individual interpretation 1071 Individuativum 888 infinitive 1372, 1458, 1703, 1727 infixation 528, 750, 934, 1045, 1330 ⫺ deictic infix 993 ⫺ infix 545, 1853 ⫺ infixal reduplication 545 inflection (s. Flexion) information retrieval 1897 ingressive 1172 Inhalt (s. Significatum) Inhärenz 1091 inheritance 166 initialism 957 initialization 1615 Inkorporation/incorporation 284, 527, 889, 916, 1118, 1428, 1451 ⫺ modifying incorporation 1451 ⫺ noun incorporation 916 ⫺ saturating incorporation 1451 innate categories and principles 1250 instantiation 261 institutionalization 837, 1233, 1626 instruction 1879 instrument 979, 1121 instrument, negative 1122 inte´grateur 1288 intensive 1513, 1718 interclausal relation 1202 Interferenz/interference 1644 Interfix 891, 934 interfixation 528 Interjektion/interjection 1316, 1492 interlinear morphemic gloss 1834 Interlinguistik 1563
International Auxiliary Language Association 1563 interrogative 1191 intersegmental 428 intonation 1198 intransitive 1413, 1434 intraparadigmatischer Ausgleich 1690 intrasegmental 428 invariance 618 Invariante 234, 238 inverse form of the verb 1010, 1151, 1414 Inversion 1608 irrealis 1440 irreality 1747 isolating language 254, 1222, 1247 Isomorphie 1694 item and arrangement 26, 409, 627 item and process 26, 409, 627 item familiar 834 item familiarity 1627 Iteration 1550 iterative 1169
J jargon 1653, 1757 Jargonaphasie 1819 juncture 341 junggrammatisch 1680 Juxtaposition 1093, 1202, 1214 Juxtapositum 1308
K Kardinalzahl 1315 Kasus/case 1011, 1073, 1304, 1331, 1568, 1595, 1699, 1713, 1747, 1862 ⫺ abessive 1085 ⫺ ablative 1494, 1514 ⫺ accusative 1493 f. ⫺ case form 1074 ⫺ case language 1074 ⫺ case marker 1074 ⫺ case marking 737 ⫺ case morphology 1812 ⫺ case relation 1074 ⫺ Dativ/dative 1092, 1494, 1862 ⫺ direct case 1076, 1080, 1723 ⫺ ergative 1081 ⫺ genitive 744, 1514, 1748 ⫺ independent case 604 ⫺ indirect case 1080 ⫺ instrumental 1085, 1514 ⫺ Kasusnivellierung 1691 ⫺ Kasussynkretismus 1766
1991
Sachregister ⫺ ⫺ ⫺ ⫺
Kasussystem 1316 Lokalkasus 1092 nominative 1493 f., 1514 oblique case 738, 1076, 1080, 1086, 1723 ⫺ obviative 1005 ⫺ primary case 604 ⫺ secondary derivative case 604 ⫺ Trennbarkeit von Kasusund Numerusflexion 1268 ⫺ vocative 1086, 1444 Kayne’s Generalization 398 Klammerform 894 Klasse/class ⫺ class-free suffix 1456 ⫺ closed class 249, 1112 ⫺ complementary class 634 ⫺ default class 635 ⫺ Deklinationsklasse/declension class 735 f., 1316 ⫺ distributional class 733 ⫺ form class 677 ⫺ Klassenwechsel 1690 ⫺ macroclass 632, 634 ⫺ morphological class 1507 ⫺ open class 249, 1112 ⫺ unmarked class 635 Klassifikator/classifier 735, 1070, 1094, 1467 ⫺ general classifier 1017 ⫺ genitive classifier 1021 ⫺ incorporated classifier 1022 ⫺ mensural classifier 1020, 1070 ⫺ noun classifier 1021 ⫺ numeral classifier 1019 ⫺ possessive classifier 1862 ⫺ sortal classifier 1020, 1070 ⫺ specific classifier 1017 ⫺ typology of classifiers 1018 ⫺ unique classifier 1017 ⫺ verbal classifier 1022 klassifikatorischer Schlüssel 1243 Klassifizierung 1271 Kognitionswissenschaft 1766 koine 1741 Kombinationsfähigkeit 677 Komparation/comparison 1212, 1306, 1569 ⫺ absolute comparative 1213 ⫺ comparative construction 1212 ⫺ comparee/topic of comparison 1212 f. ⫺ comparison of equality 1213 ⫺ comparison of majority 1213 ⫺ comparison of minority 1213 ⫺ komparativ/comparative 1213, 1257, 1306 f., 1717 ⫺ marker of comparison (s. comparison, pivot of) ⫺ pivot of comparison/marker of comparison, 1212
⫺ standard of comparison 1212 f. Kompatibilität 1565 Kompositabildung 1307 f. Komposition/composition/compounding 344, 370, 524, 527, 833, 1262, 1275, 1325, 1416, 1428, 1511, 1567, 1675 f., 1693, 1759, 1801, 1852 ⫺ bahuvrihi compound 1631, 1675 ⫺ composite word/complex word 405 ⫺ compound lexeme 1627 ⫺ compound phrase 889 ⫺ denominales Kompositum 1566 f. ⫺ deverbales Kompositum 1566 ⫺ Distanzkompositum 1277 ⫺ eigentliches Kompositum 1683 ⫺ endozentrisches Kompositum 444 ⫺ exozentrisches Kompositum/ exocentric compound 444, 1276, 1631 ⫺ fossilized compound 1627 ⫺ imperative compound 1631, 1677 ⫺ Kompositum/compound 888, 1308, 1677, 1704, 1717, 1728, 1813, 1890, 1926 ⫺ Konfixkompositum 895 ⫺ Kopulativkompositum 887, 1276 ⫺ phrasal compound 166, 1657 ⫺ Possessivkompositum 1276 ⫺ präpositionales Kompositum 889 ⫺ Pseudokompositum 890 ⫺ stub compound 1256 ⫺ syntactic compound 1627 ⫺ synthetic compound 166, 891 ⫺ verbales Pseudokompositum 1277 ⫺ Verbalkompositum 1277 ⫺ Verb-Ergänzung-Kompositum 890 Kompositionalität, synchrone 1273 konditional/conditional 1204, 1458, 1569, 1703, 1715 Konditionierung 239 ⫺ distributionelle phonologische Konditionierung 463 ⫺ formale phonologische Konditionierung 463 ⫺ grammatische Konditionierung 474 ⫺ lexikalische Konditionierung 474 ⫺ morphologische Konditionierung 474
