DIE ALTSUMERISCHE VOKALHARMONIE UND DIE VOKALE DES SUMERISCHEN Jan Keetman (Istanbul)
Der einzige groß angelegte Versuch die Vokalharmonie bei den Konjugationspr ä ˜xen e- und ì- im Altsumerischen zu deuten und damit auch den Vokalismus des Sumerischen insgesamt, liegt schon mehr als 70 Jahre zurück. Arno Poebels kühner Interpretation1 folgte nur noch die Dissertation von Samuel Noah Kramer2 über die Prä˜xe bé- und bí- einige Jahre später.3 Dabei bewegte sich Kramer ganz auf den Bahnen Poebels, um dann das ganze Thema für immer zu verlassen. Zu den Gründen für die geringe Beschäftigung mit der von Poebel und Kramer untersuchten Vokalharmonie, die wir kurz als altsumerische Vokalharmonie bezeichnen und damit von anderen Vokalangleichungen abgrenzen wollen, gehört die Ansicht, dass es sich lediglich um ein Dialektproblem handele.4
Geschrieben wird diese Vokalharmonie in den altsumerischen Texten aus Lagas und Umma, es gibt auch Beispiele aus Ur und Uruk, in Texten aus dem Norden wird sie nur in Fara geschrieben. Ob wir es dabei mit zwei Schrifttraditionen des Sumerischen oder mit Dialekten zu tun haben, ist unklar. Die ungleiche Verteilung der Textfunde lässt kein abschließendes Urteil zu, doch da es südlich von Umma keine Stadt gibt, für die wir die Nichtbeachtung der altsumerischen Vokalharmonie nachweisen könnten, sei die Vermutung gewagt, dass es sich um eine Erscheinung des Sumerischen des Südens handelt und dass ihr Fehlen im Norden mit frühem semitischem Ein˘uss zusammenhängt. Dabei kann oˆen bleiben, ob dieser Ein˘uss die Schreibweise oder die Aussprache des Sumerischen betraf. Während der Akkad-Zeit verschwindet diese Vokalharmonie bei dem Konjugationsprä˜x ì-/e-, so wie den übrigen Elementen, die den Wechsel i/e zeigen. Genauer gesagt, die “e-Zeichen” e-, bé-, -me-, -sè- werden in diesem Zusammenhang nicht mehr geschrieben, während die “i-Zeichen” ì-, bí-, -mi-, -si- beibehalten werden. Dies kann ein reines Phänomen der Schrift sein. D. h. die Schrift kodiert wie bei den Wortwurzeln auch bei den
Verschiedene Fragen im Zusammenhang mit diesem Artikel habe ich mit Bram Jagersma diskutiert. Er hat mir auch großzügig Teile aus seiner noch nicht abgeschlossenen Dissertation zur Verfügung gestellt. Sowohl aus der Diskussion als auch aus Jagersmas Manuskript habe vieles gelernt und es ergaben sich erhebliche Veränderungen. Hierfür möchte ich mich bei Bram Jagersma herzlich bedanken. Die Verantwortung für diesen Artikel liegt natürlich nur beim Autor. 1. Arno Poebel, The Sumerian Pre˜x e- and i- in the Time of the Earlier Princes of Lagas, AS 2 (Chicago: University of Chicago, 1931). 2. S. N. Kramer, The Sumerian Pre˜x Forms be- and bi- in the Time of the Earlier Princes of Lagas, AS 8 (Chicago: University of Chicago, 1936). 3. Einzelne Bemerkungen zum Thema sind hier nicht mitgezählt. Beachte auch Anm. 37. 4. Vgl. das distanzierte herangehen von Marie-Louise Thomsen, The Sumerian Language: An Introduction to its History and Grammatical Structure, Mesopotamia 10 (Kopenhagen: Akademisk Forlag, 1984), 39. Dem Autor ist nicht ganz
einsichtig, warum der lokalen Beschränkung der Vokalharmonie (s. auch unten) oˆenbar so viel Gewicht beigemessen wird. Die Vokalharmonie ist jedenfalls da, sie ist ein auˆallendes, für agglutinierende Sprachen typisches Phänomen. Diesen Umständen wird eine Einordnung unter irgendwelche Dialektverschiedenheiten kaum gerecht.
1
JCS 57 (2005)
2
JAN KEETMAN
Prä˜xen die Morpheme, nicht ihre Aussprache. Allerdings werden andere rein phonetisch bedingte Veränderungen innerhalb der Prä˜xkette auch weiter durch eigene Zeichen ausgedrückt, z. B. ha-/hé-, -da-/-dì-, -di-. Es kommen sogar neue phonetisch bedingte Wechsel wie ha-/hu- später noch hinzu, während auf die e-Zeichen nie wieder zurückgegriˆen wird. Interessant ist auch das Beispiel -si- bei dem das zugehörige Su¯x die wohl ursprüngliche Form -sè beibehält, während das Pr ä ˜x auf die durch Angleichung an die Vokalharmonie entstandene Variante -si- reduziert wird.5 Es ist daher gut m ö glich, dass diese Vokalharmonie tats ä chlich w ä hrend der Akkad-Zeit aufgegeben wurde, obwohl die Angleichung von Vokalen an andere Vokale innerhalb des Sumerischen auch später noch zu beobachten ist. Es fällt auch auf, dass die Ausdehnung der morphemanalytischen Schreibweise auf die betreˆenden Elemente der Prä˜xkette zu einem Zeitpunkt erfolgt wäre, zu dem die herrschende Schicht das bereits weitgehend phonetisch bzw. phonologisch geschriebene Akkadische gebrauchte.6 Dietz Otto Edzard übergeht in seiner Grammatik bei der Besprechung der Vokale Poebels Ansatz, erwägt ihn dann aber doch bei der Erörterung der Pr ä ˜xe b é -/b í - und e-/ ì -. 7 Obwohl
Edzard zu keinem Ergebnis kommt, führt er doch zwei bisher vernachlässigte Aspekte in die Diskussion ein: In agglutinierenden Sprachen sind Vokalharmonien üblich und es wäre verwunderlich, wenn dieses Phänomen nur auf einen kleinen Bereich der Sprache beschränkt wäre. Dem muss allerdings hinzugefügt werden, dass dies nicht bedeutet, dass die gleiche Vokalharmonie überall anzutreˆen ist. Sprachen wie Türkisch und Ungarisch zeigen, dass verschiedene Vokalharmonien nebeneinander bestehen können. Allgemein ist ein gewisses Misstrauen gegen Poebels Lösung zu beobachten, und dies besteht nicht ohne Grund. Schließlich postuliert Poebel sein Ergebnis einfach. Weder hat er eine andere Sprache zum Vergleich herangezogen, noch Alternativen diskutiert. Er zeigt lediglich, dass die von ihm angenommene Lösung tatsächlich fast alle Belege erklären kann. Doch das ist kein Beweis, denn schließlich könnte dies eine andere Theorie eventuell genauso tun. Die hohe Zahl seiner Treˆer täuscht, denn im Grunde würde jede Theorie, die die Vokale in zwei Gruppen teilt und dafür sorgt, dass e- normalerweise vor a steht, zu vielen Treˆern führen. Abgesehen von dieser generellen Kritik an der Beweisführung, teilen wir aber nicht den Einwand gegen Poebels Theorie, dass die von
5. Auˆallend ist, dass -si- zur allgemeinen Form wird, obwohl wir historisch die Vokalharmonie im Norden nicht nachweisen können. Dies ist immerhin ein Indiz dafür, dass diese Vokalharmonie irgendwann weiter verbreitet war. Beachte auch, dass die Inschriften Sargons ebenfalls der altsumerischen Vokalharmonie folgen. 6. Für eine Analyse des sumerischen Schriftsystems und die Unterschiede zur akkadischen Schrift vgl. Miguel Civil, „The Sumerian Writing System: Some Problems,“ OrNS 42 (1973) 21–34. 7. D. O. Edzard, Sumerian Grammar, HdO 71 (Leiden: Brill, 2003), 13–14, 110 und insb. 99–100. Skeptisch bin ich, was die von Edzard eingeführten langen Vokale im Sumerischen betriˆt. Eine Durchsicht aller in AHw gebuchten sumerischen Lehnworte ergibt fast ohne Ausnahme, dass es keine langen Vokale außer in der letzten Silbe gibt und dass die letzte Silbe mit sumerischem Anteil entweder einen langen Vokal hat oder auf einen verdoppelten Konsonanten endet. D. h. wenn man das Antreten der akkadischen Endung berücksichtigt, sind diese Lehnworte im Akkadischen alle auf der letzten sumerischen Silbe betont. Die einfachste Erklärung hierfür ist, dass diese Worte auch im Sumerischen auf der letzten Silbe betont waren und das dies die Struktur der Silbe mit langem Vokal oder geschlossen bedingte. Davor
mag es schwächere Betonungen, insbesondere auf der letzten Silbe von Bestandteilen zusammengesetzter Worte gegeben haben, die sich durch Konsonantenverdoppelungen, aber nie durch die Pleneschreibung von Vokalen anzeigen, wie z. B. in gi-izi-lá > gizillû, „Fackel“. Würde man nun eine Bedeutungsunterscheidende Einteilung in lange und kurze Vokale im Sumerischen annehmen, so wäre es auˆallend, dass diese Möglichkeit auf die letzte Silbe beschränkt bliebe und bei Zusammensetzungen, d. h. selbst wenn nur ein so farbloses Adjektiv wie gal „groß“ antritt, aufgegeben würde. Bei der Materialsammlung für diese sehr geraˆte Darstellung wurde berücksichtigt, dass die Funktion von Pleneschreibungen im Südaltbabylonischen auch im Akkadischen unklar ist (dies betriˆt allerdings nur wenige sumerische Worte) und dass Vokallängen auch am Anfang längerer akkadischer Worte sehr selten geschrieben werden. Doch kann dies nicht das praktisch totale Ausbleiben von Pleneschreibungen auf der vorletzten sumerischen Silbe erklären. Kein Gegenbeispiel ist hebr. hêkal < é-gal. Es ist sehr gut möglich, dass der sumerische Vokal in einer semitischen Sprache am ehesten mit ai wiedergegeben wurde und sich die Länge hieraus entwickelt hat. Vgl. arab. haykal.
One Line Long
DIE ALTSUMERISCHE VOKALHARMONIE UND DIE VOKALE DES SUMERISCHEN
ihm konstruierten 6 Vokale im Widerspruch zur Diˆerenzierung der Silbenzeichen nach lediglich 4 Vokalen stünden.8 Es gibt nämlich keinerlei Beweis dafür, dass die Silbenzeichen die sumerischen Phoneme vollständig und auch eins zu eins wiedergeben. Die Existenz einer Vokalharmonie spricht ja gerade für eine Tendenz zur Vereinfachung des Vokalismus bei den A¯xen, deren Wiedergabe wiederum die Hauptaufgabe der Silbenzeichen ist. Darüber hinaus lässt sich an einzelnen Beispielen sogar nachweisen, dass Silbenzeichen nicht immer nur für eine Silbe gebraucht wurden. Siehe dazu das Beispiel -NI-, zu dem wir gleich kommen.9 Noch aufschlussreicher ist das Beispiel TI: Von gis ti „Pfeil“ wurde der Lautwert entliehen, der zur Verwendung des Zeichens zur Schreibung von ti(-l) „leben“, su ti „empfangen“ und schon alt ì-ti „Mondlicht“10 führte. Vgl. jedoch su nu-TI-nébal-e „sie werden ihm daran nichts ändern“,11 obwohl wir mit hoher Wahrscheinlichkeit den Lautwert /de/ erwarten würden. 12 D. h. die Schreiber gingen gleichzeitig beim Vokal und beim Konsonanten einen Kompromiss ein. Wenn dies bei su nu-TI-né-bal-e geschah, so sind ähn-
8. Vgl. M.-L. Thomsen, Sumerian Language, 39. 9. Die Terminologie ist zugegeben etwas schwammig, denn man kann sich fragen, ob NI ein Silbenzeichen ist, das für verschiedene Silben /ni/ und /ne/ gebraucht wurde oder das Zeichen für ein Morphem der Prä˜xkette mit zwei lautlichen Realisierungen /ni/, /ne/ steht. Allerdings würde es vom eigentlichen Thema zu weit wegführen, solche Diskussionen bis auf den Grund zu führen. Also gebrauchen wir unkritisch den Begriˆ „Silbenzeichen,“ im Falle von NI auch mit gutem Grund, denn die Silbe /ni/ kann mit diesem Zeichen auch geschrieben werden, wenn sie nicht auf das entsprechende Morphem der Prä˜xkette zurückgeht. 10. OIP 99, 278 vi 10u. Vgl. dazu Manfred Krebernick, Die Beschwörungen aus Fara und Ebla, Texte und Studien zur Orientalistik (Hildesheim: Georg Olms, 1984), 204–5. 11. Ukg. 34, 1. 12. Ein Lautwechsel d > t, dessen Erscheinen zumindest in der Schrift immer mit a > e, i an dieser Stelle gekoppelt wäre, erscheint ziemlich unwahrscheinlich. Ein Wechsel von da/ta ist zwar belegt, aber fast ausschließlich tritt da für ta ein, insbesondere nach ba-, wo es auch die Variante -ra- für -ta- gibt. Vgl. Gene B. Gragg, Sumerian Dimensional In˜xes (Neukirchen-Vluyn: Butzon & Bercker, 1973), 40–43, 47–51; zu -ta-/-ra- 93–98, Thomas Balke, „Anmerkungen zum Terminus bar-ra(-)kar-ra in den neusumerischen Wirtschaftstexten,“ in dubsar anta-men, Fs. W. H. Ph. Römer, M. Dietrich, O. Loretz (Hrsg.), AOAT 253 (Münster: Ugarit-Verlag, 1998), 1–16.
3
liche Abweichungen auch bei den übrigen Beispielen für TI möglich. Das Prinzip „bildhaftes Zeichen > Lautwert > Schreibung eines anderen Wortes oder Wortbestandteiles“ konnte auch unmöglich zu einer auch nur annähernd genauen Wiedergabe der Lautgestalt der sumerischen Sprache führen, denn der Bestand an wirklichen Homophonen ist in den meisten Sprachen sehr beschränkt. Zwar gibt es auch Lautwerte deren Herkunft wir bislang nicht kennen, sie lassen jedoch keinen systematischen Versuch erkennen, den Mangel an Lautwerten auszugleichen. Was nun Poebels raschen Lösungsweg betriˆt, so teilen wir aus obigen Gründen das Unbehagen. Eine gründlichere Diskussion soll hier nachgeholt werden. Poebels Feststellungen und Lösung Folgende Feststellungen können von Poebel und Kramer übernommen werden: Die altsumerische Vokalharmonie bewirkt den Wechsel in der Schreibung des Konjugationsprä˜xes zwischen e- und ì-. Später bleibt ausschließlich ì- übrig. Es gibt Prä˜xe, die selbst unveränderlich sind und den Wechsel auslösen. Dazu gehören insbesondere alle Prä˜xe die /a/13 als Vokal enthalten.14 Andere Prä˜xe wechseln analog zu eund ì-, insbesondere -me-/-mi-. Weniger verlässlich ist der Wechsel von -sè-/-si-. Bei -NI- wird das Zeichen nicht getauscht, doch die Vokalharmonie wird auch von -NI- nach vorne weitergeleitet. Deshalb setzten Poebel und Kramer die Lesungen -né-/-ni- an. Das Prä˜x BI-/BÍ- wechselt vor Prä˜xen und Wurzeln ebenso wie e-/ì-. Es gibt aber eine größere Zahl von Ausnahmen, so dass etwa aus einer Schreibung mit bí- nicht 13. Schrägstriche sollen hier immer den Vokal von seiner Schreibweise isolieren. D. h. zu a „Wasser,“ á „Kraft,“ -ra Dativsu¯x etc. gehört /a/, ohne dass damit ausgesagt ist, dass es sich um den gleichen Laut oder das gleiche Phonem handeln soll. In klarem Zusammenhang lassen wir die Striche auch gelegentlich weg. 14. Eine Ausnahme stellt der bereits erwähnte Übergang -da-ni- > TI-ni/né- dar. In jedem Fall ist dies keine rein phonetisch bedingte Abhängigkeit, denn vor der Wurzel, wo -da- unveränderlich erscheint, steht es sicher vor allen möglichen Vokalen.
4
JAN KEETMAN
unbedingt geschlossen werden kann, dass das Konjugationsprä˜x vor der Wurzel ì- lauten würde. Weil die Zahl der Übereinstimmungen doch deutlich überwiegt, hat Kramer die Lesungen bé-/bíangenommen. Was die Vokale betriˆt, so steht, wie bereits erwähnt, vor allen Prä˜xen, die /a/ enthalten, nur e-: e-na-, e-ma-, e-da- usw. Übereinstimmung besteht, von ganz wenigen Ausnahmen abgesehen, mit den Vokalen von -me-/-mi-, -sè-/-si-.15 Die Beziehung zwischen dem Wurzelvokal und dem Gebrauch von ì-/e- in der Prä˜xkette ist in folgender Tabelle zusammengefasst. Veränderungen durch neue Belege und wo es notwendig erschien, Diskussionen zu den Lesungen ˜nden sich in den Fußnoten. Eine aktualisierte Sammlung aller Belege und ausführliche Diskussion wären sicher wünschenswert, doch dies war vom Autor nicht zu leisten. Andererseits ist bei der Auswertung ohnehin mit Zufällen zu rechnen, so dass Veränderungen, die sich durch eine andere Einordnung des einen oder anderen Verbums ergeben, für die Analyse nicht relevant seien sollten. Die bedeutendste Veränderung gegenüber Poebel ergibt sich aus einer Theorie von Bram Jagersma, nach der das dem Lokativ entsprechende Prä˜x ni-/né- auch am Anfang stehen kann. Es ist dann graphisch von ì- nicht zu unterscheiden. Ich überlasse es Bram Jagersma, seine These, die Teil seiner Dissertation ist, selbst darzustellen und zu begründen. Bei der Durchsicht der Belege ergab sich vor allem, dass sich die bei Poebel als unregelmäßig auftauchenden Formen vor /a/, wie „ì-gaz, su ì-bal-e, ì-Gar“ leicht als Lokative erklären ließen. Andererseits ˜ndet sich neben häu˜gem e-Gál kein *ì-Gál (bzw. *né-Gál), weil hier nicht né-, sondern die Variante /n/ zu erwarten ist. ì- . . . (i) e- . . . (i)
616 0
15. Besser nicht einbeziehen sollte man Formen mit -ne-, da das Pluralelement des Dativs wahrscheinlich auf -e-ne(-a)zurückgeht, so dass der Vokal vor -ne- kontrahiert sein könnte. 16. gi4, gíd, íl, si, ti, zi. Zur Lesung von ÍL vgl. G. Selz, Altsumerische Wirtschaftsurkunden aus amerikanischen Sammlungen, FAOS 15/2 (Stuttgart: Franz Steiner, 1993), 555–56. Von
One Line Long
ì- . . . (e) e- . . . (e) ì- . . . (a) e- . . . (a) ì- . . . (u) e- . . . (u)
617 5 (davon 3 ziemlich unsicher)18 0 2019 2220 621
den möglichen Lesungen íl, gùr, ga6(-G) scheidet die letzte ziemlich sicher aus, da sonst e- zu erwarten wäre. Ansonsten schließe ich mich den von Selz vorgebrachten Argumenten an, die für eine Lesung íl im Kontext der Wirtschaftstexte sprechen. Die Form e-ÍL Urn. 49 iv 4 wurde als Ausnahme/ Fehler gewertet, wie sie in geringer Zahl vorkommen. Eine andere Möglichkeit wäre natürlich die Lesung ga6(-G) an dieser Stelle. 17. dé, de6, e, è, kése(-DR), na de5. Für e vgl. ì-e Ukg. 6 ii 14u und den Eigennamen a-ba-di-ì-e „Wer wird den Prozess führen?“ Dazu gibt es auch die (Schreib-)?Variante a-ba-di-ì-bé (DP 195 ivu 1u, vgl. Selz, FAOS 15/2, 100). Vgl. auch si-e Ean. 1 Rs. x 25. Da das Präformativ sonst altsumerisch immer sè- lautet, lässt auch diese Form auf ì- schließen. Die Form e-na-né-és-a Ensak. v. Uruk 1, 7 spricht hingegen für den gleichen Vokal, wie in e-né- . . . Allerdings gibt es in den Texten aus Uruk eine Reihe von Abweichungen zu Lagas, die auf Dialektunterschiede oder einen weniger genauen Gebrauch der Zeichen zurückgehen könnten. Beachte außerdem eine mögliche Wirkung der Endung. Neusumerisch wird n + e (sprechen) ebenfalls né geschrieben, z. B. bí-né-es NG 84, 14. Eine andere Möglichkeit, den Widerspruch zu klären wäre ein weiteres „e“ s. u. Wegen des Auslautes auf DU (rá, re6) scheidet die Lesung sìr für EZEN im Sinne von rakasu „binden“, kasaru „zusammenfügen“ aus, denn Zeichen, die altsumerisch auf das DRPhonem enden haben später niemals Glossen, die auf /Vr/ enden. Vgl. Walther Sallaberger, The Sumerian Verb na de5(-g) “To Clear,” in An Experienced Scribe who Neglects Nothing, Fs. Jacob Klein, Y. Sefati et al. (Hrsg.), (Bethesda: CDL, 2005), 229– 53 u. ders. „ ‚bringen‘ im Sumerischen. Lesung und Bedeutung von de6 (DU) und túm (DU),“ in Von Sumer bis Homer, Fs. Manfred Schretter, R. Rollinger (Hrsg.), AOAT 325 (Münster: Ugarit-Verlag, 2004), 557–76. 18. me, Gen, †ID, sè, se12 (SIG7 ). Davon sind die letzten drei Beispiele ziemlich unsicher. Es stehen sich gegenüber e-né-sèga-a Luzag. 1 i 45 und TÙN.†È ì-ni-sè Ent. 28 iii 14 = 29 iv 5. Eannatum schreibt in den meisten Inschriften TÙN.†È bé-sè doch in Ean. 11 durchgehend TÙN.†È bí-sè. Für SIG7 ist e- nun auch durch BIN 8, 359 (AWAS 86) iii 9; viii 43 gesichert. Unsicher ist aber die Lesung des Zeichens selbst. Für die altsumerischen Belege ist die Gleichung (w)asabu sa mad¿ti „wohnen in Bezug auf mehrere“ heranzuziehen. Diese ist nur durch den neubabylonischen SipparText CT 17, 12 belegt. CAD B (1965) 90f. ergänzt die beiden Zeilen davor (nach Sb I 361): [sa-a] [SIG7] = ba-nu-ú sa rama-ni, MIN sa ki-da-tum EME.SAL (nach MSL ist hur-da-tum zu lesen). Gestützt wird diese Interpretation auch durch sa6/ sa6(-g) = banû ebd. und sa7/sa6 = banû Adj. ebd. 81a. In CAD A II (1968) 386b u. M I (1977) 20b wird die fehlende Lautwertangabe als si-e ergänzt, AHw 1481a schreibt oˆenbar unter
DIE ALTSUMERISCHE VOKALHARMONIE UND DIE VOKALE DES SUMERISCHEN
5
Wir werden einige Einzelheiten dieser Tabelle noch an verschiedenen Stellen kommentieren.
Jedenfalls zeigt die Liste, dass zwischen der Wahl von e- oder ì- als Prä˜x und dem Vokal der
dem Eindruck des 1979 erschienen Bandes MSL 14 [SIG7] und ignoriert damit die Textlücke, die si-i[SIG7] von den Einträgen vor a-sá-bu sá ME† trennt (cf. MSL 14, 425). Würde man von den vorhergehenden beiden Zeilen ausgehen, so wäre mit hoher Wahrscheinlichkeit sa7 zu lesen. Doch davor ˜nden sich noch die Einträge la natilu, „blind“ und la banû, „nicht gut(aussehend)“ (?). Man denkt hier am ehesten an eine Augenkrankheit und die Verfärbung des Gesichts unter Schock. Dazu passt die Gleichung sig7 (bzw. seg7 ) = warqu, „grün, gelb.“ Letztendlich kann man wohl nur sagen, dass es zu der Gleichung SIG7 = asabu sa mad¿ti keinen Text gibt, der die sumerische Aussprache sicher festlegt. Zur statistischen Verteilung von e- und ì- in den altsumerischen Texten würde eine Lesung sa7 am besten passen. Da wir den Mangel an Wurzeln mit e- . . . (e) als wichtiges Argument gebrauchen, soll hier die zu unseren Theorien schlechter passende, aber nicht unmögliche Lesung se12 bevorzugt werden. Mit der gleichen Begründung ist †ID einbezogen. Zum Akkadischen Abstraktum minûtu „Zählung, Berechnung; Rezitation“ ˜nden sich für †ID die Glossen se-ed (am häu˜gsten), si-ti, su-ud, ut-tu (CAD M II 98b). Man könnte daran denken, die Glossen su-ud, ut-tu abzutrennen und sie wegen der Bedeutung „Rezitation“ von minûtu zu sùd, sùdu = ikribu, karabu „Gebet“ zu stellen. Doch sind Form und Bedeutung so ähnlich, dass man eine Etymologie von sùd aus †ID „rezitieren, zählen“ erwägen muss. Kommt hinzu, dass sich zu ikribu auch die Gleichung sita ˜ndet. 19. ak, ba, babbar, bala, bar, dab5 (so konventionell, doch vgl. unten), Gar, Gál, lá, nag, pà, ra, sa4, sa10, tag, taka4, ha-la, lah4/5. Mit der von Jagersma vorgeschlagenen Lösung für NIlassen sich noch die Verben gaz und ha-lam hinzufügen. Das Verbum pad (DR) „zerbrechen, niederbrechen, herausbrechen“ (nicht bux „ausreißen“ wie in PSD B 162–66.) seberu, kasapu hat nur NI-. Doch dies wäre die einzige Wurzel mit ì- . . . (a), so dass sicher né-pad zu lesen ist. Damit fällt auch ein Argument für die Vermischung mit bu = nasahu, „ausreißen,” weg. Wenn man sich die lexikalische Tradition ansieht, so wurden pad und bu von den akkadischen Lexikographen immer auseinander gehalten (PSD B 165–66). Die Trennung der Auslaute in DR für pad und r für bu etc. ist ebenfalls eindeutig. Nur in den Abschriften altakkadischer Inschriften kommt eine Vermischung vor (PSD B 165b 6). Die Ursache dürfte eine akkadische Wendung isdam nasahu, „den Sockel ausreißen“ sein, die sich ab Gudea auch im Sumerischen belegen lässt. Dagegen ist in Ukg. 16 vii 7–9 gána dninGír-su-ka en-na uru4-a se-bé né-pad wohl nicht gemeint, dass Lugalzagesi seine Armee umständlich alles Korn herausziehen ließ, sondern dass das Heer absichtlich über die Felder zog: „Auf den Feldern Ningirsus, soweit sie bebaut waren, ließ er das Getreide niedertreten.“ Wegen des Auslautes hat bereits P. Steinkeller, Rez. zu PSD B, JNES 46 (1987) 55–59 die Trennung von pad und bu verlangt. Das Verb pad(-DR) kann auch die Arbeit in einem Steinbruch bezeichnen. Siehe B. Foster, Umma in the Sargonic Period, MCAAS XX (Hamden: Archon, 1982), 27. Foster führt die geringen Dimensionen an, in denen
diese Arbeit vorging. Dies passt nicht zum Ausgraben von Steinen, sondern zu ihren herausbrechen in einem regelrechten Steinbruch. Dies kann man sich unter den technischen Bedingungen der Zeit wohl so vorstellen, dass man den Fels erhitzte, abschreckte und dann zerschlug. So könnte man Obwohl auch Stehlen pad(-DR) zerstören. „zerbrechen, niederbrechen, herausbrechen“ Obwohl gegenüber pad(-DR)bu „zerbrechen, niederbrechen, vorzuziehen ist, herausbreist diese x „ausreißen“ chen“ Alternative gegenüber nicht bu immer sicher auszuschließen. vorzuziehen ist, G. ist Selzdiese bex „ausreißen“ Alternative merkt zu BIN nicht 8, 369, immer diesicher Bedeutung auszuschließen. „ausreißen“ G. Selz sei ziemlich bemerkt zu sicher. BINHier 8, 369, bezeichnet die Bedeutung e-PAD„ausreißen” eine Tätigkeit, sei ziemlich die der Genersicher. Hier alverwalter bezeichnet (nu-bànda) e-PADineine Bezug Tätigkeit, auf eine die spezielle der GeneralverZwiebelwalter sorte vornahm, (nu-bànda) bzw.invornehmen Bezug auf ließ. eineNormalerweise spezielle Zwiebelsorte werden vornahm, Zwiebeln bzw. ausgegraben vornehmen (ba-al). ließ.Doch Normalerweise bei den sum-GUD werden„FrühZwiebeln lingszwiebeln“ ausgegraben (?)(ba-al). handeltDoch es sich beiwohl den sum-GUD um eine Sorte „Frühlingsgrüner zwiebeln“ Zwiebeln, die (?) auch handelt eingesammelt es sich wohl (RI-RI) umund eineinSorte Nippur grüner auch Zwiebeln, tatsächlichdie ausgerissen auch eingesammelt wurden (bu, (RI-RI) bù, nicht und in inNippur PSD). Folgt auch tatsächlich man diesemausgerissen Ansatz, so käme wurden man(bu, zu ebù, . . .nicht (u). Allerdings in PSD). überFolgt man wiegen diesem die Schwierigkeiten Ansatz, so käme bei man diesem zu Ansatz. e- . . . (u). Es Allerdings gäbe zwei überwiegen Begriˆe für die dieErnte Schwierigkeiten dieser Zwiebeln, bei diesem RI-RI und Ansatz. PAD. EsDabei gäbe zwei kann Begriffe man sichfür für dieden Ernte Bedeutungsansatz dieser Zwiebeln,„ausreißen“ RI-RI und nicht PAD. Dabei auf diekann obenman erwähnten sich für altsumerischen den Bedeutungsansatz Belege „ausreißen“ stützen, da nicht diese durch auf dieihren obenAuslaut erwähnten klar altsumerischen von bu(-r) „ausreißen“ Belegegetrennt stützen, da sinddiese und durch eine plausiblere ihren Auslaut Interpretation klar von bu(-r) durch„ausreißen“ kasapum/ getrennt seberum möglich sind und ist. eine Das Verb plausiblere käme auch Interpretation in die seltenedurch Katekasapum/seberum gorie e- . . . (u), anstatt möglich in die gut ist. belegte Das Verb Gruppe kämee-auch . . . (a). inVom die seltene Kontext Kategorie her könntee-man . . . (u), raten, anstatt dassinder dieGeneralverwalter gut belegte Gruppe die eZwiebelfelder . . . (a). Vom vor Kontext der Ernte her könnte inspiziert manhat raten, (pà(-d)), dassdoch der dafür Generalverwalter dieinhaltliche Zwiebelfelder der Ernte inspiziert hat kann ich weder nochvor orthographische Parallelen (pà(-d)), nennen. doch dafür kann ich weder inhaltliche noch orthographische Parallelen Selz, FAOS 15/2,nennen. 603 weist noch auf PAD.DU = petû erseti Selz,859b), FAOS 15/2, 603 weist noch PAD.DUbildlich = petû erseti (AHw „urbar machen“ hin.auf Vielleicht „den (AHw „urbar machen“ Vielleicht bildlich „den Boden 859b), aufbrechen“ oder parshin. pro toto „(Schollen) zerBoden brechen.“ aufbrechen“ Vgl. die damit oder verbundenen pars pro Tätigkeiten toto „(Schollen) in KH §zer44. brechen.“ Die Bedeutung Vgl. die passt damit jedenfalls verbundenen nicht in Tätigkeiten den Kontext in KH von §BIN 44. Die 8, 369 Bedeutung und hat passt schonjedenfalls wegen des nicht Auslautes in den nichts Kontextmit vonbu(-r) BIN 8, „ausreißen“ 369 und hat zu tun. schon wegen des Auslautes nichts mit bu(-r) „ausreißen“ Zu den Zwiebelp˘anzen zu tun. und den Verben vgl. H. Waetzold, „Knoblauch Zu den Zwiebelp˘anzen und Zwiebeln nach undden denTexten Verbendes vgl.3.H. Jt.,“Waetzold, Bulletin „Knoblauch on Sumerianund Agriculture Zwiebeln3nach (1987) den23–56. Texten des 3. Jt.,“ Bulletin on Nun Sumerian kommt Agriculture der sichere 3 (1987) Beleg 23–56. für e-PAD (BIN 8, 369 = AWAS Nun96, kommt iii 5) von der dieser sicherelexikalisch Beleg fürschwierigen e-PAD (BINStelle. 8, 369 De= AWAS shalb wird 96, iii das 5)Verbum von dieser trotzlexikalisch starkem Verdacht schwierigen auf e-Stelle. hier nicht Deshalb berücksichtigt. wird das Verbum trotz starkem Verdacht auf e- hier nicht Zu berücksichtigt. babbar vgl. PSD B 30b. Die Variante bi-bi(-r) ist bé-bé(-r) zu Zu lesen. babbar Dasvgl. Zeichen PSD Bbé 30b. bleibt Die Variante wie e, me, bi-bi(-r) sè auch ist bé-bé(-r) im späzu teren lesen. Sumerischen Das Zeichen auf /e/ bé festgelegt, bleibt wie so e, wie me, bí, sè ì,auch mi, siim aufspä/i/ teren . Damit Sumerischen ist BABBARauf ein/e/ Beispiel festgelegt, für den so wie a/e-Wechsel. bí, ì, mi, si auf /i/. Damit Zu lah ist4/5 BABBAR : s. Nik. 164 ein Beispiel iv 1: e-la-lah für den oder e-lalah5-he. 5-hea/e-Wechsel. la Zu Mitlah der Gleichung : s. Nik. 164 nabû iv 1: „nennen“ e-la-lah -he sind oder diee-Lesungen lah5-he. sa4 4/5 5 (amMit häu˜gsten), der Gleichung se21 und nabû sa22 „nennen“ belegt (CAD sind die N ILesungen 32a). Wieder sa4 a/e-Wechsel. (am häu˜gsten), Wiese bei und sa22 und belegt dab5 (CAD bleiben N wir I 32a). hierWieder vorerst 21 babbar a/e-Wechsel. bei den herkömmlichen Wie bei babbar und gut undbelegten dab5 bleiben Lesungen wir hier mit vorerst a, weil bei wir den meinen, herkömmlichen dass die Wanderungen und gut belegten von Lesungen Beispielenmit füra,e-weil (e) wir zu e-meinen, (a) einedass wichtige die Wanderungen Erkenntnis unserer von Beispielen Statistik ist. für e- (e) zu e- (a) eine wichtige Erkenntnis unserer Statistik ist.
6
JAN KEETMAN
Wurzel ein Zusammenhang besteht. Dies ist schon eine bemerkenswerte Tatsache, denn die von uns für die Wurzeln eingesetzten Vokale stammen im wesentlichen aus lexikalischen Listen, die Jahrhunderte später abgefasst wurden und die wir mit den für das Akkadische gültigen Lautwerten lesen. Poebel versuchte den Zusammenhang zu erklären, indem er die vier Vokale der Umschrift so ergänzte, dass immer ein Paar aus einem Vokal 2021 mit mehr oˆener und mehr geschlossener Mundöˆnung entstand. Dazu nahm er eine Vokalharmonie an, die geschlossene Vokale auf ì-, oˆene auf e- abbildet. D. h. zu /i/ gibt es ein oˆenes ‰, geschrieben als /e/. Zu /a/ gibt es ein geschlossenes ê (Poebel schreibt e), geschrieben /e/ oder /a/. Zu /u/ gibt es ein mehr oˆenes o, geschrieben als /u/. Für die weitere Diskussion ist bei dieser Konstruktion folgendes zu beachten: Wie man sofort sieht, könnte Poebels o in einer Wurzel nicht entdeckt werden, wenn es in den Lautwertangaben lexikalischer Listen durch /a/ wiedergegeben wurde, da das betreˆende Wort, dann einfach unter /a/ gelistet würde und das Prä˜x e- nicht au¯ele. Analoges gilt auch für andere Vokale, 20. dù, du7, du8, du10, du11, dub, dul, durunx (TU†.TU†), gu7, gub, gul, kús, mú, su, sù, su8, sub, sus, tuku, tus, ús, uru4. Zu du10 vgl. die Eigennamen sùd-sùd Gá ì-du10 „Ich habe die Gebete gut gemach“ (Gá oder Ge26 für Gá-e) DP 133 vi 6 und mu dnanse ì-du10 „Das Jahr hat Nanse gut gemacht“, DP 175 ii 5. Bauer, „Altsumerische Wirtschaftsurkunden in Leningrad,“ AfO 36/37 (1989–90) 76–91, interpretiert im ersten Namen G als (sonst nicht belegten) Auslaut zu †ÙD. Dem stimmt Selz, FAOS 15/2, 186 zu, obwohl er den zweiten Namen als Gegenbeispiel interpretiert. mú lässt sich auch ma4 lesen. Die beiden Lesungen wechseln jedoch nicht bei der Gleichung uddusu „erneuern,“ die für ì-ni-mú Ukg. 14 iiiu 6u einzusetzen ist. 21. hul, luh, sur, túm, ur4, zu. Da die Abschriften der zweisprachigen Sargon-Inschriften die Vokalharmonie genau wie die Lagas-Texte gebrauchen, wenn man statt ì-luh mit Lokativ né-luh liest, wurde e-hul aus ihnen übernommen. Für túm vgl. Nik. 1, 201 ii 1: e-sè-túm u. e-[sè]-túm Sar C 1, 30. Zur Unterscheidung von de6/túm vgl. Anm. 18 Ende. Die Form e-zu ˜ndet sich in dem schwierigen Text MAD 4 33 (Umma), der auch e-Gen, e-na-sum enthält, sich also nach den Regeln der altsumerischen Texte aus Lagas richtet. Zu diesem Text Foster, USP, 67 (nach Vorschlägen von Th. Jacobsen). Demnach ist in En. I 29 xiii 6 entgegen J. S. Cooper, RA 76 (1982) 202 u. danach H. Behrens, H. Steible, Glossar zu den altsumerischen Bau- und Weihinschriften, FAOS 6 (Wiesbaden: Franz Steiner, 1983), 371 ì-su-su nicht in ì-zu!zu! zu verbessern. Da der Kontext beschädigt ist, kann die Stelle selbst nicht geklärt werden.