⫺ phonologische Konditionierung 463 ⫺ primäre phonologische Konditionierung 465 ⫺ reziproke Konditionierung 475 ⫺ sekundäre phonologische Konditionierung 465 ⫺ semantische Konditionierung 477 ⫺ syntaktische Konditionierung 480 Konfix 893, 895, 1276 Konfixderivat 895 Konfixkompositum 895 Kongruenz/agreement 365, 1002, 1270, 1514, 1725, 1799, 1811 Konjugation 1569 Konjunktion 675 Konjunktiv 1274, 1303, 1569 Konnektionismus 1766 Konnektor 1093 konsekutiv 1521 Konsonant/consonant ⫺ appendix consonant 384 ⫺ consonant gradation 568, 1337 ⫺ consonant mutation (s. konsonantische Veränderung) ⫺ konsonantische Veränderung/ consonant mutation 568, 1272 Kontamination 1819 Konversion/conversion 27, 165, 527, 932, 944, 1275, 1660, 1709, 1759, 1927 ⫺ categorical conversion 530 ⫺ paradigmatic conversion 530 ⫺ principe d’orientation des conversions 871 ⫺ rectional conversion 530 ⫺ zero conversion 531 Kopiesprache (replica language) 1646 Kopulativkompositum 887, 1276 Körperteil 1557 Korrelationskoeffizient 1242 Kreolsprache 1646 Kunstsprache, internationale 1563 Kürzungssuffigierung 894 Kurzwort/truncation 529, 583, 958, 1286, 1289, 1295, 1390, 1537
L language language language language language
death 1653, 1655 external 1043 internal 1042 island 1741 processing 1251
1992 language system 1870 language type 1224 language use 253 langue analogue 1222 langue transpositive 1222 lapse (s. word, portmanteau) Lautsymbolik 418 Lautverschiebung, althochdeutsche 1681 Lautverschiebung, zweite 1681 Lautwechsel 1685 leaner 391 lect-marking 307 Lehnübersetzung/loan translation/calque 833, 1645, 1649 Lehnwortbildung 895 lemmatization 1896 Lernen, deduktives 1933 Lernen, induktives 1932 letter try 1894 level 189 level ordering 165 leveling 1587 ⫺ paradigm leveling 1581 ⫺ analogical levelling 1613 lexeme 152, 248, 596, 868, 1106 ⫺ lexeme-class 596 ⫺ lexemic domain hypothesis 197 ⫺ phrasal lexeme 249 lexical category, major 196 lexical content 373 lexical coverage 1897 lexical decision task 1781 lexical enrichment 306 lexical entries, fully-specified 346 lexical entries, impoverished 346 lexical expansion 1656 lexical expression 370 lexical formative 345 lexical generality 1251 lexical hypothesis, generalized 199 lexical hypothesis, strong 199 lexical hypothesis, weak 199 lexical insertion 196 lexical integrity 197, 397 lexical item 250, 596 lexical meaning 257 lexical representation 346 lexical retrieval 1807 lexical rule (s. lexikalische Regel) lexical split 364 lexical unit 250, 596 lexicalism, strong 197, 346, 367 lexicalism, weak 197, 367 lexicalist hypothesis 165, 188, 195 lexicalization 300, 346, 524, 834, 837, 1577, 1597, 1625 f., 1641 lexicalization, synchronic 933 lexicography 1857, 1882
Verzeichnisse und Sprachenkarten lexicon, mental 364 lexicon, permanent 345 lexicon, potential 345 lexifier language 1757 f. lexikalische Entscheidung 1770 lexikalische Regel/lexical rule 201, 1767 liaison 393 lingua franca 1752, 1758 linguistic type 1873 linguistic union 511, 524 link, empty 183 linkage, asyndetic 1202 linkage, tail-head 1506 list fallacy 297 listeme 249, 596 loan translation (s. Lehnübersetzung) loan word 833, 1528 loan, foreign (s. Fremdwort) local 1081, 1083 localistic 1076 location 1413, 1469 locative 1433, 1445, 1494, 1514 logographisch 1916 logophoric 1005 logophoric, second-level 1006 Lokalkasus 1092
M macroclass 632, 634 macroparadigm 632, 634 malapropism 1615, 1790 manner 1121 mark 492, 608 ⫺ Hauptmarker 1609 ⫺ marker paradigm 1878 ⫺ Marker/marker 10, 616 ⫺ marking, double 1491 masdar 1750 masking 1010 maskulin 1305 material (semantic feature) 1016 materielle Gleichheit 1914 meaning (s. Bedeutung) means 1121 media tantum 1150 megamorph 511 megamorph suppletion 515 Meillet principle 1664 mensurative 747 metagrammaticography 1858 f. meta-lexicography 1857 metanalysis 1615 metaphony 567, 751 metaphor, grammatical 977 metaphorical extension 1596 Metathese/metathesis 467, 549, 576, 936, 1390, 1465 ⫺ contact metathesis 577 ⫺ morph metathesis 397
⫺ morphological metathesis 576 ⫺ morphophonological metathesis 577 methodological reflection 1869 metonymy 1631 Minimalzeichen/minimal sign 406, 417, 511 mirror principle 407 Mischsprachen/mixed languages 1646 mitigation 1192 mixed languages (s. Mischsprachen) mixed processing 1778 modality 1190 ⫺ deontic modality 1193 ⫺ doxastic modality 1196 ⫺ dubitative modality 1196 ⫺ epistemic modality 1193 ⫺ event-oriented modality 1193 ⫺ evidential modality 1193, 1197, 1414 ⫺ facultative modality 1193 ⫺ hypothetical modality 1196 ⫺ inferential modality 814, 1197 ⫺ modal distinction 1087, 1414, 1743 ⫺ modal function ⫺ modal suffix 1459 ⫺ reportative modality 1197 Modalverb 1274 model, process 572 Modellsprache/source language/ model language 1646 Modifikation/modification 527, 888, 938, 1451 ⫺ internal modification 751, 932, 935 ⫺ modifikatorische morphologische Regel 1607 ⫺ modifying incorporation 1451 ⫺ nominal modifier 1511 module 1790 Modus/mood 1190, 1198, 1303, 1318, 1509, 1673, 1703 ⫺ superordinate clause mood 1392 ⫺ subordinate clause mood 1392 Monoflexion, kooperierende 1270 monogenetic origin 1741 monolinguales Konzept 1930 monosystemic 494 Morph/morph 1225 ⫺ cranberry morph 893 ⫺ diskontinuierliches Morph 1270, 1275 ⫺ empty morph 183, 626 ⫺ leeres Morph 447
1993