One Line Long
wenn sie bei ihrer Wiedergabe im Akkadischen im Mund im wesentlichen nur vor oder zurück gewandert sind. Wird aber bei der Wiedergabe vor allem die Mundöˆnung verändert, so kann wenigstens dies nachgewiesen werden. Das verblüˆendste Beispiel hierzu ist ì-e, das man so auslegen kann, dass bei der Übernahme des bildhaften Zeichens von e „Graben“ für das Prä˜x eund das Verbum e „sagen“ (marû) in mindestens einem Fall eine Abweichung der Aussprache toleriert wurde.22 Andere Möglichkeiten Bevor wir uns weiter mit Poebels Idee befassen, sollen erst einige allgemeine Fragen behandelt werden. Eine Vokalharmonie bedeutet, dass ein Vokal mit einem benachbarten Vokal23 in bestimmten Eigenschaften übereinstimmen muss. D. h. ein Vokal sendet nach bestimmten Regeln eine Art Signal aus, das von einem anderen Vokal empfangen wird. Wir betrachten zunächst nur den Fall, dass alle Vokale solche Signale geben können.24 In Eurasien ist diese Form der Vokal-
22. Jagersma wendet die Möglichkeit einer Dissimilation als Erklärung von ì-e statt *e-e ein. Das lässt sich natürlich nicht ausschließen. Andererseits ist ì- . . . (e) gut belegt. Ebenso die Mehrdeutigkeit der Zeichen in Bezug auf ihre Vokale. Man vergleiche etwa das im Akkadischen durchweg für /a/ verwendete Zeichen A, das so weit wir es sagen können auch als Silbenzeichen im Sumerischen für /a/ steht. Trotzdem kommt es von A „Wasser“ wofür auch andere Aussprachen, insbesondere /e/ angegeben werden (vgl. PSD A I 29). Es fällt auch auf, dass es zu /e/ eine ganze Reihe von Homophonen gibt, nämlich e „sprechen“, e „Graben“, é „Haus“, è „hinausgehen“, e4 (A) „Wasser,“ e11 „herauf-, herabsteigen.“ Auch dies lässt den Gedanken an mehr als einen Lautwert wahrscheinlich erscheinen. Entgegenhalten kann man insbesondere wegen hebr. hekal < é-gal die Möglichkeit eines ausgefallenen Konsonanten. Vgl. Edzard, HdO 71, 19 und ablehnend Verf. „Der Verlust der „Kehllaute“ und der Lautwandel a > e,“ AoF 31 (2004) 5–14 und demnächst ausführlich in BuB 3 (2006). 23. Einige Sprachen lassen neutrale Vokale zu, die von der Vokalharmonie quasi übersprungen werden. Im Sumerischen ist ein solcher neutraler Vokal nicht zu erkennen. 24. Nicht diesem Gesetz folgen die Bantu-Sprachen, die wir später besprechen. An einer Stelle gibt es wohl auch im Sumerischen eine Vokalangleichung, die nicht diesem Gesetz folgt, nämlich der Übergang e > u nach bestimmten Verben, die nach Umschrift u enthalten, wie sum + e > sum-mu. Das
DIE ALTSUMERISCHE VOKALHARMONIE UND DIE VOKALE DES SUMERISCHEN
harmonie weit verbreitet. Ihre Darstellung orientiert sich hier am Türkischen, Ungarischen, Mongolischen und Finnischen. Es lässt sich zeigen, dass wenn das Sumerische diesem Typ entsprach und der bei diesen Sprachen nachweisbaren Tendenz zu einem symmetrischen Aufbau der Vokale entsprechend der für die Vokalharmonie relevanten Eigenschaft folgte, nur Poebels Vokale oder eine darauf aufbauende Erweiterung in betracht kommen. Bei diesen Vokalharmonien werden die Vokale, je nachdem wie viele Eigenschaften relevant sind und wie viele Abstufungen zugelassen werden, in Gruppen eingeteilt, denen jeweils ein Vokal entspricht. Dabei ergibt sich alles aus einfacher Kombinatorik. Da wir nur von zwei Vokalen ausgehen,25 ist eine Eigenschaft gesucht, die zwei gegensätzliche Zustände erlaubt. Z. B. große und kleine Mundöˆnung (keine Mittelstellung).26 Bevor wir die verschiedenen Eigenschaften durchgehen, die für eine Vokalharmonie relevant sein könnten, wollen wir noch kurz eine weitere Bemerkung zu Vokalharmonien einschieben, auf die wir später wieder zurückkommen. Man kann Vokalharmonien in solche einteilen, die für eine Wurzel oder ein Wort nur Vokale zulassen, die in einem oder mehreren Punkten übereinstimmen, z. B. nur vordere (helle) oder nur hintere (dunkle) Vokale. Das zweite Phänomen besteht darin, dass eine Gruppe von Vokalen in einem Wort den Vokal in einem A¯x bestimmt oder dass A¯xe in
Beispiel stellt aber wohl eine beschränkte Ausnahme dar. Die Untersuchung ist an dieser Stelle auch besonders schwierig. Ein Übergang e > a wurde vielleicht übergangen um Verwechslungen zu vermeiden. Soll bei e > i die Möglichkeit einer Verschmelzung ausgeschlossen sein, so bedarf es eines konsonantischen Auslautes. Bei Silben KV wird aber zwischen e und i häu˜g nicht unterschieden. 25. Ganz theoretisch besteht natürlich auch die Möglichkeit, dass e- und (oder) ì- noch weiter aufzuteilen sind. Die von uns gefundene Lösung würde dann halt eine Feinheit unterschlagen. 26. Die Artikulation eines Vokals kann durch verschiedene Umstände bedingt sein, denn der Mundraum ist kein Schachbrett mit völlig gleichen Feldern. Aber man kann annehmen, dass versucht wird, innerhalb der betreˆenden Eigenschaft maximale Diˆerenzen anzustreben und andere Diˆerenzen zu minimieren.
7
gleicher Weise aufeinander wirken.27 Hier wird einer Gruppe von Vokalen ein spezieller Vokal bei den A¯xen zugeordnet, z. B. im Türkischen der hinteren Gruppe der Vokal a, der vorderen Gruppe der Vokal e. Zunächst wollen wir uns vor allem mit dieser Erscheinung beschäftigen, denn sie entspricht der altsumerischen Vokalharmonie. Vokale sind primär durch ihre Artikulationsstelle, die Größe der Mundöˆnung und die Haltung der Lippen (gerundet oder gespreizt) charakterisiert. Dazu können weitere Eigenschaften treten, insbesondere Länge, Nasalierung, Tonhöhe. Theoretisch könnten natürlich auch solche Eigenschaften als Auswahlkriterien für die Vokalharmonie in Frage kommen. Allerdings konnten die Akkader mit dem Wechsel der Zeichen e/ì, me/mi, sè/si in ihrem eigenen System etwas anfangen und ihn als Vokalwechsel deuten, wenn auch vermutlich mit etwas veränderten Vokalen. Damit scheiden die sekundären Elemente für die Vokalharmonie aus. Zu einem möglichen Ein˘uss einer Nasalierung wäre auch noch zu sagen, dass sowohl e- als auch ì- vor nasalem Konsonanten stehen können. Geht man die von uns als primär bezeichneten Eigenschaften durch, so ergibt sich sofort eine Alternative zu Poebels Vorschlag, nämlich die Einteilung in vordere und hintere Vokale. Diese Einteilung benutzen auch das Finnische, Ungarische und Türkische für ihre Vokalharmonien. Wobei das Türkische und Ungarische gelegentlich noch eine feinere Einteilung gebrauchen. Die Verschiebung der Artikulationsstelle wirkt sich vor allem auf die zweite Formantfrequenz aus, während die Mundöˆnung für den ersten, niedrigeren Formanten Bedeutung hat.28 Dabei ist anzunehmen, dass der zweite Formant eine Vokalharmonie klarer hervortreten lässt als der sich nach der Mundöˆnung ändernde erste 27. Der Fall, dass Su¯xe auf Worte wirken ist sehr selten bezeugt. Vgl. Martin Krämer, Vowel Harmony and Correspondence Theory (Berlin: Mouton de Gruyter, 2003). In der Beschreibung wird hier teilweise von Krämer abgewichen. 28. Für diese und die folgenden Ausführungen zur Wirkung der Artikulationsweise auf die Frequenzen vgl. allgemein H. G. Tillmann, Phonetik, Lautsprachliche Zeichen, Sprachsignale und lautsprachlicher Kommunikationsprozeß (Stuttgart: Klett, 1980).
8
JAN KEETMAN
Formant, denn er bewegt sich in einem weiteren Frequenzbereich (ca. 0,6 bis 2,4 kHz gegenüber ca. 0,15 bis 0,7 kHz) und erlaubt daher größere Frequenzunterschiede. Dieser Vorteil des zweiten Formanten wird durch die feinere Diˆerenzierungsfähigkeit des Gehörs bei tieferen Frequenzen nur zum Teil ausgeglichen. Zugleich liegt der zweite Formant auch näher an der maximalen Emp˜ndlichkeit des Gehörs bei ca. 3,5 kHz. Möglicherweise sind dies einige der Gründe für die weite Verbreitung dieses Typs von Vokalharmonien. Es gibt einen gewissen Zusammenhang zwischen einer Vokalharmonie nach vorderen und hinteren Vokalen und der Bildung gerundeter Vordervokale. Das soll kurz erläutert werden. Die beiden unten aufgeführten Graphen geben die Artikulationsstellen von 2549 Vokalen aus 317 Sprachen an.29 Je weiter links ein Eintrag erfolgt, desto weiter vorne ist der Vokal zu sprechen. Oben stehen Vokale mit geringer unten mit großer Mundöˆnung. Dabei ergibt sich das charakteristische V, bei dem die Kardinalvokale i und u links bzw. rechts oben und a in der Mitte unten steht. Da Merkmale wie die Vokallänge und kleinere Unterschiede bei der Artikulationsstelle in der Aufteilung nicht berücksichtigt sind, kann eine Sprache an der gleichen Stelle mehrere Einträge haben. Ungerundete Vokale: 252 55 31 425 100 19 81 392 13
Gerundete Vokale: 29 10 417 32 8 448 0 1 36
Wenn man in einer Sprache mit Vokalharmonie zwischen vorderen und hinteren Vokalen gerundete Vordervokale vermeiden will, so müssen die hinteren Vokale entgegen der in dem zweiten Graphen klar erkennbaren Tendenz alle ungerundet sein. Es gibt überhaupt keine gerundeten Vokale mehr und die Lippenrundung scheidet als Mittel bei der Bildung von Vokalen ganz aus. Die Lippenrundung wird aber für die Bildung von Vokalen derart häu˜g gebraucht, dass
29. Nach I. Maddieson, Patterns of Sound (Cambridge: Cambridge University Press, 1984).
One [Body] Line Long
man sie als eines der wichtigsten Mittel ansehen kann. Vokale mit gerundeten Lippen lassen sich vorne problemlos artikulieren, sie werden nur weniger gebraucht als bei den hinteren Vokalen, wo die zusätzliche Absenkung der zweiten Formantfrequenz durch die Lippenrundung eine klarere Diˆerenzierung gegenüber a schaˆt. Gerundete Vordervokale sind also gegenüber der völligen Aufgabe der Lippenrundung bei der Artikulation sicher das kleinere Übel. Schaˆen wir aber nur zu u ein gerundetes ü und begnügen uns mit ungerundeten Paaren zu i und a so bleibt ausgerechnet die Stelle mit den meisten Vokalen, die hintere Mitte frei. Ausgewogen wird das System, wenn man noch ein Paar o, ö hinzufügt. Dieses System haben das Türkische und ursprünglich das Mongolische,30 das Finnische und Ungarische haben einen ganz ähnlichen Aufbau,31 der ebenfalls ü und ö einschließt. Wir werden diesen Zusammenhang zwischen einer Vokalharmonie nach vorderen und hinteren Vokalen nicht direkt gebrauchen. Zum Verständnis des Zusammenhangs zwischen Vokalharmonien und bestimmten Ausprägungen des Vokalsystems mögen unsere Argumente trotzdem hilfreich sein. Außerdem mag es geschehen, dass unter verschiedenen Vokalharmonien im Sumerischen eines Tages auch der reine Gegensatz zwischen vorderen und hinteren Vokalen auftaucht. Auswahl unter den Artikulationsparametern Ausschließen lässt sich mit Sicherheit eine Vokalharmonie, die gerundete und ungerundete 30. Vgl. Juha Janhunen, “Proto-Mongolic,” in The Mongolic Languages, J. Janhunen (Hrsg.) (London: Routledge, 2003). 31. Theoretisch lassen sich die Systeme des Ungarischen und Finnischen ebenso wie die heute gesprochenen mongolischen Dialekte aus dem Vokalsystem des Türkei-Türkischen ableiten: Der „schwache Punkt“ dieses Systems ist der ungerundete Vokal über a (türkisch „ı“), wie man aus der geringen Belegung dieser Stelle in unseren Graphen sehen kann. Im Finnischen und dialektisch im Ungarischen ist ein Vokal nachweisbar, der auf einen solchen geschlossenen Mittelvokal zurückgehen könnte, im Mongolischen kann sein Ein˘uss im Konsonantismus beobachtet werden (s. vorige Anm.). Außerdem bietet ein ausgefallenes „hinteres i“ eine Erklärung dafür, warum im Ungarischen und Mongolischen i als „neutraler Vokal“ sowohl mit den hinteren als auch mit den vorderen Vokalen gebraucht werden kann.
DIE ALTSUMERISCHE VOKALHARMONIE UND DIE VOKALE DES SUMERISCHEN
Vokale einander gegenüberstellt. Schon die Annahme eines gerundeten Vokals in der Nähe von a und eines ungerundeten in der Nähe von u führt zu sehr seltenen Vokalen. Das eigentliche Problem ist jedoch die Ergänzung zu i. Die Rundung der Lippen bewirkt eine Senkung des zweiten Formanten.32 Allerdings ist der Eˆekt nicht so stark, dass der Wert des 2. Formanten von a unterschritten wird. Wir hätten also eine Vokalharmonie, die sich auf den 2. Formanten auswirkt ohne dass sich die gebildeten Gruppen bezüglich des zweiten Formanten jeweils im gleichen Feld befänden.33 Das zu i gebildete ü läge oberhalb von a, u und der zu a zu bildende gerundete Vokal aber unterhalb von a. Außerdem gibt es noch einen ungerundeten Vokal zu u. Dieses u* sollte hinter a liegen und ist damit u. a. durch ü von i getrennt. Auch ohne alle weiteren Probleme, wie die enorme Rückverlagerung von i, im einzelnen durchzugehen, sieht man, dass die Konstruktion scheitert. Einiges spricht hingegen für eine Vokalharmonie nach vorderen und hinteren Vokalen. Es
32. Da das Deutsche ebenfalls gerundete Vordervokale besitzt, kann es zum Vergleich herangezogen werden. Kurzes ü und wenig tiefer langes ü, sowie kurzes und langes ö liegen etwa um 1500 Hz, a noch tiefer, etwa 200 Hz unter ü. Die absoluten Zahlen schwanken natürlich schon wegen der Körpergröße und individuellen Gewohnheiten. Aber es erscheint doch als Unwahrscheinlich, dass auch die relative Verteilung so weit verändert wird, dass eine ganz andere Ordnung deutlich hervortritt. Zu den Frequenzen siehe A. Rausch, „Untersuchungen zur Vokalartikulation im Deutschen,“ Beiträge zur Phonetik von Heinrich Kelz und Arsen Rausch, IPK Forschungsberichte 30 (1972) 35–82. 33. Zu beachten ist, dass wir eine Vokalharmonie nur nach den Lippen konstruieren wollen, weil wir nur zwei Zustände brauchen. Es ist aber möglich, mit Hilfe der Lippen und der ebenfalls auf den zweiten Formanten wirkenden Artikulationsstelle eine Vokalharmonie für 4 Zustände zu konstruieren, wie es sie als eine von zwei Vokalharmonien des 2. Formanten im Türkischen gibt (große Vokalharmonie, s. u.). Eine weitere Mischform hält das Ungarische bereit, das neben seiner Vokalharmonie nach vorderen und hinteren Vokalen noch eine Dreiteilung kennt, nämlich eine Zweiteilung der vorderen Gruppe in ungerundete und gerundete und eine hintere Gruppe aus a und gerundetem o, u. Entscheidend für die Möglichkeit solcher Bildungen ist aber ganz oˆensichtlich, dass die Vokale einer Gruppe jeweils einen klar von den anderen Gruppen geschiedenen Abschnitt auf dem zweiten Formanten besetzen. Was im Türkischen und Ungarischen selbstverständlich der Fall ist.
9
gibt jedoch ein Argument an dem eine solche Vokalharmonie als Erklärung für e-/ì- ebenfalls scheitert. Sehen wir uns dazu die Kardinalvokale i, a, u an, von denen wir annehmen können, dass sie das Sumerische ebenso wie das Akkadische besaß. Ob wir nun eine Vokalharmonie nach vorderen und hinteren Vokalen oder nach der Mundöˆnung annehmen, es wird immer ì- . . . (i), e- . . . (a) herauskommen, jedoch nur bei Poebels Ansatz ì- . . . (u). Schauen wir uns noch einmal unsere Tabelle an, so haben wir 6 Verben mit e- . . . (u) gegen 22 Verben mit ì- . . . (u). Der Kardinalvokal u würde sich also bei einer Vokalharmonie nach vorderen und hinteren Vokalen nur in 6 Fällen verraten. Das ist zu wenig für die insgesamt 65 Verben und vor allem zu wenig im Verhältnis zu den 22 Verben mit ì- . . . (u). Damit ist die Entscheidung zu Gunsten einer Vokalharmonie nach der Mundöˆnung gefallen. Es stellt sich die Frage, ob diese Vokalharmonie nun auch die sumerischen Vokale eindeutig auf Poebels System festlegt. Geht man davon aus, dass die Mundöˆnung, bzw. der 1. Formant immer übereinstimmen müssen, so kommt in der Tat nur Poebels System oder eine Erweiterung in Frage. Man kann nämlich annehmen, dass das Sumerische mindestens die drei Kardinalvokale i, a, u besaß. Poebels Vokale sind dann die minimale Erweiterung, zu einem Vokalsystem, das auf die altsumerische Vokalharmonie maßgeschneidert ist, weil es zu jedem Kardinalvokal genau einen im wesentlichen nur durch die Mundöˆnung unterschiedenen Vokal enthält. Nehmen wir an, dass das System zwar aus zwei Gruppen von Vokalen bestand, dass aber nicht alle Vokale über ein Gegenüber in der anderen Gruppe verfügten. Aus der oben gegebenen Statistik sehen wir, dass Poebels geschlossenes ê, das mehr oder weniger in der Mitte über a zu denken ist, kein häu˜ger Vokal ist. Ein Verzicht auf ê würde auch die Symmetrie weniger stören, als ein Verzicht auf o, zu dem es außerdem einige Verben gibt, deren Lesung durch den Auslaut gesichert erscheint und die e- zu /u/ in der Wurzel haben. Zählt man nun alle Verben mit ì- und mit e- zusammen, so kommt man auf 34 (ì-) zu 31 (e-), mit etwas mehr Unsicherheiten bei den letzteren.
10
JAN KEETMAN
Ohne ê hätten wir aber ein umgekehrtes Verhältnis bei den Vokalen von 2 zu 3. Das ist keine Abweichung, die ganz ausgeschlossen erscheint, aber doch eher unwahrscheinlich ist. In Poebels System lässt sich auch ein Problem unserer Statistik lösen, welches Poebel bei seiner Konstruktion nicht im Auge hatte, die Seltenheit von Verben mit e- . . . (e) und e- . . . (u). Überraschend ist vor allem, dass es für e- . . . (e) nur zwei wirklich sichere Verben und bei großzügiger Zuteilung vielleicht 5 Verben gibt. Andererseits kannte das Akkadische ein e. Schon dieses grobe Missverhältnis beweist, dass unser Umschriftsystem mit den vier Vokalen i, e, a, u unmöglich stimmen kann. Auch eine Erweiterung auf 5 Vokale lässt e- . . . (e) noch immer stark unterbesetzt erscheinen. Dieser Mangel an entsprechenden Wurzeln mit /e/ kann nicht auf die teilweise unsichere Wiedergabe von e und i in der Keilschrift zurückgeführt werden, denn es gibt keine Beispiele für e- . . . (i). Die fehlenden Wurzeln können nur in e- . . . (a) stecken. Wenn wir die Vokale so anordnen, dass das für die Vokalharmonie relevante Merkmal Mundöˆnung betont wird, so rückt ‰ näher an a. Akkadisches e dürfte aber der Tendenz gleichmäßiger Verteilung im Mundraum folgend, mehr in der Mitte zwischen i und a gestanden haben. Daher könnte ‰ in manchen Fällen im Akkadischen zu a geworden sein. Für einen Wechsel von a und e gibt es viele Beispiele, nur die Vorliebe der Sumerologen für den Vokal i lässt ihn als i/a-Wechsel erscheinen. Z. B. in díb/dab5 „packen“. Obwohl die Glosse di-ib genauso de-eb zu lesen ist,34 wurde díb gewählt.35 Nach dieser willkürlichen Bevorzugung
34. Proto-Ea 19 bietet da-ab, ebenso YBC 7158, 31 (MSL 14, 30 u. 141). Dagegen haben Ea I 156 und I/3 117 DI.IB (MSL 14, 184 u. 219). Proto-Aa 65 liest ebenfalls DI.IB (MSL 14, 91). Für das Altsumerische gibt es einen weiteren Anhaltspunkt für eine Lesung /dib/ bzw. /deb/, nämlich die Schreibungen pa5-ti-bir5-ra (Ent. 79 iv 8, v 6) und bàd-DÍB-ra (RGTC 1, 23) für den Ortsnamen Badtibira Dabei legt TI, wie wir gesehen haben, den Vokal nicht unbedingt auf i fest. 35. Vgl. sìg/sàg, sig 11 /ság, bí(r)/bar 7, sig 5/sa 6(-g); limmu, límmu/lammu, lámmu „4 “. Syllabische Schreibungen si-im für saG in Ur-Namma B 52, Text B. Vgl. Esther FlückigerHawker, Urnamma of Ur in Sumerian Literary Tradition,
One Line Long
von i passte díb nicht mehr zum altsumerischen Prä˜x e-36 und die Glosse da-ab bzw. die Umschrift dab5 wurden bevorzugt. Auch die anderen Auˆälligkeiten unserer Statistik lassen sich in Poebels System erklären. Wie ‰ konnte auch o Beispiele an a abgeben. Die relative Häu˜gkeit von ì- . . . (u) wird zumindest zum Teil erklärt, wenn man annimmt, dass sumerisches i wegen der Opposition zu ‰ etwas weiter hinten artikuliert wurde. Akkadisches i könnte aber weiter vorne gestanden haben und damit weiter entfernt von ê, das daher Beispiele an u abgab. Jedoch ist eine Erweiterung nicht auszuschließen. Wegen der Betonung der Mundöˆnung sollte das vordere Vokalpaar i/‰ nahe an a heranrücken. Damit wird der Platz für eine Erweiterung eng. Wir können aber an die gleiche Stelle gerundete Vordervokale ü/ö setzen, da sie sich durch die Lippenrundung genügend unterscheiden. Dieses System lässt die Häu˜gkeit von ì- . . . (u) noch etwas plausibler erscheinen und wird auch dem Problem der Homophone im Sumerischen eher gerecht (s. u.).
Silbenzeichen Man kann sich nun fragen, ob sich mehr als 4 Vokale vielleicht auch durch die Silbenzeichen rechtfertigen lassen. Nun stellen die sumerischen
OBO 166 (Freiburg: Biblical Institute of the University of Fribourg u. a., 1999). Es fällt auf, dass in allen diesen Beispielen im Akkadischen keine Zeichen wie si, mi vorliegen, die weitgehend auf /i/ festgelegt sind. Einen Grund, warum bei diesem Wechsel der von a stärker abweichende Vokal i einseitig bevorzugt wird, kann ich nicht erkennen. Fraglich erscheint mir das Beispiel des früher me-lám, heute meist me-li9-(-m) geschriebenen Schreckensglanzes. Für beide Aussprachen gibt es Belege, doch besteht die Möglichkeit einer Entwicklung me-lám > melammum > melemmum im Babylonischen, die sich auch auf das Sumerische ausgewirkt haben könnte. Für e/a nach s/m im Sumerischen vgl. musen-dù/usandu > us/sandû „Vogelfänger,“ urudusen-del > sandalu aB Mari (ein Behälter), me/ma6 „Me“ und se21/sa22 als Varianten zu sa4 „nennen.“ 36. Ukg. 1 iv 11; 14; 17; 4 iii 3; 6 ; 8; 10; 13; AWL 2 iii 2f. VS 25, 81, Rs. Iii 2; VS 25, 105 ii 3; Rs. iii 2 Nik. 1, 30 Rs. viii 3 Nik. 1, 32 ii 9; Rs. iv 3; v 2 usw.
DIE ALTSUMERISCHE VOKALHARMONIE UND DIE VOKALE DES SUMERISCHEN
Silbenzeichen kein System dar, das ersonnen wurde, um einen beliebigen Ausdruck im Sumerischen phonetisch oder phonologisch korrekt wiedergeben zu können. Außerdem besteht die Möglichkeit, dass ein Wechsel in der Schreibweise keinen Wechsel der Aussprache, sondern des zugrundeliegenden Morphems meint. Vgl. englisch I und eye. Deshalb ist eine Argumentation, die von den Silbenzeichen ausgeht, um auf die Phoneme zu schließen, grundsätzlich mit Vorsicht zu betrachten. Wir wollen uns auf ein Beispiel beschränken.37 Besonders deutlich für eine Erweiterung spricht altsumerisch das Paar ne/né: Vgl. e-ne-ba „sie hat es ihnen zugeteilt,“38 e-né-ba-e „er wird es (ihr) zuteilen,“39 a-ne „er/sie,“ -a-né „sein/ihr,“ -e-ne Pluralsu¯x,40 an-né, en-né für den Ergativ. Dabei werden weder -a-né noch -e-ne verändert, wenn sie in einen mit /-e/ markierten Kasus treten, 41 obwohl man eine Kontraktionslänge 37. Stephen J. Lieberman hat bereits aus Betrachtungen zu Proto-Ea (MSL II) auf einen Vokal o im Sumerischen geschlossen. Er kommt zu ú = [u], u, ù, u4 = [o]. Siehe Lieberman, „The Phoneme /o/ in Sumerian,“ in: M. A. Powell u. Ronald H. Sack (Hrsg.), Studies in Honor of Tom B. Jones, AOAT 203 (Neukirchen-Vluyn 1979) 21–28. Marie-Louise Thomsen, Sumerian Language, 40, wendet ein, es könnte sich auch um einen Re˘ex auf ein akkadisches Phonem handeln. Letztlich könnten beide Recht haben. Schon dass Lieberman drei Vokalzeichen für [o] reklamiert, spricht sogar für noch mehr gerundete Vokale. Zu /u/ gehören dann noch mindestens u5, u6, u8, wobei es wohl Überschneidungen in der Aussprache gibt und Diphtonge bei Wortzeichen möglich sind. Trotzdem ist die Zahl von 7 Zeichen und nur zwei bisher angenommenen Lesungen verdächtig. 38. DP 130 xv 10. Beachte, dass e- nicht einfach für e-/ìsteht, sondern aus *ì- + e-ne(-a) entstanden ist. 39. DP 133 i 6. 40. Die Pluralformen a-ne-ne (jünger e-ne-ne) „sie“ -a-nene „ihr“ sehen so aus, als wären sie aus a-ne + e-ne bzw. -a-né + e-ne entstanden. Doch die 2. Person Plural -zu-ne-ne verträgt sich schlecht mit dieser Deutung. Eine einfache Addition von -e-ne hätte auch zu dem Problem geführt, dass nicht klar ist, ob sich der Plural auf das Su¯x oder das Nomen bezieht: dumu-zu-ne heißt „deine Kinder“ und die gleiche Form für „euer Kind“ wäre verwirrend. Deshalb wohl die Bildung -zu-ne-ne. Die dritte Person wird altsumerisch mit (-a)-ne-ne gebildet, belegt durch gi-gù-na-ne-ne mu-ne-dù „er hat ihnen ihre Ziqqurrat gebaut“ En. I 29 v 1–2, ob aus a-né + e-ne oder a-né + (e-)ne-ne ist nicht zu entscheiden. Weil das Problem der Verwechslung in der 3. Person nicht weniger bestand, ist 41. So altsum. wohl immer. Bei Gudea gibt es die Ausnahme lugal-ki-áG-né-e „sein geliebter König“ (Erg.) Statue B v 24. anzunehmen, dass die Form mit drei /ne/ einfach zu lang
11
erwarten könnte. Unterschiedlich verhalten sich hingegen die beiden Endungen vor dem -a(k) des Genitivs: é-a-na-ka „im Hause seines Vaters,“42 GÁNA sa6-ga dingir-ré-ne-ka „auf den fruchtbaren Feldern der Götter.“43 Gegen eine rein graphische Diˆerenzierung spricht, dass weder ne noch né auf eine besondere Form beschränkt sind und dass sich das „e“ tatsächlich unterschiedlich verhält. Eine Unterscheidung der Vokallänge erklärt den Gebrauch vor /-e/ nicht. Vokalharmonie innerhalb der Wurzeln Der Umstand, dass der gleiche Vokal in Wurzeln nicht immer, aber auˆallend oft wiederholt wird, hat G. Gragg dazu geführt, dies als Vokalharmonie zu beschreiben: „ . . . polysyllabic words rarely contain both a low and a high, or a back and front vowel.“ 44 Die auˆallende Wiederholung des gleichen Vokals bei der Wiedergabe einer sumerischen Wurzel mit akkadischer Keilschrift kann auch kaum als Au˘ösung von Doppelkonsonanz erklärt werden. Zum einen macht diese Tendenz oˆenbar nicht vor Worten halt, bei denen die Au˘ösung einer Doppelkonsonanz nicht anzunehmen ist, wie aga „Tiara“, eme „Zunge“, izi „Feuer“, udu „Schaf “ usw. zum andern wäre bei Au˘ösung von Doppelkonsonanz am ehesten ein Murmelvokal, geschrieben als i oder e zu erwarten.45
erschien. Es wäre also auch eine Verkürzung aus *-a-né-ne-ne denkbar. Falls -a-né + e-ne vorliegt, ist die Veränderung zu ne jedenfalls nicht als Beweis dafür zu werten, dass der Wechsel des Zeichens eine Vokallänge ausdrückt, da né + „e“ sonst keine Schreibung als ne verursacht. Dies spricht dafür, dass das „e“ des Lokativ-Terminativ und des Ergativ von dem „e“ in e-ne- verschieden war. Ab Ur III ist die Kombination von a-né und e-ne direkt bezeugt: dumu-ne-ne „ihre (3. Sing.) Kinder“ NG 75, 22; 171, 5; 212, 15; 19; 43; dumu-n[itá dumu-mí]-ne-ne „seine Söhne und Töchter“ NG 83, 12; nam-dumu-ne-ne-sè, Inannas Gang zur Unterwelt 230. 42. AnUr 17, 4u 43. Ukg. 4 iv 13f. 44. Gragg, In˜xes, 41. 45. Beachte dass dies auch der akkadischen Praxis entspricht. Vgl. GAG § 8d. Der Zusatz GAG § 65d*, wonach auch vor Su¯x eingeschobenes a als Murmelvokal gesprochen wurde (ohne Begründung), überzeugt nicht. Es gibt keine
12
JAN KEETMAN
Aber angenommen, es stünde beispielsweise bára für /bra/ und solche Doppelkonsonanten wären im Sumerischen nicht selten. Dann könnten wir nicht sicher, aber doch mit einiger Wahrscheinlichkeit auch von solchen Worten abgeleitete Silbenzeichen erwarten. Diese würden von uns als K1VK2V-Zeichen wahrgenommen. Das Fehlen solcher Silbenzeichen ist also ein weiteres Argument gegen verbreitete Doppelkonsonanz im Sumerischen. Auch akkadische Lehnworte im Sumerischen und Lehnworte, die aus anderen, meist unbekannten Sprachen in beide Sprachen übernommen wurden, zeigen im Sumerischen eine Tendenz zur Vereinfachung der Vokale, wobei meistens ein vorderer an einen hinteren Vokal angeglichen wird. Z. B. na-gada < naqidum, „Hirte,“ ma-al-ga/Galga < milkum, „Rat,“ isib < wasipum, „Beschwörer,“ niGir < nagirum, „Herold,“ ugarum ~ a-gàr, wohl verwandt mit gr. agrovÍ und anderen idg. Wörtern für „Acker“;46 siparrum ~ zabar „Bronze;“ pilaqq/kkum ~ Gisbala, „Spindel,“47 Gisù-suh5 ~ as¿hum, „Tanne?.“ Es stellt sich die Frage, ob man nicht einfach von identischen Vokalen ausgehen kann. Ein Lautgesetz, das für alle Wurzeln vorschreibt, dass jeweils nur ein Vokal gebraucht werden darf, würde aber eine extreme Einschränkung bedeuten und ist nicht belegt.48 Wir müssen also nach einer Vokalharmonie suchen, die die Auswahl von Vokalen für die Wurzeln gestattet. Die von uns bereits für die Prä˜xketten gebrauchte Vokalharmonie nach der Mundöˆnung scheidet aber prinzipiell aus. Das liegt an dem
Verlauf des 1. Formanten, auf den die Mundöˆnung wirkt. Dieser steigt von u über o nach a an und fällt dann über e nach i ab. Egal ob man die Mundöˆnung wie bei Poebels Vokalharmonie in zwei oder in drei Stufen zerlegt, immer kommen die so unterschiedlichen Vokale u und i in eine Gruppe. Der Sinn einer Vokalharmonie besteht aber oˆenbar darin, ähnliche Vokale zusammenzuführen. 49 Daher sollte eine Vokalharmonie innerhalb der Wurzel ihre Vokale unbedingt nach dem 2. Formanten gruppieren.50 Dieser steigt von u nach i kontinuierlich an. D. h. Vokale, die auf ihm benachbart sind, sind sich auch in ihren übrigen Eigenschaften ähnlicher. Dies dürfte ein weiterer Grund sein, warum Vokalharmonien nach dem 2. Formanten weit verbreitet sind. Im Sumerischen scheidet für die Wurzeln eine Vokalharmonie nach der Mundöˆnung schon deshalb aus, weil Kombinationen von u und i erheblich seltener sind als Wurzeln, die i oder u wiederholen. Nun kann man die Vokale Poebels entsprechend dem 2. Formanten und das heißt zugleich weitgehend entsprechend ihrer Lage vorne und hinten im Mund, in drei Paare ordnen: {u, o}, {a, ê}, {E, i}.51 Bei der Transformation in akkadische Vokale werden die ersten beiden Paare meistens (u, u) und (a, a) liefern. {E, i} sollte hingegen (e, e), (i, i), (e, i), (i, e) ergeben, denn hier hat das Akkadische einen für eine Alternative günstigen Vokal e. Wegen der nicht ganz gesicherten Unterscheidbarkeit von e und i in der Keilschrift sticht die Abweichung in den letzten beiden Paaren aber weniger ins Auge. Wie in modernen Sprachen mit Vokalharmonie mag es Ausnahmen gegeben haben. Allerdings
Erklärung für die unterschiedliche Schreibweise und warum dieser Vokal der assyrischen Vokalharmonie unterlag. Die Vokale a und u bedeuten auch, durch die weiteste Mundöˆnung bei a und durch die gerundeten Lippen bei u eine erhebliche Veränderung gegenüber einer neutralen Mundstellung für einen Murmelvokal. 46. Vgl. Gonzalo Rubio, „On The Alleged ‘Pre-Sumerian Substratum’,“ JCS 51 (1999) 10. 47. Das Akkadische geht vielleicht auf eine ältere sumerische Form zurück. Vgl. nukaribbum < nu-kiri6 „Gärtner.“ 48. Siehe C. Boisson, “The Phonotactics of Sumerian,” in Indo-European, Nostratic and Beyond, Fs. V. V. Shevoroshkin, I. Hegeds et al. (Hrsg.), Journal of Indo-European Studies, Monograph 12 (Washington: Institute for the Study of Man, 1997), 30–50; speziell 41.
49. Selbst wenn das Gegenteil, sei es nun sinnvoll oder nicht, nämlich möglichst heterogene Vokalkombinationen erreicht werden sollte, geht das mit einer Vokalharmonie nach der Mundöˆnung nicht. Zwar sind die Vokale mit niederem 1. Formanten sehr heterogen, die mit hohem aber eher homogen. 50. Beachte dass in diesen Schluss unsere Voraussetzung eingeht, dass nicht nur einzelne Vokale an der Vokalharmonie beteiligt sind. 51. Wir verwenden die Schreibweise für mathematische Mengen mit geschweiften Klammern { }, wenn wir Vokale lediglich zusammenfassen wollen. Legen wir aber Wert darauf, dass sie in einer bestimmten Reihenfolge als Paar auftreten, weil sie in dieser Reihenfolge in einem Wort vorkommen, dann benutzen wir ebenfalls analog zur mathematischen Schreibweise runde Klammern ( ) für geordnete Paare.
One Line Long
DIE ALTSUMERISCHE VOKALHARMONIE UND DIE VOKALE DES SUMERISCHEN
ist kaum zu entscheiden, ob etwa das Wort agrig „Hausverwalter“ als agrig, „agreg“ bzw. „agrêg“ zu interpretieren ist. Damit lässt sich auch kein Urteil über regelmäßige und unregelmäßige Formen fällen. Es fällt immerhin auf, dass {a, i} besonders häu˜g52 erscheint, so dass in diesem Schema auch regelmäßige Formen zu erwaten sind. Wieder funktioniert die Vokalharmonie mit Poebels 6 Vokalen und man kann sich damit zufrieden geben. Nun zeigt aber die sogenannte große Vokalharmonie des Türkischen, dass in einer Vokalharmonie nach dem 2. Formanten zwischen {a, ê} und {E, i} noch sehr gut Platz für ein Paar {ö, ü} ist. Dies führt dann auch zu einer etwas anderen, aber ebenfalls sinnvollen Beschreibung der Gruppenbildung. Während man die Aufteilung aufgrund von Poebels Vokalen als Gruppierung an den Stellen der drei Kardinalvokale beschreiben kann, kann man die vier Gruppen als Aufteilung gemäß vorderen und hinteren und gerundeten und ungerundeten Vokalen verstehen. Damit wird auch das Prinzip, einem Parameter bei der Artikulation eines Vokals, zwei entgegengesetzte Stellungen zuzuordnen, das wir bereits auf die Mundöˆnung angewendet haben, erneut aufgegriˆen. Wieder stellt sich die Frage, warum an dieser Stelle erneut die gerundeten Vordervokale ins Spiel kommen. Der Grund ist ganz einfach. Die Vokalharmonie schränkt die Möglichkeiten, mit den Vorhandenen Vokalen zweisilbige Wurzeln zu bilden erheblich ein. Mit 6 Vokalen ließen sich theoretisch 36 Paare bilden. Mit der Einschränkung durch die Vokalharmonie nur noch 12. Das legt eine Erweiterung nahe und wir können feststellen, dass ü, ö eine Erweiterung darstellt, die keine völlig neuen Bildungsweisen erfordert und zugleich gut möglich ist. Damit stiege die Zahl der möglichen Kombinationen wenigstens auf 16. Zwei Vokalharmonien in einem Mund? Wir haben festgestellt, dass für die auf die Prä˜xketten der Verben wirkende altsumerische 52. Vgl. Piotr Michalowski, “Sumerian,” in The Cambridge Encyclopaedia of the World’s Ancient Languages, R. D. Woodard, (Cambridge: Cambridge University Press, 2004), 19–59, 30.
13
Vokalharmonie nur die Mundöˆnung als Auswahlkriterium in Frage kommt. Eine Übertragung dieses Prinzips auf die Auswahl der Vokale für die Wurzeln ist aber auszuschließen. Man kann sich nun fragen, ob das noch als ein in sich stimmiges Ergebnis anzusehen ist. Einmal mussten die Sumerer auf den ersten Formanten achten und im nächsten Moment auf den zweiten. Der Wechsel ist jedoch keineswegs willkürlich, sondern korrespondiert mit anderen Merkmalen des Sumerischen. Anführen kann man einen fundamentalen Unterschied. An Nomen können nur Su¯xe angefügt werden, Verben werden zwar auch mit Su¯xen verbunden, zugleich werden sie aber noch mit Prä˜xketten ausgestattet, die häu˜g aus mehreren Gliedern bestehen. Die Vokalharmonie innerhalb der Wurzeln wirkt sich aber vor allem oder sogar ausschließlich bei den Nomen aus. Verbale Wurzeln sind häu˜g einsilbig. Von den zweisilbigen dürften, wenn man die Vokalharmonie auf sie anwendet, 50% den gleichen Vokal aufweisen. Darüber hinaus wäre es möglich, dass es um den Wechsel zweier Prinzipien zu vermeiden, die Tendenz gab, neu eingeführte Verbalwurzeln zu monovokalischen zu vereinfachen. D. h. soweit überhaupt neue Verbalwurzeln eingeführt werden, denn im Sumerischen gibt es eine Alternative, die Bildung zusammengesetzter Verben, wie ki . . . áG „lieben,“ á . . . áG, „befehlen,“ saG . . . rig7 „schenken“ (< sarakum !). Das reduziert natürlich die Zahl notwendiger Verbalwurzeln und erleichtert es, mit einsilbigen und monovokalischen auszukommen. Es liegt also eine besondere Situation vor. Die Verbalwurzeln sind extrem einfach, entweder bestehen sie nur aus einer Silbe oder es liegen zwei Silben vor, die aber häu˜g oder immer mit dem gleichen Vokal gebildet sind. Daher kann auf die Prä˜xkette ein Typ von Vokalharmonie wirken, der zur Auswahl von Vokalen für die Wurzel ungeeignet wäre. 53 Der Wechsel der beiden Typen von Vokalharmonie geschieht also
53. Implizit ergibt sich so auch der vermutlich wichtigste Grund für die weite Verbreitung von Vokalharmonien nach dem 2. Formanten. Nur solche Vokalharmonien können sowohl auf die A¯xe wirken und zugleich zur Auswahl der Vokale der Wurzeln gebraucht werden.
14
JAN KEETMAN
nicht innerhalb eines Wortes und in einem Satz höchstens an einer, ohnehin herausgehobenen Stelle. Andere Vokalharmoniegesetze Wir haben uns am Anfang auf Vokalharmonien eingeschränkt, bei denen das ganze Vokalsystem beteiligt ist. Diese Systeme sind in Eurasien weit verbreitet. Innerhalb der entsprechenden Sprachen spielen die Vokalharmonien eine große Rolle, was man auch vom Sumerischen sagen kann, wenn man zu den Prä˜xen noch die Vokalharmonie in den Wurzeln hinzunimmt. Eine Vokalharmonie nach der Mundöˆnung ist mir aus diesem Zusammenhang aber nicht bekannt. Es ist schwer zu sagen, welche Bedeutung dieser Unterschied hat. So verwendet das Türkische für die A¯xe beide Parameter, die sich im Sumerischen nachweisen lassen, nur genau umgekehrt. Die altsumerische Vokalharmonie bildet Gruppen nach der Mundöˆnung, die durch ihre vorderen Vokale vertreten werden. Das Türkische bildet Gruppen von vorderen und hinteren Vokalen (bei der „großen Vokalharmonie“ noch in gerundete und ungerundete unterteilt), die bei der „kleinen Vokalharmonie“ durch ihre oˆenen (ungerundeten) Vokale e, a vertreten werden und bei der „großen Vokalharmonie“ durch ihre geschlossenen Repräsentanten i, ü, ı, u. Z. B. Izmir’e „nach Izmir,“ Izmir’e mi? „nach Izmir?“ In den Wortwurzeln berücksichtigen alle bisherigen Beispiele ohnehin den 2. Formanten. Eine Zuordnung nach der Mundöˆnung haben dagegen die Bantu-Sprachen.54 Diese in Afrika weit verbreitete Sprachgruppe, der weit über 100 Einzelsprachen zugezählt werden, besitzt ein auˆallend einheitliches Vokalsystem. Wenn man zusätzliche Parameter, wie Länge und Nasalierung außer Acht lässt, so gibt es ein vermutlich ursprüngliches System von 7 Vokalen und ein reduziertes von 5 Vokalen, selten 9 Vokale.55 Die 54. Ich verdanke den Hinweis auf die Bantu-Vokalharmonien Bram Jagersma, der damit wesentlich zur Vervollständigung dieses Artikels beigetragen hat. 55. Hierfür und für das folgende vgl. Larry Hyman, “The Historical Interpretation of Vowel Harmony in Bantu,” in Bantu Historical Linguistics: Theoretical and Emperial Perspectives, Jean-Marie Hombert u. L. Hyman (Hrsg.), CSLI
One Line Long
Vokale sind symmetrisch zu a, die hinteren gerundet, die vorderen ungerundet. Nach der Höhe kann man die 5 Vokale in drei Stufen (i/u, e/o, a) und entsprechend die 7 Vokale in vier Stufen einteilen. Die Vokalharmonien in den BantuSprachen bedeuten, dass ein Vokal an eine bestimmte Stufe angeglichen wird, an die anderen jedoch nicht. Z. B. wird im Luganda i vor e/o zu e, bleibt aber vor u, i und vor a erhalten. Dazu gibt es den weiteren Fall, dass nur an den hinteren Vokal angeglichen wird, der dann auch folgerichtig eingesetzt wird. Z. B. bleibt im Nyamwezi u nach 6 der 7 Vokale einschließlich e erhalten, wird aber nach o ebenfalls zu o. Die Vokalharmonien der Bantu-Sprachen gleichen Su¯xe an die Wurzel an, seltener Prä˜xe. Auch im Stamm gibt es Vokalharmonien. Diese sind aber weit weniger ausgeprägt als im Sumerischen. Das System der Bantu-Vokalharmonien lässt sich nicht direkt auf das Sumerische übertragen. Da e- vor /a/, also eindeutig vor einer anderen Stufe steht, wäre e- der ursprüngliche Vokal. Also stünde ì- nur vor einer, e- vor mindestens zwei Stufen, im Gegensatz zur Häu˜gkeit von e- und ì-. Dieses Problem wird umso gravierender je mehr Stufen wir annehmen. Andererseits ist im einfachsten Bantu-System mit 5 Vokalen die Seltenheit des Schemas e- . . . (e) nicht zu erklären. Das Bantu-System vor Augen, bei dem quasi immer nur eine Stufe ein Signal weitergibt, die anderen aber wirkungslos bleiben, lässt sich theoretisch auch ein anderes System denken: Beiden Vokalen werden mehrere Stufen zugeordnet. Ein Beispiel für ein solches System aus einer lebenden Sprache ist mir nicht bekannt.56 Tauscht man aber die Einteilung in vordere und hintere Vokale gegen eine Einteilung in obere
Lecture Notes Number 99 (Stanford: Center for the Study of Language and Information, 1999), 235–97 (Literaturhinweis Jagersma). 56. Scheinbare Beispiele aus afrikanischen und amerikanischen Sprachen haben in Wirklichkeit wohl die Pharyngalisierung nicht die damit häu˜g korrespondierende Mundöˆnung als Kriterium ihrer Vokalharmonien. Für das Sumerische müsste man dies eigens diskutieren. Allerdings sind die meisten Argumente bei der Mundöˆnung die gleichen, so dass wir uns auf diese beschränken können.
DIE ALTSUMERISCHE VOKALHARMONIE UND DIE VOKALE DES SUMERISCHEN
und untere, so ergibt sich der Vorschlag als einfache Umformulierung. Z. B. kann man die „kleine Vokalharmonie“ des Türkischen als Modell benutzen. Die vorderen Vokale werden auf e, die hinteren auf a abgebildet, oder anders formuliert: die vordere und die hintere Gruppe werden jeweils auf ihren untersten ungerundeten Vokal abgebildet. Entsprechend umformuliert könnte es für das Sumerische lauten: Die obere und die untere Gruppe werden jeweils auf ihren obersten ungerundeten Vokal abgebildet. Auf diese Weise lässt sich eine Alternative zu Poebels System, bzw. dessen Erweiterung formulieren. Ein System mit 7 Vokalen und vier Stufen, von denen die oberen beiden ì-, die unteren e- verlangen, könnte den Mangel bei e- . . . (e) vielleicht erklären, schon weil theoretisch nur noch 9,3 Beispiele auf einen Vokal entfallen und weil ein unteres „e“ existiert, das zwischen akkadischem e und a anzunehmen ist. Bei 9 Vokalen rückt das entsprechende „e“ noch näher an a und das Ergebnis für e- . . . (e) wird einsichtiger. Wir müssten drei obere und zwei untere Stufen annehmen, denn bei nur zwei oberen Stufen würde ì- . . . (u) für nur zwei von 9 Vokalen stehen, obwohl 22 von 65 Verben zu diesem Schema gehören. Mit diesem System lässt sich, anders als bei 7 Vokalen, auch ein ausgewogenes Schema für die Vokalharmonie in den Wurzeln aufbauen: Wir nehmen an, dass in einer Wurzel nur entweder die hinteren drei, die mittleren drei oder die vorderen drei Vokale geduldet werden.57 D. h. wie bei Poebel nehmen wir auch in diesem System zusätzlich eine Vokalharmonie nach dem 2. Formanten für die Wurzeln an. Die scheinbare Tendenz zur Wiederholung des gleichen Vokals in zweisilbigen Wurzeln, bzw. in mehrsilbigen Worten kann bei der Wiedergabe durch akkadische Silbenzeichen bzw. als Folge einer Angleichung ans Akkadische erklärt werden. Zur Erklärung der vielen Homophone in den Umschriften58 liefert dieses Modell reichlich Vokale.59
57. Wie bereits bemerkt lassen wir Ausnahmen zu. 58. Es ist nicht damit getan, dass man durch Langlesungen, die Annahme eines geschwundenen h oder in ähnlicher Weise die Zahl der Homophone etwas reduziert. Eine Sprache, die ihre phonologischen Möglichkeiten zur Wortbildung häu˜g bis an ihre Grenzen strapaziert, ist ebenfalls auˆallend. Rein
15
Dieser von Poebel nicht betrachtete Lösungsweg hat nur einen gravierenden Nachteil, es gibt keinen Grund für die Verschiebung von Beispielen zwischen der vorderen und hinteren ì-Gruppe. Theoretisch sollten alle drei Gruppen jeweils ca. 22 Verben enthalten. Nach unserer Tabelle sind es 12 : 31 : 22. Der Mangel an Verben mit vorderen Vokalen ist auˆallend aber nicht so gravierend, dass man deswegen dieses System prinzipiell ausschließen könnte. Ebenfalls auˆallend, aber schwer zu bewerten, ist folgende Beobachtung: Der erste Vokal hinter a steht diesem viel näher als u. Er würde daher im Akkadischen kaum als u wiedergegeben, wenn er sich von u auch noch dadurch unterscheiden würde, dass er nicht gerundet war. D. h. aus der Existenz von 6 Verben, die zum Schema e- . . . (u) gehören, kann darauf geschlossen werden, dass die Grenze zwischen gerundeten und ungerundeten Vokalen sowohl von der Vokalharmonie der Prä˜xe als auch der Wurzeln ignoriert wurde. Bei der Bedeutung, die die Lippenrundung für die Bildung der Vokale hat, überrascht dies. Eine andere Möglichkeit zwischen den beiden Lösungswegen zu entscheiden, wäre natürlich der direkte Nachweis gerundeter Vordervokale im Sumerischen. Gelegentlich wurde ü wegen unterschiedlicher Angaben zum Lautwert als i/e oder u vorgeschlagen. 60 Bei dieser Argumentation besteht natürlich die Gefahr, dass man Dialektvarianten aufsitzt, auch wenn sie wegen der grundsätzlich unveränderlichen Wortwurzeln im Sumerischen weniger wahrscheinlich sind. D. h. eine Quelle für Dialektvarianten, unterschiedliche Wortbildungen für den gleichen Begriˆ, wie im Akkadischen gelegentlich bei pirs und purs, scheidet aus. Ein weiteres Argument stützt sich nur auf das Wort nu-gig „Hierodule“ und seine altbabylonische Variante nu-u8-gig. Es ist schwer, für das rechnerisch ist aber die Vermehrung der Vokale der eˆektivste Weg, um die Wurzelsbildungsmöglichkeiten des Sumerischen zu erweitern. 59. Vgl. Anm. 8 zu der Einschätzung, dass Vokallänge vermutlich eine untergeordnete Rolle spielt und also auch nicht ein Mittel war mit dem Homophonie vermieden werden konnte. 60. Z. B. G. Selz, Fs. Kienast, AOAT 274 (Ugarit-Verlag, Münster 2003), 507 Anm. 37.
16
JAN KEETMAN
zusätzliche Zeichen eine Erklärung zu ˜nden. Moderne Sprachen mit Vokalharmonie würden eine Form /nugig/ möglicherhweise irgendwann in /nügig/ oder wenn die Ausgangsform /nogig/ war, in /nögig/ verändern. Nun steht u8 gelegentlich für ein in Nachbarschaft von /u/ abgewandeltes e61 und daher ist nuu8-gig = /nügig/ eine denkbare Erklärung. Ein Gegenargument ist, dass Vokalharmonie zwischen den Bestandteilen zusammengesetzter Wörter im Sumerischen kaum nachzuweisen ist. Sie ist allerdings wegen der Schreibweise auch schwer nachzuweisen. Ein Beispiel stellt Gis-ùr > gus¿ru „Balken“ dar. Insgesamt erscheinen mir jedoch die Argumente für ü bzw. ö bislang nicht ausreichend. Resümee Die altsumerische Vokalharmonie bei den Prä˜xen war mit Sicherheit eine Vokalharmonie nach der Mundöˆnung. Ihr muss aber eine weitere Vokalharmonie innerhalb der Wortwurzeln an die Seite gestellt werden, die die Vokale in vordere, mittlere und hintere einteilte oder in eine vordere und hintere Gruppe, die noch jeweils in ungerundete und gerundete unterteilt wurden. Die von Poebel symmetrisch zur Mundöˆnung konstruierten Vokale i, e, ê, a, o, u bzw. eine Erweiterung dieses Systems um ü, ö sind die nach unserer Analyse wahrscheinlichsten Vokalsysteme. Aber auch ein System mit 7 oder 9 Vokalen, in V-Form mit a in der Mitte ist nicht völlig auszuschließen. Diese Vorschläge sind aus dem Versuch entstanden, alle möglichen Lösungen zu ermitteln und zu bewerten. Bei dieser Bewertung sind auch Faktoren, wie Symmetrieeigenschaften, die Fähigkeit zur Bildung von Wortwurzeln und die statistische Häufung des Vokals u in den Umschriften eingegangen. Diese Faktoren beziehen sich aber nicht notwendig auf den Vokalbestand am Ende von ED III, d. h. kurz vor dem völligen Verschwinden der Prä˜xharmonie aus der Schrift. 61. Eine eingehende Untersuchung über den Gebrauch von u8 wäre wünschenswert.
Es besteht die Möglichkeit, dass sich das System der sumerischen Vokale bereits im Verfall befand62 und diese Frage kann für die Beurteilung des Gebrauchs der Silbenzeichen Bedeutung haben. Daher erhebt sich die Frage, welche minimalen Erkenntnisse sich aus der noch ziemlich regelmäßigen Schreibung der Vokalharmonie ableiten lassen. Nicht zu bestreiten ist ein Vokal zwischen a und u, denn völlig andere Lesungen für die Zeichen der sechs Verben mit e- . . . (u) sind wegen der Kontinuität der Auslaute und der Semantik höchst unwahrscheinlich und eine Interpretation als /hal/, /lah/, /sar/, /tam/, /ar/, /za/ erklärt ihre spätere Lesung mit /u/ nicht. Sehr wahrscheinlich wurden auch noch zwei verschiedene „e“ gesprochen. Allerdings ist das Paradigma ì- . . . (e) wegen der Möglichkeit der Dissimilation etwas weniger belastbar als e- . . . (u). Dass der aus der Nähe eines „e“ an a erklärbare Lautwechsel e/a noch aus späteren Glossen hervorgeht, spricht ebenfalls gegen eine bereits früh erfolgte Verschmelzung der sumerischen Vokale zwischen i und a. Jedenfalls folgt aus der Statistik der Prä˜xharmonie ebenso wie aus dem Umstand, dass eine Vokalharmonie, die zur Wiederholung der Vokale a und u in den Wurzeln zwingt, nicht belegt ist,63 dass unser System der Umschriften mit 4 Vokalen und ebenso eine Erweiterung auf nur 5 Vokale im Sumerischen höchstens als Ergebnis seines Aussterbens möglich sind. Es wäre auch eine große Überraschung gewesen, wenn eine Sprache mit so strengen Regeln für den Bau ihrer meist kurzen Wurzeln, ein so armes Inventar an Vokalen gehabt hätte und dieses überdies dem Akkadischen so ähnlich sehen würde. Auch wenn es nicht gelungen ist, ein bestimmtes Vokalsystem mit hoher Wahrscheinlichkeit als einzig mögliche Lösung nachzuweisen, so hoˆt der Autor doch, weiteren Untersuchungen Nahrung gegeben zu haben.
62. So ein berechtigter Einwand von Bram Jagersma. 63. Ob sich e und i abwechseln können oder nicht, lässt sich nicht eindeutig sagen.
DIPLOMACY AND THE RITUALS OF POLITICS AT THE UR III COURT T. M. Sharlach (Oklahoma State University)
In the ancient Near East, as in more recent times, diplomacy served to link states together, creating a forum to air and resolve disagreements. Diplomacy was largely a man’s world. Its aim was to establish and maintain the state of brotherhood between great kings. 1 On one level, “brotherhood” represented what in modern times would be called cordial relations; on another level, marriage alliances between dynastic families had the result of creating actual kinship relations. No treaties or correspondence among rulers have been preserved from third-millennium Sumer, but a wealth of information about diplomacy survives in the Ur III administrative archives. Such texts suggest that religious observances were also political events. Studies on ancient Near Eastern diplomacy have largely drawn on two rich sources of information
from the second millennium, namely the letters from Amarna and Mari. In her 1956 article, “Diplomacy in Western Asia in the Early Second Millennium B.C.,”2 Munn-Rankin compared the Amarna evidence to the data from Mari and concluded that “relations between heads of states were governed by a common code of manners which in essentials was identical with that of the Amarna period.” Should we then assume that diplomatic relations in the third millennium were also governed by a similar set of protocols? A number of important studies have appeared in recent years on diplomatic relations, many focusing on the foreign policy of the Ur III court and on reconstructing the evidence for dynastic marriages between royal houses.3 Yet, surprisingly little has been written
2. M. Munn-Rankin, “Diplomacy in Western Asia in the Early Second Millennium BC,” Iraq 18 (1956) 96. 3. Such studies include J. Boese and W. Sallaberger, “Apil-kin von Mari und die Könige der III. Dynastie von Ur,” AoF 23 (1996) 24–39; R. Cohen and R. Westbrook, eds., Amarna Diplomacy (Baltimore: The Johns Hopkins University Press, 2000); B. Lafont, “Messagers et ambassadeurs dans les archives de Mari,” in La circulation des biens, des personnes et des idées dans le Proche-Orient ancien, ed. D. Charpin (Paris: Editions Recherche sur les Civilisations, 1992), 167–83; T. Maeda, “The Defence Zone during the Rule of the Ur III Dynasty,” ASJ 14 (1992) 135–72; P. Michalowski, “The Bride of Simanum,” JAOS 95 (1975), 716–19; “The Men from Mari,” in Immigration and Emigration within the Ancient Near East, ed. K. van Lerberghe (Leuven: Peeters, 1995), 181–88; “The Ideological Foundations of the Ur III State,” in 2000 v. Chr. Politische, wirtschaftliche und kulturelle Entwicklung im Zeichen einer Jahrtausendwende. 3. Internationales Colloquium der
A version of this paper was delivered in July 2001 at the 47th Rencontre Assyriologique Internationale in Helsinki, Finland. I would like to thank D. I. Owen, W. Sallaberger, and P. Jones for their helpful comments. (Any errors or omissions are of course my own.) Thanks are also due to the Pennsylvania Sumerian Dictionary for permission to quote from the electronic databases. W. W. Hallo and B. Foster generously permitted the publication of certain tablets from the Yale Babylonian Collection here (see Appendix II). Abbreviations follow T. Gomi and M. Sigrist, The Comprehensive Catalogue of Published Ur III Tablets (Bethesda: CDL, 1991), updated by W. Sallaberger, “Ur III-Zeit,” Mesopotamien, OBO 160/3 (Fribourg: Editions Universitaires Fribourg, 1999), 393–402. Ur III month names are (unless otherwise speci˜ed) in the state calendar (R). 1. M. Liverani, Prestige and Interest (Padova: Sargon, 1992), 197–202.
17
JCS 57 (2005)
18
T. M. SHARLACH
about the topic of diplomacy itself in the last centuries of the third millennium B.C. and the intersection between diplomacy and ritual has hardly been explored. The Diplomatic Corps Foreign emissaries appear in a variety of Ur III documents, which tell us something of their lives in Babylonia and suggest the shape of the organized code of manners for diplomacy in existence at this time. Delegates came to Babylonia from all directions, north, west, south, and east; from neighboring lands which were often allied with the court but which were not necessarily vassals.4 Clearly the kingdom of Ur belonged to the ˜rst circle of importance, while some of the emissaries represented countries of a lesser rank. The Ur III material thus stands in contrast to the diplomatic activities recorded in the letters of Mari or Amarna, both of which record diplomatic activity aimed at adjusting or maintaining a precarious balance of power in a period of political fragmentation.
Deutschen Orient-Gesellschaft 4.–7. April 2000 in Frankfurt/ Main und Marburg/Lahn, eds. W. Meyer and W. Sommerfeld (Saarbrücken, 2004), 219–35; and “Iddin-Dagan and his Family,” ZA 95 (2005) 65–76; D. I. Owen and R. Veenker, “MeGum, the the First Ur III Ensi of Ebla,” in Ebla 1975–85, ed. L. Cagni (Napoli: Istituto Universitario Orientale di Napoli, 1987), 276– 91; D. I. Owen, “Syrians in Sumerian Sources from the Ur III Period,” in New Horizons in the Study of Ancient Syria, ed. M. Chavalas (Malibu: Undena, 1992), 108–75; W. Sallaberger, “Ur III–Zeit,” Mesopotamien, OBO 160/3 (Fribourg: Editions Universitaires Fribourg, 1999); M. Sigrist, “Noveaux noms géographiques de l’Empire d’Ur III,” JCS 31 (1979) 166–70 and “Les Courriers de Lagash,” in Fragmenta Historiae Elamicae, ed. H. Gasche et al. (Paris: Editions Recherches sur les civilisations, 1986), 51–62; P. Steinkeller, “More on the Ur III Royal Wives,” ASJ 3 (1981) 77–92, and “The Question of Marhasi,” ZA 72 (1982) 237–65. 4. Places attested in the texts include: Dudduli Mukis Ebla Simanum Gubla Simurrum Harsi †imaski Iabibu Talmus Iabrat Tikitinhi Iamadium Tuttul Mari Ursu Marhasi Zidanum Mahili Zidahri
One [Body] Line Long
The Ur III state contained a branch of government whose business was foreign policy both at home and abroad. In charge of this organization was the sukkal-mah, a term which can be translated literally as “chief secretary” or perhaps better as “secretary of state.”5 Included under his purview was a foreign service, which consisted of trusted emissaries, sent by the king of Ur into foreign lands. These were known in Sumerian as lú-kingi4-a lugal, “royal emissaries.” Indeed, the most famous sukkal-mah, Arad-Nanna served as a lúkin-gi4-a lugal, that is, a royal emissary who went abroad, prior to his promotion to the o¯ce of his father. There was also a domestic branch, which consisted of civil servants who held the title sukkal, “secretary.” In these contexts, sukkal denotes a state employee in the diplomatic corps. Although one frequently ˜nds the word sukkal translated as “messenger,” this is perhaps an inaccurate translation, which has no support in lexical material. According to lexical databases compiled by the Pennsylvania Sumerian Dictionary, sukkal was never translated mar sipri, “messenger.”6 Its usual translation was sukkallu, de˜ned by CAD as “a court o¯cial,” and also a title of the ruler of Elam. We also ˜nd sukkal translated as sarru, “a king,” no doubt referring to the royal title of Elam.7 Lafont came to a similar conclusion, that sukkal in Old Babylonian Mari referred to a “ministre des aˆaires étrangères.”8 5. The translation “grand vizier” is frequent, but unfortunate, since it inaccurately describes the duties of this o¯cial. A discussion of the o¯ce of sukkal-mah can be found in W. Sallaberger, “Ur III-Zeit,” in Mesopotamien. OBO 160/3 (Fribourg: Editions Universitaires Fribourg, 1999), 188–90, with previous bibliography. 6. The lexical evidence is as follows: sukkal = su-kal-lu4 [Lu I.87 = MSL 12 95] su-uk-kal sukkal = su[k-kal-lum] [Aa iv/2 118 = MSL 14 376] su-kal sukkal su-uk-kal-lum [Ea iv 80 = MSL 14 358] su-uk-kal sukkal suk-kal-lum [Sb Voc. II. 75 = MSL 3 135] sukkal = MIN (LUGAL) [Group Vocab. CT 51 58–63 No. 168 iv 22] 7. We also ˜nd in certain late traditions the equivalent sukkal = pasÿsu, “salve priest”; e.g., su-kal sukkal pa-si-s[u] [Ea iv 81 = MSL 14 358]; sukkal = MIN (pa-si-sú) [S1G7. ALAN 23.335 = MSL 16 221]. This equation is opaque. 8. See B. Lafont, “Messagers et ambassadeurs dans les archives de Mari,” in La circulation des biens, des personnes et des idées dans le Proche-Orient ancien, ed. D. Charpin (Paris: Editions Recherche sur les Civilisations, 1992), 183, note 94.
DIPLOMACY AND THE RITUALS OF POLITICS AT THE UR III COURT
The sukkals of Ur III times saw to the needs of foreign emissaries visiting Babylonia and acted as intermediaries between the latter and the royal bureaucracies.9 They also could serve as interpreters when foreign emissaries did not speak Sumerian or Akkadian. For instance, emissaries from Marhasi were provided with translating sukkals.10 Some of the sukkals bore foreign names— Rabhuti, Huziri, and so on. If they were of foreign origin, their language pro˜ciency and translating abilities are easily explained. Often we ˜nd that the sukkal was assigned to a speci˜c foreigner and that where he occurs we also ˜nd his sukkal.11 For example, we often ˜nd the sukkal Rabhuti with Marhuni of Harsi, the sukkals †u-†ulgi or Lu-Damu with Libanas-gubi of Marhasi. Chart 1 displays these correspondences. Emissaries were generally identi˜ed as so-andso lú-kin-gi4-a of so-and-so of such a place, for instance Zubus lú-kin-gi4-a Iabrat †imaski12 or Amur-DINGIR lú-kin-gi4-a Libanuk-sabas énsi Marhasi.13 Alternatively, envoys could be identi˜ed as simply, so-and-so of such a place, for example Budur lú Urshu14 or Pusam lú Simanum.15 Of course, the term “man of a certain region” is a very imprecise designation, which, as Michalowski showed, covered a whole range of statuses from a foreign ruler or his representative to an expatriate.16 Even among the emissaries, not all would have occupied the same rank. Based on parallels from later periods, we could well imagine that foreign dignitaries in Ur III court circles might include ambassadors, detained ambassadors,
9. At Mari, the sukkal had a similar role, shepherding foreign diplomats into the court, see Lafont “Messagers et ambassadeurs dans les archives de Mari,” 174–75. 10. In two very similar texts, namely Buccellati, Amorites 22 (†S6) and JCS 7 106 = †A LXXI 85 (und.), wood and other items were allocated to various emissaries, including Banana lú Marhasi and to sukkal eme-bal ki lú mar-ha-siki gub-bame-sè, “sukkals, translators, who are stationed with the Man of Marhasi.” 11. The number of sukkals was not large; sometimes we ˜nd the same sukkal reassigned at a later date to a diˆerent emissary, as Chart 1 shows. 12. CST 466. 13. MVN 5 111. 14. PDT 594. 15. PDT 2 1092. 16. P. Michalowski, “The Men from Mari,” 187.
19
exiled rulers ˜nding asylum in Babylonia, and possibly even heirs apparent sent to “grow up in the palace like a puppy.” The discussion of the evidence for these various types of dignitaries at the Ur III court is beyond the scope of the present essay. Some emissaries appear only once or twice in the records we have; perhaps they came to the court to deliver a message or gift and departed again rapidly. Others occur repeatedly, week after week, month after month, even year after year; these were the delegates of foreign rulers sent to conduct negotiations on behalf of their king.17 Comparisons from the Mari and Amarna archives suggest that such emissaries were usually highranking courtiers or even princes in their native lands.18 Far from being merely messengers, these were, as Munn-Rankin characterized them, “men of consequence.”19 According to the Amarna correspondence, failure to treat these persons of consequence with the proper respect could be taken as a political message, indicating a breach of good relations between the two states. The Pharaoh wrote to the Babylonian king: “Now we are brothers, you and I, but I have quarreled because of your messengers, since they report to you, saying, ‘nothing is given to us who go to Egypt.’ Those who come to me, has a single one of them ever come [and not] received silver, gold, oil, solemn garb, every sort of ˜nery, [more than] in any other country?”20 17. For instance, Ili-Dagan of Ebla occurs in AS 2 iii 13, AS 2 iii 20, AS 2 vi 3, AS 2 vi 4, AS 2 vi 8, AS 2 vii 21, AS 2 viii-, AS 3 iv 14 (references collected by D. I. Owen, “Syrians in Sumerian Sources,” in New Horizons, 117–19). Amur-DINGIR of Marhasi appears in AS 3 xii 28 (CST 286), AS 4 i 6 (TCL 2 5508), AS 4 i 27 (SAT 2 806), AS 4 ii 25 (AUCT 2 278), AS 4 iii 22 (TLB 3 24), AS 4 U6 (AnOr 7 99), AS 4 v 9 (MVN 5 111) and AS 4 v 15 (BCT 90). 18. For instance, Mari sent the governor of the Sagaratum province to the court at Babylon—see M. Munn-Rankin, “Diplomacy in Western Asia,” 99. In the Amarna period, princes were often emissaries; see S. Meier, The Messenger in the Ancient Semitic World (Atlanta: Scholars Press, 1988), 19–20. But courtiers could also serve in this function: according to EA 17, Tusratta sent his chief minister to Egypt; see W. Moran, The Amarna Letters (Baltimore: The Johns Hopkins University Press, 1992), 42. 19. Munn-Rankin, “Diplomacy in Western Asia,” 99. 20. This is EA 1, see Moran The Amarna Letters, 2; see also EA 20, p. 42; EA 16, p. 38.
20
T. M. SHARLACH
As Owen noted, evidence in the surviving administrative records for exchanges of gifts between ambassadors and the Ur III court is scant indeed;21 nevertheless we must assume gift exchange formed an important part of diplomatic contacts. Almost nothing is known of gifts received by the Ur III kings, with the exception of a speckled cat, perhaps a leopard, which Ibbi-Sin received from Marhasi.22 Slightly more is known about gifts the foreign emissaries received from the crown. Ur III documents tell us that foreign guests were provided with food rations, servants, transportation, livestock and even luxury goods like silver rings and textiles.23 Envoys must also have been provided with places to live. Did foreign emissaries reside in the palace itself ? Or did they live near the palace in residences supported by the king?24 The latter seems more consonant with what we know of later practices, but regrettably, the Ur III evidence hardly addresses this topic. We do know that the foreign emissaries did not necessarily remain in the same place—the emissaries could apparently take trips to other cities. One tablet records a tour, which men of Marhasi took to Isin.25 Sometimes it is stated that the foreign emissaries were living in villages (é-duru5-ne-ne) or
21. Owen, “Syrians in Sumerian Sources,” 116. 22. P. Steinkeller, “The Question of Marhasi,” ZA 72 (1982) 253. 23. Servants were provided to foreign guests, e.g., Nesbit iii and MVN 16 880; they also received livestock as a royal gift, níg-ba lugal, according to NBC 704 (Appendix II, Text One). N. Koslova, Ur III – Texte der St. Petersburger Eremitage Text 127 records the gift of silver rings to emissaries, as does TrDr 84; on silver ring texts, see P. Michalowski, “The Neo-Sumerian Silver Ring Texts,” SMS 2/3 (1978) 43–58. One diplomat, Hulibar, received a chariot according to NBC 2900 (Appendix II Text Two). Textiles given as a gift appear, for example in DAS 317. It is interesting that most of these texts (with the exception of the silver ring texts) come from the Umma archives. 24. A Mari letter quoting the king’s instructions about Simah-ilani, the ruler of Kurda, reads “if more than 1000 men are in his entourage, they should camp outside the town. But if (only) 200 or 300 soldiers are in his entourage, let them enter the outer wall precinct and give them quarters (bÿt naptarim). As for Simah-ilani . . . give him quarters which are suitable, lest he be oˆended.” See DEPM I letter 266, 415–17 and CAD N/1 325. 25. This tablet is Hirose 403 (AS 5).
One [Body] Line Long
in ˜elds (a-sag4).26 Sigrist hypothesized that such villages or encampments were the normal place of residence for the foreign emissaries, that “le é-duru5 pouvait avoir été une large zone en dehors de la ville, . . . dans laquelle se retrouvaient tous ceux qui n’étaint pas citoyens de la ville, donc les marchands et les messagers des pays étrangers.”27 Yet when we consider the data, we see that only a very small percentage of tablets refer to the villages. This suggests that the residence in the countryside was unusual; often, it can be correlated with the time of the two Akiti festivals around Ur and Uruk. Festivals and Foreign Emissaries Many of the trips can be linked with the cultic calendar and major religious festivals that the state sponsored. It seems likely that the foreign emissaries were following the king and his court to various festivals throughout the year. Our main sources of information on this topic are PuzrisDagan texts. Particularly important for the present discussion are a group of oˆering lists, most of which are dated to the reign of Amar-Sin. In such documents, animals sacri˜ced to the statesupported central shrines are enumerated, then sometimes we also ˜nd animals for princes or courtiers and for foreign emissaries.28 In other words, the animals expended to the diplomats occur at the end of an oˆering list, as we see from the sample text, published as SACT 1 160. 1. 1 udu niga 2. den-líl dnin-líl 3. 1 udu niga dnanna 4. á-gi6-ba-a 5. 4 udu niga 6. den-líl dnin-líl 7. 1 más-gal niga dnanna 26. It is possible that we should read é-a-ne-ne, “their households,” not é-duru5-ne-ne, “their villages,” though the variation with a-sag4 may suggest that the latter is the more likely reading. 27. M. Sigrist, Drehem 371; see also 364. 28. In MVN 13 695 (v 27), we ˜nd the reverse of the usual pattern of many animals for the gods and a few for the emissaries: here one lamb is allocated to Nintin[ugga], then 36 sheep to named Elamites, ending with the phrase ezem lú mar-hasiki-ke4-ne, “festival (for/of) the men of Marhasi.”
DIPLOMACY AND THE RITUALS OF POLITICS AT THE UR III COURT 8. á-ud-te-na 9. lugal ku4-ra 10. 4 udu niga a-tu5-a 11. dnin-tin-[ug5-ga] 12. gìri a-tu 13. 1 udu niga 1 más-gal 14. sul-gá-a-ad lú zi-da-ah-riki 15. gìri su-dsul-gi sukkal 16. árad-mu maskim 17. iti ud-1 ba-zal 18. sag4 tum-ma-al 19. ki en-dingir-mu-ta 20. ba-zi 21. iti su-es5-sa 22. mu hu-úh-nu-riki ba-hul l.e. 13 1 fattened sheep – (for) Enlil, Ninlil, 1 fattened sheep – Nanna, night-time, 4 fattened sheep – Enlil, Ninlil, 1 fattened big-goat – Nanna, dawn, when the king entered, 4 fattened sheep – lustration of Nintinugga, via Atu. 1 fattened sheep, 1 big-goat – †ulgad, the man of Zidahri, via †u-†ulgi, sukkal. Arad-mu, requisitioner, the ˜rst day having passed from the month in Tummal expended from the o¯ce of En-dingirmu. Month 8, Year Huhnuri was destroyed (Amar-Sin Year 7). (Total sheep) 13. The appearance of foreign emissaries at the end of an oˆering list is not in and of itself proof that they were present at the festival. PuzrisDagan tablets may group together various unrelated expenditures on the basis of the date of the transaction and the disbursing o¯cial. Caution is therefore necessary. But when texts such as the one translated above are supplemented with other texts, which allow us to place the emissaries at the Akiti in Ur or in Tummal in the eighth month, we can be sure that in these cases there was a direct connection between the list of oˆerings for the festival and the foreign emissaries.
21
Emissaries attended a number of religious celebrations, namely the Akiti Festival in Ur at the New Year29 and also the second Akiti Festival in the fall.30 In addition to these festivals, foreign diplomats may have been present at royal lustrations (a-tu5-a lugal)31 and perhaps other celebrations as well. It would be wrong to suggest that the presence of foreign emissaries was characteristic for all the important religious festivals; there were many celebrations that the foreign emissaries did not attend. The emissaries do not appear in oˆerings at provincial shrines. And in the central shrines, there were some oˆerings at which they never appear, for instance various lunar festivals patronized by the queen, including the ès-ès or UD.SAKAR rites. Interestingly, they did not participate in the state-supported cults of imported gods.32 In fact, the data suggest that although the cults of imported gods received signi˜cant royal support at this time, foreign emissaries were involved with the cults of the traditional gods, chie˘y Enlil, Ninlil, Nanna and Inanna, and their involvement primarily coincided with major festivals. The Tummal and Akiti Festivals The Tummal Festival The best-attested and probably the most important cultic event that the foreign emissaries
29. E.g., SET 59, AS3 i 3; TCL 2 5508, AS 4 i 6, MVN 3 228, AS4 i 14 and UDT 92, AS8 xii 29. 30. E.g., BCT 68, †47a vii 5, CTMMA 17, AS 4 viii 1. 31. Representatives from Simurrum, Pasime and the Amorite territories held by Naplanum occur together with the a-tu5-a lugal; see NBC 631, in Appendix II, Text Ten. Furthermore, messenger texts from Lagash sometimes refer to men associated with royal lustrations traveling from the sea; e.g., ITT 3 6207. 32. There are a few tablets in which both foreign emissaries and oˆerings to foreign gods are listed, though, as stated above, the two transactions were not necessarily linked, for instance, †ulgad of Zidahri and a Dagan oˆering, MLC 93 (see Appendix II, Text Five) and Hulibar with Allatum, CST 415, †S 3a i 19. For a more general discussion of foreign gods in the Ur III period, see Sallaberger, Kultische Kalender and my own “Foreign In˘uences on the Religion of the Ur III Court,” SCCNH 12 (2002) 91–114.
22
T. M. SHARLACH
attended was the Tummal festival in the 8th month.33 As Chart 2 shows, a series of PuzrisDagan texts record the expenditure of sheep from the Tummal o¯cial En-dingirmu. Most of the livestock was intended for sacri˜ces to Enlil, Ninlil and Nanna, but foreign emissaries also received animals. To understand why, a closer look at the Tummal festival itself is needed. Using a variety of sources, Sallaberger has succeeded in piecing together much information about these celebrations. At some level, both the Tummal and Akiti were considered to be New Year celebrations (zag-mu), despite the fact that only one took place in the ˜rst months of the year.34 Each celebration was a ritual renewal of the monarchs’ legitimacy and authority. The goddess Ninlil played the central role in the Tummal festival: she went back and forth to Nippur by boat and decreed a suitable fate for the king in Tummal.35 Since foreign emissaries entered the country at various times of the year, but regularly departed immediately after the Tummal Festival, in the third week of the eighth month,36 it is di¯cult to escape the conclusion that their presence at the festival was obligatory.
33. E.g., SRD 17, AS 7 viii 2; JCS 39 (1987) 122 No. 6, and Steinkeller, “The Question of Marhasi,” note 16, †48 viii 2. 34. On the Tummal festival, see W. Sallaberger, Kultische Kalender, 142 and note 669 and 670, and S. Oh’e, ASJ 8 (1986) 121. 35. Literary evidence for these aspects of the Tummal festival seem to be found in lines 82–90 of Shulgi Hymn R, “Shulgi and Ninlil’s Barge,” translated by the Oxford Electronic Text Corpus of Sumerian Literature (ETCSL), www-etcsl. orient.ox.ac.uk/. 36. Arrivals of emissaries are recorded for a number of months, in month iv – from Simanum, TrDr 84 (†42 or AS 6 iv), from Mari, JCS 7, 104 (AS iv 1); in month vi – from Tilmun, CST 254 (AS 2 vi 3), TRU 305 (AS 2 vi 4) and in month xii – from Marhasi, CST 286 (AS 3 xii 28). With the exception of an undated text published by D. I. Owen (text “K” in “MeGum,” Ebla 1975–1985), in which an envoy departed to the Martu-land (AS 1–26), all departures occurred in month viii : on day 1 to Iamadium, Ebla and Mari (Amorites 21, 556), on day 20, Nawir-AN-A.ZU (the reading of which is still uncertain) to his city (MVN 15 244, AS 7); see JCS 39 (1987) 122 (#6), and on day 21 to Dudduli (MVN 15, 190, AS 9).