Sachregister ⫺ morph metathesis 397 ⫺ Quasimorph 419 ⫺ replacive morph 568, 573 ⫺ Submorph 417 ⫺ zero morph 1852 Morphem/morpheme 22, 1225, 1564, 1852, 1864 ⫺ autonomous morpheme 258 ⫺ bound morpheme 1864 ⫺ dependent morpheme 257 ⫺ Flexionsmorphem/inflectional morpheme 370, 1423, 1236 ⫺ free grammatical morpheme 371 ⫺ freies Morphem 1062 ⫺ grammatiches Morphem 1302 ⫺ morpheme boundary 1852 ⫺ morpheme structure constraint 429 ⫺ morpheme structure rule 428 ⫺ morpheme-based 197 ⫺ morpheme-oriented morphology 1864 ⫺ Nullmorphem/zero morpheme 438, 1268, 1675 ⫺ obligatorily bound morpheme 406 ⫺ Quasimorphem 893 ⫺ submorphemic 1045 ⫺ unbounded morpheme 1469 morphemically irrelevant material 183 morphemicon 1867, 1874 morpholexical alternant 176 morpholexical variation 337 morphological asymmetry 842 morphological description 1863 morphological domain 307 morphological expression 370 morphological means 523 morphological model, dynamic 489, 499 morphological model, static 490 morphological network 164 morphological operation (s. morphologische Operation) morphological pattern 1865 morphological process, zero 531 morphological realization 200 morphological rule, productive 298 morphological spelling operation 266 morphological structure 1864 Morphologie/morphology 4 ⫺ autosegmental morphology 573 ⫺ derivational morphology 370, 833, 1248, 1426, 1801, 1867 ⫺ finite-state morphology 1894
⫺ inflectional morphology (s. Flexionsmorphologie) ⫺ level-ordered morphology 367 ⫺ lexeme-based morphology 197, 868 ⫺ lexical morphology 165, 411, 833 ⫺ model of morphology 1886 ⫺ morphology-syntax interface 1594, 1865, 1871 ⫺ nominal morphology 1719 ⫺ paradigmatic morphology 164 ⫺ position-class morphology 412 ⫺ prosodic morphology 342 ⫺ split morphology 364, 366 morphologisch determiniert 1300 morphologische Operation/morphological operation 25, 1255 morphologische Regel 1605 ⫺ morphologische Regel, additive 1607 morphologisches Prinzip 1686 Morphologisierung/morphologization 13, 936, 1582, 1584, 1588, 1594, 1600, 1685 ⫺ morphologisch initiierte Morphologisierung 1605 ⫺ Parameter der Morphologisierung 1604 ⫺ phonologisch initiierte Morphologisierung 1605 morphology (s. Morphologie) Morphophonemik/morphophonemics/morphophonology/morphonology 335, 337, 470, 485 ⫺ morphonologische Anpassung 893 ⫺ morphonologische Anpassungsregel 891 ⫺ morphophoneme 490, 1225 ⫺ morphophonemic level 490 ⫺ morphophonemic representation 490, 1875 ⫺ morphophonemic rule 497 ⫺ morphophonology, generative 496 morphosyntax 204 ⫺ morphosyntactic category 266 ⫺ morphosyntactic feature 266 ⫺ morphosyntactic property 266, 616 ⫺ morphosyntactic representation 346 motion (s. Bewegung) motivation 1122, 1633 movement, bounded 1469 Movierung 444
MS-operation 266 multifunctionality 1660 multisegmental 894 Mutation/mutation 888, 1670
N nahe Zukunft 1521 naturalness 1253 Nebenmarker 1609 negation 1207, 1374 ⫺ constituent negation 1210 ⫺ narrow scope negation 1210 ⫺ sentential negation 1210 ⫺ wide scope negation 1210 neologism/neology 835, 1861 Neologismus, morphologischer 1819 neue Wurzel 1275 neutral 1305 neutralisation 485, 493, 497, 1210 ⫺ neutralization, absolute 189 Nomen/noun 674, 733, 746, 1114, 1257, 1391, 1433, 1466, 1528, 1668 ⫺ abstract noun 747 ⫺ action noun 1708, 1721, 1727 ⫺ agent noun 1707, 1721, 1738 ⫺ collective count noun 1068 ⫺ collective mass noun 1068 ⫺ common noun 746 ⫺ compound noun 1450, 1531 ⫺ concrete noun 747 ⫺ corporate noun 1071 ⫺ count noun 747, 1067 ⫺ deadjectival abstract noun 1708 ⫺ deverbal noun 1536, 1707 ⫺ expressive noun 1128 ⫺ inalienable noun 740 ⫺ instrumental noun 1707 ⫺ locative noun 1707 ⫺ mass noun 747, 1067 ⫺ Nomina-agentis Bildung 1684 ⫺ noun of multitude 737 ⫺ noun stripping 920 ⫺ nouniness squish 976 ⫺ personal noun 1707 ⫺ possessed noun 1434 ⫺ proper noun 746 ⫺ relationales Nomen/relational noun 1098, 1595 ⫺ relator noun 1077 ⫺ short noun 1441 ⫺ unit noun (s. Singulativum) ⫺ verbal noun 748, 1750 nominal 1114, 1329, 1412, 1492 ⫺ derived nominal 196, 1716 ⫺ gerundive nominal 196 ⫺ location nominal 1494
1994 ⫺ nominal modifier 1511 ⫺ nominal system 1713 nominalization 747, 939, 974, 1260 ⫺ nominalizer 1737 ⫺ future nominalizer 1458 ⫺ stative nominalizer 1458 ⫺ generalized nominalizing markers 978 Nominalklasse 477 nominative 1493 f., 1514 non-continuous 1725 non-human 747 non-pro-drop 1002 non-realis 1481 non-sensory evidence 1197 non-suppletive allomorphy 336 norm 834 noun (s. Nomen) noun class 735, 1007, 1031, 1367 ⫺ noun class system 1019 ⫺ noun classification 1528 Nullableitung/zero derivation 444, 932, 1275, 1531, 1660, 1801 Nullallomorph/zero allomorph 183, 438 Nullausdruck 27 Nullmorphem/zero morpheme 438, 1268, 1675 Nulloperation 437 Nullsuffix 894 Nullzeichen/zero sign 438, 531 Numeral (s. Zahl) Numerus/number 1304, 1316, 1331, 1371, 1508, 1568, 1673, 1699, 1702, 1713, 1747 ⫺ Trennbarkeit von Kasusund Numerusflexion 1268 ⫺ distributive number 1071, 1734 ⫺ mass number 1069 ⫺ Numerusablaut 1680 ⫺ Numerusprofilierung 1691 ⫺ Handlungsnumerus 1058 ⫺ collective number 1071 nunation 1747