One [Body] Line Short
The Akiti Festival Foreign emissaries also appear to have been present at the Akiti celebrations in Ur. Ur observed two Akiti festivals, but the second one, lasting perhaps ten or twelve days, was the most important, giving its name to the seventh month of the Ur calendar.37 This second Akiti celebration in the fall, like the Tummal ceremonies, featured a divine journey by boat; in this case Nanna went from his sanctuary in Ur to the Akiti house near Ga’es, located slightly to the northeast of the city. While in the countryside, the king ritually guaranteed the land’s fertility by beginning the ploughing himself. It has been argued that an Old Babylonian literary text, “The Contest between the Hoe and the Plough,” may refer to such an event.38 The Plough addressed the Hoe: “I am the Plough . . . I am mankind’s faithful farmer. To perform my festival in the ˜elds . . . the king slaughters cattle and sacri˜ces sheep, and he pours beer into a bowl . . . . The king takes hold of my handles, and harnesses my oxen to the yoke. All the great high-ranking persons walk at my side. All the lands gaze at me in great admiration. The people watch me in joy. 39
This passage may suggest that the presence of the courtiers and the foreign emissaries (all the lands who watch admiringly) was a signi˜cant element of this religious event. Indeed, their presence at such a gathering may also have served as the visible reminder of the king’s power and legitimacy. These major festivals were occasions for large-scale oˆerings for the gods and for banquets for the men involved—a social occasion for the priests, the king and his guests.40 At the same 37. W. Sallaberger, “Königtum und Kult in der Hauptstadt Ur,” in Von Babylon bis Jerusalem, ed. J. Westenholz (Milano: Skira, 1999), 258. 38. See W. Sallaberger, “Königtum und Kult,” 258. 39. For the translation, see www-etcsl.orient.ox.ac.uk/ 40. The expenses incurred were substantial, but the court had help in paying for them. Akiti celebrations were partially underwritten by contributors from courtiers, as shown by texts like TCL 2 5508, AS 4a i 6, while the Tummal festival was paid for out of the bala fund of the provincial governor (according to NBC 10794, Appendix II, Text Three).
DIPLOMACY AND THE RITUALS OF POLITICS AT THE UR III COURT
23
time, such gatherings provided opportunity for ambassadors and courtiers to have access to the king, the “party politics” of ancient Babylonia. In all likelihood, the attendance of foreign emissaries at major religious festivals was not a practice unique to the Ur III period. The Amarna correspondence between great kings indicates that it was customary to invite allied rulers or their envoys to major festivals or inaugurations.41 An interesting parallel to the Ur III material can be found in a Mari letter, ARM 1 50, written from †amsi-Addu to his son. “The current month is Addaru (the eighth); when the 16th day arrives, the New Year’s Festival will be performed. The envoys of the ruler of the city of Esnunna are already present.”42 We may also compare a series of tablets from Larsa, which list oˆerings for
temples along with a group of dignitaries from other city-states.43 Ur III administrative documents are a rich source for understanding the structures of ancient near eastern diplomacy. Foreign emissaries in Babylonia were assigned sukkals, “secretaries,” who served as liaisons to the court and, if needed, as interpreters. Because of the nature of the Ur III archives, we have little information about gift exchange or the substance of diplomatic negotiations. Such sources show that foreign diplomats attended the major religious celebrations along with the royal court. “All the lands” witnessed the Akiti ceremonies at Ur and the Tummal festival of the eighth month. These were religious, but also political occasions, suggesting the extent to which diplomacy, politics and ritual were intertwined in the ˜nal centuries of the third millennium B.C.
41. We may compare EA3 (“When you celebrated a great festival, you did not send your messenger to me, saying, ‘Come t[o eat an]d drink.’ No[r did you send me] my greeting-gift in connection with the festival”) and EA34 (“As to your having written me ‘Why did you not send your messenger to me?’ The fact is that I had not heard that you were going to perform a sacri˜ce”), W. Moran, The Amarna Letters, 7 and 105.
42. ARM 1 50 was also translated by A. L. Oppenheim, Letter 55 in Letters from Mesopotamia (Chicago: Chicago University Press, 1967), 109, and J.-M. Durand, DEPM III.965 (p. 112). 43. See A. Goetze, “Sin-iddinam of Larsa,” JCS 4 (1950), 91– 95. I am grateful to M. Fitzgerald for bringing this reference to my attention.
APPENDIX I: CHARTS Chart 1. Foreign envoys and their sukkals A. The envoy Marhuni, lú Harsi Date AS 7 viii 16 AS 7 viii 14 AS 7 viii 7 AS 7 viii 5 AS 8 i — AS 8 ix 14 AS 8 ix 16
Text HUCA 29 (1958) 77 6 ASJ 4 (1982) 67 14 PDT 2 1170 MVN 20 28 MVN 13 636 BIN 3 173 BCT 83
Sukkal (gìri o¯cial) Rabhuti Rabhuti Rabhuti Ha-NI [ ] Ribhuti Rizuhuti
B. Budur, lú Ursu AS 1 – 26 AS 1 iii 4
Owen, “Megum,” text Ka PDT 594
a. D. Owen, “MeGum,” 289.
Lugal-KA-gina Lugal-KA-gina
24
T. M. SHARLACH
C. Gura’a, lú Ursu TLB 3 25 MVN 5 111 AnOr 7 99 MVN 3 384
AS 4 iii 22 AS 4 v 9 AS 4 v 6 Und. iii 7
Huziri Pusam Huziri Huziri
D. Nanau, lú Ursu RA 9, SA 79 JCS 7 (1953) 104, 61
AS 2 iii 13 AS 2 iii? 20
Lugal-magure maskim Lugal-magure sabra
E. Pusam, lú Simanum AS 2? i 16 AS 2 i 30 AS 5 viii 12
Sigrist 1795 = MCS 7 24 PDT 2 1092 MVN 15 199
Huziri Huziri —
F. Arbatal, lú Simanum AS 9 ix 11 †S 1 iv 1 †S 1 vi 21
MVN 13 639 AUCT 3 294 BCT 95
Kirbatal Erra †u-estar
G. †u-Sin-wuzum-isar lú Simanum †S 2 ix 9 †S 2 ix 17
PDT 2 904 Princeton 83
†u-estar †u-estar
H. Kirib-ulme, lú Simurrum AS 8 ix 16 AS 8 ix 14 AS 8 x 13 AS 9 ii 26 AS 9 xi 3 AS 9 xi 4 AS 9 xi 7 †S 1 i 24 †S 2 ix 24
BCT 83 BIN 3 173 PDT 1147 SET 66 BibMes 25 150 44 TCND 260 SACT 1 169 BIN 3 217 Sigrist 874
Nanna-NI-GI Huziri Nanna-kam Baza Huziri Huziri Huziri Ili-beli †u-kubum
I. Amur-DINGIR, lú-kin-gi4-a Libanuk-sabas énsi Marhasi AS 4 i 14 AS 4a i 27 AS 4 ii 25 AS 4 v 9 AS 4 v 15 — iii 7
MVN 3 228 SAT 2 806 AUCT 2 278 MVN 5 111 BCT 1 90 MVN 3 384
Lugal-KA-gina Lugal-KA-gina Lugal-KA-gina Pusam Lú-damu Lú-damu
DIPLOMACY AND THE RITUALS OF POLITICS AT THE UR III COURT
25
J. Libanas-gubi, lú-kin-gi4-a Libanuk-sabas énsi Marhasi AS 5 v 9 AS 5 viii 12 AS 5 viii 24 AS 5 ix 1 AS 6 ii 24 AS 6 iv 1 AS 6 iv 1 AS 6 iv 5 AS 6 iv 6 AS 6 iv 23 AS 6? iv 26 AS 6 viii 3 AS 6 viii 5 Undated Undated
PDT 1171 MVN 15 199 MVN 1 142 MVN 11 154 CST 466 JCS 7 (1953) 104 CST 468 MVN 13 635 TRU 344 Hirose 231 MVN 11 146 RA 8 No. 111 A 30 MVN 3 384 MVN 11 144
Ha-NI — Lú-Enki — †u-†ulgi †u-†ulgi †u-†ulgi †u-†ulgi †u-†ulgi Lu-Damu Lu-Damu Lu-Damu Lu-Damu Lu-Damu †u-†ulgi
Chart 2. The Tummal Festival Date †47 viii 28
Text MVN 15 189
Expended From 19 udu Lu-dingirra 1 gud
†48 viii 2
Steinkeller, ZA 72 (1982) note 16
26 udu
AS 1 viii 17 MVN 1 124
12 udu
In For DNs, PNs Tummal 1. Ninlil 2. Enki 3. sízkur Inanna sag4 Unug 4. Isin 5. Amir-†ulgi, †ulgi-palil lú Mari 6. ud Ahutiki a-sag4 dab5-dè . . . 7. níg-mussa PN . . . En-dingirmu Tummal 1. Enlil, Ninlil 2. Nanna 3. giskak gu-la 4. é-muhaldim 5. é †elhaha lú 5. †azibi 6. é Taram-†ulgi dumu-munus lugal dam †udda-bani lú Pasime En-dingirmu Tummal 1. Enlil, Ninlil 2. Nanna 3. Inanna of Girsu 4. lú-kin-gi4-a Arbikukbi lú Marhasi 5. lú-kin-gi4-a dumu Arbikukbi lú Marhasi 6. Marhuni lú Harsi
26
T. M. SHARLACH
AS 5 viii 12 MVN 15 199
13 udu
AS 7 viii 1
SACT 1 160
13 udu
AS 7 viii 2
Nesbit 17
11 udu
AS 7 viii 5
MVN 20 28
14 udu
AS 7 viii 7
PDT 2 1170
15 udu
AS 7 viii 14 ASJ 4 (1982) p. 67 no.14
12 udu
En-dingirmu Tummal 1. Enlil, Ninlil 2. Nanna 3. sízkur ki-dSin 4. sízkur Inanna of [Uruk] 5. Libanas-gubi lú-kin-gi4-a Libanuk-sabas énsi Marhasi 6. Pusam lú Simanum 7. Adda-gina énsi Harsi En-dingirmu Tummal 1. Enlil, Ninlil 2. Nanna 3. Enlil, Ninlil 4. Nanna 5. a-tu5-a Nintinugga 6. †ulgad lú Zidahri En-dingirmu Tummal 1. Enlil, Ninlil 2. Nanna 3. Enlil, Ninlil 4. Nanna 5. Inanna Unugki-sè 6. †ulgad lú Zidahri En-dingirmu Tummal 1. Enlil, Ninlil 2. Nanna 3. sízkur ki-dSin 4. Inanna sag4 é-gal 5. †ulgad lú Zidahri 6. Marhuni lú Harsi En-dingirmu Tummal 1. Enlil, Ninlil 2. Nanna 3. Enlil, Ninlil 4. Nanna 5. Dada gala . . . 6. †ulgad lú Zidahri 7. Marhuni lú Harsi En-dingirmu Tummal 1. Enlil, Ninlil 2. Nanna 3. Enlil, Ninlil 4. Nanna 5. Inanna Unugki-ke4 6. [†ulgad lú Zidahri?] 7. [Mar]huni lú Harsi
DIPLOMACY AND THE RITUALS OF POLITICS AT THE UR III COURT
AS 7 viii 16 HUCA 29 (1958) 17 udu pp. 77–78 no. 6
AS 7 viii 20 MVN 15 244
63 udu
AS 7 viii 23 JCS 39 (1987) p. 122 no. 6
16 udu
En-dingirmu Tummal 1. Enlil, Ninlil 2. Nanna 3. Enlil, Ninlil 4. Nanna du6-babbar 5. Enmesara 6. gúrum é Zazi sagina 7. †ulgad lú Zidahri 8. Marhuni lú Harsi En-dingirmu Tummal 1. Enlil, Ninlil 2. Nanna 3. Enlil, Ninlil 4. Nanna 5. gúrum é Zazi-sè 6. kas4 7. é-muhaldim, gardu 8. Na-wi-ir-AN A.ZU ud uru-ni-sè ì-gen-na-a En-dingirmu Tummal 1. Enlil, Ninlil 2. Nanna 3. A-aba énsi Talmus 4. gurum2 é Zazi-sè 5. Na-wi-ir-AN A.ZU ud uru-ni-sè ì-gen-na-a
APPENDIX II: Unpublished Texts from the Yale Babylonian Collection Text One (NBC 704) AS 5 iii U Umma 1. 10-lá-1 udu 2. 1 ùz 3. su-da-gan 4. lú ma-ríki 5. níg-ba lugal R 6. sukkal-mah PA.ªKA†4º 7. iti se-kar-ra-gál-la 8. mu en-unu6-gal dinanna ba-hun Seal: damar-dEN.ZU nita kalag-ga lugal urim5ki-ma lugal an-ub-da-limmu Text Two (NBC 2900) AS 7 — Umma 1. 1 gisgigir su-du7-a 2. hu-lí-bar elam
Ur-dli9-si4 énsi Ummaki árad-zu
27
3. R 4. 5. 6.
su ba-ti (erasure) ki a-kal-la-ta kisib lú-kal-la mu hu-úh-nu-riki ba-hul
Seal: lú-kal-la dub-sar dumu ur-e11-e kus7 Text Three (NBC 10794) AS 9 ix R 3 Puzris-Dagan 1. 2 gud niga 3-kam-ús 2. á-gi6-ba-a 3. den-líl dnin-líl 4. sag4 tum-ma-alki 5. 1 gud niga den-líl 6. 1 gud niga dnin-líl 7. má dnin-líl-lá tum-ma-alki-ta ku4-ra 8. sag4 nibruki R 9. á-ud-te-na
28
10. 11. 12. 13. 14. 15. 16. edge:
T. M. SHARLACH
lugal ku4-ra iti ud-3 ba-zal bala ur-dnin-kur-ra énsi suruppakki ki lú-sag5-ga-ta ba-zi gìri ur-su-ga-lam-ma dub-sar iti ezem-mah mu en dnanna kar-zi-da 4 gud
Text Four (NBC 10804) AS 1 ii R 23 Puzris-Dagan 1. 2 gud niga 2. é-muhaldim-sè 3. mu-kas4-e-ne-sè 4. sag4 nibruki 5. 1 más-gal niga 6. ì-lí-ddagan lú eb-laki 7. 1 más-gal niga 8. bu-du-úr lú ur-suki R 9. 1 más-gal niga 10. is-me-dda-gan lú ma-ríki 11. gìri lugal-KA-gi-na ªsukkalº 12. árad-mu maskim 13. iti ud-23 [ba]-zal 14. ki a-hu-ni-ta ba-zi 15. iti ses-da-kú 16. mu damar-dEN.ZU lugal Text Five (MLC 93) AS 8 ix R 12 Puzris-Dagan 1. 4 udu [niga?] 2. 6 udu 3. e-ru-ba-tum é dda-gan ù dis-ha-ra 4. den-[líl]-zi-sag4-gál maskim 5. 2 udu [niga] gìri da-hi-is-a-ªtalº 6. 1 udu ªnigaº [x +] 20 ùz más ªhi?º-a niga mu-DU [tu?-r]a-am-é-a 7. é-uz-ga-sè ba-zi gún-sè 8. gìri Ur-dEN.ªZUº muhaldim 9. ME-dKA.[DI] é-ud-15 [ ] 10. 1 udu [ ] 11. 1 [más?] 12. sul-gá-ad lú z[i]-da-ah-riki 13. gìri la-la-mu sukkal 14. árad-mu maskim 15. iti ud-12 ba-zal 16. ki zu-ba-ga-ta 17. ba-zi
18. 19. 20. edge:
gìri ad-da-kal-la dub-sar iti ezem-mah mu en eriduki [ba-hun] 5 udu
Text Six (NBC 10790) 47 xii R 23 Puzris-Dagan 1. 1 udu niga 2. E2-NI-da-gú lú kin-gi4-a 3. lú ur-kiski-sè 4. gìri lugal-KA-gi-na sukkal 5. árad-mu maskim R 6. iti ud-23 ba-zal 7. sag4 a-sag4 8. zi-ga 1. ki lú-dingir-ra 2. iti se-ªKINº-[kud] 3. mu-ús-sa ki-maski ba-hul Text Seven (MLC 36) AS1 v R 28 Puzris-Dagan 1. 3 udu niga 2. 1 más-gal niga 3. é-muhaldim-sè 4. mu-kas4-ke4-ne-sè 5. 1 udu niga 6. lú-kin-gi4-a ar-wa-lu-uk-bi lú mar-ha-siki 7. 1 udu niga R 8. lú-kin-gi4-a dumu ar-wa-lu-uk-bi lú mar-ha-siki 9. é-duru5-ne-ne-sè 10. gìri lugal-KA-gi-na 11. lugal-má-gur8-ri sukkal mask[im] 4. iti ud-28 ba-zal 5. [k]i lú-dingir-ra-ta 6. [ba]-zi 7. it[i] ezem-dnin-a-zu 8. mu damar-dEN.ZU lugal Text Eight (NBC 12549) Undated, from Umma or Lagas 1. [ ] kas sig5 2. 4 kas DU 3. 1 (bán) zíd-ba-ba 4. 6 (bán) zíd-gu [s]ig5 5. 6 (bán) ninda-GI†-A† 6. 1 (PI) 1 (bán) zíd-gu DU R 7. 6 sìla ì-gis
DIPLOMACY AND THE RITUALS OF POLITICS AT THE UR III COURT
8. 9. 10. 11. 12. edge:
6 udu niga ab-du-sa lúkin-gi4-a ì-ab-ra-[at] é-ta gin-na-as gìri da-a-a lú[gist]ukul ra-gaba?
Text Nine (NBC 4317) Undated. Provenience unclear. 1. 21 gurus 1 sìla kas 1 sìla ninda sig5-ta 2. lú-TUG2 x gisgi ku4-ra 3. gìri dug4-ga-ni 4. dUTU-DÙG sukkal maskim 5. ki-é-A-AN 6. 1 (bán) kas 1 (bán) [ninda?] sig5 ur-GIR3 lú-kin-gi4-
a-b[í]-zi-im-ti ªNINº unugki-ta gin-na R 7. ªkiº-é-A sukkal maskim 8. 1 (bán) ninda musen-dù ki ur- dsul-pa-è 9. 1 (bán) ninda musen-dù ki nir-ì-da-gál 10. 1 (bán) kas 1 (bán) ninda su-es4-dar lú-kin-gi4- sar-ì-lí énsi unugki HU.GAR3 sukkal 11. A.AN.NA.TI sukkal PA.[KA†4] Text Ten (NBC 631) Undated. Provenience unclear. 1. 3 (bán) zíd-gu zíd BULUG.SÁ 2. ud-24-kam 3. gìri lú-TUR.TUR 4. 1 kas dida DU 5. 2 (bán) dabin lú-ama-ga-ke4? si-mu-ru-um 6. 1(bán) zíd-gu níg-dirig 7. gìri nagar 8. 1(bán) 5 sìla zíd ur-dEN.ZU 9. 2 lú-ús-sa-ni 1(bán)-ta 10. lú pa-sim-e-me R 11. 3 kas dida DU 12. 3 kas dida sig5 13. 1(PI) zíd-gu 2 (PI) dabin 14. na-ap-ªlaº-[núm] MAR.TU 15. 1 [ka]s dida DU 2 (bán) dabin 16. na-ap-ªtaº-núm lú-kin-gi4-a 17. na-ap-ra-núm (erasure -núm) 18. kusA.GÁ.LÁ kés-rá ud-25-kam 19. 2 (bán) zíd-gu 5 sìla esa 20. a-tu5-[a] lugal 21. gìri l[a]-la-ba 22. 3 sìla esa
29
23. 7 !- sìla 2 zíd-gu 24. 2 sìla ninda é a-tu5-a-sè Text Eleven (NBC 5020) AS 8 i U U Umma 1. 1 udu niga 4 udu ú 2. 8 kas dida sig5 1 (bán) 5 sìla 3. 15 kas dida DU 2 (bán) 4. 3 (bán) zíd-KALAG 4 (bán) zíd-gu sig5 5. 3 ninda sig5 1 (gur) 3 (bán) ninda DU gur 6. 2 (PI) zíd-gu DU 1 (gur) dabin gur 7. 1 sìla ì-dùg-ga 8. 1 (bán) ì-gis 9. 1 (PI) se sag4-gal anse 10. 10 gigur kisib 1-ta 11. 10 gikaskal 1 (PI)-ta 12. 15 (gur) gú KALAG gisasál 13. hu-lí-bar 14. 1 udu ú 15. 1 kas dida sig5 1 (bán) 5 sìla 16. 2 kas dida DU 3 (bán) 17. 1 (bán) zíd-gu sig5 1 (PI) dabin 18. 2 sìla ì-gis R 19. dam hu-lí-bar 20. 1 udu 1 kas dida sig5 1 (bán) 5 sìla 21. 2 kas dida DU 2 (bán) 22. 1 (bán) zíd-gu sig5 1 (PI) dabin 23. 2 sìla ì 24. ba-la-la 25. 1 kas dida sig5 1 (bán) 5 sìla 26. 1 kas dida DU 3 (bán) 27. 1 sìla ì 1(PI) dabin 28. lú-kin-gi4-a 29. gìri ì-ti-su-NI-su 30. iti nesag mu en eriduki ba-hun edge uninscribed Concordance: MLC 36 = Text 7 MLC 93 = Text 5 NBC 631 = Text 10 NBC 704 = Text 1 NBC 2900 = Text 2 NBC 4317 = Text 9 NBC 5020 = Text 11 NBC 10790 = Text 6 NBC 10794 = Text 3 NBC 10804 = Text 4 NBC 12549 = Text 8
AN UR III TABLET FROM IRAN Piotr Michalowski (University of Michigan, Ann Arbor) and Parsa Daneshmand (Tehran)
The tablet published here is presently located in the Tehran Museum. It was con˜scated by the authorities from a gentleman who claimed to have found it in a sack at the Tehran bus station. The Museum authorities turned it over to P. Daneshmand for publication, who in turn solicited the cooperation of P. Michalowski. The two worked together by email and the present publication is the result of this unique form of intercontinental communication. A number of people were of invaluable assistance in this project; the authors wish to thank Sh. Razmjou, director of the Achaemenid Center who informed P. Daneshmand that such a tablet has been found, Mohammad R. Kargar, the director of the Tehran Museum, and Mr. Hassan Gharekhani, former chief of smuggled goods’ o¯ce of the museum. We must also thank Miss Firoozeh Babaee for her help with photography and other assistance, and to Robert Englund, Wolfgang Heimpel, and David Owen for valuable suggestions. It is di¯cult to establish the original provenance of the tablet. One would hope that it came from Iran, as although there is ample documentation for contact between Mesopotamia and Iran during the Ur III period, no administrative texts from this time have been found outside of Susa. This tablet is unique in style and content, and does not seem to belong to any known Ur III archive, that much is certain. It shares certain features with texts from Garsana, but does not belong together with the tablets that come from that site.1 Most
probably this account was written at a hitherto unknown site in the Umma province. The general area of origin is suggested by references to the (old) Tigris, and to the Idlugal canal, which was also in the Umma region.2 The general provenience is supported by the use of the “imperial calendar,” which was not used in accounts from Umma, but was utilized in smaller centers in the province. Where it was discovered is another matter, as it may have been taken elsewhere in antiquity, and therefore an Iranian provenance cannot be completely ruled out. The tablet is a summary of various disbursements made during the whole calendar year of †uSin 3, summarizing a number of smaller accounts. This is apparent from the calculations presented in i 5–9 and ii 14–15, in which monthly expenditures are recorded, followed by a multiplication by twelve, that is the number of months involved. The accounts are referred to as kisib3, followed by a number and the name of the o¯cial involved. In view of the unusual structure of the document, and the lack of archival context, it is di¯cult to discern the level of administration that it belongs to. It is obvious, however, that the bureau from which it originates was responsible for the support of a broad range of economic activities. 1. 36.0.0 se gur numun mur-gud 2. 18.0.0 a2 al-ak u2 ze2-a 3. 54.0.0 gur 2. On the “old Tigris,” see Piotr Steinkeller, “New Light on the Hydrology and Topography of Southern Babylonia in the Third Millennium,” ZA 91 (2001) 22–84.
1. These will be edited and discussed in D. I. Owen, The Garshana Archives, in press.
31
JCS 57 (2005)
32
PIOTR MICHALOWSKI AND PARSA DANESHMAND
4. gud-apin 3-kam 5. 2.4.5 se-ba geme2 arad2 6. iti mas-du3-gu7 -ta 7. iti se-kin-kuru5-se3 8. iti-bi 12-am3 9. se-bi 35.3.0 gur 10. 13.3.5 dabin ªgurº 11. kisib3 4 i-sar-ªliº-bur 12. 116.0.0 se gur 13. 5.3.4 imgaga3 gur 14. 1.0.0 gis-i3 gur 15. 7.0.0 gu2-gal ªgurº 16. 12.2.2 gu2-tur ªgurº 17. kisib3 4 i3-li2-as-ra-ni 18. 19.1.4 dabin gur 19. kisib3 1 sar-ru-um-ba-ni ii 1. 162.0.0 gur 2. kisib3 1 nu-ur2-dIM 3. 180.0.0 gur 4. kisib3 1 su-dma-lik ma2-gid2 5. 41.3.0 gur 6. kisib31 su-dma-lik sabra unugki-ga 7. 1.2.0 dabin gur 8. se-ba ma2-gid2-e-ne 9. kisib3 1 u3-ze2-na-pi-ir dub-sar 10. 0.1.4 dabin gur 11. se-ba aga3-us-e-ne 12. kisib3 1 ta2-at-tu-ri 13. 291.1.1 gur 14. kisib3 1 pi2-la-ah-DINGIR dumu ab-ba-ti 15. giri3 i-mi-id-a-hi nu-banda3 (vacat) 16. 12.1.5.@- sila 3 3 se gur 17. a2 sahar zi-ga iii 1. e agar4 ur-dnin-a-ªzuº 2. 4.1.0 gur 3. a2 sahar zi-ga 4. e gu2 bar-ra gu2 id2-idigna-ka 5. giri3 er3-ra-nu-id lu2 kin-gi4-a lugal 6. 6.2.0. gur a2 sahar zi-ga 7. u3 gi IMxTAK4.IMxTAG4-se3 8. gu2 id2-lugal-ka
9. giri3 ze2-er-ze2-ra-num2 ku3-gal2 10. 13.0.0 gur 0.2.0.gu2-tur 11. numun mur-gud al-ak 12. u3 se-ba gurus sa3-gud 13. su-dis-ha-ra engar 34 asa5-kam 14. 7.0.5 gur a2 ma2 hun-ga2 15. gu2 86.0.0 gur-kam 16. 2.0.0 gur se-ba a-wa-at-i3-li2 dah-hu iti 10-kam 17. 10 dug.ku.kur-du3 dida 0.0.3-ta 18. se sa10 dug 2.0.0 gur 19. kisib3 i3-li2-as-ra-ni sabra 20. ª1.0.0 zid2º munu4 gur 1 ªudu u2º 21. lugal-an-ne2-se3 22. kisib3 su-ªes4º-tar2 dub-sar iv 1. su-nigin2 924.2.4 ªse gurº 2. su-nigin2 35.4.1.dabin gur 3. su-nigin2 5.3.4 imgaga3 ªgurº 4. su-nigin2 ª7.0.0º gu2-gal ªgurº 5. su-nigin2 12.3.2 gu2-tur gur 6. su-nigin2 1.0.0 gis-i3 1 udu u2 7. numun mur-gud al-ak se-ba u3 zi-ga-didli 8. ki ta-ni-a-ta 9. ba-zi 10. ªmu si-maº-num2ki ba-ªhulº 1. 36 gur of barley for seed and fodder for oxen, 2. 18 (gur), wages for hoeing and pulling weeds, 3. —(altogether) 54 gur (of barley) 4. for three teams of plow-oxen, 5. 2 (gur), 4 (barig), and 5 (ban2), grain rations for male and female “slaves” 6. from month 1 7. to month 12, 8. that is for twelve months— 9. its barley (total) is 35 (gur) and 3 (barig). 10. 13 gur, 3 (barig), and 5 (ban) of ˘our, 11. (from) four sealed (tablets) of Isar-libur, 12. 116 gur of barley, 13. 5 (gur), 3 (barig) and 4 (ban) of emmer, 14. 1 gur of oil, 15. 7 gur of broad beans, 16. 12 (gur), 2 (barig), and 2 (ban) of lentils, 17. (from) four sealed (tablets) of Ili-asranni. 18. 19 (gur), 1 (barig), and 4 (ban) of ˘our,
AN UR III TABLET FROM IRAN 19. (from) one sealed (tablet) of †arrum-bani. ii 1. 162 gur (of barley) 2. (from) one sealed (tablet) of Nur-Adad. 3. 180 gur (of barley) 4. (from) one sealed (tablet) of †u-Malik, the boat hauler. 5. 41 gur and 3 (barig of barley) 6. (from) one sealed (account) of †u-Malik, the sabra of Uruk. 7. 1 (gur) and 2 (barig) of ˘our— 8. grain rations of the boat-haulers 9. (from) one sealed (account) of Uze-napir, the scribe. 10. 1 (barig) and 4 (ban) of ˘our— 11. grain rations of the soldiers 12. (from) one sealed (account) of Tat(t)uri. 13. 291 (gur), 1 (barig) and 1 (ban of barley) 14. (from) one sealed (account) of Pilah-ili, son of Abbati. 15. via Imid-ahi, the sergeant. 16. 12 (gur), 1 (barig), 5 (ban), and 2/3 (sila) of barley— 17. wages for earth work iii 1. (at the) ditch of the arable tract of Ur-Ninazu. 2. 4 (gur) and 1 (barig of barley), 3. wages for earth work 4. on the outer ditch on the bank(s) of the Tigris, 5. via Erra-nuååid, the royal courier. 6. 6 (gur) and 2 (barig) of barley, wages for earth work 7. and for cutting reeds. 8. on the bank of Idlugal, 9. via Zerzeranum, canal inspector. 10. 13 gur (of barley) and 2 (barig) of lentils— 11. seed and fodder for oxen involved with hoeing 12. as well as grain rations for sagud workers (oxdrivers); 13. †u-Ishara, farmer for 34 ˜elds. 14. 7 gur and 5 (barig of barley)—(monthly) wages for (the crews) of hired boats. 15. —(12 month) subtotal: 86.0.0 gur (of barley).
33
16. 2 gur, grain rations for Awat-ili, an additional (boatman) for 10 months. 17. 10 kurku vessels for 3 litres of dida-beer each, 18. the 2 gur of barley is for the purchase of the jars. 19. from the sealed (tablet) of Ili-asranni, the sabra. 20. 1 (gur) of ˘our and malt and 1 fattened sheep 21. for Lugal-ane— 22. (from) one sealed (account) of †u-Estar, the scribe. iv 1. Total: 924 gur, 2 (barig), and 4 (ban) of barley, 2. Total: 35 gur, 4 (barig), and 1 (ban) of ˘our; 3. Total: 5 gur, 3 (barig), and 4 (ban) of emmer; 4. Total: 7 gur of broad beans; 5. Total: 12 gur, 3 (barig), and 2 (ban) of lentils; 6. Total: 1 gur of oil, 1 grazed sheep, 7. seed and fodder, hoe work, grain rations, and various withdrawals, 8. from (the accounts) of Tani’a, 9. withdrawn. 10. Year that Simanum was destroyed (Shu-Sin year 3). Commentary Some of the names mentioned in the text have not been attested in any Ur III text published to date. These are: †u-Malik, Pilah-ilum, Imid-ahi, and Zerzeranum. Column i 6–7. Note that these are months one and twelve of the imperial calendar. 11. The name Isar-libur is otherwise attested only in two Drehem texts: CTMMA 1 17:65 (AS4.vii.0) and PDT 1 561:10 (AS9.vi.6) 17. The name Ili-asranni is well attested at Garsana. It is also known from the “SI.A.A archive,” Ur, Drehem, and Umma. A person by that name, who served as sabra, is known from a Garsana text dated †S 7. See David I. Owen, The Garsana
34
19.
PIOTR MICHALOWSKI AND PARSA DANESHMAND Archives, forthcoming (courtesy of the author). This may very well be the general †arrumbani. The name occurs in Garsana.
Column ii 6. This is the ˜rst time that we encounter a sabra of Uruk. 8. To our knowledge, ma2-gid2-e-ne is only known from Garsana. 9. u3-ze2-na-pi-ir dub-sar must be compared to u2-zi-na-pi-ir, who is attested in Garsana, once in IS1.vii and once in IS2.v, also designated as dub-sar. 12. David Owen kindly informs me that the name Ta2-at-tu-ri, written ta-tu3-ri, appears frequently in Garsana documents dated to †S6. Column iii 4. See perhaps u3 bar-ra id2-lugal-ka (Or 47/9 361:8, Umma) and u3 bar-ra id2-EGIRd nanna (NRVN 1 264 5–6, Nippur). 5. This is very likely the same person as Erra-nuååid sukkal lugal in the Umma text SAT 2 606:21, viii † 48 and sukkal lú kin-gi4-a lugal in an unpublished tablet from iri-sag-rig7 dated to AS 9 (courtesy of David Owen). 7. The verb IMxTAK4, which seems to mean, “to cut,” has been discussed by M. Civil, Farmer’s Instructions 81. There are only a few attestations, primarily from Girsu, where it usually refers to trees, and only once to grass. This is the ˜rst time it is documented with reeds. 8. A watercourse id2-lugal, and a place by the same name, is attested many times in texts from Umma.
13.
15.
16.
17.
†u-Ishara is not a common name in Ur III. Most attestations document the same person, from Mari, who was present in Sumer in the year AS 6 and is documented in Drehem texts (CST 468:9; MVN 11 146:14; MVN 13 635:7; Hirose 231:2; JCS 7 104:2, possibly CTMMA 1 21:20). A messenger (lu2-kas4) by that name is known from Girsu (TCTI 1 715:9, †S 9.x). None of the attested Ur III usages of gú seems to be pertinent here. Because the number in this line is equivalent to twelve times the wages mentioned in line 14, is seems probable that the latter is the monthly expenditure, and that this is the yearly total. One thinks of the older Sumerian gú-an-sè, used for subtotals, but it is not attested after the Old Akkadian period. Only one person named Awat-ili is otherwise known, a builder (sitim) who receives two sheep in AnOr 7 95:18. dug-ku.kur-du3 is a writing of the well attested pythos dug-kur-ku-du3 (W. Sallaberger, Der babylonische Topfer und seine Gefasse ach Urkunden altsumerischer bis altbabylonischer Zeit sowie lexikalischen und literarischen Zeugnissen [Ghent: University of Ghent, 1996], 102). The order of constituents encountered here is also known from a tablet from Girsu (MVN 5 268:3u) and from one from Garsana.
Column iv 4–5. The grain total is lightly oˆ; the ˘our total combines the ˜gures for dabin and zi3-munu4.
Reverse
35
Obverse
AN UR III TABLET FROM IRAN
36
PIOTR MICHALOWSKI AND PARSA DANESHMAND
Top
Bottom
Right Edge
QUELQUES TEXTES SCOLAIRES PALÉO-BABYLONIENS RÉDIGÉS PAR DES FEMMES Brigitte Lion (Université de Paris) and Eleanor Robson (University of Cambridge)
À l’époque paléo-babylonienne, on peut répertorier plusieurs cas de femmes ayant pratiqué l’art scribal, surtout dans les milieux auxquels les hommes avaient un accès restreint. Deux corpus sont bien connus: d’une part, les tablettes trouvées dans les appartements des femmes du palais de Mari;1 d’autre part, la documentation de Sippar, où l’activité de femmes scribes était liée à la présence du gagûm («cloître»), institution qui abritait les nadÿtum consacrées au dieu †amas et limitait probablement leurs contacts directs avec la population masculine de la ville.2 Cependant il ne reste que peu de traces attestant la façon dont ces femmes ont pu apprendre leur métier. Seules, quelques rares tablettes d’exercice portant mention d’une rédactrice sont parvenues jusqu’à nous.3 Cette particularité n’a jamais vraiment retenu l’attention des chercheurs, probablement parce que les textes en question sont des plus courants dans le cursus scolaire.4 Cependant,
les études récentes consacrées à la formation des scribes permettent désormais de mieux les situer. 1. Les textes Les textes sont présentés dans cette première partie suivant l’ordre dans lequel ils ont été publiés. 1.1. Le Chant de la Houe Une tablette scolaire rédigée par une femme a été publiée dès 1913: VS 10 207 (VAT 6457 + 6492 + 6493). Elle est fragmentaire, avec deux colonnes sur la face et deux au revers, et contient la composition sumérienne dite Chant de la Houe,5 copiée dans son intégralité, ce qui représente plus d’une centaine de lignes. À la ˜n du revers, dans la colonne de gauche, un colophon d’une seule ligne porte la mention su munus dub-sar, «écrit par une femme scribe». Cette scribe n’indique pas son nom. Le Chant de la Houe présente une structure complexe.6 La première partie, de loin la plus longue, commence par une évocation mythologique de la création de la Houe, et se poursuit par
Nous adressons nos remerciements à D. Charpin, B. Lafont, J. Marzahn, C. Michel, C. Proust, S. Richardson, M. Tanret, N. Veldhuis et C. B. F. Walker. 1. Ziegler 1999. 2. Harris 1975 et 1990, Meier 1991, Lion 2001. 3. Ce fait est signalé par Sjöberg 1976: 177, n. 66. 4. L’emploi du terme «scolaire» dans cet article renvoie à la notion d’apprentissage s’eˆectuant selon un cursus bien établi; en revanche, il ne fait pas référence à un cadre institutionnel dans lequel cette formation serait dispensée, une telle institution n’existant probablement pas à l’époque paléobabylonienne (Tanret 2002: 168).
5. Il ne s’agit donc pas d’une «Tafel mit einem Liede wohl an Nisaba», comme l’indiquait Zimmern 1913: XII; il a vraisemblablement été trompé par la mention classique dnisaba zà-mí, qui ˜gure à la ˜n du texte. 6. Wilcke 1972–1975: 36–38.