Verzeichnisse und Sprachenkarten obligatory elements in the sentence 1118 oblique form 604 obscuration 1615, 1627 obsolescence 1655 obstruent, initial 1732 obviation 1413 obviative 1005 omission 1010 omnipredicative language 1433 one-form-one-meaning principle 840 onomasiology 1870 ⫺ onomasiologische Basis 887 ⫺ onomasiologisches Merkmal 888 onomatopoetic formation 965 opaque 1216 opaqueness 1615 operateur de construction de sens 1290 Opposition 234 ⫺ bilateral opposition 492 ⫺ constant opposition 492 ⫺ exclusive opposition 493 ⫺ phonological opposition 492 Oppositionsdual 1057 optativ/optative 1191, 1440, 1458, 1521, 1716 optimality theory 1613 optional elements in the sentence 1118 order ⫺ order of morphemes 1249 ⫺ conjunct order 1416 ⫺ independent order 1416 ⫺ linear order 1010 ⫺ scope order 1880 ⫺ sequential order 1880 ⫺ template order 1880 Ordinalzahl 1315 ordinary 1724 orientation 1481 orthographic representation 1875 overgeneralization 1807 overt analogue criterion 363, 443, 871
O P object 1009, 1405 ⫺ genitively marked object 1069 ⫺ indirect pronominal object 1747 ⫺ morphological object 596 ⫺ object marking 1508 ⫺ object pronoun 1533 objective conceptualisation 1346 obligatory contour principle 1613
Paradigma/paradigm 27, 171, 596, 630, 808, 1867, 1878 ⫺ defective paradigm 1067 ⫺ dominant paradigm structure condition 635 ⫺ implicative paradigm structure condition 635 ⫺ inflectional paradigm 647 ⫺ inte´grateur paradigmatique 1289, 1294
⫺ intraparadigmatischer Ausgleich 1690 ⫺ macroparadigm 632, 634 ⫺ marker paradigm 1878 ⫺ paradigm cell 596 ⫺ paradigm economy 635 ⫺ paradigm leveling 1581 ⫺ paradigm symmetry 656 ⫺ paradigmatic conversion 530 ⫺ paradigmatic deviation 618 ⫺ paradigmatic gap 363 ⫺ paradigmatic morphology 164 ⫺ paradigmatic relation 494 ⫺ paradigmatic uniformity 1613 ⫺ potential paradigm 808 ⫺ subparadigm/slab/screeve 597, 633 ⫺ specimen paradigm 1878 Paragrammatismus 1817 f. parameter 1212 f. parameter marker 1213 Paraphasie 1819 paraphrase 1001 parasitic formation 526 parasitic word 605 Parasynthese/parasynthesis 890, 1706 parataxis 1202 Paritätsdual 1056 parole 834 parser 1893 part of speech 938, 1233, 1864 participant 1146 ⫺ participant role 998 ⫺ participant-oriented modality 1192 participle (s. Partizip) particle (s. Partikel) particularism 1858, 1862 Partikel/particle 258, 1198, 1316, 1330, 1412, 1433, 1492 partitive 1048, 1085 Partizip/participle 674, 1362, 1373, 1441, 1569, 1703, 1747 ⫺ attributive participle 1728 ⫺ past participle 1671 ⫺ present participle 1672 passe´ (s. past) Passiv/passive 1414, 1448, 1703, 1727, 1742 ⫺ bekommen-Passiv 1273 ⫺ eigentliches Passiv 1304 ⫺ internal passive 1746 ⫺ passivische Verwendung der Verben 1304 past 1182 ⫺ habitual past 1458 ⫺ hesternal past 1187 ⫺ passe´ ante´rieur 1703 ⫺ passe´ compose´ 1702 ⫺ passe´ simple 1702
1995
Sachregister ⫺ passe´ surcompose´ 1703 ⫺ recent past/hodiernal past 1186 f. ⫺ remote past 1187 path 1120 f. patient 979, 1081, 1146 pattern 173 Paukal 1057 pejorative 1413 perception, multiple 1346 perception, single 1346 percolation 860 percolation, feature 200 perfect (s. Perfekt) perfective (s. perfektiv) perfectivity 1166, 1168, 1746 Perfekt/perfect 1177, 1185, 1274, 1414, 1521, 1702, 1715, 1732 ⫺ doppelte Perfektform 1273 ⫺ future perfect 1187 ⫺ intransitive de-resultative perfect 1186 ⫺ perfect stem 1715 ⫺ perfect subjunctive 1703 ⫺ possessive de-resultative perfect 1186 perfektiv/perfective 1166, 1302, 1481, 1568, 1725, 1734, 1736, 1747 ⫺ perfektives Futur 1303 peripherality constraint 478 Periphrase/periphrasis 750, 1684 ⫺ periphrastic form 600, 808 ⫺ periphrastische Konstruktion, grammatikalisierte 1273 permissive 1138 permutation 529 person 998, 1304, 1317, 1508, 1673, 1702 ⫺ active personal prefix 1425 ⫺ consciousness model person 998 ⫺ first plural exclusive person 986, 1457, 1465 ⫺ first plural inclusive person 986, 1457, 1465 ⫺ fourth person 1005 ⫺ Personalpronomen/personal pronoun 716, 784 f., 1270, 1369, 1412, 1701 ⫺ second person pronoun 986 ⫺ third person pronoun 986 perspective 999, 1174 phase 1172 Phonem/phoneme 22, 1225 phonematic unit 149, 495 phonemic representation 1875 phonestheme 953 phonetic representation 1875 phonetic rule 497 Phonologie/phonology ⫺ automatic phonological rule 336
⫺ autosegmental phonology 496 ⫺ generative phonology 499 ⫺ lexical phonology 346, 411 ⫺ natural generative phonology 496 ⫺ phonological restriction 1047 ⫺ phonologisch determiniert 1300 ⫺ phonologische Regel/phonological rule 428, 499, 1605 ⫺ prosodic phonology 494 phonosymbolism 963 phonotactics 427 phrase marking 1491 phrase-grammar 201 phraseme 249 pidgin 1653, 1757, 1759, 1761, 1764 pidginization 1752 place 1121 Plansprache 1563 pluperfect (s. Plusquamperfekt) ⫺ pluperfect indicative 1702 ⫺ pluperfect subjunctive 1703 Plural/plural 1055, 1234, 1465, 1736, 1749 ⫺ Abundanzplural 1057 ⫺ kollektiver Plural 1056 ⫺ assoziativer Plural 1056 ⫺ broken plural 750, 1749 ⫺ distributiver Plural 1056 ⫺ indefinite plural 1736 ⫺ Plurale Tantum/plurale tantum 364, 737, 1061, 1969 ⫺ plurality 1486 ⫺ sortaler Plural/sortal plural 747, 1056 ⫺ sound plural 1749 Plusquamperfekt/pluperfect/past perfect 1187, 1274, 1702 f. Plusquamperfektform, doppelte 1273 point 1169 point state 1169 Polyfunktionalität/polyfunctionality 238, 841 f. Polymorphie 238 Polysemie/polysemy 238, 250, 259, 526, 638, 842, 1841 polysynthetic language 254, 916, 1247, 1389, 1465 polysynthetic structure 1223 polysystemic 150, 494 polysystemicity 150 Ponästhem 418 Portmanteau/portmanteau ⫺ portmanteau affix 1455 ⫺ portmanteau pronoun 1010 ⫺ portmanteau word/lapse 953 ⫺ Portmanteau-Morph (portmanteau morph) 183, 439, 573, 622
Position 676 Positiv/positive 1213, 1306 Possession/possession 743, 1091, 1413 ⫺ alienable possession 1748 ⫺ attributive Possessivkonstruktion 1093 ⫺ inalienable possession 1748 ⫺ inhärente Possession 1091 ⫺ Possessivaffix 1099 ⫺ possessive adjective 743 ⫺ possessive case 1085 ⫺ possessives Prädikat 1091 ⫺ Possessivkompositum 1276 ⫺ Possessivsuffix 1360 ⫺ Possessor/possessor 1091, 1442 ⫺ Possessum 1091 postbase 1468 potential 1440, 1458 potential deobjective 1132 potentive 1483 Prädikation/predication 1104 f. ⫺ possessives Prädikat 1091 ⫺ predicate 1104, 1106, 1112 ⫺ predicate determiner 1492 ⫺ predicate, basic 1400 prädikativ/predicative 1301, 1433 Präfigierung/prefixation 528, 934, 1045, 1260, 1308, 1320, 1685 ⫺ cross-referencing prefix 1425 ⫺ deictic prefix 993 ⫺ Halbpräfix 1276 ⫺ Präfix/prefix 535, 1062, 1218, 1236, 1275, 1465, 1674 f., 1677, 1705, 1926 ⫺ Präfixoid 1276 ⫺ Präfixtausch 893 ⫺ prefix-stripping 1779 ⫺ Pseudopräfix 1770 pragmatic 999 pragmatic effect 963 pragmatische Frequenzuntersuchung 1236 Präposition/preposition 675, 1511, 1737 Präsens 1274, 1303 Präteritopräsentien 1685 Präteritum 1521 Präverb/preverb 892, 1135, 1277 precative 1716 precision (reliability measurement) 1897 precision principle 1928 predicable 1433 predication (s. Prädikation) predicative (s. prädikativ) prefix (s. Präfix) present 1185, 1372 ⫺ present in the past 1175 ⫺ present indicative 1702
1996 ⫺ present subjunctive 1703 presentation of language 1869 preterite 1372 preterite participle 1671 preverb (s. Präverb) Primärsuffix 1914 Priming/priming 1770, 1781 ⫺ cross-modal priming 1781 ⫺ masked priming 1781 ⫺ repetition priming 1781 privative 492, 1085 process, diachronic 653 process, formal 523 process, morphological 523 proclitic (s. Proklise) pro-drop 1002 productivity 163, 298, 1577, 1636 f., 1641, 1802 Profilierung 1684 pro-form 783 progressiv/progressive 1521, 1737 progressive Steuerung 1685 prohibitive 1191 Proklise 1271 ⫺ proclitic 392, 1002, 1465, 1532 Pronomen/pronoun 675, 746, 783, 983, 1099, 1301, 1305, 1313, 1330, 1392, 1465, 1492 ⫺ bound pronoun 783 ⫺ demonstrative pronoun 716, 788, 1370, 1713 ⫺ dummy pronominal 1007 ⫺ free pronoun 783 ⫺ impersonal pronoun 1006 ⫺ incorporated pronoun 922 ⫺ indefinite pronoun 716, 790, 1713 ⫺ independent pronoun 1002 ⫺ interrogative pronoun 790, 1370, 1702 ⫺ intimate pronoun 1724 ⫺ Personalpronomen/personal pronoun 716, 784 f., 1270, 1369, 1412, 1701 ⫺ portmanteau pronoun 1010 ⫺ possessive pronoun 788, 1002, 1701 ⫺ reciprocal pronoun 790 ⫺ reflexive pronoun 789, 1370 ⫺ Relativpronomen/relative pronoun 716, 791, 1270, 1702 property 1105 ⫺ permanent property 1169 ⫺ property concept 1112 ⫺ temporary property 1169 propositional content 1104 proposition-oriented modality 1193 propositus 1003 proprietive 744, 1085
Verzeichnisse und Sprachenkarten prosody 149, 495 ⫺ prosodic analysis 494 ⫺ prosodic contrast 1735 ⫺ prosodic inversion 397 prototype 611 prototypical meaning 845 proximate 1005 pseudoaffix 931 Pseudoderivat 894 pseudo-dual 1749 punctual 1168, 1169, 1734 purpose 1122, 1205 purposive 1861
Q Quadral 1057 quadruplication 529 Quasimorph 419 Quasimorphem 893 quasi-representable linguistic unit 511
R readjustment rule 198 realis 1481 Realisation/realization 235, 617 reanalysis 936 reason (circumstantial notion) 1122 recall (reliability measurement) 1897 recategorialization 1594 recent past/hodiernal past 1186 f. recipient 1122 reconstruction 1661 reconstruction, internal 1661 rection (s. government) recursiveness 1389 reduction 1742 redundancy rule, lexical 188 Redundanz/redundancy 271, 346, 427, 1567 Reduplikation/reduplication 467, 529, 676, 751, 935, 970, 1045, 1061, 1217, 1320, 1415, 1477, 1540, 1549, 1659, 1678, 1735, 1758 f., 1763, ⫺ deictic reduplication 994 ⫺ expressive reduplication 1440 ⫺ inexact reduplication 558, 1217 ⫺ infixal reduplication 545 ⫺ intensive reduplication 1440 ⫺ partial reduplication 558 ⫺ reduplicative system 1512 ⫺ total reduplication 558 reference 1122 ⫺ frame of reference 984 ⫺ general reference 1924