37
JCS 57 (2005)
38
BRIGITTE LION AND ELEANOR ROBSON
l’énumération des usages qu’en font divers dieux. La deuxième partie présente des spéculations philologiques sur le nom de la Houe. La troisième recense les emplois de cet outil, et une doxologie termine l’œuvre. Assez di¯cile à classer en termes de «genre» littéraire, ce poème a reçu des titres variés dans les langues modernes.7 Son titre antique était, selon l’usage, son incipit, en en5-du7-e, «Le Seigneur (a vraiment fait apparaître) tout ce qui convient». Cette composition a fait l’objet d’une abondante littérature critique.8 Le Chant de la Houe est connu par près d’une centaine de manuscrits, dont 70 proviennent de Nippur.9 Quelques autres sites, comme Ur10 ou Babylone, en ont fourni aussi des copies. Dans les recensions des diˆérents manuscrits, VS 10 207 a reçu l’appellation de manuscrit H.11 H. Zimmern, qui a copié le texte, a noté: «Große, dicke Tafel mit großer Schrift».12 Il n’est pas certain que la scribe doive être tenue pour responsable de l’aspect peu ra¯né de sa tablette (165 x 135 x 40 mm, d’après le catalogue). Une photographie permet de mieux apprécier son écriture,13 qui est très lisible, avec des signes bien formés. Quant au texte, si on le compare à la version standard,14 il comporte de nombreuses variantes qui ne sont pas forcément fautives. Par exemple, aux l. 8 (= i 6 de VS 10 207) et 25 (= i 22), elle ajoute le déterminatif Gis devant al, alors qu’il n’est pas du tout question de la houe (al), mais que al est un pré˜xe verbal dans le premier cas et le début du verbe composé al . . . dug4 dans le second; cela 7. Par exemple Lehrgedicht / Gedicht / Lied von der Hacke, Creation of the Pickax, Creation / Song of the Hoe, Hymne à la Houe . . . 8. Voir p. ex. Farber 1999, qui donne une bonne partie de la bibliographie antérieure, ainsi que la bibliographie ˜gurant sur le site développé par Black et al. 1998–. 9. Edzard 2000 donne une liste des manuscrits. Pour un compte global, voir Robson 2001: 53, qui se fonde sur les données ˜gurant dans Black et al. 1998–. 10. Charpin 1986: 428, 442 et 449. 11. Edzard 1966: 87; cette désignation «H» lui est restée dans les publications postérieures. 12. Zimmern 1913: XII. 13. Cette photographie, ainsi que celle de VAT 6574 (cidessous § 1.2), nous a été fournie par J. Marzahn, que nous remercions vivement. 14. Établie par Black et al. 1998–.
peut être tenu pour un jeu graphique, attesté par plusieurs autres manuscrits de la composition.15 Quelques erreurs peuvent cependant être relevées: à la l. 10 de la version standard (= i 8), la scribe oublie la racine verbale; elle saute les l. 21 et 93, et inverse les l. 90 et 91 (= iv 5 et 4).16 Elle semble donc, pour le moins, distraite . . . .17 1.2. Proto-Aa VAT 6574 est une liste lexicale fragmentaire. Landsberger 1967: 148–49, en a donné une transcription et une traduction d’après une photographie. Puis Civil 1979: 135–36 l’a à nouveau transcrite. Il n’existe pas de copie publiée. Là encore, une photographie communiquée par les services du Musée de Berlin permet de préciser la situation: il s’agit d’un fragment dont toute la face est perdue, mais ce qui subsiste du revers est parfaitement lisible et correspond à la ˜n des trois dernières colonnes. M. Civil le décrit comme le «lower left corner of a multicolumn tablet», et
15. Wilcke 1972–1975: 36 et Edzard 2003: 87. 16. Nous remercions vivement Bertrand Lafont d’avoir bien voulu examiner la copie et de nous avoir fait part de ses remarques. La photographie permet de préciser les points suivants: l. 8 (= i 6), la racine verbale Gar est bien lisible, juste après la cassure; au revers les l. 100–101 n’ont probablement pas été oubliées, mais l’une au moins, sinon les deux, se placent dans la cassure entre deux fragments de la tablette, après la l. 99 (= iv 11). 17. Qu’une composition copiée sur une grande tablette comporte des fautes n’est pas rare. Cooper 1983: 46 a noté ce phénomène à propos des diˆérents manuscrits de la Malédiction d’Agade: «Manuscript quality at Nippur cannot be correlated with format. The worst manuscript in terms of omissions and errors is unquestionably Y2, a complex six-column edition, and the likewise complete G (prism) and S1 are not especially impressive. In fact, the necessity of crowding and anxiety about whether there would be enough room to complete a line or the composition may have had a deleterious eˆect on the texts of Y2 and S1. Thus, it would be incorrect to characterize any of the complete editions as master exemplars as opposed to school exercises represented by the excerpt tablets». Veldhuis 1997: 31 renvoie à ces observations de J. Cooper; il suggère également que les prismes puissent avoir été écrits à l’occasion des examens. Si jamais les grandes tablettes de type I avaient une fonction de ce genre, on comprendrait d’autant mieux les erreurs des scribes en formation; cela pourrait aussi expliquer la proximité des dates de rédaction de EPHE 495 et BM 96950 (ci-dessous § 1.3 et 1.4), mais il s’agit d’une pure hypothèse.
QUELQUES TEXTES SCOLAIRES PALÉO-BABYLONIENS
B. Landsberger comme «a tablet fragment with part of the last three columns of the reverse». Selon la classi˜cation établie par M. Civil, 18 il s’agit d’une tablette de «type I», c’est-à-dire d’une grande tablette contenant une bonne partie ou la totalité d’une liste lexicale, à l’exclusion d’un autre texte.19 Ici, seule une vingtaine de lignes subsiste. Le coin gauche conservé donne la ˜n du revers, avec un colophon dont les deux dernières lignes sont [su be]-ªelº-ti-re-mi-ni, ªmunusº dubsar, «[main de] Beltÿ-reminni, femme scribe».20 La liste copiée est Proto-Aa, précurseur de la série Aa = nâqu. Il s’agit d’une liste de signes à quatre entrées;21 la première (0) fait commencer chaque ligne par un clou vertical; vient ensuite la prononciation (1) du mot sumérien ˜gurant dans l’entrée suivante (2); en˜n celle de droite (4) donne une traduction akkadienne du mot sumérien. Dans VAT 6574, les colonnes 0 et 1 sont inversées: le clou vertical ˜gure entre la lecture du mot sumérien et l’idéogramme de ce mot.22 Les exemplaires les plus nombreux de cette série proviennent de Nippur. Cependant, il existe des recensions diˆérentes, trouvées soit à Nippur, mais ne correspondant pas à la tradition générale, soit sur d’autres sites comme Kis, Ur et Sippar. 18. Civil 1969: 27; 1995: 2308. 19. Les tablettes de type «II», de taille moyenne, dont la face (II/1) porte un exercice préparé par le maître et recopié par l’élève, et le revers (II/2) une composition souvent différente de celle de la face, répartie sur plusieurs colonnes, n’ont pas de colophon. 20. Cette mention exceptionnelle, soulignée par l’éditeur du texte, a aussi été relevée par Labat 1970: 187: «Du fragment additionnel VAT 5674, étudié p. 148–49, nous noterons, à titre épisodique, qu’il est signé d’une femme, Bêlti-remÿni». 21. La numérotation des entrées, proposée par Landsberger 1967: 125, a ensuite été adoptée pour la description des listes lexicales. La diˆérence entre les séries Proto-Aa et Proto-Ea tient à l’entrée (4): Proto-Ea ne donne pas les traductions akkadiennes, ou alors n’en donne qu’une seule, alors que Proto-Aa oˆre en général plusieurs équivalents akkadiens pour un même mot sumérien. Sur ces listes, voir Cavigneaux 1980–1983: 620–21 et Civil 1995: 2308–309. 22. L’ordre des colonnes est donc 1-0-2-4, comme l’a bien vu Landberger 1967: 148, mais Civil 1979: 108 et 135 a indiqué par erreur 0-1-2-4. Sur les 22 exemplaires de cette liste rassemblés par M. Civil 1979, seules trois autres tablettes présentent les entrées dans l’ordre 1-0-2-4: no 10 (p. 127–31) provenant de Sippar, no 19 (p. 140–41) de provenance inconnue et no 22 (p. 143–44), une tablette médio-babylonienne d’Ugarit.
39
VAT 6574 pourrait avoir Sippar pour origine, selon B. Landsberger, pour deux raisons: d’une part, son numéro d’inventaire la rattache aux lots publiés dans VS 2 et 10, dont certains textes viennent de Sippar;23 d’autre part, Sippar est la seule ville de Babylonie où la documentation témoigne du travail de nombreuses femmes scribes. Cette proposition a été acceptée par Civil 1979, qui a publié aussi deux autres recensions de la liste Proto-Aa provenant de Sippar; mais le nom de leur rédacteur (ou rédactrice), si jamais il a été indiqué, n’y est pas conservé. Ces deux exemplaires de Sippar sont des fragments plus grands et conservent donc des extraits plus longs que VAT 6574, qu’ils recoupent en partie.24 Quant au nom de la scribe, Beltÿ-reminni, il demeure cependant inconnu à Sippar non seulement parmi les femmes scribes,25 mais même parmi la population féminine de la ville.26 Landsberger 1967: 148 a en outre suggéré que les tablettes VS 10 207 et VAT 6574, toutes deux dues à des femmes, pourraient être de la même main, selon des critères paléographiques («paleographic considerations») qu’il n’a malheureusement pas explicités. Il semble di¯cile de trancher, même en examinant attentivement les photographies;27 on peut noter que Beltÿ-reminni a indiqué son nom à la ˜n de VAT 6574, alors que la scribe VS 10 207 ne l’a pas fait, même si elle disposait de toute la place nécessaire. 23. Zimmern 1912: V. 24. Civil 1979: 122–27 (no 9) et 127–131 (no 10); il s’agit de deux tablettes du British Museum. Sur les 22 exemplaires publiés par M. Civil, ce sont les deux seuls à oˆrir des parallèles avec l’extrait conservé sur VAT 6574. 25. Harris 1975: 196–97 donne une liste des femmes scribes de Sippar, à compléter par Stol 1976: 152. Depuis 1976, les publications de textes de Sippar ont révélé quelques autres femmes scribes, mais aucune ne s’appelle Beltÿ-reminni. 26. D’après les index ˜gurant dans les publications récentes des textes de Sippar: Lerberghe et Voet 1991, Dekiere 1994a et b, 1995a et b, 1996, 1997, Al-Rawi et Dalley 2000, ainsi que Richardson 2002. 27. Par exemple les signes ni et sar semblent diˆérents sur les deux tablettes, comme nous le fait remarquer J. Marzahn. Mais d’autres signes sont très semblables. En outre, sur une même tablette, la graphie des signes peut montrer de légères variantes. Sur les deux tablettes, l’écriture présente des graphies paléo-babyloniennes très classiques, ce qui rend di¯cile le repérage de diˆérences signi˜catives.
40
BRIGITTE LION AND ELEANOR ROBSON
1.3. Ur5-ra Parmi les documents conservés à l’École Pratique des Hautes Études à Paris et copiés par J.-M. Durand,28 le no 495 (pl. 123) peut être ajouté à ce petit corpus. Sa provenance n’est pas connue.29 Le catalogue p. 31 donne ses dimensions: 6,8 x 6,5 cm et le décrit comme un «texte lexical (laine et habits)». Il s’agit d’un extrait de la grande liste lexicale ur5-ra.30 Il ne reste qu’un fragment du revers de la tablette: deux colonnes, dont celle de gauche porte, après la dernière ligne de texte, deux traits horizontaux, puis le colophon. Il s’agit là encore d’une tablette de «type I». 31 Dans le colophon
EPHE 495 revers iu 1u siki [. . .] 2u siki [. . .] 3u siki ªxº [. . .] 4u siki sal-ªla?º 5u túg-bar-[si4] 6u túg-bar-si4-sal-[la] 7u túg-bar-si4-sig7? 8u túg-níG-lá[m] 9u túg-níG-lám-ªsig7?º 10u túg-níG-lám-ªsal-la?º 11u túg-níG-eren2-a-ªak?º 12u túg-a-ªak?º 13u túg-su-a-ªak?º 14u [túg . . .]-ªguzº-[za] 15u [. . .]
˜gure la mention su munus dub-sar, «main d’une femme scribe»32 et, après un blanc, la date. La transcription de cette tablette est donnée ci-après, ainsi que les comparaisons avec les «précurseurs» paléo- et médio-babyloniens connus pour cette section, tels qu’ils ˜gurent dans Landsberger, Civil et Reiner 1970. Un autre extrait de la série des textiles, originaire de la maison d’Ur-Utu à Sippar, a été publié par Tanret 2002: 64–66, no 24, mais il ne correspond que très partiellement à cette section, et ne recouvre pas non plus avec exactitude les autres précurseurs de la série. En fait certaines sections de ur5-ra, à l’époque paléo-babylonienne, ne sont pas encore canonisées et sont sujettes à des variations.33
précurseur Nippur MSL 10, p. 146–48
précurseur OB tardif MSL 10, p. 143
précurseur Ugarit MSL 10, p. 149–51
109 túg-bar-si 110 túg-bar-si-gal 111 túg-bar-si-hus-a 43 túg-níG-lám 44 túg-níG-lám-hus-[a]
22 siki sal-la 27 túg-bar-si 28 túg-bar-si-sal-la 29 túg-bar-si-hus 30 túg-níG-lám 31 túg-níG-lám-hus
44 siki sal-la 68 túg-bar-dul 69 túg-bar-dul-sal-la 70 túg-bar-dul-hus-a 74 túg-níG-lám-ma 78 túg-níG-lám-ma-hus-a 79 túg-níG-lám-ma-sal-la 122 túg-a-ak-a
60 túg-guz-za
28. Durand 1982. 29. Un premier catalogue de cette collection a été établi par Scheil 1932–1933; il décrit 361 tablettes, ainsi que des briques et estampages, de la collection «formée en 1910–1914» (p. 3). La présente tablette ne ˜gure pas dans ce catalogue et a donc été acquise plus tard par l’EPHE. 30. Landsberger, Civil et Reiner 1970: 125–56. A Nippur, à l’époque paléo-babylonienne, ce passage correspond à la ˜n de la section 4 de la liste (Veldhuis 1997: 47). 31. Veldhuis 1997: 31, n. 86, qui note par ailleurs (p. 53): «The last section of division 4 (wool and clothing), on the other
hand, is rarely copied on the reverse of a type II tablet»; il en cite cependant quelques exemples, pp. 50–51. A Sippar, voir Tanret 2002: 64–66, no 24, la liste ˜gure sur la face d’une tablette de type II, où elle est copiée à gauche par le maître alors que le côté droit a été eˆacé; le revers est anépigraphe. 32. Comme l’a souligné Durand 1982: 38; le fait a également été remarqué par Veldhuis 1997: 31, n. 86. 33. Veldhuis 1997: 53.
QUELQUES TEXTES SCOLAIRES PALÉO-BABYLONIENS
revers iiu 1u [gada? . . .] ªaº 2u [gada? . . .] ªxº a 3u [gada? . . .] ªxº la? bi 4u [gada? . . .] ªxº la? 5u [gada? . . .] ªxº la? Gìr? na 6u su munus dub-sar 7u iti du6-kù u4 21–kam 8u mu sa-am-su-i-lu-na ªlugalº-[e] 9u ªlugal níº-[gi gú-bar-ra] Remarques: Revers col. iu l. 5u–7u: EPHE 495 note si’ là où les autres versions paléo-babyloniennes notent si et celle d’Ugarit dul (U+SI4). l. 7u et 9u: là où les autres recensions ont hus, EPHE 495 a ªsig7º, ou peut-être ªgalº ou ªsagº. l. 8u–10u: les premières lignes de la tablette de la maison d’Ur-Utu, Tanret 2002: 64–66, no 24, sont [túg]-ªníGº-lám, ªtúg-níGº-lám-sag, túg-níG-lámhus-a. l. 11u: des túg-níG-bara2 ˜gurent dans le précurseur de Nippur l. 58–59 et dans Tanret 2002: 64– 66, no 24, l. 8–10; s’agirait-il ici d’une variante avec le signe eren2? Revers col. iiu l. 1u–5u: les cinq lignes conservées sont trop abîmées pour être compréhensibles, mais elles devaient probablement enregistrer des types de tissus en lin (gada) que l’on trouve à la ˜n de cette section de ur5-ra. Elles ne semblent avoir de parallèle dans aucun des précurseurs connus. l. 7u–9u: la tablette daterait du 21—vii-Si 14. Le nom de l’année est cassé, on pourrait aussi choisir pour la l. 9u d’autres restitutions: ªlugal sagº-[kal . . .] (an 20), ªlugal nam!º-[. . .] (an 24), voire bàd ªìº-[siin . . .] (an 15), bàd ªan!º-[da . . .] (an 16), bàd ªdidliº [gal-gal . . .] (an 17) (Horsnell 1999). La proposition de dater ce texte de l’an 14 vient de la comparaison avec un autre exercice, voir ci-dessous § 1.4. 1.4. Le «Silbenalphabet A» Un quatrième document, BM 96950, conservé au British Museum, est inédit. 34 Il s’agit d’une 34. S. Richardson nous a indiqué ce dernier document, que C. B. F. Walker lui avait signalé. Que tous deux soient ici vivement remerciés.
41
grande tablette (19 x 13,5 x 3,5 cm) à quatre colonnes de chaque côté; au revers, seule la première colonne est inscrite, ainsi que la dernière où ne ˜gure que le colophon. Elle porte le Silbenalphabet (ou Syllable Alphabet) A, en entier, bien écrit, avec quelques variantes par rapport au texte habituel. Encore une fois, c’est une tablette de «type I». Les Silbenalphabet sont des listes de syllabes, qu’A. Cavigneaux présente ainsi: «Ces listes énumèrent des combinaisons syllabiques—en général 2 ou 3 syllabes—elles-mêmes réparties en groupes de 2 ou 3 (. . .). C’est à Landsberger que revient le mérite d’avoir identi˜é la nature de cette liste mystérieuse: il s’agit de noms propres ou d’éléments de noms propres». 35 Ce dernier point a été nuancé par N. Veldhuis qui, même s’il admet que certains éléments sont bien des noms propres, pense que cette liste est en premier lieu un exercice d’écriture et d’apprentissage des signes.36 On distingue le Silbenalphabet B, typique de Nippur, et le Silbenalphabet A, copié en dehors de Nippur,37 ce qui exclut déjà une provenance de Nippur pour le présent texte. B. Landsberger a signalé l’existence de plusieurs exemplaires paléobabyloniens du Silbenalphabet A provenant de Sippar.38 On doit désormais ajouter neuf, peutêtre même onze tablettes d’exercices trouvées dans la maison d’Ur-Utu et publiées par Tanret 2002: 31–50 (no 2 et 6 à 15); elles contiennent divers passages de ce syllabaire, qui est écrit jusqu’à deux ou trois fois de suite sur une même tablette. D’autres exemplaires proviennent d’Ur, Uruk, Larsa, Isin, Kis, Tell Harmal et Ishchali.39
35. Cavigneaux 1980–1983: 618–19. L’interprétation selon laquelle il s’agit de noms propres est exprimée dans Landsberger 1959: 101–14. 36. Veldhuis 1997: 41–42. 37. La version standard du Silbenalphabet A a été établie par Landsberger 1933 et reprise dans Landsberger 1959: 100–101. Sollberger 1965: 21–28 a également édité un Silbenalphabet A paléo-babylonien, de provenance inconnue. 38. Landsberger 1933: 170: il s’agit des textes publiés par Scheil 1902. 39. Veldhuis 1997: 43, n. 134, avec la bibliographie.
BM96950 (L’édition est par E. Robson.)
42 BRIGITTE LION AND ELEANOR ROBSON
QUELQUES TEXTES SCOLAIRES PALÉO-BABYLONIENS L’étude la plus récente du Silbenalphabet A ˜gure dans Tanret 2002: 31–50, qui suit l’édition de Sollberger 1965.40 Les remarques ci-dessous
43
se reportent également aux diˆérentes versions connues de cette série.
Face: col i
col ii
col iii
col iv
1. me-me 2. pap-pap 3. a-a 4. a-a-a 5. ku-ku 6. lu-lu 7. mas 8. mas-mas 9. mas-kak 10. mas-ni 11. si-kak 12. si-ni 13. si-a 14. u-bar 15. bar-bar 16. lá-lá 17. si-bar 18. bar-ªsiº 19. si-ªsiº 20. si-[si]-ªsiº 21. a-ªsiº 22. a-s[i]-si 23. [m]e-a 24. [me]-ni 25. [as]-ni 26. [as]-ur
27. nun-[ni] 28. nun-[ur] 29. a-[ku] 30. kil!-[a] 31. a-[pap] 32. pap-[a] 33. a-ªanº 34. ªan-aº 35. kur-[ba] 36. kur-u-[ta] 37. me-[zu] 38. me-[wa-zu] 39. ì-ªzuº 40. a-ªzuº 41. ªzu-zuº 42. ªì-baº 43. [ì-ba-ba] 44. [a-ba] 45. [a]-ªba-baº 46. ba-ba 47. ba-ba-a 48. ba-za 49. ba-za-za 50. ni-a 51. a-ni 52. tab-ni
53. [kas]-ni 54. ni-ur-ba 55. [b]e-lí 56. [G]is-be 57. [n]u-nu 58. [a]-nu 59. [saG]-kud 60. [saG]-kud-da!-a 61. [saG]-an 62. ªsaG-an-tukº 63. [saG]-ªGu10º 64. [saG]-ªkurº 65. ªsaG-kur-taº 66. ªkud-daº-[a] 67. ªkudº-[da] 68. [gab]-ªgabº 69. ªnin-gabº 70. ªnin-ezenº 71. ªnin-sukkalº 72. nin-ªsukkal-an-kaº 73. pú-ta 74. sila-ta 75. é-ta 76. é-gud 77. an-dùl 78. an-an-dùl 79. an-Gá
80. an-kal 81. tam-ma 82. tam-tam-ma 83. ug4-ga 84. ug4-ug4-ga 85. an-gàr 86. gàr-an 87. an-ás 88. ás-an 89. ªanº-ba 90. an-ba-ni 91. an-ni 92. an-ni-zu 93. dùg-ga 94. dùg-dùg-ga 95. me-dùg 96. me-dùg-ga 97. si-ªbaº-ni 98. si-ba-ur 99. hu-hu 100. hu-ba 101. hu-ur 102. hu-ru 103. an-ú 104. ú-ªaº 105. ú-d[a] 106. pa-pa 107. pa-ªGáº
Remarques:40 Col. ii, l. 29–30: pour ces lignes, certains manuscrits de Sippar donnent a-ba, ba-a; un manuscrit aurait a-me-[. . .] à la place de a-ba (Scheil 1902: 38, l. 32–33 et n. 2); en revanche Tanret 2002: 35, no 2, iii l. 17u-18u et 46, no 13, ii, 3u-4u donne a-ku, kil-a. 40. La provenance du texte publié par Sollberger 1965 est inconnue, la tablette est conservée au British Museum.
Col. ii, l. 35–36: entre ces deux lignes, certains manuscrits de Sippar insèrent: kur-ta (Scheil 1902: 38, l. 39), mais pas ceux trouvés dans la maison d’Ur-Utu (Tanret 2002: 35, no 2, R iii l. 1–2 et 46, no 13, R iu l. 5–6). Col. ii, l. 42–48: variante dans un manuscrit de la maison d’Ur-Utu, Tanret 2002: 46, R iiiu, 4–8: ì-ba, ì-zu, a-ba, a-ba-[ba], ªbaº-[za]. Col. ii, l. 52: ni-ni dans certains manuscrits de Sippar (Scheil 1902: 38, l. 56), mais pas dans celui
44
BRIGITTE LION AND ELEANOR ROBSON
de la maison d’Ur-Utu (Tanret 2002: 36, no 2, R iv l. 4). Col. iii, l. 54: ni-ur-ni dans certains manuscrits de Sippar (Scheil 1902: 38, l. 58); mais [ni]-ªurº-ba dans Tanret 2002: 39, no 8, iiiu l. 5. Col. iii, l. 55: ba-ú-lí dans certains manuscrits de Sippar (Scheil 1902: 38, l. 59); mais be-lí dans Tanret 2002: 39, no 8, iiiu l. 6. Col. iii, après la l. 55: la ligne be-lí-tabx manque (Landsberger 1959: 100, l. 26b, Sollberger 1965: 23, l. 56). Mais cette ligne manque aussi dans les manuscrits de Sippar (Scheil 1902: 38, entre les l. 59 et 60). Ce passage n’est conservé dans aucun des exemplaires de Tanret 2002. Il peut donc s’agir soit d’omissions des scribes, soit d’une version standard de Sippar. Col. iii, l. 57: nu-be dans Scheil 1902: 38, col. l. 61. Col. iii, l. 58: a-nu également dans Scheil 1902: 38, l. 62 et Sollberger 1965: 23, l. 59, mais a-nu-nu dans d’autres manuscrits (Landsberger 1959: 100, l. 27b). Col. iii, l. 60: saG-kud-da ou saG-kud-da-a dans Scheil 1902: 38, l. 64–65 et n. 3 et dans Landsberger 1959: 100, l. 28b; mais saG-kud-dar-a dans Sollberger 1965: 23, l. 61. Col. iii, l. 62: cette ligne manque dans Scheil 1902: 38, entre l. 66 et 67. Col. iii, l. 65: saG-ga-ta dans Scheil 1902: 39, l. 69; la ligne suivante commencerait également par saG-[. . .]. Mais saG-kur-ta dans Landsberger 1959: 100, l. 30b. Ligne cassée dans Sollberger 1965.
Col. iii, l. 66–67: ces lignes, endommagées, sont inversées (Landsberger 1959: 100, l. 31b, 31a; Sollberger 1965: 23, l. 68, 67). Dans Scheil 1902: 39, elles devraient correspondre aux l. 71–72, mais la l. 71 est complètement perdue, et ce qui reste de la l. 72 (un clou vertical) ne ressemble pas à un kud. Elles manquent complètement dans Tanret 2002: 44, no. 12. Col. iii l. 68: dans Scheil 1902: 39, les l. 73 et 74 montrent des signes ou débuts de signes qui n’ont rien à voir avec gab. Col. iii l. 72: nin-sukkal-an-ta dans Scheil 1902: 39, l. 78. La ˜n de la ligne est cassée dans Tanret 2002: 40, no 9, iiiu, l. 3. Col. iii l. 76: Scheil 1902: 39, l. 82 donne é-ta, exactement comme à la l. précédente. Col. iii l. 78: variante an-dùl-dùl dans Landsberger 1959: 101, l. 36 b. Col. iv, l. 85–88: l’ordre de ces lignes dans Tanret 2002: 42, no 10, iiu l. 1u–4u, est [gar]-ªanº, ªan-garº, ªásº-an, an-ás. Col. iv, l. 93: avant cette ligne, Scheil 1902: 39, l. 99 a une ligne supplémentaire, dùg-ni-ba. Col. iv, l. 96: cette ligne manque dans Scheil 1902: 39, entre les l. l. 102 et 103. Col. iv, l. 97: l’un des manuscrits de Sippar aurait za-ba-ni, Scheil 1902: 39, l. 104 et n. 2. Col. iv, l. 102: cette ligne manque sur l’un des manuscrits de Sippar, Scheil 1902: 39, l. 108 et n. 3. Col. iv, l. 105: même chose dans Scheil 1902: 39, l. 111 et dans Sollberger 1965: 24, l. 106, mais ú-da dans Lansberger 1959: 101, l. 49c.
Revers: col v
col vi
col vii
col viii
108. pa-pa-Gá 109. a-a-ur 110. an-ur 111. ni-ni 112. ni-ni-ªaº 113. ni-ni-ni 114. ni-ni-ni-a 115. ab-ba 116. ab-ba-Gu10
(non inscrite)
(non inscrite)
(haut de la colonne non inscrit) su munus dub-sar iti du6-kù u4-13–kam mu sa-am-su-i-lu-na lugal-e lugal ní-gi gú-bar-ra
QUELQUES TEXTES SCOLAIRES PALÉO-BABYLONIENS
45
117. ab-ba-ni 118. ab-ba-a 119. ab-ba-uru3 120. ªigiº-sa11 121. igi-sa11-s[a11] 122. sa11-s[a11] 123. sa11-sa11-ªaº
Remarques: Col. v, l. 108: même chose dans Scheil 1902: 39, l. 114. En revanche Landsberger 1959: 101, l. 48c, donne pa-Gá-Gá. Les deux premiers signes sont perdus dans Sollberger 1965: 24, l. 109 et dans Tanret 2002: 40, no 9, R iu, 1u. Col. v, l. 112: cette ligne est omise dans Tanret 2002: 40, no 9, R iu.
BM 96950
Scheil 1902
ªigiº-sa11 igi-sa11-s[a11]
sa11-s[a11] sa11-sa11-ªaº
sa11 [. . .]-sa11 sa11-sa11 sa11-sa11-sa11 sa11-sa11-a
Col. v, l. 119: le ki ˜nal est absent de tous les autres manuscrits. Scheil 1902: 40, l. 125, donne ab-ba-Gìr. Col. v, l. 120–123: la ˜n du Silbenalphabet A comporte de nombreuses variantes. Elle ne subsiste dans aucun des manuscrits de Tanret 2002. Nous présentons les diˆérentes versions sous forme de tableau:
Landsberger 1933
Sollberger 1965
igi-sa11
[igi]-sa11 [igi]-sa11-sa11
igi-sa11-aa sa11
sa11
sa11-sa11-ab
[sa11-s]a11-a
a. Landsberger 1959: 101 propose aussi pour cette ligne une variante igi-sa11-sa11. b. Landsberger 1959: 101 donne en revanche pour cette ligne sa11-sa11.
Col viii, colophon: «main d’une femme scribe, le 13–vii-Samsu-iluna 14».41 Ce colophon, par son contenu et sa disposition sur la tablette, ressemble beaucoup à celui de la tablette de l’EPHE. Tous deux datent en outre du règne de Samsu-iluna, du même mois, et il est tentant de suggérer qu’ils auraient pu être rédigés la même année, donc à huit jours d’intervalle, par une même femme.
41. Horsnell 1999 vol. 2: 199–200.
2. La formation scribale: comparaison avec les données de Nippur Des études récentes portant sur l’éducation à Nippur,42 où ont été retrouvés les lots les plus importants de textes scolaires, ont permis de dé˜nir d’une façon assez ˜ne le cursus suivi par les élèves. Même si les tablettes présentées ci-dessus ne proviennent pas de cette ville, elles s’intègrent par-
42. Pour une présentation générale du cursus mésopotamien, voir Pearce 1995, en particulier pp. 2270–72. A Nippur: Veldhuis 1997, Tinney 1999 et Robson 2001.
46
BRIGITTE LION AND ELEANOR ROBSON
faitement dans cette formation. Elles permettent, par comparaison avec les données de Nippur, de voir à quel niveau de connaissances la ou les femmes scribes ayant copié ces textes étaient parvenues. Une éducation que l’on a quali˜ée d’«élémentaire»43 s’organise à Nippur en quatre phases: 1o) Acquisition des techniques de l’écriture. Le Silbenalphabet, dans sa version B à Nippur, correspond à un travail de débutant, puisqu’il s’agit d’une liste de signes assez simples, destinée à faire mémoriser par l’élève leur forme correcte. Si le cursus des écoles ailleurs qu’à Nippur était organisé de la même manière, le Silbenalphabet A se situerait aussi à ce niveau. 2o) Copie de listes lexicales, en particulier la série ur5-ra, qui s’organise selon des regroupements thématiques de mots. Les versions paléobabyloniennes ne comportent que les mots sumériens, sans traduction akkadienne. La liste d’étoˆes correspond à la section qui devient, au Ier millénaire, la dix-neuvième tablette de la série. Si on fait l’hypothèse que la tablette de l’EPHE portant cette liste lexicale a été écrite par la même femme que le Silbenalphabet A, seulement quelques jours, plus tard, il ne faudrait pas comprendre ce phénomène comme le passage très rapide d’un niveau à un autre. Lorsque les élèves progressent, ils continuent à réviser les connaissances déjà acquises. Veldhuis 1997: 35–6 et 41 a souligné ce phénomène, en étudiant les corrélations entre la face et le revers des tablettes de type II: le revers porte en général des exercices correspondant à un stade moins avancé que ceux de la face, car l’étudiant acquiert de nouvelles connaissances (face) tout en reprenant des exercices plus faciles, qu’il a déjà bien assimilés (revers).
43. Veldhuis 1997: 40–63 et Robson 2001: 45–50; nous résumons ici ces travaux, en insistant sur les points qui intéressent plus particulièrement les tablettes présentées ci-dessus au §1. Une comparaison de la situation observée à Sippar dans la maison d’Ur-Utu avec le cursus de Nippur ˜gure dans Tanret 2002: 156–61 et 166–67.
One Line Short
3o) Listes de signes plus complexes, comme Proto-Ea (et Proto-Aa),44 avec notation de la lecture sumérienne des signes. C’est donc ici que se placerait le travail présenté par VAT 6574. 4o) Rédaction de textes en sumérien: modèles de contrats ou proverbes. Exercices de calcul. La copie de compositions littéraires sumériennes témoigne d’une deuxième étape du cursus. S. Tinney45 a distingué deux groupes de textes très fréquemment copiés par les apprentis scribes. L’un, qu’il appelle «tétrade», comprend quatre hymnes et correspond à un premier niveau d’apprentissage de la littérature sumérienne. Le second, la «décade», compte dix pièces plus élaborées, hymnes et textes mythologiques, parmi lesquelles ˜gure le Chant de la Houe. La rédactrice de VS 10 207 aurait donc atteint ce niveau avancé. Les quatre textes certainement rédigés par des femmes s’insèrent donc parfaitement dans la formation canonique des scribes telle qu’on peut la dé˜nir à Nippur. Cela montre d’une part, que ce cursus, sans doute sous des formes un peu diˆérentes, est susceptible d’avoir existé dans d’autres villes,46 et, d’autre part, que la formation d’une femme n’a rien de spéci˜que, mais participe exactement du même schéma que la formation masculine. En termes de niveau, les quatre tablettes retrouvées vont du plus simple au plus complexe et re˘ètent bien toutes les étapes du cursus, jusqu’aux plus avancées. Le bagage des femmes
44. Veldhuis 1997: 54, n. 174 considère cependant que ProtoAa correspond à un stade du cursus supérieur à celui auquel était enseigné Proto-Ea. Il signale également, p. 66, une tablette de Nippur portant au recto Proto-Aa et au verso, précisément, le Chant de la Houe: CBS 9856 (Civil 1979: 87). 45. Tinney 1999 et Robson 2001: 50–59. 46. Le phénomène était déjà connu pour Ur: les découvertes faites dans la maison no 1 de Broad Street témoignent bien de ce qui correspondrait à la deuxième étape de formation, alors que l’instruction élémentaire est peu représentée, et surtout pour sa phase ˜nale; le corpus de textes littéraires sumériens retrouvé est abondant et varié, et une large part relève des mêmes grandes compositions classiques qu’à Nippur (Charpin 1986: 449–59). En revanche les découvertes faites dans la même ville, dans la maison no 7 de Quiet Street, oˆrent un ensemble plus atypique; mais il s’agit alors d’un milieu très particulier, qui peut re˘éter la tradition d’Eridu plutôt que celle d’Ur (Charpin 1986: 420–34).
QUELQUES TEXTES SCOLAIRES PALÉO-BABYLONIENS ne se limitait donc pas à des formules stéréotypées utiles pour la rédaction de contrats; elles maîtrisaient également, du moins pour certaines d’entre elles, les compositions littéraires en langue sumérienne. 3. La formation scribale à Sippar Si l’on conserve l’hypothèse que les textes scolaires écrits par des femmes ont été trouvés à Sippar, on doit examiner de plus près les données relatives à la formation des scribes dans cette ville.47 La documentation sur les lieux de découverte des tablettes est loin d’être précise, du moins en ce qui concerne les énormes lots exhumés à la ˜n du XIXème siècle et au début du XXème. Il est même souvent di¯cile d’établir ce qui provient de Tell ed-Der (Sippar d’Annunÿtum, SipparAmnanum) ou de Tell Abu-Habbah (Sippar de †amas, Sippar Yahr¿rum).48 Or parmi ces tablettes ˜guraient de nombreux exercices scolaires paléobabyloniens (textes lexicaux par exemple, notamment des précurseurs de ur5-ra)49 et des textes littéraires qui peuvent aussi relever de la production des scribes en formation; ils n’ont pas fait l’objet d’études systématiques. Dans certains cas cependant, il paraît possible de repérer l’origine et le contexte du matériel scolaire, même s’il convient de rester très prudent quant à la provenance exacte de ces textes. 3.1. Tell Abu-Habba (Sippar Yahr¿rum) 3.1.1. Un catalogue d’oeuvres littéraires. Plusieurs catalogues paléo-babyloniens donnent des titres d’œuvres littéraires sumériennes. E. Robson en a examiné sept: un ou peut-être deux trouvés à 47. Sur les scribes de Sippar, voir Harris 1975: 284–302, mais le problème de leur formation n’est pas abordé. 48. Renger 1986, Walker 1988: xi–xxv, Driel 1989, Kalla 1999. Sur les deux villes de Sippar, voir Charpin 1988. 49. Deux exemples sont cités dans le catalogue de Leichty et Grayson 1987: 389; il ne s’agit cependant pas de la tablette XIX. Plus généralement, on peut se reporter aux mentions de textes scolaires paléo-babyloniens dans le catalogue de Leichty, Finkelstein et Walker 1988. On peut citer aussi les exemplaires de Proto-Aa provenant de Sippar (ci-dessus § 1.2. et n. 24).
47
Nippur, deux autres à Ur, deux de provenance inconnue et un septième exhumé à Sippar.50 Ils portent souvent les mêmes titres, parfois dans le même ordre, en particulier les textes de la «décade» dé˜nie par S. Tinney. Le catalogue de Sippar doit provenir des fouilles faites par V. Scheil à Tell Abu Habbah en 1894. 51 Il donne dans ses premières lignes neuf des dix textes de la «décade», quasiment dans le même ordre que l’un des catalogues de Nippur. Le Chant de la Houe, qui ˜gure à Nippur en troisième place, se trouve en quatrième position à Sippar. On peut donc supposer qu’au moins pour cette partie du cursus, la copie des œuvres de la «décade», la tradition scribale de Sippar est très proche de celle de Nippur;52 alors que dans les catalogues d’Ur, par exemple, le Chant de la Houe est absent, et le choix et l’ordre des compositions sont assez diˆérents de ceux de Nippur.53 3.1.2. Les tablettes scolaires exhumées par V. Scheil. Les résultats des fouilles menées à Tell Abu-Habbah en 1894 sont publiés par Scheil 1902. Un chapitre, intitulé «L’école à Sippar», est consacré aux découvertes faites dans une maison située «vis-à-vis et à proximité du temple»; dans l’une des pièces de cette habitation se trouvait «une masse énorme, compacte et cohérente de tablettes de toutes sortes, hymnes sumériens, listes métrologiques, syllabaires, contrats; tous documents appartenant à l’époque de Hammurabi».54 V. Scheil a copié certains de ces documents, d’autres sont mentionnés dans le catalogue qui termine l’ouvrage: 55 on y trouve des listes de 50. Une transcription de ce catalogue, Si 331, a été donnée par Dijk 1989: 447–48 (Robson 2001: 55). 51. Ce texte, Si 331, n’est pas mentionné dans le catalogue établi par Scheil 1902: 95–141, mais ce catalogue ne concerne que les «principales tablettes provenant de Sippar». D’après son numéro, il doit néanmoins faire partie du même lot. 52. Tinney 1999: 159, n. 5 indique que les deux catalogues de Nippur et celui de Sippar «should probably be considered curricular catalogues»; voir aussi p. 168. 53. Charpin 1986: 453–59. 54. Scheil 1902: 30–54; les citations ˜gurent p. 33. Un commentaire sur les trouvailles de V. Scheil et leur comparaison avec les tablettes de la maison d’Ur-Utu se trouve dans Tanret 2002: 162–66. 55. Pp. 95–141.