⫺ generalized referent 1000 ⫺ homogeneous reference 1068 ⫺ non-participant referent 998 ⫺ reference point 1182 ⫺ referent conflation 1596 reference form 631, 636 reflex (result of a shortening process) 957 reflexive marker 1437 Reflexivmedium 1304 regelmäßig 236 Regelordnung, extrinsische 471 regressiv 1685 regularization 1577 f. regulärkompositionell 888 reification 1426 reinforcement (s. Verstärkung) reinterpretation 1579, 1587, 1615 relation (type of entity) 1105 relational 748 relational form 1413 relational meaning 1222 relational rule/via-rule 337, 498 Relationierung 1098 Relationsmorphen 1685 relative clause 1113, 1498 Relativpronomen/relative pronoun 716, 791, 1270, 1702 relevance 203, 600, 1140, 1251 relevant feature 493 remoteness 1187 remotivation 1615 repeater 1038 replacement 529 replacive signified 526 replacive signifier 526 replica 527 replica language (s. Kopiesprache) replication 527 representable (linguistic unit) 511 representation, mental 1788 resegmentation 1580 resultative 1133, 1177, 1747 reversal 1210 rhematic 1434 rhotacism 1612 rhyme 971 root (s. Wurzel) rote-learning 1797 Rückbildung/back formation/ back derivation 532, 583, 932, 937, 957, 1256, 1275, 1579, 1587, 1615, 1638, 1678, 1926 rule 1893 ⫺ rule change 1575 ⫺ rule creation 1575 ⫺ rule discovery 1575 ⫺ rule fallacy 297 ⫺ rule generalization 1578, 1587
1997
Sachregister ⫺ rule of construal 1067 ⫺ rule of referral 617 S Sachbedeutung 677 salience of the parts 1049 same subject subordinator 1459 sandhi 251, 383, 389 saturation 1434, 1451 Satzname 894 schema 633, 1797 Schriftsystem, Tiefe 1917 screeve (s. subparadigm) secondary motivation 1615 secretion 1584 segmental 524 segmental geometry 574 Sekundärsuffix 1914 selected type 261 selection 182 semantic domain/frame 260 semantic field 260 semantic integrality 1350 semantic map 1877 semantic potential 611 semantic restriction 1047 semantic shift 1596 semasiological 1870 semasiology 1870 semelfactive 1169 semi-indirect discourse 1001 sensory evidence 1197 sentence type 1191 sentence, conditional 1725 sentential marker 1423 serial verb 1558, 1595 serial verb construction 815, 1538 serialization 1592 serialization, verb 1595 set interpretation 1071 setting of the language 1869 shape (semantic feature) 1016 sign (s. Zeichen) signal 210 Significans/Form/signifier/signifiant/form/expression 1, 173, 210, 378, 404 f., 1349 Significatum/Inhalt/signified/signifie´/content 2, 210, 378, 405, 1349 Silbenschrift 1916 Silbenwort 894 similative 1213 simplicity 1802 Singular/singular 1055, 1465, 1749 Singulare Tantum/singulare tantum 737, 1061, 1069 Singulativum/singulative/unit noun 737, 888, 1055, 1071, 1509
situated type 261 situation 611, 1050 skeletal tier 496 slab (s. subparadigm) slip of tongue 1782 sonority hierarchy principle 550 sound symbolism 570 source 210, 957, 1120 f. source, language external 1042 source, language internal 1042 source, material 1122 span 1005 speaker-and-spouse 1004 speaker-hearer centric 985 special language 1924 special reference 1924 specified quantity 1171 speech error 1788 speech level 1004 speech point 1182 speech role 998 speech situation 984 speech-act participant 1412 spelling checking 1897 spezifisch onymische Morphe 1904 spezifische Sprachentwicklungsstörung (SSES) 1820 split-S language 1082 spoonerism 576 Sprachbequemlichkeit 1567 Sprachebene 22 Sprachentwicklungsauffälligkeit 1816 Spracherwerbsforschung 1769 Sprachklassifikation 1241 Sprachlehre/reference grammar 1860 Sprachökonomie/economy 1252, 1567, 1928 Sprachstörung 1769, 1816 Sprachunterricht 1929 Sprachverfall/language decay 1646 Sprachwechsel/language shift 1646 Sprechfehler 1768 stabilisation 1757 Stamm/stem 250, 257, 523, 605, 869, 1687, 1914 ⫺ Aspektstamm 1319 ⫺ discontinuous stem 1853 ⫺ final stem component 1417 ⫺ initial stem component 1416 ⫺ medial stem component 1416 ⫺ oblique stem 1080 ⫺ Stammalternation 1062 ⫺ Stammklasse 1687 ⫺ stem allomorphy 363 ⫺ stem form 363 ⫺ stem-forming element 605, 637 Standard Average European 1563
standard marker 1213 standard theory 196 standardisation 1927 Standardsprache 1267 state 1165, 1168 state of affairs 1104, 1165 state, absolute 362 statistical approach, deductive 1898 statistical approach, inductive 1898 stativity 1115 status constructus 741, 1095 status suffix 1707 ⫺ honorific status 1003 ⫺ humble status 1003 stem (s. Stamm) stemming 1896 Stichprobengröße 1237 stratum 165, 189, 1586 stress 383, 587 ⫺ phonemic stress 1232 ⫺ stress assignment 1477 ⫺ stress-shift 1658 strict cyclicity 340 structural relationship 172 structure, conceptual 845 structure/configuration/pattern 173, 494 Strukturalismus 1267 subgender 1034 subjacency 199 subject 1009, 1405, 1514 ⫺ indirect subject 1085 ⫺ subject index 1875 subjectification 263 subjective conceptualisation 1346 subjunctive 1200, 1862 ⫺ imperfect subjunctive 1703 ⫺ perfect subjunctive 1703 ⫺ pluperfect subjunctive 1703 ⫺ present subjunctive 1703 sublanguage 1924 Submorph 417 submorphemic 1045 subordination 1202, 1737 subordinator, different subjects 1459 subparadigm/slab/screeve 597, 633 Substantiv/substantive 1114, 1301, 1441 ⫺ Artikelflexion des Substantivs 1270 ⫺ Substantiv, schwaches 1684 Substitution 676 Substrat/substratum 1646 substratal influence 1741, 1763 substratum 1764 subsystem 494 Subtraktion/subtraction 475, 581, 751, 936
1998 ⫺ subtractive signified 525 ⫺ subtractive signifier 525 Suffigierung/suffixation 528, 934, 1045, 1307, 1319, 1465 ⫺ class-free suffix 1456 ⫺ cohering suffix 341 ⫺ deictic suffix 993 ⫺ detransitive suffix 1371 ⫺ diminutive suffix 1707 ⫺ directional suffix 1460 ⫺ grammatical suffix 1534 ⫺ Halbsuffix 1276 ⫺ hypocoristic suffix 969 ⫺ independent suffix 1456 ⫺ instrumental suffix 1707 ⫺ interrogative suffix 1374 ⫺ lexical suffix 1534 ⫺ non-cohering suffix 341 ⫺ Nullsuffix 894 ⫺ participial suffix 1439 ⫺ Primärsuffix 1914 ⫺ pronominal suffix 1747 ⫺ Sekundärsuffix 1914 ⫺ sentential suffix 1456 ⫺ suffix stripping 1896 ⫺ Suffix/suffix 535, 1061, 1218, 1236, 1258, 1275, 1477, 1674, 1676 f., 1926 ⫺ Suffixaufnahme 1491 ⫺ Suffixerweiterung 891 ⫺ suffixing preference 941 ⫺ Suffixkombination 891 ⫺ Suffixoid 1276 ⫺ Suffixtausch 893 ⫺ Suffixvariante 891 ⫺ Suffixwechsel 893 ⫺ unconventional suffix 967 ⫺ verbal suffix 1709 ⫺ verbalizing suffix 1371 superclassing 1034, 1039, 1041 superfix, deictic 993 superlativ/superlative 1213, 1257, 1306 f., 1717 ⫺ absolute superlative 1213 ⫺ relative superlative 1213 superstratum 1764 supervised machine learning algorithm 1898 supine 1703 Suppletion/suppletion 465, 469, 510, 874, 935, 1062, 1727 ⫺ affixal suppletion 515 ⫺ derivational suppletion 515 ⫺ inflectional suppletion 515 ⫺ morph suppletion 515 ⫺ radical suppletion 515 ⫺ suppletive 512, 605 ⫺ suppletive allomorphy 336 ⫺ suppletive Form 1685 ⫺ suppletive stem 1716 ⫺ Suppletivismus 1302 suprafixation 527 ⫺ accentual suprafixation 529
Verzeichnisse und Sprachenkarten ⫺ suprafix 527 ⫺ tonal suprafixation 529 suprasegmental 524, 587 ⫺ suprasegmental apophony 529 ⫺ suprasegmental process 751, 936, 1415 switch-reference 1008, 1203, 1459, 1501 syllable 1512, 1731 ⫺ syllable structure constraint 430 ⫺ syllable word 958 symbolic framework 1806 symbolization type 370 Synaffix 891 synaffixation 934 synchronic analysis 1591 Synkretismus/syncretism 363, 602 f., 618, 812, 1300, 1842 synonym 240 synonymy 842 syntactic category 264 syntactic expression 370 syntactic reanalysis 1596 syntactic restriction 1046 syntagma-interne Beziehung 1565 syntagmatic deviation 618 syntagmatic relation 494 syntaktische Funktion 676 syntax, lexical 346 synthesis 176, 1223 ⫺ interrupted synthesis 1400 ⫺ synthetic language 1698 ⫺ synthetic structure 1223 ⫺ synthetisch-analytische Sprache 1301 ⫺ synthetische Bildungsweise 1306 ⫺ synthetische Ersatzkonstruktion 1271 ⫺ synthetische Wortform 1271 ⫺ synthetischer Prozeß 1307 ⫺ synthetischer Zug 1268 ⫺ Synthetismus 1238 system 173, 494, 614, 834 system, pronominal 1713, 1720 systematic gap 834 systematische Bedeutung 1911
T tactic of sign 379 tagmemics 179 takeover 639 target language (s. Zielsprache) telic 1168 telicity 1166 temporal relation 1202 temporality 1501
Tempus/tense 1181, 1302, 1414, 1673, 1702, 1867 ⫺ absolute tense 1182, 1188 ⫺ perfect tense 1458 ⫺ relative tense 1188 ⫺ sequence-of-tenses 1189 ⫺ sudden discovery tense 1458 ⫺ Tempusform 1274 ⫺ tempusneutrale Form 1274 ⫺ tense logic 1182 ⫺ tense stem 1715 ⫺ tense-aspect-modality 1423, 1533 ⫺ tense-mode 1405 terminative 1171 terminativity 1168 terminologische Kombinatorik 895 terminology 1861 ⫺ conventional term 1801 ⫺ innovative term 1801 ⫺ term 1924 textbook 1860 Textfrequenz 1272 Themavokal/thematic vowel 447, 751 theme 1401 Tilgung 436, 468 time-stability 733, 1422 token blocking 347 Ton/tone 587, 937, 1062, 1232, 1538, 1731 ⫺ floating tone 1539 ⫺ tonal morph 588 ⫺ tonal morphology 588 ⫺ tone-shift 1658 top-down progression 1872 tracking 1000 Transfer/transfer 1644 transfixation 528, 750 ⫺ transfix 552 ⫺ transfixational process 934 transitivity 1496 ⫺ transitive 1413, 1434 ⫺ transitivity marker 1752 transkortikale Aphasie 1819 Transnumeral 1054 transparency 814, 932, 1615, 1802 Transposition/transposition 361, 888, 938, 1278 transpositional inflection 362 tree diagram 1879 triadic kin term 1493 Trial 1057 triangular kin term 1003 trilinear representation 1881 triplication 529, 559 truncation (s. Kurzwort) tun-Periphrase 1273 type 261 type blocking 347 type familiar 834 type-familiarity 1627
1999
Sachregister U Übergeneralisierung 1766 Umlaut/umlaut 567, 1272, 1275, 1582, 1685 unaccusative 813 uncompleted 1440 unconventional word-formation process 963 underlying form 489, 499 underlying representation 336, 499 underlying segment 428 underspecified 270 unidimensional 492 unikal 418 unikales Restelement 893 unisegmental 894 Univerbierung 889, 1277, 1694 universal 1764 universal grinder 1067 universal packager 1067 universal sorter 1067 universal, implicational 1249 universalism 1858, 1862 Universalsprache 1563 unmarked 608 unregelmäßig 1272 Unregelmäßigkeit 1272 unsupervised machine learning algorithm 1898 utterance, minimum 379
V Valenz/valence/valency 1118, 1130, 1402, 1565 ⫺ morphological valence 164 ⫺ valence class 1879 ⫺ valency pattern 1130 ⫺ valency shifter 1395 validator 1461 vantage point, egocentric 985 vantage point, sociocentric 985 Variation 234, 1860 ⫺ asymmetrische Beziehung der Varianten 236 ⫺ automatische Beziehung von Varianten 236 ⫺ conditioned variant 484 ⫺ freie Variation/freie Variante/ fakultative Variante/free variation/free variant 180, 240, 484 ⫺ funktionale Beziehung von Varianten 242 ⫺ graduelle Beziehung zwischen Varianten 237 ⫺ kombinatorische Variante 241 ⫺ kontextuelle Beziehung von Varianten 242