48
BRIGITTE LION AND ELEANOR ROBSON
signes, de mots, de noms propres, des textes mathématiques et métrologiques concernant par exemple les mesures de super˜cie et de capacité. On notera la présence de plusieurs copies du Silbenalphabet A56 (voir ci-dessus § 1.4). C’est certainement aussi de ce tell que sont issus plusieurs extraits d’ur 5-ra, dont un petit texte de huit lignes, l’un des précurseurs de la tablette XIX (correspondant à la quatrième tablet du précurseur Nippur), publié par le P. Scheil comme «provenant des dernières fouilles à Sippara» et conservé à Istanbul.57 Le passage copié est cependant diˆérent de EPHE 495. 3.1.3. Les tablettes scolaires trouvées par la mission irakienne. Une mission irakienne a repris des fouilles à Tell Abu-Habbah entre 1978 et 1983, dans deux zones d’habitations datant de l’époque paléo-babylonienne. 136 textes trouvés lors de ces campagnes ont été récemment publiés par F. N. H. Al-Rawi et S. Dalley. 58 Parmi eux ˜gurent plusieurs tablettes scolaires, certaines en forme de lentilles,59 réparties dans plusieurs maisons, et des tablettes non inscrites.60 Le groupe le plus cohérent provient de la maison 11, pièce 2, niveau III, où des textes scolaires étaient associés aux archives des enfants d’Ilÿ-kÿma-abiya, Taribilisu et sa soeur, {umta-Adad, une prêtresse nugig: il comprend une lentille fragmentaire (le no 74, dont seuls subsistent les signes ma-hi-ir), une liste de noms d’années d’Apîl-Sîn et Sîn-muballit (no 72), 61 et un hymne à Marduk inconnu par ailleurs (no 38);62 d’après les auteurs, d’autres 56. Ci-dessus n. 38. Scheil 1902: 38–40 publie une copie du Silbenalphabet A portant le numéro S. 204+687, mais indique dans les notes des variantes provenant des tablettes 213 et 689. Tanret 2002: 164, n. 76 ajoute à cette liste Si 350, 202 et 657. 57. Scheil 1895: 213, copie p. 214; une transcription de ce texte ˜gure dans Landsberger, Civil et Reiner 1970: 154. 58. Al-Rawi et Dalley 2000. Sur les tablettes scolaires, voir p. 5–6. 59. Les no 74, 78, 79, 84, 85, 91, 116, 124 sont des lentilles et il faut leur ajouter les no 44 et 45, qui ont une forme rectangulaire. 60. No 3, 48, 49, 50 et commentaires p. 35 et 69. 61. Al-Rawi 1993 considère les listes de noms d’années comme des textes scolaires. Cette liste est reprise dans Horsnell 1999 vol. 1: 281–86. 62. Al-Rawi 1992.
textes scolaires, et notamment des textes mathématiques, s’y trouvaient aussi. Diverses tablettes ont été exhumées en plusieurs points du site: des listes de noms propres (no 116, no 44?), des textes mathématiques et métrologiques (no 45, 79, 84, 124), et peut-être un texte en sumérien (no 45). La tablette lenticulaire no 85 ne comporte que trois courtes lignes sur la face: me-ªmeº, pap-pap, a-a, qui correspondent au début du Silbenalphabet A. Une autre lentille, no 91, également de trois lignes, note les mesures de poids de 30, 20 et 10 mines, ce qui correspond à un précurseur paléo-babylonien de la série ur 5-ra, tablette XVI (correspondant à la quatrième tablet du précurseur Nippur).63 Ces documents épars re˘ètent donc dans l’ensemble, comme les archives d’Ur-Utu (voir cidessous § 3.2.2), un niveau élémentaire d’éducation. Les zones fouillées ont livré des archives relevant des nadÿtum,64 mais sans qu’un lien entre ces dernières et les textes scolaires puisse être établi. 3.2. Tell ed-Der (Sippar Amnanum) Les fouilles Tell ed-Der, l’antique Sippar Amnanum, ont mis au jour au moins deux groupes de textes scolaires; l’un a été exhumé par les fouilles irakiennes en 1941, le second par celles de l’équipe de Gand, dans la maison d’Ur-Utu. 3.2.1. Les tablettes du bâtiment administratif. En 1941, les archéologues irakiens ont fouillé un grand bâtiment administratif, dont le niveau III correspond au début de l’époque paléobabylonienne et le niveau II à la ˜n de cette même époque (II A) et à l’époque cassite (II B). D. O. Edzard indique que 21 tablettes scolaires (Übungstexte) y ont été retrouvées65 (niveau II ou III), mais elles n’ont pas encore fait l’objet d’une publication et leur contenu n’a pas été décrit. D. O. Edzard souligne néanmoins que «Der Fund der Übungstexte ist insofern wichtig, als er die Existenz einer Schreiberschule an Ort und Stelle erschließen läßt».
63. Al-Rawi et Dalley 2000: 94. 64. Al-Rawi et Dalley 2000: 15–16. 65. Edzard 1970: 15.
QUELQUES TEXTES SCOLAIRES PALÉO-BABYLONIENS 3.2.2. La maison d’Ur-Utu. Les fouilles belges ont mis au jour, dans la maison d’Ur-Utu, le grand lamentateur de la déesse Annunÿtum, 80 tablettes scolaires publiées et étudiées par M. Tanret.66 Ces textes relèvent de l’éducation élémentaire. La première étape du cursus est bien représentée par des exercices de calame pour apprendre à former les signes, des syllabaires, ainsi que des noms de personnes et de divinités, qui correspondraient au premier stade dé˜ni par N. Veldhuis pour Nippur. Parmi les syllabaires, le Silbenalphabet A est connu par une dizaine de manuscrits. Huit autres tablettes, avec des extraits de la liste lexicale ur5-ra, s’inscrivent dans la deuxième étape du cursus; l’une d’elles contient un passage relatif aux étoˆes. La troisième phase est également représentée, par un extrait de la liste de signes Proto-Diri et surtout par des tables métrologiques notant des mesures de capacité et de poids. En revanche, il n’y a pas trace de la liste Proto-Aa. Quant au second niveau d’éducation, la copie de textes littéraires sumériens, il n’est pas du tout présent dans cette maison. L’histoire de la maison est documentée à la fois par les textes et par l’archéologie. Les archives retrouvées indiquent qu’une nadÿtum de †amas, Lamassani, avait acheté la maison en l’an 7 d’Ammi-ditana, pour la vendre à Inanna-mansum, le père d’Ur-Utu, en l’an 28 du même roi;67 Inannamansum ˜t raser cette maison pour la reconstruire à sa convenance. Plusieurs phases d’occupation du site peuvent être distinguées. M. Tanret a proposé, en fonction de critères archéologiques, d’attribuer à l’époque où la maison appartient à Lamassani (phase IIIe essentiellement) une douzaine de documents scolaires, et 68 autres à l’époque où la maison appartient à Inannamansum, le père d’Ur-Utu (phases IIId à IIIa). M. Tanret, soulignant la présence de ce type de textes dans la maison d’une nadÿtum, s’est demandé si ces exercices pouvaient être le fait de Lamassani elle-même, ou éventuellement de
66. Tanret 2002. Le développement qui suit s’appuie sur cet ouvrage. 67. Jannsen, Gasche et Tanret 1994, spécialement pp. 93– 103 et les tableaux récapitulatifs pp. 109 et 122.
49
l’une de ses nièces: les nadÿtum de †amas, qui ne devaient pas avoir d’enfants, adoptaient en eˆet fréquemment des membres de leur famille plus jeunes qu’elles pour leur léguer leurs biens. La question peut di¯cilement être tranchée car, comme l’a souligné l’auteur, les nadÿtum louaient parfois les maisons qu’elles possédaient en ville. Lorsqu’elles étaient riches, ces femmes étaient propriétaires de plusieurs habitations et pouvaient résider dans leur maison du gagûm, leurs biens en ville étant transformés en immeubles de rapport. S’il faut distinguer la propriétaire de la maison de ses occupants, les textes scolaires ne renseignent que sur les activités de ces derniers. On peut ajouter une observation chronologique: à partir de l’époque d’Ammi-ditana, c’est-à-dire sous les trois derniers rois de la première dynastie de Babylone, plus aucune femme scribe n’est attestée à Sippar. Le travail des femmes scribes, bien documenté dans cette ville au début et au milieu de l’époque paléo-babylonienne, semble disparaître ensuite; les tablettes rédigées par des femmes et datées des règnes de Hammu-rabi et de Samsu-iluna sont encore nombreuses, mais une seule scribe est attestée sous le règne d’Abi-esuh, la dernière dont au puisse à ce jour dater les activités: Amat-Mamu, qui rédige CT 6 24b, la mise en location d’un champ appartenant à une nadÿtum de †amas. Néanmoins, cela ne signi˜e pas que les femmes n’avaient plus accès à l’écrit. L’apprentissage de l’écriture pouvait permettre aux élites de Sippar, hommes ou femmes, de contrôler la gestion de leurs biens, sans avoir pour but la formation de professionnels de l’écrit. En outre, certains métiers comme celui de lamentateur (gala-mah), qu’exerça Ur-Utu, devaient nécessiter ce type de savoir. 3.3. Bilan Malgré le grand nombre de documents encore inédits, ce rapide survol montre au moins que l’enseignement, sur les deux sites de Sippar Amnanum et de Sippar Yahr¿rum, était largement dispensé. Le matériel scolaire retrouvé y est abondant et très diversi˜é. Même si les tablettes publiés correspondent surtout au niveau élémen-
50
BRIGITTE LION AND ELEANOR ROBSON
taire d’éducation, plusieurs compositions littéraires sumériennes sont attestés. L’ensemble témoigne de la vivacité de l’activité intellectuelle de la ville, qui n’est guère surprenante si l’on pense aux milliers de textes exhumés sur les deux sites. Les tablettes scolaires retrouvées sont les vestiges de la formation reçue par les très nombreux scribes des deux villes, 68 ainsi que, probablement, par une élite urbaine dont une partie au moins pouvait maîtriser l’écriture. 4. Le problème du lieu de provenance et de la date des textes Le lieu de provenance des quatre textes scolaires rédigés par des femmes demeure inconnu. Pour VAT 6574 et VS 10 207, Sippar reste la meilleure possibilité; la relative proximité des numéros de musée de ces tablettes, VAT 6457 + 6492 + 6493 pour l’une, VAT 6574 pour l’autre, peut faire penser qu’elles sont issues d’un même lot ou d’un même secteur. Quant à la tablette BM 96950, elle fait partie de la collection BM 1902-10-11. Plusieurs documents de cette série auraient été trouvés à Sippar. C’est le cas par exemple des exercices mathématiques BM 96954+ et BM 96957+, publiés par E. Robson,69 qui indique à leur sujet «probably from Sippar». L’absence de contexte archéologique est, comme d’habitude, à déplorer. Il aurait été essentiel de savoir si ces tablettes provenaient des maisons des intéressées et pouvaient en quelque manière être liées à des documents autres que scolaires; ou éventuellement si elles avaient été retrouvées chez un maître, ou encore avaient été utilisées en remploi, ou mises au rebut.70 Des quatre tablettes rédigées par des femmes, seule BM 96950 est précisément datée, de l’an 14 de Samsu-iluna. La tablette de l’EPHE date également du règne de Samsu-iluna, peut-être de la même année. Plusieurs femmes scribes sont
68. Harris 1975: 284–302. 69. Robson 1999: 218–30 (Appendix 3) et 231–44 (Appendix 4). 70. Sur ces problèmes, voir Charpin 1986: 419–86 et 1990, Wilcke 2000: 7–22, Robson 2001 et Tanret 2002: 3–24 et 131–53.
actives à Sippar sous le règne de ce roi.71 Si l’on admet l’hypothèse d’une origine sippariote pour les deux autres tablettes, il faudrait, en l’état actuel de la documentation, les placer entre le règne de Sabium et celui d’Abi-esuh. Avant Sabium, il y a des femmes scribes à Sippar, mais elles ne notent jamais leur profession sous la forme munus dub-sar, préférant dub-sar ou dubsar munus. Après Abi-esuh, plus aucune scribe n’est documentée. Pour cette période, dix-neuf femmes scribes sont attestées à Sippar et connues par leur nom; il est logique de penser qu’elles ont reçu une éducation standard et qu’elles ont rédigé de nombreuses tablettes scolaires. Il n’en resterait alors que ces rares témoignages. Cela s’explique aisément. D’une part, le matériel scolaire, de par sa nature même, n’est pas destiné à l’archivage mais plutôt à la destruction. D’autre part, il est relativement rare qu’un scribe donne, dans un colophon, son nom, ou quelque indication que ce soit sur sa personne, comme ici la précision que ce travail est le fait d’une femme; par exemple dans la maison d’Ur-Utu, aucune des 80 tablettes retrouvées n’a de colophon, du moins lorsque l’emplacement de celui-ci est conservé, ni ne porte de nom de scribe. En˜n, nombreuses sont les tablettes scolaires fragmentaires, pour lesquelles le colophon, s’il a existé, est perdu. Pour ces deux dernières raisons, il est possible que, dans le matériel scolaire attribuable à Sippar, il y ait d’autres documents rédigés par des femmes, mais non identi˜ables comme tels.72 71. Voir la liste donnée par R. Harris 1975: 197, à laquelle il faut maintenant ajouter Taribatum, MHET II, 3 453, active en Samsu-iluna 30. 72. On peut confronter ces remarques à celles faites récemment par A. Beach à propos de femmes scribes à l’époque médiévale: «It is impossible to determine the sex of a scribe based on any supposed inherent diˆerence in the handwriting of men and women. Female hands have been described variously as delicate, irregular, nervous, and light—judgments based on attitudes towards women rather than on any sound paleographical evidence. Women at a particular scribal center may have been trained to write a highly distinctive book hand, but only its demonstrated use by female copyists marks that hand as feminine . . . . The only de˜nitive evidence of a scribe’s sex is a self-identi˜cation in a contemporary colophon» (Beach 2004: 5–6).
QUELQUES TEXTES SCOLAIRES PALÉO-BABYLONIENS
Conclusion La mention munus-dub-sar sur quatre tablettes scolaires fait penser que la ou les femmes qui les ont rédigées se considéraient comme des professionnelles de l’écrit, ou se destinaient à le devenir. À l’époque paléo-babylonienne, les scribes n’étaient apparemment pas les seuls à savoir lire et écrire. Cet apprentissage faisait partie de la formation des prêtres. 73 Il est probable aussi qu’une élite sociale pouvait juger nécessaire de connaître su¯samment de cunéiforme pour contrôler la gestion de ses biens. Rien n’interdit de penser qu’il en allait de même des femmes, surtout à Sippar, où certaines nadÿtum se comportaient en femmes d’aˆaires conscientes de leurs intérêts.74 Cependant la mention explicite dub-sar sur les textes scolaires semble faire référence au cas particulier des femmes scribes, tel qu’on en connaît de nombreux exemples dans cette ville. En outre le niveau atteint par une de ces femmes, celui de la copie de textes sumériens, suggère une formation très poussée, probablement à visée professionnelle. Pourquoi cette femme, ou ces femmes, ont-elles éprouvé le besoin d’inscrire dans les colophons munus dub-sar, indiquant leur sexe et leur métier alors même que dans trois des quatre cas répertoriés elles omettent de se nommer? Peut-être cette a¯rmation exprime-t-elle une certaine ˜erté, peutêtre aussi souligne-t-elle la rareté, voire l’étrangeté que ce phénomène pouvait revêtir aux yeux des rédactrices elles-mêmes. Les deux hypothèses vont d’ailleurs de pair: le nombre des femmes scribes connues, même à Sippar, est très inférieur à celui des scribes de sexe masculin; plus leur situation devait paraître exceptionnelle, plus grande pouvait être la satisfaction de celles qui accédaient à cette formation. Les textes scolaires rédigés par des femmes, puisqu’ils relèvent de la tradition la plus banale, montrent du moins que la formation et la culture des femmes scribes étaient identiques à celles des 73. Voir p. ex. à Ur le cas du no 7 Quiet Street, habité par le puri˜cateur Ku-Ningal et ses ˜ls (Charpin 1986: 419–86), et à Sippar-Amnanum celui du lamentateur Ur-Utu (Tanret 2002: 155). 74. Tanret 2002: 168–69.
51
hommes. On peut donc, à propos du contenu de l’éducation, parler de «mixité». La production des femmes scribes de Sippar, d’un point de vue typologique, est similaire à celle des hommes: contrats de prêt, de vente, de location, d’héritage ou comptes rendus de procès,75 tous les genres de textes qu’elles rédigent auraient aussi bien pu l’être par des hommes. Le choix d’une femme comme scribe semble tenir aux parties concernées par les transactions, non au contenu des documents. Si une formation identique des hommes et des femmes paraît logique pour aboutir à la rédaction des mêmes documents, un tel constat laisse cependant plusieurs questions en suspens. On ignore tout des conditions dans lesquelles les femmes recevaient leur formation. Faut-il croire que ˜lles et garçons étaient instruits par les mêmes professeurs, éventuellement dans les mêmes lieux et en même temps, ou imaginer deux circuits d’éducation séparés mais parallèles? Puisqu’à Sippar plusieurs femmes scribes sont connues, il est possible qu’elles aient assuré l’éducation des jeunes ˜lles, donc que l’éducation des ˜lles ait été dispensée dans un cadre féminin. On peut aussi avancer l’hypothèse d’une formation à l’intérieur même de la famille; à Sippar, la scribe Inannaama-mu a pu apprendre son métier auprès de son père Abba-tabum, s’il s’agit bien du scribe du même nom, et non d’un homonyme.76 Une autre question demeure sans réponse. Il est bien connu que la formation des scribes ne répond pas à des besoins strictement utilitaires (copier des textes littéraires sumériens n’apprend pas à rédiger des contrats ou des lettres en akkadien) mais plutôt à des usages sociaux.77 L’accès à un certain type de culture dé˜nit l’appartenance à une communauté, unie par la même éducation et les mêmes références. À quelle communauté pouvaient alors appartenir les femmes lettrées? À une élite intellectuelle mixte, qui regroupe hommes et femmes partageant un savoir commun? Ou à un 75. Elles rédigeaient probablement aussi des lettres, mais la correspondance ne portant jamais le nom du scribe, ce point ne peut être précisé. 76. Lion 2001: 15–16. 77. Veldhuis 1997: 142–46 fait le point sur cette question. Voir aussi Tanret 2002: 170.
52
BRIGITTE LION AND ELEANOR ROBSON
groupe spéci˜que de femmes scribes, que l’on peut imaginer, à Sippar, proche du milieu des nadÿtum de †amas, voire lui appartenant,78 formant une élite exclusivement féminine, coexistant avec les cercles savants masculins, mais sans qu’il y ait
78. Ce point n’est pour l’instant pas prouvé: aucune des femmes scribes connues à Sippar n’est clairement désignée comme nadÿtum. Cependant plusieurs d’entre elles portent des noms répandus dans le milieu des nadÿtum, comme Ayakuzub-matim, Amat-†amas ou †at-Aya par exemple. A Mari, les femmes scribes connues dans le palais de Zimrî-Lîm sont
de véritables relations entre ces deux ensembles, qui partagent pourtant une même formation et, peut-être, les mêmes origines sociales? Les réponses se trouveront peut-être un jour dans les textes de Sippar, qui devraient permettre d’avancer vers une meilleure connaissance de la société de cette ville, et de la place que pouvaient y tenir les femmes.
intégrées aux listes du «harem» royal tel que l’a dé˜ni N. Ziegler et appartiennent à la population féminine du palais.
Références Al-Rawi, F. N. H. 1992 A New Hymn to Marduk from Sippir. RA 86: 79–83. 1993 A New Old Babylonian Date List from Sippir with Year Names of Apil-Sin and Sin-muballit. ZA 83: 22–31. Al-Rawi, F. N. H., et Dalley, S. 2000 Old Babylonian Texts from Private Houses at Abu Habbah Ancient Sippar. Londres: Nabu Publications. Beach, A. I. 2004 Women as Scribes: Book Production and Monastic Reform in Twelfth-Century Bavaria. Cambridge: Cambridge University Press. Black, J.A., Cunningham, G., Ebeling, J., FlückigerHawker, E., Robson, E., Taylor, J., et Zólyomi G. 1998– The Electronic Text Corpus of Sumerian Literature (http://www-etcsl.orient.ox.ac.uk/). Oxford. Cavigneaux, A. 1980– Lexikalische Listen. RlA 6: 609–41. 1983 Charpin, D. 1986 Le Clergé d’Ur au siècle d’Hammurabi (XIXe– XVIIIe siècles av. J.–C.). Genève et Paris: Droz. 1988 Sippar: deux villes jumelles. RA 82: 13–32. 1990 Un quartier de Nippur et le problème des écoles à l’époque paléo-babylonienne (suite). RA 84: 1–16. Civil, M. 1969 Materials for the Sumerian Lexicon 12. Rome: Ponti˜cal Biblical Institute. 1979 Materials for the Sumerian Lexicon 14. Rome: Ponti˜cal Biblical Institute. 1995 Ancient Mesopotamian Lexicography. Pp. 2305–14, in Civilizations of the Ancient Near East, ed. J. Sasson. New York: Charles Scribner’s Sons.
Cooper, J. S. 1983 The Curse of Agade. Baltimore: The Johns Hopkins University Press. Dekiere, L. 1994a Old Babylonian Real Estate Documents, PreHammurabi Documents. MHET II,1. Gand: University of Ghent. 1994b Old Babylonian Real Estate Documents, Documents from the Reign of Hammurabi. MHET II,2. Gand: University of Ghent. 1995a Old Babylonian Real Estate Documents, Documents from the Reign of Samsu-iluna. MHET II,3. Gand: University of Ghent. 1995b Old Babylonian Real Estate Documents, PostSamsu-iluna Documents. MHET II,4. Gand: University of Ghent. 1996 Old Babylonian Real Estate Documents, Documents without Date or with Date Lost. MHET II,5. Gand: University of Ghent. 1997 Old Babylonian Real Estate Documents, Documents from the Series 1902–10–11 (From Zabium to Ammi-aaduqa). MHET II,6. Gand: University of Ghent. Dijk, J. van 1989 Ein spätbabylonischer Katalog einer Sammlung sumerischer Briefe. OrNS 58: 441–52. Driel, G. van 1989 The British Museum ‘Sippar’ Collection: Babylonia 1882–1893. ZA 79: 102–17. Durand, J.-M. 1982 Documents cunéiformes de la IV e Section de l’Ecole pratique des Hautes Etudes, Tome I. Catalogue et copies cunéiformes. Genève et Paris: Droz. Edzard, D. O. 1966 Compte rendu de C. J. Gadd et S. N. Kramer, UET 6/1 (Londres 1963). AfO 21: 86–89.
QUELQUES TEXTES SCOLAIRES PALÉO-BABYLONIENS 1970 Altbabylonische Rechts- und Wirtschaftsurkunden aus Tell ed-Der im Iraq Museum, Baghdad. Munich: Bayerische Akademie der Wissenschaften. 2000 U 7804 // UET VI/1 26: “Gedicht von der Hacke.” Pp 131–35 in Wisdom, Gods and Literature, Studies in Assyriology in Honour of W. G. Lambert, ed. A. R. George et I. L. Finckel, Winona Lake: Eisenbrauns. 2003 Zum sumerischen Verbalprä˜x a(l)-. Pp. 87–98 in Literatur, Politik und Recht in Mesopotamien. Fs für C. Wilcke, ed. W. Sallaberger, K. Volk et A. Zgoll. Wiesbaden: Harrassowitz. Farber, G. 1999 “Das Lied von der Hacke”, ein literarischer Spaß? Pp. 369–73 in Landwirtschaft im alten Orient, CRRAI 41 (Berlin, 1994), BBVO 18, ed. H. Klengel et J. Renger. Berlin: Dietrich Reimer. Harris, R. 1975 Ancient Sippar: A Demographic Study of an Old Babylonian City (1894–1595 BC), Istanbul: Nederlands Historisch-Archaelogisch Instituut te Istanbul. 1990 The Female “Sage” in Mesopotamian Literature. Pp. 3–17 in The Sage in Israel and in the Ancient Near East, ed. J. G. Gammie et L. G. Perdue. Winona Lake: Eisenbrauns. Horsnell, M. J. A. 1999 The Year Names of the First Dynasty of Babylon. Volume 1: Chronological Matters: The Year-Name System and the Date-List. Volume 2: The Year-Names Reconstructed and Critically Annotated in Light of their Exemplars. Hamilton: McMaster University Press. Jannsen, C., Gasche, H. et Tanret, M. 1994 Du chantier à la tablette. Ur-Utu et l’histoire de sa maison à Sippar-Amnanum. Pp. 91–123 in Cinquante-deux ré˘exions sur le ProcheOrient ancien oˆertes en hommage à Léon de Meyer, MHEO II, ed. H. Gasche, M. Tanret, C. Jannsen et A. Degraeve. Louvain: Peeters. Kalla, G. 1999 Die Geschichte der Entdeckung der Altbabylonischen Sippar-Archive. ZA 89: 201–26. Labat, R. 1970 Compte rendu de B. Landsberger, Materials for the Sumerian Lexicon 9 (Rome 1967). OrNS 39: 185–88. Landsberger, B. 1933 Die angebliche babylonische Notenschrift. Pp. 170–78 in Aus fünf Jahrtausenden morgenländischer Kultur, Festschrift Max Freiherrn von Oppenheim zum 70. Geburtstage, ed. E. F. Weidner. Berlin: E. F. Weidner.
53
1959 Zum “Silbenalphabet B.” Pp. 97–116 in M. Çig, H. Kizilyay et B. Landsberger, Zwei altbabylonische Schulbücher aus Nippur. Ankara: Türk Tarih Kurumu Basimevi. 1967 Materials for the Sumerian Lexicon 9. Rome: Ponti˜cal Biblical Institute. Landsberger, B., Civil, M., et Reiner, E. 1970 Materials for the Sumerian Lexicon 10. Rome: Ponti˜cal Biblical Institute. Leichty, E., et Grayson, A. K. 1987 Catalogue of the Babylonian Tablets in the British Museum, Volume VII, Tablets from Sippar 2. London: British Museum Publications. Leichty, E., Finkelstein, J. J. et Walker, C. B. F. 1988 Catalogue of the Babylonian Tablets in the British Museum, Volume VIII, Tablets from Sippar 3, London: British Museum Publications. Lerberghe, K. van, et Voet, G. 1991 Sippar-Amnanum. The Ur-Utu Archive, vol. 1. MHET I,1. Gand: University of Ghent. Lion, B. 2001 Dame Inanna-ama-mu, scribe à Sippar. RA 95: 7–32. Meier, S. A. 1991 Women and Communication in the Ancient Near East. JAOS 111: 540–47. Pearce, L. E. 1995 The Scribes and Scholars of Ancient Mesopotamia. Pp. 2265–78 in Civilizations of the Ancient Near East, ed. J. Sasson. New York: Charles Scribner’s Sons. Renger, J. 1986 Zu den altbabylonischen Archiven aus Sippar. Pp. 96–105 in Cuneiform Archives and Libraries, CRRAI 30 (Leiden, 1983), ed. K. R. Veenhof. Istanbul: Nederlands HistorischArcheologisch Instituut. Richardson, S. F. C. 2002 The Collapse of a Complex State: A Reappraisal of the End of the First Dynasty of Babylon, 1683–1597 BC. PhD Thesis. Columbia University. Robson, E. 1999 Mesopotamian Mathematics, 2100–1600 BC. Technical Constants in Bureaucracy and Administration. OECT 14. Oxford: Clarendon Press. 2001 The Tablet House: a Scribal School in Old Babylonian Nippur. RA 95: 39–66. Scheil, V. 1895 Fragments de syllabaires assyriens. ZA 10: 193–221.
54
BRIGITTE LION AND ELEANOR ROBSON
1902 Une saison de fouilles à Sippar. Le Caire: Institut français d’archéologie orientale. 1932– Catalogue sommaire de la collection des tab1933 lettes cunéiformes de l’Ecole. Ecole Pratique des Hautes Etudes. Section des sciences historiques et philologiques, Annuaire: 3–27. Sjöberg, A. W. 1976 The Old Babylonian Eduba. Pp. 159–79 in Sumerological Studies in Honor of Thorkild Jacobsen on his Seventieth Birthday, AS 20, ed. S. J. Lieberman. Chicago: University of Chicago Press. Sollberger, E. 1965 A Three Column Silbenvokabular A. Pp. 21– 28 in Studies in Honor of Benno Landsberger on his Seventy-Fifth Birthday, ed. H. Güterbock et T. Jacobsen. Assyriological Studies 16. Chicago: University of Chicago Press. Stol, M. 1976 On Ancient Sippar. BiOr 33: 146–54. Tanret, M. 2002 Per aspera ad astra. L’apprentissage du cunéiforme à Sippar-Amnanum pendant la période paléobabylonienne tardive. MHET I,2. Gand: University of Ghent.
Tinney, S. 1999 On the Curricular Setting of Sumerian literature. Iraq 61: 159–72. Veldhuis, N. 1997 Elementary Education at Nippur, The Lists of Trees and Wooden Objects. Groningen: Rijksuniversiteit Groningen. Walker, C. B. F. 1988 Introduction. Pp. xi–xxv in Leichty, Finkelstein et Walker 1988. Wilcke, C. 1972– Hacke—B. Philologisch. RlA 4: 33–38. 1975 2000 Wer las und schrieb in Babylonien und Assyrien. Überlegungen zur Literalität in Alten Zweistromland. Munich: Bayerische Akademie der Wissenschaften. Ziegler, N. 1999 Le Harem de Zimrî-Lîm. Florilegium marainum IV = Mémoires de N.A.B.U. 5. Paris: SEPOA. Zimmern, H. 1912 VS 2. Leipzig: J. C. Hinrichs’sche Buchhandlung. 1913 VS 10. Leipzig: J. C. Hinrichs’sche Buchhandlung.
ANOTHER OLD BABYLONIAN FORERUNNER TO HAR-ra={UBULLU XX–XXII Mark E. Cohen (University of Maryland)
YBC 16318 is a prism containing an Old Babylonian forerunner of the lexical series HAR-ra= hubullu XX–XXII,1 which corresponds to tablet 5 of the series in the Old Babylonian scribal tradition.2 The major portion of the preserved pieces of the prism contains a forerunner to Hh XX (a-sà), including at least thirty-three entries unattested elsewhere in the lexical tradition. Just the ˜rst ten entries of a forerunner to Hh XXI are preserved; however, the order of the geographical entries is unique. Finally, just a few entries from the forerunner to Hh XXII are preserved, but they include
six previously unattested entries for terms with és, “rope.” The provenance of the prism cannot be established. However, most likely, it is not from Nippur, since neither Nippur nor localities associated with Nippur are in the ˜rst position in this forerunner to Hh XXI.3 Our prism was originally four-sided (23 x 12 cm) with three columns per side, of which only three large pieces survive. The surface area is either totally or seriously eroded, but despite its horrible condition, some sections are legible.
YBC 1631841234 Nippur fore.
OB fore.
Other fore.
Hh
col. i 1u. 2u. 3u. 4u. 5u. 6u. 7u.* 8u.*
(break) [a]-sà-[gi]skiri6 [a]-sà-KA-ªxº-IGI-d[a] [a-sà]-KA-x-x-x-a [a-sà . . .]-GÁx ªXº [a-sà . . .]-ªxº a-sà-bala-ri [a-s]à-a-bala [a-s]à-ªé-a?-la?-hulº
V
Bo
x
3. See the Nippur forerunner in MSL 11 101 161, wherein Nibruki is in ˜rst position. 4. In our edition of YBC 16318, an asterisk (*) after the line number indicates that the entry does not occur in the edition of the series in MSL 11. The following designations are used: x = occurs without a signi˜cant variant; V = occurs with a signi˜cant variant.
1. I thank Benjamin R. Foster, Curator of the Yale Babylonian Collection, for his kind permission to publish YBC 16318. I also thank W. W. Hallo, who, while curator, gave me permission to copy the prism. 2. The assignment of a tablet number for the Old Babylonian series is according Niek Veldhuis, Religion, Literature, and Scholarship: The Sumerian Composition Nanse and the Birds. (Leiden: Brill, 2004), 87.
55
JCS 57 (2005)
56
9u. 10u. 11u. 12u. 13u. 14u. 15u. 16uu. 17uu.* 18uu.* 19uu.* 20uu.* 21uu.* 22uu.* 23uu. 24uu. 25uu. 26uu. 27uu. 28uu. 29uu. 30uu. 31uu.
MARK E. COHEN
a-sà-é-ªkiº-ág a-sà-é-kur a-sà-é-ninnu [a-sà-é-am]ar-ra [a-sà-é- . . .]-ªhulº [a-sà-é- . . .]-ªtaº [a-sà-é- . . . -b]i/ga (break) a-sà-é-d[. . .] a-sà-é-d[x]-ki a-sà-é-dgestin-an-na a-sà-é-dnin-é-ªanº-ki a-sà-é-dlugal-mára-da a-sà-til-til a-sà-túl-túl a-sà-suku a-sà-ZUKUM-se a-sà-nì-gál-la a-sà-nì-ku5-da [a-sà]-dùn-suhur [a-sà-ma-t]a-mu-un-g[i4?] [a-sà]-x-sig [a-sà-zi?-ib?-n]a?-tum [a-sà-ké]s-kés
Nippur fore.
OB fore.
Other fore.
x x
x x
Bo
x x
x
x
x V
x x
x V x x
col. ii 1u. 2u. 3u. 4u. 5u. 6u. 7u. 8u. 9u. 10u.* 11u.* 12u. 13u. 14u. 15u.* 16u. 17u. 18u.* 19u.*
(break) a-sà-[. . .] a-sà-[. . .] a-sà-du6-d[x] a-sà-du6-d[x] a-sà-du6-d[x] a-sà-du6-d[x] a-sà-du6-dx-x-ªkùº a-sà-du6-ªha-ba-iaº a-sà-ªdu6-x-a?-xº a-sà-du6-é-x-x a-sà-du6-mas-mas a-sà-du6-mas-mìn a-sà-du6-ùz a-sà-du6-sipa a-sà-du6-ªUD?.NUN?-xº-kur a-sà-du6-sal-la a-sà-du6-ur-mah a-sà-du6-x-UD a-sà-du6-gisma-nu
Long
x
x x RS x x
Hh V V
ANOTHER OLD BABYLONIAN FORERUNNER TO HAR-ra={UBULLU XX–XXII
Nippur fore. 20u.* a-sà-du6-túl-túl 21u.* a-sà-du6-gul-gul (break) 22uu. a-s[à-du6 . . .] 23uu.* a-sà-du6-é-x-[x] 24uu.* a-sà-du6-é-a-na-da 25uu.* a-sà-du6-lugal-a-ni 26uu.* a-sà-du6-lugal-pa-è 27uu.* a-sà-igi-ní-íl 28uu.* a-sà-a-zi-da 29uu.* a-sà-ta?-igi-ì-dub 30uu.* a-sà-munus-e-ne 31uu.* a-sà-nar-e-ne 32uu.* a-sà-gala-e-ne 33uu. a-sà-aga-us-e-ne 34uu.* a-sà-nin-e-ne 35uu. a-sà-túg-ba13(ME)-gar-ra 36uu. a-sà-burux(GÁNA)buru3-ma 37uu. a-sà-buru14-mah 38uu. a-sà-mah col. iii 1u. a-sà-[. . .] 2u. a-sà-x-[. . .] 3u. a-sà-x-[. . .] 4u. a-sà-x-[. . .] 5u. a-[sà- . . .] 6u. a-sa-[. . .] 7u. a-sà-[. . .] 8u. a-sà-[. . .] 9u. a-sà-[. . .] 10u. a-sà-d[. . .] 11u. a-sà-d[. . .] 12u. a-sà-dnin-[x] 13u. a-sà-x-x-[. . .] 14u. a-sà-x-[. . .] 15u. a-[sà]-x-[. . .] 16u. a-[sà- . . .] (break) 17uu. a-sà-[. . .] 18uu. a-sà-a-gàr 19uu. a-sà-a-gàr-gibil 20uu. a-sà-a-gàr-x 21uu. a-sà-a-gàr-duru5 22uu. a-sà-a-gàr-kù 23uu. a-sà-a-gàr-gisasal
OB fore.
Other fore.
x V V x x
V x x
x x
x
x x x
x
x
57
Hh
58
24uu. 25uu. 26uu.* 27uu. 28uu. 29uu. 30uu. 31uu. 32uu.
MARK E. COHEN
a-sà-a-gàr-tur a-sà-a-gàr-gu-la a-sà-a-gàr-IGI.x a-sà-al-ªakº a-sà-kin-ªakº a-sà-TÚG.KIN-ªakº a-sà-igi-bar-bar-r[e] a-sà-ki-[x] a-sà-ki-[x]
col. iv (break) 1u. a-sà-[. . .] 2u. a-sà-ú-sal 3u.* a-sà-gú-íd-da 4u.* a-sà-ka/sag-íd-da 5u.* a-sà-kun-íd-da 6u. a-sà-a-má-ru-ru 7u. a-sà-a-má-su-su 8u.* a-sà-a-ùr-ru 9u.* a-sà-a-íl-lá 10u. a-sà-hu[r]?-x-s[ag]? 11u.* a-sà-kar-x 12u. a-sà-[. . .] 13u. [a-sà]-x-[. . .] (break)
Nippur fore. x x
OB fore. x x
Other fore.
x x x x
x x x
RS
x
x
x
V x
V V/x
x x x x x x
x x x x x x x x x x?
V V
Hh
V V
col. v 1u. 2u. 3u. 4u. 5u. 6u. 7u. 8u. 9u. 10u. 11u. 12u. 13u.
(break) a-sà-[. . .] a-sà-[. . .] Ú[riki] Nibruki Ì-si-inki Unugki Ararmaki Adabki Zabalamki Kèski Éreski Nina[ki] x [. . .] (break)
col. vi (break) 1u. ªídº-[. . .]
One Line Long
x x
x x x x x
x
ANOTHER OLD BABYLONIAN FORERUNNER TO HAR-ra={UBULLU XX–XXII
2u. 3u. 4u. 5u. 6u.
Nippur fore.
OB fore.
-ma/TÚG-
-max
Other fore.
59
Hh
ªídº-[. . .] ªídº-[. . .] ªídº-[. . .] ªídº-[. . .] ªídº-[. . .] (break)
cols. vii–x destroyed col. xi (break) 1u. [. . .]-x 2u. [. . .]-x 3u. ªpa5º-x (x)-[raº 4u. ªpa5-ma/zì?º-da 5u. ªpa5º-gu-la 6u. ªeº-dal-ba-na 7u. mul 8u. [mul]-mul 9u. [mul-mu]-ªsírº 10u. [. . .]-x (break)
x x x x
col. xii (break) 1u. és-[. . .] 2u. és-[. . .] 3u. és-[. . .] 4u. és-gu-kés 5u.* és-gu-bal 6u.* és-hi-dam 7u.* és-ga-SIKIL?-UR-x 8u.* és-us-x 9u.* és-á-zi-d[a] 10u.* és-á-ªgùb-buº 11u. és-su-an-ªtaº-(x) 12u. és-ªú-ri-inº 13u. és-x-[. . .] 14u. és-gí[d-da] (break of the few last lines of the prism) Comments i 1u. The Nippur forerunner (MSL 11 98) 41 has the entry: a-sà giskiri6 lú-ús-gi-na. i 9u–10u. The Neo-Babylonian text from Ur, UET 7 127 (MSL 11 6), contains two lines (3–4) that are phonetic variants of our text (i 9u–10u). Note
x x V
-zix x x
x
x x x
x
that the subsequent entry (5) in UET 7 127 is the same as two entries further (12u) in YBC 16318, thus strongly supporting the observation that UET 7 127 3–4 and YBC 16318 i 9u–10u are indeed variants.