⫺ optional variant 484 ⫺ selbständige Beziehung von Varianten 236 ⫺ symmetrische Beziehung der Varianten 237 ⫺ Variante 234 Varietätendifferenzierung 1682 variety of language 1861 Vendler class 1168 venitiv 1521 Verb/Verbum/verb 674, 1301, 1370, 1433, 1466, 1492 ⫺ adjectival verb 1113 ⫺ analytische Verbform/zusammengesetzte Verbform 1271 f. ⫺ aspectual verb 1172 ⫺ auxiliary verb 1257, 1495 ⫺ denominal verb 1531 ⫺ dependent clause verb 1497 ⫺ derived verb 1508 ⫺ direct verb 1010, 1151 ⫺ generic verb linkage 1506 ⫺ impersonal verb 1448 ⫺ lexical verb 1257 ⫺ main clause verb 1497 ⫺ main verb 1495 ⫺ narrative verb 1187 ⫺ ornative verb 744 ⫺ Partikelverb/particle verb 166, 1277 ⫺ phrasal verb 889 ⫺ privative verb 744 ⫺ reciprocal verb 1134 ⫺ reduplizierendes Verb 1689 ⫺ reflexive verb 1133, 1157 ⫺ rückumlautende Verbflexion 1685 ⫺ schwaches Verb 1271, 1685 ⫺ sound verb 1743 ⫺ starkes Verb 1272, 1684 ⫺ stative verb 1113, 1520 ⫺ synthetische Verbform/einfache Verbform 1271 ⫺ trennbares Verb 1277 ⫺ vector verb 1729 ⫺ verb class 1168, 1412 ⫺ verb morphology 1532 ⫺ verb prefix 1436 ⫺ verbal derivational suffix 1459 ⫺ verbal measure 1744 ⫺ verbal morphology 1400 ⫺ verbalization 939, 1107 ⫺ verbum finitum 1317 ⫺ verbum infinitum 1317 ⫺ Verbzusatz 892, 1277 ⫺ weak verb 1744 Verbreitungsgleicheit 1914 Verdeutlichung 1913 Verdichtung 1278 Verfremdung 1275 Verknüpfung 1914 Verlaufsform 1273
Verschmelzungsform 1271 Verstärkung/reinforcement 1192, 1217, 1648 Verteilung, inklusive 241 Verteilung, komplementäre 241 Verteilung, überlappende 241 via-rule (s. relational rule) viewpoint 1166 voice 1146, 1703 ⫺ active voice 1146, 1148 ⫺ conveyance voice 1480 ⫺ inverse voice 1148 ⫺ middle voice 1149, 1508 ⫺ passive voice 1146, 1148 ⫺ patient voice 1480 ⫺ undergoer voice 1480 Vokal/vowel ⫺ epenthetic vowel 1512 ⫺ Vokalharmonie/vowel harmony 251, 466, 495, 568, 1359, 1512 ⫺ Vokalwechsel 1685 ⫺ vowel gradation 568 ⫺ vowel lengthening 994 ⫺ vowel mutation 567 ⫺ vowel pattern 1746 volitional 1746 volitiv/volitive 1193, 1569 Vollformenlexikon 1771 Vorform 1910 vowel (s. Vokal)
W Wechsel, e-i- 1272 Welthilfssprache 1563 werden 1274 Wernicke-Aphasie 1817 Wort/word 5, 152, 377, 595, 1231 f., 1864, 1924 ⫺ actual word 197, 833 ⫺ content word 249, 258 ⫺ echo word 570 ⫺ Entstehung der Wort 1910 ⫺ full word 1474 ⫺ generalized word class 1867 ⫺ grammatical word 152, 248, 378, 596, 868 ⫺ lexical word 248 ⫺ morphosyntactic word 248 ⫺ orthographic word 251, 377 ⫺ phonological word 251, 378 ⫺ possible word 833 ⫺ potential word 197, 833 ⫺ stump word 958 ⫺ syllable word 958 ⫺ synthetisches Wort 895 ⫺ word and paradigm 26, 409, 627 ⫺ word blend 1790 ⫺ word boundary 377 f., 1852
2000 ⫺ word class 394, 596, 938, 1864 ⫺ word manufacture 833 ⫺ word marking 1491 ⫺ word order 391, 1198 ⫺ word stress 251 ⫺ word substitution 1790 ⫺ word-and-paradigm 1887 ⫺ word-based morphology 197 f., 868 ⫺ word-form paradigm 647 ⫺ word-grammar 201 ⫺ word-oriented morphology 1864 ⫺ Wortform/word form 22, 152, 248, 378, 523, 595, 868 Wortart 1301, 1565 Wortbildung/word formation 198, 344, 676, 833, 1325, 1801, 1867, 1887 ⫺ analytische Wortbildung 1307 ⫺ identifikatorische Wortbildung 1307 ⫺ innere Wortbildung 1683 ⫺ Lehnwortbildung 895 ⫺ mutatorische Wortbildung 1307 ⫺ word formation rule 198, 932 ⫺ Wortbildungsaktivität 889 ⫺ Wortbildungsart 887, 1274 ⫺ Wortbildungsmittel 1274 ⫺ Wortbildungsmodell 1274 ⫺ Wortbildungspotenz 889 ⫺ Wortbildungssystem 1307 ⫺ Wortbildungstyp 1275
Verzeichnisse und Sprachenkarten Wortform/word form 22, 152, 248, 378, 523, 595, 868 Wortgeschichte 1910 Wortgleichung 1910 Wortgleichung, reine 1914 Wortklasse, grammatische 1551 Wortkreuzung 1275 Wortkürzung 1275 Wortlehre 1682 Wortprägung 1911 Wortschöpfung 886, 1275 Wortstammvariante 1564 Wortstand 1278 würde-Konjunktiv 1274 Wurzel/root 257, 406, 524, 869, 1236, 1401, 1914 ⫺ empty root 258 ⫺ full root 258 ⫺ root creation 965 ⫺ root noun 752 ⫺ root word 752 ⫺ Wurzeletymologie 1914 ⫺ Wurzelflexion 1683 ⫺ Wurzelnomen 1914 ⫺ Wurzelverb 1914 X X-bar theory 408 Z Zahl/numeral ⫺ collective numeral 1071 ⫺ Kardinalzahl 1315
⫺ Ordinalzahl 1315 Zeichen/sign 378 ⫺ elementary linguistic sign 511, 523 ⫺ grammatical sign 524 ⫺ leeres Zeichen 447 ⫺ linguistic sign 404, 523 ⫺ morphological sign 523 ⫺ Nullzeichen/zero sign 438, 531 ⫺ quasi-elementary sign 511 ⫺ quasi-sign 406 ⫺ sign vehicle 210 ⫺ unique sign 406 zero allomorph (s. Nullallomorph) zero marking 745 zero sign (s. Nullzeichen) zero suffixation 531 zeroing 1010 Zielsprache/target language 1646 Zirkumfigierung/circumfixation 528, 934, 1308 Zirkumfix/circumfix 891, 535, 1215, 1275, 1747, 1853 Zitierform/citation form 248, 442, 604, 869, 1887 zugrundeliegend 234 Zusammenbildung 891, 1276 zusammengesetztes Wort 1547 Zusammenrückung 250, 889, 1277 Zusammensetzung 250 Zweisprachigkeit 1647
Die Register wurden unter Mitarbeit von Silke Fließ, Margret Schulze und Isabel Schlote erstellt. The indices were produced with the assistance of Silke Fließ, Margret Schulze and Isabel Schlote.