60
MARK E. COHEN
UET 7 127
YBC 16318
3. a-sà-a-ki-ág-gá 4. a-sà-a-gur4-ra 5. a-sà-é-amar-ra
9u. a-sà-é-ki-ág 10u. a-sà-é-kur 11u. . . . 12u. [ a-sà-é-am]ar-ra
i 14u. See MSL 11 97 11: a-sà-da-ab-ta. i 21uu–22uu. Note the juxtaposed homophonous entries: 21uu. a-sà-til-til 22uu. a-sà-túl-túl For references to a-sà-túl-túl in economic texts, see G. Pettinato, Neusumerischen Landwirtschaft (Naples: Istituto Orientale di Napoli, 1967), no. 681: a-sà pú-pú “Brunnenfeld.” i 26uu. Note the Old Babylonian forerunner 1 (MSL 11 130) 24: ª a-sàº-nì-tar/ku5-ra. i 28uu. This line may well be a variant for the Nippur forerunner (MSL 11 99) 81: a-sà-ma-da mu-un-gi-en. i 29uu. Note MSL 11 101 142: a-sà-se-sig and MSL 11 136 6: a-sà-igi-sig. The traces of our text do not seem to support either restoration. i 30uu. The Nippur forerunner (MSL 11 99) 83 has the entry a-sà-zi-ib-na-tum directly before a-sàkés-kés. Some variation of this would seem to be a likely restoration in our line, which is also directly before the entry a-sà-kés-kés. ii 14u. Note Ras Shamra forerunner 1 (MSL 11 169) ii 2u: a-sà-du6-lúsipa-e-ne. ii 19u. For references to a-sà-du6-gisma-nu in economic texts, see Pettinato, Untersuchungen, no. 281: “Feld ‘Hügel des Kornelkirschbaumes’.” ii 20u–21u. Note the juxtaposed homophonous entries: 20u. a-sà-du6-túl-túl 21u. a-sà-du6-gul-gul ii 24uu. See Pettinato, Untersuchungen, no. 90: a-sà-a-na-da-hul-engar “Feld des Anadahul, des Bauern.” ii 35uu. Our line, which immediately precedes the entries with /buru/, may be a phonetic variant of the Nippur forerunner (MSL 11 97) 23: a-sà-tu musen -gar-ra, which is three entries after the entries with /buru/.
ii 36uu–37uu. Note the juxtaposed homophonous entries: 36uu. a-sà-burux(GÁNA)bùru-ma 37uu. a-sà-buru14-mah For the /buru/ value of the GÁNA sign, see M. E. Cohen, The Cultic Calendars of the Ancient Near East (Bethesda: CDL, 1993), 43. Note Old Babylonian forerunner 1 (MSL 11 130) 32: a-sà-burux(GÁNA)bùru. The signi˜cance of the ˜nal -ma in our entry ii 36uu is unclear. ii 37uu. Note, e.g., M. Sigrist, Documents from Tablet Collections in Rochester, New York (Bethesda: CDL, 1991), no. 174 11: a-sà-burux(GÁNA)-mah. iv 6u–7u. Note the variants: Nippur forerunner (MSL 11 101) 143: a-sà-a-má-ùr-ùr; Hh (MSL 11 5) 5 1–2: a-sà-gismá-ru-r[u] a-sà-gismá-su-su; Old Babylonian forerunner 1 (MSL 11 130) ii 8–9: a-sà-a- gismá-ùr-ùr a-sà-a- gismá-sú-sú; Ras Shamra forerunner 2 (MSL 11 169) i 6–7: a-sà-má-KUMru-ru a-sà-má-má-su-su. v 3u–12uu. The order of the ˜rst ˜ve cities agrees with IM 51143 (MSL 11 56), but diverges thereafter: YBC 16318
IM 51143
Úriki Nibruki Ì-si-inki Unugki Ararmaki
Úriki Nibruki Ì-si-inki Unugki Ararmaki La-ga-aski Kisal?(PÀR)ki Gír-suki Adabki
Adabki (rest diˆers)
xi 4u. Note that according to the footnote on MSL 11 107, the second sign in CBS 6074+ (text A of the Nippur forerunner) for this entry “looks like TÚG,” as in our text. Note also Hh XXII section 9, 2u: pa5-zi-da. xi 9u. See Old Babylonian forerunner 8 (MSL 11 143) x 15: mul-mu-sír-kesda.
61
xi
xii
i
YBC 16318
ii
iii
iv
v
vi
ANOTHER OLD BABYLONIAN FORERUNNER TO HAR-ra={UBULLU XX–XXII
AN OLD ASSYRIAN TREATY FROM KÜLTEPE Veysel Donbaz (Istanbul)
The tablet published here, Kt 00/k 6, was excavated during the 2000 ˜eld season at Kültepe.1 It was found together with eighteen other tablets, including Kt 00/k10, a two-column treaty with Hahhum that was assigned for publication to Prof. Cahit Günbatti.2 Despite the abundance of documentation on the mercantile relationships between Anatolians and Assyrians, treaties between the Assyrians and their trading partners are rare. To date we know only the fragmentary treaty tablet found at Tell Leilan (Eidem 1991) and a text with treaty language, perhaps a letter, discovered at Kültepe (Bilgiç 1992; Çeçen and Hecker 1995). The tablet is di¯cult to read, especially on the obverse, and, despite several collations made since its discovery, can only be presented here in a highly preliminary manner, with the help of prominent Old Assyrian specialists from Leiden and Paris.3 The subject of Old Assyrian treaty making has been discussed thoroughly by Jesper Eidem (1991) in his study of the Leilan treaty, and therefore I do not discuss such matters here, but restrict myself to the problems concerning the text at hand. The
treaty, almost ninety lines long, seems to record bilateral conditions agreed to by both parties, the Assyrians and the local Anatolians. This contrasts with what we know from the other two Old Assyrian treaties, which present only the unilateral conditions relating to the Assyrians. Kt 00/ k6 This tablet, which measures 157x7x2.4 cm, now consists of two pieces that have been glued together. It is broken at the top and bottom: the left upper portion of the obverse and the left lower part of the reverse are badly broken. 1. [. . . DU]MU? dA-sùr sí-im su-mi sa [. . .] 2. [. . . DUMU? K]à-ni-ªis hiº-a-su-kà-am-ma 3. [. . . K]à-ni-ªis?º dsú-en6 dUTU 4. [. . .] dKu-ba-ba-ªatº [. . .] 5. [. . .]-ªmaº x [. . .]-ªaº-ni [. . .] 6. [. . .]-ªkà i-dáº-gu5-lu [. . .] 7. [x]-sa [. . .]-kà-ni a-ba-[. . .] 8. [. . .] 9. [. . .] x [. . .] 10. [. . .] x x [x] ªniº-is-ha-tí-[kà] sa-nim 11. [x] x x x x ªtaº-lá-qé-ú-ni 12. [x (x)]-ku-ni [. . .]x x bi4-tim 13. [a-n]a a-ªli-kà ta-liº-qé-ú-ni 14. a-na e-né-kà x? ta-su-ru-ni 15. ªiº-na e-mu-qí-e lá té-té-ru-ni 16. a-na ªsí-mìº-im ba-at-qí-im lá ta-lá-qé-ú-ni 17. [a]-na TÚG.ªHI.Aº ku-ta-nim is-tù ni-is-ha-tí-kà 18. ta-lá-qé-ªú-niº su-ma TÚG.HI.A ku-ta-ni-im 19. a-na si-[mí-im ta-lá]-ªqé-ú-niº [. . . ?]ªTÚG.HIº.A ku-ta-ni
1. I would like to thank Prof. Tahsin Özgüç† for permission to publish this text. I discussed it in my presentation to the RAI in Leiden in July 2002 and submitted the paper for publication in JCS in February of 2003. In the meantime Prof. Günbatti also received publication permission and his edition of this tablet, together with that of Kt 00/k10, has now appeared (2004). 2. See note above. 3. I am grateful to Klaas R. Veenhof and Cécile Michel for valuable suggestions. I am indebted to Dr. Michel for taking the photographs, which were of great help during my work on the tablet.
63
JCS 57 (2005)
64
VEYSEL DONBAZ
20. [t]a-lá-q[é ku-t]a-ni [. . .] KÙ.BABBAR 21. i-na [. . .]-im i-nu-mì 22. DUMUd]A-sùr i-na a-li-kà [. . .]-pá-lá-ni 23. [. . .] 24. [. . . T]ÚG 10.T[A ku-ta]-ni 25. [. . .]-tim ku-t[a-ni tù]-su-ru-ni 26. [lá ta-lá-]qé NA4.GÌN.ZA [. . .]-ªpìº-is-tim 27. [su-ma ta]-áb-su-um i-ªdaº-[na]-ªkuº-um 28. [su-ma lá t]a-áb-su-um a-sí-su tù-ta-ar-ma 29. [. . . K]Ù.BABBAR-pì-su ú-ba-al ú a-ta 30. [i-na] pá-ti-kà ra-bi4-ú-tim 31. x-kà sú-qá-qí-kà i-na ma-tí-kà 32. [ú a-li]-kà lá tú-sa-ªru-úº-ni-ma 33. [x] GÀR?-su ú ur-ki-su 34. [a-n]a? sí-su x x x UD 35. [. . .] ú-bi4-il5 a-sí-i-im 36. [. . .]-ni-ma a-sé-su 37. [. . .]-ú-ni i-nu-me da-mu-ªumº 38. [sa DUMUd]A-sùr i-na a-li-kà 39. [ú ma]-ªtíº-kà i-né-pu-su-ni 40. [hu-lu]-qá-um i-ba-sí-ú-ni 41. [a-n]a da-me-e sí-im-ªtámº 42. [ta]-dá-na-ni-a-tí-ma ni-du-wa-ªakº-su 43. [sa-n]i-a-am pu-ha-am lá ta-dá-na-ni-a-tí 44. [hu]-ªluº-qá-am lu tù-ta-ru-ni i-nu-me 45. x [x] x-ªsuº ú DUMUdA-ªsùrº i-na dì-ni-im 46. [. . .] ªDAMº.GÀR-su-nu i-ªnaº ke-na-tim 47. [. . .] ªaº-na ªDUMUº a-lì-ka Kà-ni-is 48. [. . .]-ªniº su-ma DUMUdA-sùr 49. [. . .]-ªsuº i-tá-ah-ba-al 50. [. . .] x x-ba i-lá-qé 51. [. . .] x x-ªsíº i-lá-qé su-ma 52. [. . .] ªDUMUºdA-sùr 53. x x [. . .] ú-ta-ªruº-su DUMU A-sùr 54. i-lá-qé-ªe DUMUdº[A-sùr] x x ur-kà-tám 55. i-lá-qé-e su-ma be-el ªhu-lu-qá-eº [hu-lu]-qá-um 56. i-na a-li-kà ú ma-ti-kà i-ªba-siº 57. ta-sí-e-ma tù-ta-ar su-ma 58. lá ta-ta-mar be-ªelº hu-lu-qá-e i-ta-ma-ma 59. ú a-ta hu-lu-qá-am tù-ma-lá a-na Ébi4-it 60. tám-kà-ri-im al-ma-tim DUMU Kà-ni-is 61. ú ha-pì-ra-am lá tù-su-ru-ni 62. a-na Ébi4-tim SIG5 ur-dim SIG5 63. am-tim SIG5 ªA†A5º SIG5 ú ki-ri-im 64. SIG5 sa DUMU A-sùr su-um-su e-ni-kà 65. lá ta-na-sí-ú-ma i-na e-mu-qí 66. lá te-te-ru-ni-ma a-na ur-dí-kà lá ta-du-nu-ni
67. i-na 10.TA ªpáº-ra-kà-ni 1 pá-ra-kà-nam 68. ni-is-ha-tí-kà ta-lá-qé su-ma TÚG 69. i-na a-bu-lim i-sa-ba-at TÚG sí-im 70. i-ta-ba-al su-ma DUMU A-sùr a-na DUMU Kà-ni-is 71. i-ta-ah-ba-al-ma a-na ma-tim 72. sa-ni-tim a-ta-lá-ak a-pu-hi-su 73. tám-kà-ra-am sa-ni-a-am a-hi-a-am 74. ú a-hu-su lá i-sa-bu-tù-ni ha-bu-lá-nam 75. sí-im-tám lu i-sé-e-ú-ni 76. i-nu-me ma-at-kà 77. ªaº-na ú-nu-sí-im ta-lá-pu-tù-ni 78. [D]UMU A-sùr su-um-su a-na ú-nu-sí-im 79. [l]á ta-lá-pu-tù-ni i-nu-me i-na a-li-kà 80. ªiº-dí-tim ªan-duº-ra-ar am-tim 81. [. . .] ªa-naº [. . .] ta-sa-ªra-ú-niº 82. [DUMU] ªA-sùrº su-um-su lá tù-sa-ra 83. [lá tù-su]-ru-[ni . . .] i-na e-ªmu-qí taº-sa-ba-tù[ni-ni] 84. [. . .] su-a-ti[m . . .] ªtù-ta-ru-ú-ni DUMUº A-sùr [. . .] 85. [a]-na-ak-ra-im ha-ªraº-nam lu tù-su-ru-ni-atí-ni [. . .] 86. [i-n]u-me DUMU A-sùr a-na ma-mì-tim ta-dunu-ni GÍR ú su-ga-ri-ªaº-[im sa A-sùr . . .] 87. [a-n]a sa ku-nu-ku-ni DINGIR-lu-ú a-ni-útum be-ªel maº- mì -tí-kà i-dá-gu5-lu-kà mamì-tum x [. . .] 88. GAL Kà-ni-sé-e-im Translation The top part of the obverse is in very bad shape, especially at the beginning, where we would expect to ˜nd a list of divine names, as part of the treaty oath, or the names of the parties involved. The ˜rst twelve lines are too fragmentary for translation. In line 1 we have provisionally restored DUMU dA-sùr. Accordingly, in line two we have restored [DUMU K]à-ni-is; this would provide us with the two parties to the treaty. However, it is more likely that one should read d IM rather than dA-sùr, and therefore interpret the opening passage as a list of gods, with the names of Sin and †amas at the end of line three. In line four after a break we read the name of the goddess Kubabat, and in line six i-ªdáº-gu5-lu, “they will
AN OLD ASSYRIAN TREATY FROM KÜLTEPE
attend/look/aim,” implying that these deities were involved in the oath that accompanied the treaty. These gods and goddesses are well attested in the Old Assyrian pantheon (Donbaz 1995: 187). The ˜rst ten lines are too fragmentary for translation, 13
to your city . . . which you shall take. 14according to your estimation . . . you shall release, 15you shall not take away by force, 16you shall not buy it at a cheap price. 17–19You shall purchase kutanu textiles at the (˜xed) import tax (nishatu). If you [purchase] kutanu textiles against its payment 20you shall tak[e] kutanu textiles [. . .] the silver. 21 [. . .] when 22[citizen(s) of As]sur in your city have been [. . .] 23[. . .] 24for each ten kutanu [. . .] 25[. . .] you shall release kutanu, 26[and you shall not get hold of it]. The lapis-lazuli4 [which is] made, 27[if he lik]es it, he shall sell it to you. 28[If he does not li]ke it, you shall return his iron, 29[. . .] he shall carry away [it]’s silver. And you (the Anatolian) 30[in] your vast territory 31[. . .] of your, at the alleys5 in your country 32[and in] your [ci]ty (when) you did not re[lea]se and 33[. . .] . . . or after him. 34 . . . 35[. . .] the carrier of the iron [. . .] 36his iron [. . .] 37–39If the blood [of a citizen] of Assur has been shed in your city [or cou]ntry 40and there are [los]ses, you will give us the one guilty6 of the blood(letting) and we shall kill him.7 43 You shall not give us [an]other person as a substitute. 44 You shall have the losses returned. When . . . 45somebody and a citizen of Assur at the court of justice. 46[. . .] their [mer]chants, truthfully 47[t]o the citizens of your city, Kanesh. 48[. . .] if a citizen of Assur 49[. . .] has taken away 50[. . .] he shall take 51[. . .]he shall take. If 52[. . .] a citizen of Assur shall get hold of it. A citi[zen of Assur] . . . later 55shall take it. If the owner [of the loss] has experienced 56[loss]es in your city or in your country 57 you shall search for it and return it. If 58you did not discover it, the owner of the lost (merchandise) shall swear an oath 59and you shall compensate (his) losses. 60–61 You shall not deliver a citizen of Kanesh or an outsider into the house of an (Assyrian) merchant or widow. 62–66 You shall not covet a ˜ne house, a ˜ne slave, a ˜ne slave woman, a ˜ne ˜eld, or a ˜ne orchard belonging to any citizen of Assur, and you will not take (any of these) by force and hand them over to your own subjects/servants. 67–68You will take one out of each ten pirikannu textiles as your import tax; if the textiles
4. See Michel 2001. 5. See suqaqu, “alley, back streets,” CAD S 398b. In NA it is attested as tehi su-qa-qi. 6. For simtum as “principal/slayer, the accused of.” See CAD † III 10. 7. See su-ma da-mu i-na ma-tí-kà i-ta-áb-si dá-i-ki lu ta-duni-a-tí-ma lu ni-du-ku (Çeçen and Hecker 1995: 35 ll. 7–11).
65
69
are seized at the gate, the textiles 70will be sold. 71–75If a citizen of Assur becomes indebted to a Kanishean and runs oˆ to another land, they shall not seize another merchant, a colleague, or his brother in his place; (instead), they will sue (the creditor) for the full price of the debt. 76–77When you inscribe (citizens of) your land (in the ledger) for hard labor (unussu), 78–79you shall not inscribe any citizen of Assur (in the ledger) for hard labor. When, in your city, 80[the matter] of the manumission of a slave girl in court 81[. . .] for [. . .] you will let free. 82If you do not set free any [citizen] of Assur, 83 in case you [did not re]lease [. . .] when you seize by force 84[. . .] him [. . .] you shall return a citizen of Assur [. . .] 85for the outsiders the roads you may open for us8 [. . .] 86[Wh]en you make the citizens of Assur take the oath on the patrum and sugari[um of Assur . . .] 87which is sealed/under seal, these gods, the lords of your oath, will look upon you, the treaty . . . . 88the chief executive of Kanesh.
Commentary Lines 77–78. The term unussum seems to have referred to a form of “hard labor” imposed by Anatolian princes upon subject citizens. Previously known references (Donbaz 1989: 78; 1991: 134) were not informative enough to allow a precise de˜nition, but the present context helps in some respects. It is obvious that this is not something one would readily desire, as the Assyrians reject it. Lines 87–88. The treaty (mamÿtum [87]) was concluded between the Assyrians, denoted as DUMU A-sùr, and a representative of the Anatolians, who appears to have been designated as GAL Kà-ni-sé-e-im, “Kanisite chief.” This title is not hitherto attested; it could be compared to GAL a-wi-li, “chief of the functionaries,” in Kt 87/k (:147). The passage that follows is broken and it is impossible to determine if the o¯cial was mentioned by name. Certainly there is enough space for a PN and for the mention of the other party, that is the Assyrians; therefore one might restore: mamÿtum [sa DUMU Assur ú PN] GAL Kaniseåim].
8. There seem to have been certain restrictions on commerce between Anatolians an Assyrians; see Sever 1990: 260–61 and Veenhof 1995: 1733–34, discussing Kt 79/k 101.
66
VEYSEL DONBAZ References
Bilgiç, E. 1992 “Ebla” in Cappadocian Inscriptions. Pp. 61– 66 in Hittite and Other Anatolian and Near Eastern Studies in Honour of Sedat Alp, ed. Heinrich Otten et al. Anadolu Medeniyetlerini Arastirma ve Tanitma Vak˜ Yayinlari, Sayi 1. Ankara: Türk Tarih Kurumu Basimevi. Çeçen, S. and K. Hecker 1995 ina matika eblum. Zu einem neuen Text zum Wegerecht in der Kültepe-Zeit. Pp. 31–42 in Vom Alten Orient zum Alten Testament: Festschrift fur Wolfram Freiherrn von Soden zum 85. Geburtstag am 19. Juni 1993. S/AOAT/ 240. Kevelaer and Neukirchen-Vluyn; Verlag Butzon und Bercker and Neukirchener Verlag. Donbaz, V. 1989 Some Remarkable Contracts of 1-B Period Kültepe Tablets I. Pp. 75–98 in Anatolia and the Ancient Near East: Studies in Honor of Tahsin Özgüç (Tahsin Özgüç Arma‹an), ed. Kutlu Emre et al. Ankara: Türk Tarih Kurumu Basimevi. 1993 Some Remarkable Contracts of 1b Kültepe Tablets II. Pp. 131–54 in Aspects of Art and Iconography. Anatolia and its Neighbours. Studies in Honor of Nimet Özgüç (Nimet Özgüç Arma‹an), ed. Machteld J. Mellink et al. Ankara: Türk Tarih Kurumu Basimevi. 1995 Kültepe Tabletleri I§ı‹ında IÖ. 2000 Yıllarında Anadoluånun Soysal Yapısı. Anadolu Medeniyetleri Müzesi Konferansları 1995: 177–96.
Eidem, J. 1991 An Old Assyrian Treaty from Tell Leilan. Pp. 185–208 in Marchands, diplomates et empereurs: Études sur la civilization mésopotamienne oˆerts à Paul Garelli, ed. Dominique Charpin and Francis Joannès. Paris: Éditions Recherche sur les Civilisations. Günbatti, C. 2004 Two Treaty Tablets Found at Kültepe. Pp. 249– 68 in Assyria and Beyond: Studies Presented to Mogens Trolle Larsen, ed. J. G. Derksen. Leiden: Nederlands Instituut voor het Nabije Oosten. Michel, C. 2001 Le lapis-lazuli des Assyriens au début du IIe millénaire av. J.-C. Pp. 341–60 in Veenhof Anniversary Volume: Studies Presented to Klaas R. Veenhof on the Occasion of His Sixty-Fifth Birthday, ed. W. H. van Soldt et al. Leiden: Nederlands Instituut voor het Nabije Oosten. Sever, H. 1990 Yeni Kültepe Tabletlerinde Geçen (Kÿma awat naruaåim) Tabiri ve Degerlendirilmesi. Ankara Üniversitesi Dil ve Tarih-Coghrafya Fakültesi Dergisi 34: 251–65. Veenhof, K. R. 1995 “In Accordance with the Words of the Stele”: Evidence for Old Assyrian Legislation. Chicago Kent Law Review 70: 1717–44.
67
Kt 00/k6
AN OLD ASSYRIAN TREATY FROM KÜLTEPE
68
VEYSEL DONBAZ
Kt 00/k6
LITERATURES IN CONTACT: THE BALA˝ ÚRU ÀM-MA-IR-RA-BI AND ITS AKKADIAN TRANSLATION UET 6/2, 403 Nathan Wasserman and Uri Gabbay (The Hebrew University, Jerusalem) In memory of Aaron Shaˆer (1933–2004), master of Ur
The substrata of Akkadian literature in the Old Babylonian period are not yet well known. Two of the underlying layers of this literary corpus should apparently be sought in the royal inscriptions and court poetry of the Sargonic dynasty in its capital Agade, and in the oral literature of the Amorite tribes scattered throughout Mesopotamia at the beginning of the second millennium B.C. Whatever its sources, this literature produced a broad range of genres: hymns, wisdom literature (parables, riddles, and dialogues), prayers of various kinds (to the gods and the heavenly host), lamentations, prophecies, love poems, as well as incantations, epic works, and various myths (Wasserman 2003a). This profusion of texts, written in the Babylonian and Assyrian dialects of the Akkadian language, stands in close proximity to another and no less developed corpus—that of Sumerian literature. Can Akkadian literature in the OB period be seen as a lively, bubbling liquid poured into the ancient vessel of Sumerian literature? Or, rather, are these two bodies of literature to be thought of as connected vessels whose contents mingle
and ˘ow into one another? In other words, what theoretical model should be implemented here: a static one that assumes a distinction between novel Akkadian literary material and traditional Sumerian literary molds, or a more dynamic model that seeks to analyze the amalgamation of two originally autonomous literary systems? This study will not try to solve these theoretical problems. Our premise is that there is bound to be a historical and functional interaction in such connected literary corpora, both in form and in content. This study aims to examine this complex interaction, focusing on a short Akkadian text, UET 6/2, 403, which has long been suspected by various scholars to have a Sumerian background, and its newly identi˜ed Sumerian equivalent, the OB BalaG Úru àm-ma-ir-ra-bi. This unique case of two parallel monolingual OB texts, in diˆerent languages on diˆerent tablets, one being a translation of the other, oˆers a new look into OB bilingualism, as well as Emesal cult songs within the curriculum of the OB Edubaåa. The starting point of this study is the Akkadian composition UET 6/2, 403. This short text was treated by Foster (1993: 91), Metzler (2002: 861–65) and Wasserman (2003b), who all recognized that this text was very likely dependent on a Sumerian prototype. Wasserman 2003b also emphasized the anomaly of the theme of the destroyed city in OB Akkadian literature. This opinio communis is now substantiated, since a
Note the following abbreviations: ETCSL = The Electronic Text Corpus of Sumerian Literature, University of Oxford, Oriental Institute, http://www-etcsl.orient.ox.ac.uk/. The Sumerian city lamentations are abbreviated after Michalowski 1989, xv–xvi: LU = Römer 2004 (= ETCSL 2.2.2); LSUr = Michalowski 1989 (= ETCSL 2.2.3); LN = Tinney 1996 (= ETCSL 2.2.4); LW = Green 1984 (= ETCSL 2.2.5); LE = Green 1978 (= ETCSL 2.2.6).
69
JCS 57 (2005)
70
NATHAN WASSERMAN AND URI GABBAY
Sumerian text, the OB BalaG Úru àm-ma-ir-ra-bi, has been identi˜ed as the word-for-word original of UET 6/2, 403. Text and Commentary UET 6/2, 403 is an eight-centimeter-long tablet, broken at the top. Only three or four double lines appear to be missing at the beginning. The tablet was examined at the British Museum and additional collations were made from photographs (see below).1 According to the label given to the tablet during the excavation, it was discovered in the house known as No. 1 Broad Street (Charpin 1986: 446, 451). Although the excavation records do not report exactly how many tablets were discovered in the rooms of this large house, they number several hundred, many of them literary (Charpin 1986: 447, see Wilcke 2000: 11–14). An administrative text found in close proximity to our tablet may provide a date for this group of tablets, between the years Sîn-iqÿsam 5 and Rÿm-Sîn 12 (Van De Mieroop 1992: 85, 87), but one should bear in mind that this corpus is not homogeneous. UET 6/2, 403 is an Akkadian translation of the second and perhaps also the third Kirugu of Úru àm-ma-ir-ra-bi. This literary unit consists of a lament of the goddess Inanna over her destroyed city.2 Below is an edition, together with the corresponding lines from the BalaG Úru àm-ma-ir-ra-bi (with parallels). UET 6/2, 403 (with correspondences from the BalaG Úru àm-ma-ir-ra-bi)3 Obverse (ca. 3–4 lines broken) Akk.: 1u [ ]x[ ] Akk.: 2u i-na ma-an-za-zi-ia ma-an-n[u-u]m/ i-zi-iz i-sa-sí Sum.: 69 A r.3 a ba-gub-bé-en ki-gub-Gu10 im-[me] 1. Courtesy of Christopher Walker and the Trustees of the British Museum. 2. That the protagonist of UET 6/2, 403 is the lamenting goddess was observed by Charpin 1986: 451, Foster 1993: 91 and Wasserman 2003b.
Long
Akk.: Sum.:
Akk.: Sum.:
Akk.: Sum.:
Akk.: Sum.: Akk.: Sum.: Akk.: Sum.: Akk.: Sum.: Akk.: Sum.:
B ii 11u a ba-gi-bé-en k[i] D r.ii 6 a ba-gub-ªbaº ki-ku-Gu10 i-me (vacat) E 13u a ba-g[ub]-bé-en ki-gub-Gu10 di-d[a ] 3u ki-ma la le-i-im/ i-na ma-a-a-li-sa i-sa-sí 71 A r.5 á-nu-Gál-gin7 ki-na! (text: DI?)-Gu10 im-me B ii 12u a-nu-Gál-g[e]-e[n]? x [ ] D r.ii 8 a-nu-Gá-fifiGIN7flfl-al-gin7 E 15u [ ]-gin7 ku?-nu-Gu10 di-d[a ] 4u i-na a-li sa-ak-nu-tim a-li it-ta/-ah-ba-as 74 A r.8 uru Gar-ra-ba! uru-Gu10 ba-gul B ii 14u úru!? Gá?-la?-ba x [ ] D r.ii 10 [ú]-ru Gá-ra-ba ú-ru-Gu10 ba-gul 5u i-na um-ma-ni-im sa-ki-tim/ a-wi-li it-te-pí 75 A r.9 ugnim!(KI.SU!.LU!.TU†!. GAR) Gar-ra-ba! mu-lu-bi ba-an-lah5 D r.ii 12 [u]g?-nu-um Gá-ra-ba [ ] x ba-an-tu-bé-es 6u i-na dA.NUN!? wa-as-bu-tim/ a-na-ku-ú-ma ah-ta-al-qu 76 A r.10 da-nun!?-na dúr-dúr-ru-na-bé me-e ba-ge16-le-èG-Ge26-dè 7u la-ah-ri i-na qá-qá-ar na-ak-ri-im i-sa-as-sí 77 A r.11 u8-Gu10 ki kúr-ta gù im-me 8u pu-ha-di (vacat) i-ra-mu-um 78 A r.12 sila4-Gu10 ki kúr-ta se àm-sa4 9u la-ah-ri ù pu-ha-sà it-ba-luo 79 A r.13 u8-Gu10 sila4-bi-ta ba-d[a]-de6 10u la-ah-ri na-ra-am i-na eo-bé-ri-s[a] 11u pu-ha-sà i-na ki-ib-ri-im i-ta-[di] 83 A r.17 u8 íd-da bala-Gá sila4 pes10 bí-in-sub B ii a+1 x-i-ªdaº? x [ ] D r.iii 1u [ ]-su? [ ]-Gu10 [ b]i-in-su
3. Kirugu 2 and 3 have been edited by Cohen 1988: 541– 53, 588–89, 600–601 (see corrections by Cavigneaux 1993: 255, included in our transliteration). The numbering of the Sumerian lines is according to Cohen 1988: 542–53. The following list of manuscripts is based on Cohen 1988: 536–41 and Volk 1989: 5. A = NCBT 688 (Cohen 1988: 840–843. Collated) (Cohen “C”). Larsa? (see Cohen 1988: 539), OB. B = PRAK B 442+C 52+C 121 (Cohen “E”+“H”+“J”). Kish, OB. C = TCL 16, 68 (Cohen “A”). Susa? (see Volk 1989: 5, n. 42), OB. D = NFT 203 (Cohen “NN”). Telloh, Early OB (parallel text). E = PRAK D 38 (Cohen “OO”). Kish, OB (parallel text).
THE BALA˝ ÚRU ÀM-MA-IR-RA-BI Reverse (ca. 3 lines broken) Akk.: r.1u k[a-al-bi ?? ] r.2u i-n[a? na-ak]-ri-i[m? ] / [(x) si?]-ªimº?-ti(-)im(-) [ ] Sum.: 91 A r.25 ur-su-ak-Gu10 lú kúr-ra nam-mu-ªunº-ªnaº-zu-zu B ii a+7 ud-su!?-a-qa- Gu10 lu ku!?-r[a ] C 3u su-ag-ga-ª Gu10º [ ] Akk.: r.3u i-sú-ra-tu-úo-a [sa? sú?-s]í ?-ªiº?-i[m?] r.4u sa la i-di-a ka-ap-pi-si-na/úo-na-ha-ta Sum.: 94 A r.28 musen-Gu10 ambar! nu-zu-zu á bí-ib-su13-e B ii a+9 musen-Gu10 ap-pa-ar nu-z[u? ] C 4u musen-Gu10 ambar nu-zu!-e-da ab mu-un-ªsu13º-[e?] Akk.: r.5u a-li bi-ti sa a-ra-mu-ú Sum.: 95 A r.30 é dím-ba-ba-re é dím-ba-ba-re é ki-áG-Gu10 me-a C 6u [ ]-re é ki-ga-áG-Gu10 me-a Akk.: rest uninscribed Translation of Akkadian Text 1u. [ ] .. [ ] 2u. “Who stood in my abode?!” she screams, 3u. Like a helpless one she screams in her bed: 4u. “Of (all) inhabited cities, (only) my city was devastated, 5u. “Of the (entire) regular army, (only) my man was taken captive; 6u. “Of (all) the Anuna-gods, it was only I who was destroyed! 7u. “My ewe calls (for help) from the enemy land, 8u. “My lamb bleats! 9u. “My ewe and her lamb they took away (from here). 10u. “My ewe, when crossing the river, 11u. “Abandoned her lambs on the shore (across). 1uu. “[My] d[og (?) . . . ] 2uu. “Fro[m the en]em[y (?) . . .] my [de]stiny (?) [ ] 3uu. “My birds [of the mar]sh (?)— 4uu. “Those who do not know clipped their wings. 5uu. “Where is my house in which I dwell (according to Sum.: which I love)?”
71
Philological Commentary Line 1u: This line, which only preserves traces of one sign, probably corresponds to Sum. 67 or 68 (see Cohen 1988: 543). Line 2u: Sum. should probably be translated: “ ‘Oh, I used to stand (there)! My abode!’ she cries out,” (similarly Cohen 1988: 589). For this line, compare SpTU III, 61: r.27–28 (note that the next two lines correspond to Sum. 70 in Cohen 1988: 543): te-e-àm a(-)ba-gub-bé-en ki-gub-Gu10 me-[à]m // mìn-su az-za-az-ma manza-az [(?)]. As understood by us, Sum. 69 and Akk. 2u have diˆerent meanings. Akk. 2u re˘ects a-ba=mannum, “who?,” thus disregarding the Sum. ˜rst person su¯x (-en), and translating the verb in the third person. All previous interpretations of Akk. 2u read i-zi-iz i-sa-sí asyndetically, as belonging to the same clause. Foster 1993: 91 understands the combination of these two verbs to express ˜nality, Metzler 2002: 864, progression of actions, and Wasserman 2003b: 127, contemporaneity. Now, Sum. solves the syntactic problem: izziz belongs to the direct speech of the goddess, whereas isassi governs the direct speech, a regular Sumerian syntax. Akk. ina manzazija understood Sum. ki-gub-Gu10 as a locative. The same translation technique is found in Akk. 3u. Line 3u: Sum.: “Like a helpless one she cries out: ‘My sleeping place!’ ” Sum. indicates that the ˜rst signs of Akk. 3u should not be read lalîm, “kid” (Metzler 2002: 864 and Wasserman 2003b: 127), but la leåîm, “weak, powerless” (so Foster 1993: 91, but Foster 2005: 153 “newborn goat”). Cohen 1988: 589 translates Sum.: “ ‘Wherever I walk, it is as if (I have) no strength!’ she says.” This translation is based on reading text A as kidi-Gu10 (Cohen 1988: 543). However, the Akkadian version and the var. ku?-nu-Gu10 in text E, indicate that text A should in fact be read as ki-na!-Gu10, standing for ki-nú=majjalum. As in the previous line, Akk. diˆers from Sum., understanding *kinú-Gu10 as a prepositional phrase (ina majjalisa), although Sum. has no locative mark. Accordingly, the possessive su¯x is changed from ˜rst to third
72
NATHAN WASSERMAN AND URI GABBAY
person (Sum. Gu10, Akk. -sa), and the entire Akk. sentence presents the goddess screaming in her bed, a scene not found in Sum. Lines 4u–5u: Sum.: “Of (all) those inhabited cities, (only) my city was destroyed. Of that regular army, its men were taken captive (Text D: ba-an-tu-bé-es, syll. for *ba-an-tu10-bé-es, “heaped up”).” Line 4u: This is the ˜rst attestation known to us of the correspondence: gul= habasum. Line 5u: Akk. indicates that the ˜rst word of Sum. 75 should be read as ugnim! (contra Cohen 1988: 543, who read: ki-tus-zu ki-gar). This is also proven by text D, which reads syllabically [u]g?nu-um. The adjective Gar-ra proves the reading sakittim (CAD †/1, 179 a, 2a; Foster 1993: 31; Metzler 2002: 864; Wasserman 2003b: 127–28) contra Mayer 1987: 205, who understood the signs to indicate sa kittim. The Akkadian scribe prefers a parallel construction, changing mu-lu-bi to awÿlÿ, “my man,” echoing alÿ, “my city,” in the previous line. The derivation from nepûm, suggested by Mayer 1987: 205 and CAD †/1, 179a, 2a) and followed by Wasserman 2003b: 127 (contra Metzler 2002: 861–62 with n. 1 and Foster 2005: 154, who argue for tebûm), is now proved to be correct by Sum. (text A). The correspondence lah5=nepû is not attested anywhere else. Nevertheless DU (de6, túm), lah 4 and lah 5 (DU.DU) are frequently equated with salalum (see CAD †/1, 196, lex), which is semantically close to nepûm. The verb nepûm is attested in OB texts only in legal and epistolary contexts, but note that ˜rst-millennium texts use this verb (in its forms napaåu and nabaåu) in the same meaning as salalu, “to plunder.” The use of the sign BI for pí, and not PI (as in rev. 4u: ka-ap-pi-si-na) is somewhat unusual for southern orthography,4 but is found in some Ur texts.5 4. Following Goetze’s criteria (see now George 2003: 272, n. 133). A survey of the Old Babylonian literary texts from Ur (Akkadian and bilingual) shows a clear predominance of /pi/ (20 attestations) over /pí/ (3 attestations). The sign /pi/ is attested in UET 6/1, 146+: ii 9(?); 6/1, 84: 14, 17; 6/2, 193: 5u, 8u, 11; 6/2, 388: ii 9 (see also 6/2, 389: 2u(?) which belongs to the same composition, bilingual Nisaba hymn); 6/2, 399: 10; 6/2:
Short
Line 6u: Sum.: “In (the presence of) those Anunagods sitting (in council), I was (ordered) to be destroyed!,” contra Cohen 1988: 589, who translates the sentence as transitive. The image of the Anunagods sitting in council and ordering the destruction of the city of the pleading goddess is portrayed in the fourth Kirugu of LU (ll. 137–69, see Römer 2004: 38–43). Here, the goddess stands before the Anuna, who are sitting (153: ba-an-da-dúr-ru-nees-àm) in council and pleads for the rescue of Ur, but the gods order the destruction of the city. In Akk. 6u we read i-na dA.NUN!?wa-as-bu-tim, which re˘ects Sum. 76: da-nun!?-na (with Cohen 1988: 543). The plural adjective wasb¿tim supports this suggestion. However, anunnakk¿ is usually written syllabically, or more rarely with the Sumerogram dA.NUN.NA, but not dA.NUN (see Kienast 1965: 142, with n. 10). Former readings of these signs were: “among the gods(?) residing there . . . .” (Foster 1993: 91 and 2005: 153); i-na DI˝IR fifia-si?flfl wa-as-bu-tim, “von den sitzenden Göttern! . . . .” (Metzler 2002: 861, 864); i-na MA† !?.GÁN !? wa-as-bu-tim, “of (all) inhabited places(?) . . . .” (Wasserman 2003b: 127–28). Again, contra Sum., Akk. builds a parallel construction with the previous lines, emphasizing the exclusiveness of the suˆering goddess among the Anuna, similarly to Akk. 4u–5u: “Of (all) . . . (only) my city” and “Of the (entire) . . . (only) my man.” As for the verb halaqum, contrary to Foster 1993: 91, 2005: 153 and Metzler 2002: 864, who understand it according to its meaning “to ˘ee, to escape,” Sum. indicates that the meaning here is “to be destroyed, to perish,” cf. Wasserman 2003b: 127. As Foster 2005: 154, we understand the form ahtalqu as an “a¯rmative subjunctive” (note that Sum. shows no emphatic form).
402: 6, 7, 11, 38; 6/2, 403: r.4u; 6/2, 404: 4, 10; 6/2, 414: 24. In addition, see UET 6/3, 421: 10(?) (Nanaja hymn); 6/3, 422: ii 14 (god list); and 6/3, 437: 12, 32(?), 36 (medical text). The sign /pí/ is attested in 6/2, 404: 6; 6/2, 414: 12, 17. All UET 6/3 texts are based on A. Shaˆer’s ms. 5. In two literary texts from Ur both /pí/ and /pi/ are attested: UET 6/2, 404: 4 (/pi/), 6 (pí); 6/2, 414: 12 (/pí/), 17 (/pí/), 24 (/pi/).
THE BALA˝ ÚRU ÀM-MA-IR-RA-BI
Line 8u: Akk., for the ˜rst time, omits a phrase which occurs in Sum. (ki kúr-ta). Note, however, that the scribe of Akk. left a wide space before the verb, which might indicate that the phrase ina qaqqar nakrim was meant here as in the previous line, in accordance with scribal practice known in Emesal cult songs. This is the only occurrence known to us of se— sa4=ramamum. Usually ramamum is equated with ur5—sa4 and se—sa4 is equated with damamum. Line 9u: Sum.: “My ewe was taken away from her lamb.” Akk. understood -bi-ta as -bi-da. At the end of the line Metzler 2002: 861 reads it-ba-ku, abakum perf. However, a photo of the edge of the tablet proves that CAD L, 43a, 1a) is correct and that the last sign of the verb is different than KU in l. 5u and should be read luo. Hence itbal¿, from wabalum-G perf., which also corresponds to de6 in the Sumerian version. Lines 10u–11u: Sumerian bala con˜rms the reading e-bé-ri-s[a] (with Mayer 1987: 204 and contra CAD L, 43a, 1a). Lines r.1u–2u: We read the beginning of Akk. r.1u as k[a-al-bi], corresponding to Sum. ur. Supporting this suggestion is SBH 57: r.14–16 (Cohen 1988: 56: 98), which parallels our lines and translates kalbum (probably due to a misunderstanding of ur-su-ak, for which cf. lú-su-a-ak = musalillum, “plunderer,” PSD A/III, 128–29). Note that in a proverb describing a kind of calamity (Alster 1997: 37:1.186), the same sequence of the verb su—ak after ú-íl occurs (see Sum. 90–91, Cohen 1988: 543 and parallel in Cohen 1988: 56: 96–99; Akk. broken). In the beginning of line r.2u we tentatively restore a form of sÿmtum, corresponding to Sum. nam, which opens the verbal form nam-mu-ªunºªnaº-zu-zu (text A), or if the form is to be read sÿmtÿ, corresponding to nam-MU. Lines r.3u–4u: This couplet corresponds to Sum. 94: “My bird does not recognize (its) marsh. It ˘ies away,” (Cohen 1988: 589).6 For the image
6. The phrase “who/which does not know” is found in descriptions of the puzzlement of the suˆerer forced out of his dwelling. In these cases the idiom “who/which does not know”
73
of birds and marshes (ambar) in Sumerian laments, see Black 1996: 33 (see also Kramer 1987: 75: 337). For the same image in a NA Akkadian lament (perhaps reviving Sumerian motifs), see Livingstone 1989: 40: r.5–7. We understand the beginning of Akk. r.4u as alluding elliptically to (nis¿) sa la ÿdia, “the men who do not know,” i.e., the enemy (Foster 1993: 91 diˆerently). Our suggestion is supported by LU: 280–81 (see Römer 2004: 63–64), where the idiom lú . . . nu-zu, “men . . . who do not know,” is used as an abridged allusion to the ignorant enemy soldiers who do not know and are entirely indiˆerent to the evil they wreak and the sacrilege they perpetuate in the sacrosanct heart of civilization. It reads: “Men ignorant of silver have ˜lled their hands with my silver, men ignorant of gems have fastened my gems around their necks.”7 Similar phrases can be found in lines 335–36 of LSUr: “Those who are ignorant of butter were churning the butter. Those who are ignorant of milk were curdling(?) the milk.”8 This is the only instance known to us where bu is equated with nuhhutu (AHw 714a, 4; note the inaccurate translation of CAD N/II, 318b b).
quali˜es the surroundings, or the road in which the exiled person ˜nds himself. So in Cohen 1981: 98: 73–75 (cf. Krecher 1967: 38–39, 57–61; note that the next line has the image of the birds in the marsh); LU: 285 (Römer 2004: 64 and 145–46); LSUr: 332 (Michalowski 1989: 56); LSUr: 65 (Michalowski 1989: 40). This idiom is also found in Akkadian: ªum-miº a-a[m-ri] ªaº-di-i[r-ti a]-na ka-ri-im sa la i-du a-na-ku e-mi-id, “My mother, look on my distress! I am leaning against a mooring place which I don’t know” (Groneberg 1997: 112: 76 with Streck 2003: 308). 7. kù-Gu10 lú kù nu-zu-ne [su]-bé ha-ba-da-ab-si / za-Gu10 lú za nu-zu-ne gú-bé ha-ba-da-ab-si (text A, see Römer 2004: 63– 64). Note that in the next line one ˜nds the image of the anguished bird. 8. ià-bi lú ià nu-zu-ne ì-dun5-dun5-ne / ga-bi lú ga nu-zune ì-im-mùs-mùs-ù-ne (see Michalowski 1989: 56; translation: ETCSL). Similar descriptions of the enemy, often described as a storm (u4), as “not knowing,” or as “not able to discern” are found in diˆerent lamentations: LU: 400–403 (Römer 2004: 81); LE: 1.19–1.20 (Green 1978: 132); LN: 64 (Tinney 1996: 100); Cohen 1988: 607: a+14–a+19. Similarly, this description is applied to evil demons as well: Dumuzi’s Dream: 111 (Alster 1972: 64); Inanna’s Descent 297 (Sladek 1974: 140); Kramer 1980: 8: 65.
74
NATHAN WASSERMAN AND URI GABBAY
Usually Sum. bu is equated with nasahu (PSD B, 164–66). In addition, bu equals kasasu when referring to wings: á bu-ra=ka-sa-su sá kap-pi (PSD B, 166a, 5). The compound verb á—su13, “to spread wide the wings, to ˘y,” is rendered by Akk. separately: á=kappum and bu, “to tear out, to pull out.” Line r.5u: Sum.: “The Edimbabar, the Edimbabar! Where is my beloved house?” (Cohen 1988: 589). Sum. ki—áG is not in accordance with the Akkadian spelling, which points to ramûm, “to dwell.” Sum. 96 con˜rms the assumption of Wasserman 2003b: 129 that a-ra-mu-ú is used polysemically, referring to râmum (ki—áG). In addition, as already noted by Wasserman 2003b: 128, the line may exhibit another word play—ali, “where” and alÿ, “my city” (the pairing of úru/ alum and é/bÿtum is most common in Sumerian laments, so already Fleming 1986: 690); hence the line may also read: “My city, my house, where I used to dwell!” 9 Akk. r.5u may summarize the entire Sum. Kirugu 3 of the BalaG Úru àm-ma-ir-ra-bi, which contains only three similar lines. Alternatively, Akk. r.5u may be the catchline to the next tablet, which began with Kirugu 3 (for such a possibility occurring in Lambert 1983: 213: 23, see below). Discussion UET 6/2, 403 and Sumero-Akkadian Bilingualism Most ˜rst-millennium Emesal cult songs are bilingual. Among OB bilingual tablets,10 only a handful are in Emesal.11 As shown above, UET 9. This compound is attested also in the OB prophetic dialogue from Uruk, where Istar-Nanaja addresses the king (probably Sîn-kasid) with the words: “I will seize the city and the house, when a true shepherd is established for his land,” (a-la-am ù bi-tam ú-sa-ba-[at] / ki-ma re-åu5-ú ki-nu a-na ma-ªti sak-nuº, UVB 18, Taf. 28: 14–15; see Metzler 2002: 279, 866). 10. See Wasserman 2003a: 179–80. According to Tinney 1999: 167–68, there are more bilingual and glossed versions in OB Ur of some compositions than from other sites. 11. E.g., TCL 16, 69, an OB BalaG manuscript with Akkadian glosses (Cohen 1988: 38). TCL 16, 68, an OB verson of Úru àmma-ir-ra-bi, has a translation of one of its lines into Akkadian (Cohen 1988: 553: 9). Both texts are perhaps from Susa (Volk
6/2, 403 is an OB Akkadian translation of a BalaG, standing alone without its Sumerian original. As such, it belongs to a special form of bilingualism, namely, Akkadian translation circulating independently of its Sumerian original. Examples for this type of bilingualism are scarce, the best known example being the twelfth tablet of the Standard Version of the epic of Gilgames and the Sumerian composition “Gilgames, Enkidu and the Netherworld,” and “Istar’s Descent to the Netherworld,” an adapted and abridged translation of the Sumerian myth “Inanna’s Descent to the Netherworld.” The same phenomenon occurs in royal inscriptions, mainly of the OB period, where Sumerian and Akkadian parallel versions appear on diˆerent objects (Galter 1995: 31–32), as well as on an OB tablet, probably from Sippar, containing a Sumerian version of part of the epilogue to the laws of Hammurabi (Sjöberg 1991). Another example of this form of bilingulism is the Akkadian version of lines 283–84 of the Sumerian King List (Sum.: a-ba-àm lugal a-ba-àm nu lugal; Akk.: mannum sarrum mannum la sarrum; this occurrs already in an Ur III manuscript, Steinkeller 2003: 272: 26u–27u and 279). An OB Sumerian text, preserved both as an Emegir composition (Lambert 1974: 291–93), as well as an OB Ersahunga found in Tell-Haddad (Cavigneaux and Al-Rawi 1993: 94), is known in the ˜rst-millennium both as a bilingual Emesal Ersahunga (Lambert 1974: 288– 91; Maul 1988: 213–15) and as a separate Akkadian Dingir-sà-dib-ba incantation (Lambert 1974: 278– 79).12 Finally, a Seleucid Akkadian lament (Lambert 1983)13 and the NA text STT 360 (Livingstone 1989: 5, n. 42). CT 44, 24 (Maul 1988: 10–15) is a bilingual Emesal prayer, featuring formulas and content of Ersahungas. 12. Note Lambert 1974: 270: “the Akkadian text is in fact a fairly close translation of a Sumerian original. This in itself is a rarity: though it is common to ˜nd Sumerian texts with interlinear Akkadian translations, the translations did not usually circulate alone.” 13. Supporting this suggestion is the enumeration of goddesses, a feature most common in Emesal cult songs but not in Akkadian, as well as the theological theme of Bel being accused of a catastrophe, a well attested motif in Emesal cult songs. Note also that the last line of the Akkadian lament (Lambert 1983: 213: 23) is very similar to the ˜rst line of another Seleucid Emesal cult song (probably BalaG) from Uruk, BaM Beih. 2, 44. This line could be a translation of the catchline to
THE BALA˝ ÚRU ÀM-MA-IR-RA-BI 1989: no. 16)14 may turn out to be dependent on Sumerian BalaGs. Akkadian translations of Emesal cult songs had a life of their own (see Volk 1992: 765). In some ˜rst-millennium bilingual literary texts, including Emesal cult songs, the Sumerian incipit is translated into Akkadian on the upper edge of the tablet.15 Cavigneaux and Ismail (1998: 6) have noted that this was done for library purposes, functioning much as the backs of books do in our day. This practice indicates an Akkadian reference to the Sumerian composition. In addition, Akkadian renderings of Sumerian texts often diˆer syntactically and stylistically from the original. This proves that the Akkadian had its own literary and stylistic preferences independent of the Sumerian text.16 Finally, apart from their literary value,17 Akkadian translations were used in cult. In the Seleucid ritual for the eclipse of the moon (BRM 4, 6 and parallels, edited by Linssen 2004: 306–20), the kalû-priest chants in Akkadian. This Akkadian performance is clearly a translation of Sumerian Emesal lamentations (Krecher 1966: 27, n. 50; Volk 1992: 765; Linssen 2004: 111).
the Sumerian BalaG section: Lambert 1983: 213, 23: ha-mi-ru mut(DAM) la-le-e i-te-fiekfl-ma-an-ni d+EN; BaM Beih. 2, 44: 1–2: [ d]am-Gu10 i-bí íl-la-Gu10 kar-ªmènº // [ ek]-me-ek mu-ut la-le-e ek-mé-ek. See the comment to line r.5u of the text edited above. 14. See the image of the deity as a merchant (tamkaru in Livingstone 1989: 39: 1–2 and dam-gàr in Cohen 1988: 102: a+196, 163: b+214), and the image of the birds and marshes in Livingstone 1989: 40 r.5–7, see the comment to ll. r.3u–4u above. 15. E.g., IVR 2, 18, 1; BL 84; BL 87; Cohen 1988: 801; 830 (BM 128083); Cavigneaux and Ismail 1998: 8. 16. See Volk 1992: 765 and Maul 1997. Note the stylistic feature of Akkadian texts changing the order of word pairs found in the Sumerian version already in the OB period, cf. Wasserman 2003a: 82. 17. Two more literary texts, excavated in Broad Street no. 1, should be mentioned here: UET 6/2, 395—a mythical description of NinGiszida’s descent to the Netherworld (Lambert 1990) and UET 6/2, 396—a mythical description of Istar and Ea (Gabbay 2002: 49–57, 125). These two short texts show various thematic links to diˆerent Sumerian compositions, and have no continuation in later literary tradition. UET 6/2, 395 is related to the two Sumerian myths concerning NinGiszida ( Jacobsen and Alster 2000; Black 2004), both known from Ur manuscripts (UET 6/1, 23 and UET 6/1, 27, the latter found in no. 1 Broad Street, as UET 6/2, 395, see Charpin 1986: 449–51).
75
The text preserved on UET 6/2, 403 can now be added to this special type of bilingualism, in which the Akkadian translation circulated independently of its Sumerian counterpart. As such, it omits the Sumerian lines which repeat themselves with only a change of the subject (Sum. 72–73, 80–82, see Cohen 1988: 543). Unlike ˜rst-millennium interlinear translations, where ˜xed equations between Sumerian and Akkadian words are usually found, UET 6/2, 403 makes its own literary choices and supplies Akkadian equivalents which are otherwise not known (Sum. 74: gul=Akk. 4u: habasum; Sum. 75: lah5=Akk. 5u: nepûm; Sum. 94: BU= Akk. r.4u: nuhhutum). In some cases, the Akkadian version of the BalaG diverges from the Sumerian and oˆers slightly new meanings. Some of these new meanings might result from misunderstandings of the Sumerian, but others may be deliberate reinterpretations of the original (e.g., the change from ˜rst to third person in Sum. 69, 71 // Akk. 2u, 3u; the parallel construction of Akk. 4u–6u, not found in Sum. 74–76; and Akk. r.5u, where the form a-ra-mu-ú shows a fancy word-play). Therefore, in this case, the way the Sumerian original is construed by the Akkadian scribe might not merely be considered a technical inim-bal, “translation” (see Seminara 2002: 246–7), but also a búr, “interpretation.”18 The BalaG Úru àm-ma-ir-ra-bi and other Emesal Cult Songs in the OB Edubaåa The BalaG composition Úru àm-ma-ir-ra-bi is one of the longest compositions of this type known to us, consisting of at least 33 Kirugus in its OB version, written upon at least twenty-two tablets in its “canonical” version of the ˜rst-millennium (Volk 1989: 25). Its OB recension is known from various sites, including Nippur, Sippar (Volk 1993: 66), Kish, Tell Haddad, Susa (Volk 1989: 5, n. 42 18. Note PBS 12, 32: 8 (Civil 2000: 106): lú èn-du ªbúrºbúr-ra, “the ones interpreting songs.” For interpretative translations of Emesal cult songs in the ˜rst-millennium, see Maul 1997. BÚR (pasaru) appears also in two scribal remarks in the Seleucid tablet SBH 35+MMA 86.11.347 featuring many interpretative translations; see Maul 2005: 25 (contra Cohen 1981: 42, n. 178).
76
NATHAN WASSERMAN AND URI GABBAY
and n. 45), and Ur (see below). It is referred to in a ritual text from Mari (Durand and Guichard 1997: 52–58; Cavigneaux 1998). In addition, it may have been mentioned in the OB catalogue VS 10, 216: 1: . . . ú[ru? à]m?-ma!-ér-ra-bi (reading according to ETCSL 0.2.07). Clearly, this BalaG was a composition of wide dissemination in the OB period. That it was translated into Akkadian is therefore not surprising. The OB Ur manuscript requires a short discussion. UET 6/2, 140 was considered to be Neo/ Late-Babylonian (Cohen 1988: 537; Volk 1989: 5; Black 1991: 33–34).19 Cavigneaux 1993: 255, however, contested this date, stressing that it is in fact an OB tablet. Cavigneaux’s proposal is supported by our OB text. UET 6/2, 140 was found at No. 1 Broad Street in close proximity to the text discussed in this study, UET 6/2, 403 (Charpin 1986: 450–51). However, UET 6/2, 140 was not the Sumerian source which was before the scribe of UET 6/2, 403: The ˜rst ˜fteen lines of UET 6/2, 140 contain parts of the ˜rst Kirugu of the BalaG, while UET 6/2, 140: r.9–r.21 contains parts of the fourth Kirugu, and the last three lines do not parallel the continuation of this section (Cohen 1988: 548–49; Cavigneaux 1993: 255–56). However, UET 6/2, 140: 16–r.8 varies from all other known Sumerian versions of this BalaG (see Cavigneaux 1993: 256). These are the lines that one would expect to contain parts of the second and third Kirugus of the BalaG, the sections translated into Akkadian in UET 6/2, 403. Note also that the Ur manuscript of the BalaG, as shown by Cavigneaux (1993: 255–56), was modi˜ed and presents a local setting: Ur replaces Uruk. The discovery of two OB texts of Úru àm-mair-ra-bi, in Sumerian and in Akkadian, in close proximity, together with other literary texts indicates that the archaeological context is not irrelevant to the understanding of the essence of this literary corpus (Wasserman 2003b: 130 and
19. Black considers it to be part of the group consisting of UET 6/2, 200–207, all Late-Babylonian excerpts of BalaGs. An examination of the photos of these tablets shows that UET 6/2, 140 does not belong to this group, as it diˆers in shape and signs from UET 6/2, 200–207.
Sallaberger 2004: 584). Their excavation spot (if accurately reported) mirrors, even indirectly, the original literary setting of No. 1 Broad Street. Charpin (1986: 484–85) argued that this house was not a library or an Edubaåa, but was probably used for storage or even for the disposal of administrative and literary tablets. Even so, contrary to Charpin’s reasoning, we still consider it plausible that the original place from which the literary tablets were taken and brought to No. 1 Broad Street was a library or a school.20 This reasoning leads to the conclusion that the BalaG was known, perhaps even studied (in the students’ courtyard or in the nearby teachers’ room) in the OB Edubaåa.21 Another indication that Úru àm-ma-ir-ra-bi was connected to the milieu of the Edubaåa, and not only performed in rituals, is the OB Sumerian literary letter PBS 12, 32, addressed to Nudimmudsiga from Igmil-Sîn. In the ˜rst half of this letter the sender asks the local schools to produce for him diˆerent texts, in exchange for which the addressee Nudimmud-siga will send back tablets of the BalaG Úru àm-ma-ir-ra-bi to the local schools (l. 12, see Civil 2000: 105). The sender Igmil-Sîn and the addressee Ibni-Ea may well be historical ˜gures, perhaps even from Ur (Civil 2000: 105, n. 3 and 107, see Charpin 1986: 459–65). In another literary letter describing the Edubaåa, Emesal cult songs are listed at the end of an enumeration of diˆerent genres: “I want them to learn the vocabularies (inim-inim-ma), the songs (èn-du), the inscriptions (na-rú-a), the Sumerian (eme-gi7 ) . . . all the way to the art of the cantor (zà nam-gala-sè)” (van Dijk 1989: 449–52,
20. Although Charpin’s demonstration of the secondary nature of the lot in No. 1 Broad Street is convincing, his conclusion regarding the literary tablets found in it seems to us overstated. The literary texts unearthed in this house ˜t perfectly into the curricular framework of post-elementary scribal education in the Old Babylonian period, as put by Charpin (1986: 452) himself: “L’image qui ressort de ces tablettes coincide par ailleurs fort bien avec le curriculum que décrivent le textes relatifs à l’instruction dans l’eduba . . . .” 21. Note the interesting presence of a fragment of a Sumerian BalaG (Di. 113) in the house of the Gala-mah Ur-Utu in Sippar-Amnanum among school tablets; see Tanret 2002: 8, 171.
THE BALA˝ ÚRU ÀM-MA-IR-RA-BI
19–20; Civil 2000: 106). The term nam-gala refers no doubt to the corpus of Emesal cult-songs. Recently, we have signi˜cantly advanced our understanding of the diˆerent stages of scribal training in the OB period with the publication of a series of studies, mainly by Tinney (1999), Veldhuis (2000a, 2000b, 2004: 58–66), Volk (2000), and Robson (2001, 2003). Common to these studies is the examination of large lots of school tablets of various types and forms found in diˆerent sites, rather than a focus on literary compositions, such as “Schooldays” (Kramer 1949), which describe life in the Edubaåa from an ideal point of view, glorifying the scribal world (on this point see now George 2005).22 Although these studies do not see Emesal cult-songs as part of the Edubaåa curriculum in the OB period (see esp. Veldhuis 2000b: 81 and Tinney 1996: 50–51), we argue that the evidence presented above may indicate that Emesal cult-songs did ˜nd their way into the OB Edubaåa, perhaps not at the level of the students’ assignments but at the level of the teachers’ scholarly interests. The OB Edubaåa was not the only source of OB literature, nor was it—with its emphasis on Sumerian and Sumero-Akkadian bilingualism (Wasserman 2003a: 179–80)—the sole valve that controlled the ˘ow of literary oeuvres from OB to later periods. Mesopotamian scribal activity was a multi-focal system, which had diˆerent sources of activity besides the Edubaåa: the palace, the priestly circles of the temples, the milieu of folk bards and diˆerent groups of singers, as well as the professional companies of men dealing with 22. Tinney (1999: 168) and Robson (2001: 51–55). Having said this, it is important to stress that the material at hand does not point to a centralized system of education in Mesopotamia in the OB period. Scribal training in this period was not organized from above (by the crown, or by temple administration), nor was there a uni˜ed curriculum implemented throughout the country. As put recently by George (2005: 131), archaeological evidence, as well as analysis of written sources “clearly show that already in the Old Babylonian period much scribal training was a small-scale activity run by private individuals and not by the state” (for this point see also Tanret 2002: 139, 155, 168). Hence, regional diˆerences in curricular usage and in educational practices are only to be expected, and indeed attested (Tanret 2002: 159–60, 161, 167 and Veldhuis 2004: 61–62; see Waetzoldt 1990).
77
magical lore and healing rituals—all intertwining into a complex creative matrix.23 Nevertheless, in the case of texts that show clear cross-linguistic links, namely Sumero-Akkadian contacts, the Edubaåa is the ˜rst candidate to be their place of origin and the most plausible locus for their literary existence. It is the setting of the Edubaåa that we consider to be the birthplace of the bilingual process that led to the existence of UET 6/2, 403. UET 6/2, 403 within its Akkadian Literary Setting Although UET 6/2, 403 is an almost verbatim translation of a Sumerian BalaG, it is not an implantation to the OB Akkadian literary system. Its generic classi˜cation does not pose a problem: it is a lament, cried out by a goddess, recounting in ˜rst person the destruction of her city. As such it shows some features that are known in the Akkadian literary system of the period, along with other features that are alien to it. Contrary to Sumerian literature, the city is not a frequent subject in OB Akkadian literature. This motif does appear in a composition describing Istar’s promiscuity. The refrain of this text hails the city: “Exaltation is the foundation for a city!” (von Soden and Oelsner 1991: 340: 1 and passim). A reference to the destruction of a temple is found in the very last line of a fragmentary mythological text involving Istar and Lillum. After an enigmatic mention of a bellicose encounter between Sumer and Subartu at the border of the Tigris, we read: “Then the gods left their cities and sat forgotten in their homes” (see Römer 1967b: 13: viii 10u). Another Akkadian text, probably related to Istar and Dumuzi, is perhaps more relevant. This text includes a detailed description of a ruined temple: “The shrine weeps, the inner sanctum laments, wherein we used to perform the rites of marriage. The courtyard is sighing, the storehouses lament, wherein we used to perform lovemaking. How full of woe is the lofty chamber, in which we used to sit joyfully. The 23. Westenholz 1974/77: 107–10.
78
NATHAN WASSERMAN AND URI GABBAY
gate is [broken], its jambs are blackened with smoke, [where] you met and spoke with me” (Lambert 1966: 54: r.12–19). Besides UET 6/2, 403, and the passages translated above, we know of no other OB Akkadian work that deals directly with the destruction of a temple or a city.24 This non-typical theme, however, is explained by the fact that UET 6/2, 403 mirrors a Sumerian text from a genre where destroyed cities are more than common. Still, although the main theme of UET 6/2, 403 is rare in OB Akkadian, other characteristics of the text are shared by other Akkadian compositions, all of which belong to the genre of personal laments. Among the published literary works that have survived from the OB period, twenty texts (among them eight bilingual) can be identi˜ed as laments.25 Five more texts can be added to this list, belonging to the genre of literary letters to gods, which show thematic resemblance to laments.26 Not all these Akkadian texts show thematic or stylistic unity, yet the following ten themes may be considered representative. (Sumerian personal laments such as Ersahunga prayers also exhibit this set of themes. An apt example is VS 10, 179, an OB Ersahunga with an Akkadian translation). UET 6/2, 403 ˜ts well within this group. Theme (1): Physical suˆering and mental distress of the suˆerer, e.g., “he is very ill, suˆering, broken (hearted), extremely frightened” (UET 6/2, 397: 16u–17u). At times this is put in terms of suffocation, caused by the god hiding his face from the man, e.g., “release me, oh my lord; I am short of breath” (VS 24, 97: 4u).27 Insomnia is also typical, 24. Later Akkadian compositions whose main motif is the city are Lambert 1983, Livingstone 1989: nos. 8, 9 and 10, and the Akkadian lament over Babylonian cities in the fourth tablet of the Erra myth. 25. The lamentations are listed by their entry number in the catalogue of OB literary texts in Wasserman 2003a: 185– 224: §§12, 38, 52, 53, 54, 64, 66, 70, 73, 107, 119, 122, 146, 192, 195, 201, 210, 217, 221, 274. 26. The letters to the gods are listed again by their entry number in Wasserman 2003a, 185–224: §§7, 153, 163, 194, 222. Non-literary letters to the gods, such as AbB 12, 99, AbB 13/2, 164, or the recently published Old Assyrian letter to the goddess Tasmetum (Kryszat 2003), are not included here. 27. See also Groneberg 1997: 114: 82 with Streck 2003: 309, and Groneberg 1997: 112: 63 with Streck 2003: 308.
One Line Long
e.g., “I am deprived of sleep and rest during the night” (Groneberg 1997: 110: 25 with Streck 2003: 306). Since the main arena of the man’s misery is his own body, a special detailed description of it is often given. Theme (2): References to the suˆerer’s body parts and inner organs, e.g., “My misfortune has increased, it attaches itself to (my) feet, [it has in˘icted] blows on my skull (?). You have made a mouth ˜lled (with food) very bitter to me . . . .” (Lambert 1987a: 190–91: 28–29). These somatic conditions represent the psychological drama that the suˆerer goes through; thus emphasis is put on emotional gestures, namely: Theme (3): Descriptions of tears and crying often involving expressive body language, e.g., “He has become weak and then bent: he is prostrate. His toil has become too heavy for him, he has drawn near to weep” (Lambert 1987a: 188–91: 4–5). In this context tears are often mentioned, e.g., “[from] the iris of my eye tear(s) are raining . . .” (Groneberg 1997: 110: 18 with Streck 2003: 305).28 From a discourse point of view, at the heart of these genres lies the intimate address of the complaining person to his estranged god. Laments and literary letters to the gods are the most personal, perhaps even the most “autobiographical” genre in the OB literary system. The suˆerer not only recounts his troubles to his god, but also shows his self-deliberations in the form of inner monologues. This is put across by: Theme (4): Extensive usage of the ˜rst person, e.g., “My lord, I have re˘ected within my reins, . . . in (my) heart. I do not know what sin I have committed. Have I . . . abomination against you? Have I . . . a very evil forbidden fruit?” (Lambert 1987a: 190–91: 12–14). Resulting from the marked communicative aspect of these Akkadian genres, emphasis is put on the oral interaction between the human and the divine, namely: Theme (5): Vocal and verbal acts, e.g., “He has gotten loud in the god’s presence, his chief. A bull is his speech, [his] voice two lamenters, his lips bear a lament to his lord . . .” (Lambert 1987a:
28. See also Lambert 1989: 327: 99.
THE BALA˝ ÚRU ÀM-MA-IR-RA-BI 190–91: 7–9). 29 Connected to this are the two next themes: Theme (6): Apostrophe, a direct address of the suˆerer to the god, e.g., “O, An¿na, may I not be destroyed [. . .]!” (Lambert 1989: 327: 129), and Theme (7): Rhetorical questions, e.g., “Why have you become (so) indiˆerent to me? Who could give you (more) than me?” (YOS 2, 141, cf. AbB 9, 141: 4–7).30 The last three themes express through external descriptions the internal havoc of the suˆerer’s situation, speci˜cally: Theme (8): Images of natural barrenness representing the social alienation of the suˆerer, e.g., “like an untouched wasteland I have no one looking at me” (Groneberg 1997: 112: 66 with Streck 2003: 308).31 Theme (9): Images of external destruction that mirror the speaker’s inner despair, e.g., “the gate is [broken], its jambs are blackened with smoke, [where] you met and spoke with me” (Lambert 1966: 54: r.18–19), and Theme (10): Anguished animals form imagery echoing harsh experience, e.g., “He brays like the weaned foal of a donkey . . .” (Lambert 1987a, 190–91: 6).32 Of the ten themes listed above, UET 6/2, 403 contains themes nos. (4), (5), (7), (9), and (10). Therefore, it shares many characteristics with the genre of Akkadian personal lament. In sum, though translated from Sumerian, UET 6/2, 403 is not alien to the OB Akkadian literary system. Conclusion The case discussed in this paper oˆers an example of a rare mode of OB bilingualism, i.e., Sumerian and Akkadian versions of the same composition on two diˆerent tablets, to be distinguished from other formats of Sumero-Akkadian bilingual texts. These include occasional explanatory glosses, interlinear word-for-word trans29. See also Lambert 1989: 326: 91–92. 30. See also the OB bilingual Ersahunga VS 10, 179: 1–2, see Maul 1988: 14. 31. See also Lambert 1989: 326, 330: 63–64. 32. See also Groneberg 1997: 112: 80 with Streck 2003: 304, 308.
79
lations, parallel column (“synoptic”) translations, complete Akkadian translation set beneath an original Sumerian text, two-sided translations, where the Akkadian translation and the original Sumerian text are found on the obverse and reverse of the tablet, and adaptations into Akkadian of Sumerian compositions.33 UET 6/2, 403, the Akkadian version of the BalaG Úru àm-ma-ir-ra-bi, had no known literary continuation in later periods, but interlinear bilingual copies of this BalaG are known from the ˜rst-millennium. Unfortunately, the exact passage translated by UET 6/2, 403 is not preserved (or not yet identi˜ed) in the late interlinear texts. Thus, the relationship between the OB separate Akkadian translation and the later interlinear Akkadian translation cannot be examined, though diˆerences between the two Akkadian translations are to be expected. The role of Akkadian within the pedagogical process of the OB Edubaåa, whose linguistic vehicle, or at least its didactic goal, was predominantly Sumerian, calls for further investigation. The OB Edubaåa—unlike Sargonic (Westenholz 1974/77), or post-OB (Gesche 2000) schools for scribes—was a center for the studies of Sumerian, and Akkadian texts were not at home in this particular context.34 It is reasonable to assume that
33. Cooper 1968: 10–14. It is clear that the phenomenon of Mesopotamian bilingualism must be investigated not only from the formal, but also from the functional point of view. In other words, distinctions should be drawn within the nebulous category “translation,” according to the diˆerent functions of translations (e.g., explicative, interpretative, exegetic, or harmonizing) in each given bilingual text. The diˆerent social and literary settings of a given bilingual text (e.g., didactic/ scholastic, propagandistic, ceremonial, or practical settings) should also be considered. On this problem see Lambert 2001 and Seminara 2002. 34. Tinney 1999: 168, n. 44 and Robson 2001: 62. A few OB school texts in Akkadian exist, as e.g., TIM 9, 53, a short collection of riddles in Akkadian. Furthermore, Akkadian letter writing was probably also taught in OB schools (Sallaberger 1999: 153–54). A possible explanation for the clear preponderance of Sumerian in OB schools is that the basic format of scribal instruction and the main components of its curriculum were established in the Ur III empire. In this highly centralized state “the central authorities . . . spread their vision of the world by many channels, including educational systems that helped inculcate elites and their literate servants into a new cultural
80
NATHAN WASSERMAN AND URI GABBAY
Akkadian was spoken while teaching,35 but this oral practice has left very few written traces (Veldhuis 2000a: 387).36 Questions regarding who
and social order” (Michalowski 2003: 195). The eˆectiveness of the Ur III bureaucrats and scribes, and the prestige which this dynasty had in later historiography, standardized its educational system and ˜xed its components (see also George 2005: 132–35). 35. Note the famous passage from “Schooldays,” where the young pupil complains: lú eme-gi7-ra-ke4 eme-uri bí-in-du11 e-se in-túd-dè-en, “The Sumerian monitor (said): ‘You spoke in Akkadian!’ He beat me” (see George 2005: 128). 36. Only in the Kassite period did Akkadian literary texts start to make their way sporadically into the apprentices’ curriculum (Veldhuis 2000b: 81–82). This process continued, and in ˜rst-millennium schools Akkadian was part of the students
chose to compose the Akkadian translation, UET 6/2, 403, from an excerpt of the Sumerian BalaG Úru àm-ma-ir-ra-bi, under what circumstances, and for what purposes, still remain open. Addendum: B. Groneberg, JCS 55 (2003), 62 n. 3, suggested that the Akkadian literary tagging amaåerakk¿tu is etymologically connected to the speci˜c BalaG Úru àm-ma-ir-ra-bi. This suggestion cannot be sustained, as amaåerakk¿tu is derived from the profession ama-ér-ra, “female wailer,” cf. Krecher 1966: 173.
curriculum, see Veldhuis 2000b: 80–81, Gesche 2000: 55 and passim, and recently also Foster 2003: 79–87.
References Alster, B. 1972 Dumuzi’s Dream: Aspects of Oral Poetry in a Sumerian Myth. Mesopotamia 1. Copenhagen: Akademisk Forlag. 1997 Proverbs of Ancient Sumer: The World’s Earliest Proverb Collections. Bethesda: CDL. Black, J. 1991 Eme-sal Cult Songs and Prayers. AuOr 9/1 (Studies M. Civil): 23–36. 1996 The Imagery of Birds in Sumerian Poetry. Pp. 23–46 in Mesopotamian Poetic Language: Sumerian and Akkadian. Cuneiform Monographs 6, eds. M. E. Vogelzang and H. L. J. Vanstiphout. Groningen: STYX. 2004 NinGiszida and Ninazimua. OrNS 73: 215–27. Cavigneaux, A. 1993 Mesopotamian Lamentations (Review of Cohen 1988). JAOS 113: 251–57. 1998 Sur le balag Uruamma’irabi et le Rituel de Mari. NABU 1998: 46 §43. Cavigneaux, A. and Ismail, B. K. 1998 Eine zweisprachige Hymne aus dem Haus des Beschwörungspriesters. ASJ 20: 1–11. Cavigneaux, A. and Al-Rawi, F. N. H. 1993 New Sumerian Literary Texts from Tell Haddad (Ancient Meturan): A First Survey. Iraq 55: 91–105. Charpin, D. 1986 Le Clergé d’Ur au siècle d’Hammurabi (XIXe– XVIIIe siècles av. J.-C.). Genève-Paris: Librairie Droz.
Short
Civil, M. 2000 From the Epistolary of the Edubba. Pp. 105– 18 in: Wisdom, Gods and Literature: Studies in Assyriology in Honour of W. G. Lambert, eds. A. R. George and I. L. Finkel. Winona Lake: Eisenbrauns. Cohen, M. E. 1981 Sumerian Hymnology: The Ersemma. HUCA Suppl. 2. Cincinnati: Hebrew Union College. 1988 The Canonical Lamentations of Ancient Mesopotamia. Bethesda: CDL. Cooper, J. S. 1968 Sumero-Akkadian Literary Bilingualism. Unpublished Ph.D. dissertation. University of Chicago. Durand, J.-M. and Guichard, M. 1997 Les rituels de Mari. Florilegium marianum 3. Mémoires de NABU 4: 19–78. Fleming, D. E. 1986 “House”/ “City”: An Unrecognized Parallel Word Pair. JBL 105: 689–92. Foster, B. R. 1993 Before the Muses. An Anthology of Akkadian Literature (˜rst edition). Bethesda: CDL. 2003 Late Babylonian Schooldays: An Archaizing Cylinder. Pp. 79–87 in Festschrift für Burkhart Kienast. AOAT 274, ed. G. Selz. Münster: Ugarit-Verlag. 2005 Before the Muses. An Anthology of Akkadian Literature (third edition). Bethesda: CDL.
THE BALA˝ ÚRU ÀM-MA-IR-RA-BI
Gabbay, U. 2002 Akkadian Mythology in the Old Babylonian Period, its Sources and Evolution: Stylistic and Linguistic Features. Unpublished MA Thesis. The Hebrew University of Jerusalem (Hebrew). Galter, H. D. 1995 Cuneiform Bilingual Royal Inscriptions. Pp. 25–50 in Language and Culture in the Near East. IOS 15, eds. Sh. Izre’el and R. Drori. Leiden-New York-Köln. George, A. R. 2003 The Babylonian Gilgamesh Epic. Oxford: Oxford University Press. 2005 In Search of the É.DUB.BA.A: The Ancient Mesopotamian School in Literature and Reality. Pp. 127–37 in “An Experienced Scribe Who Neglects Nothing.” Ancient Near Eastern Studies in Honor of Jacob Klein, eds. Y. Sefati et al. Bethesda: CDL. Gesche, P. 2000 Schulunterricht in Babylonien im ersten Jahrtausend v. Chr. AOAT 275. Münster: Ugarit-Verlag. Green, M. W. 1978 The Eridu Lament. JCS 30: 127–67. 1984 The Uruk Lament. JAOS 104: 253–79. Groneberg, B. R. M. 1997 Lob der Istar. Gebet und Ritual an die altbabylonische Venusgöttin. Tanatti Istar (Cuneiform Monographs 8). Groningen: STYX. Jacobsen, Th. and Alster, B. 2000 Ningiszida’s Boat-Ride to Hades. Pp. 315–44 in: Wisdom, Gods and Literature: Studies in Assyriology in Honour of W. G. Lambert, eds. A. R. George and I. L. Finkel. Winona Lake: Eisenbrauns. Kienast. B. 1965 Igig¿ und Anunnakk¿ nach den akkadischen Quellen. Pp. 141–58 in Studies in Honor of Benno Landsberger on His Seventy-Fifth Birthday, April 21, 1965. AS 16, eds. H. G. Güterbock and Th. Jacobsen. Chicago: The University of Chicago Press. Kramer, S. N. 1949 Schooldays: A Sumerian Composition Relating to the Education of a Scribe. JAOS 69: 199–215. 1980 The Death of Dumuzi: A New Sumerian Version. AnSt 30: 5–13. 1987 By the Rivers of Babylon: A Balag-Liturgy of Inanna. AuOr 5: 71–90.
81
Krecher, J. 1966 Sumerische Kultlyrik. Wiesbaden: Otto Harrassowitz. 1967 Die sumerischen Texte in “sylabischer” Orthographie. ZA 58: 16–65. Kryszat, G. 2003 Ein altassyrischer Brief an die Göttin Tasmetum. Pp. 251–58 in Festschrift für Burkhart Kienast, ed. G. Selz. AOAT 274. Münster: Ugarit-Verlag. Lambert, W. G. 1966 Appendix—A Fragment of a Similar Text. MIO 12: 52–56. 1974 DINGIR.†À.DIB.BA Incantations. JNES 33: 267–327. 1983 A Neo-Babylonian Tammuz Lament. JAOS 103: 211–15. 1987a A Further Attempt at the Babylonian ‘Man and his God’. Pp. 187–202 in Language, Literature and History: Philological and Historical Studies Presented to Erica Reiner, (American Oriental Series 67), ed. F. Rochberg-Halton. New Haven: American Oriental Society. 1989 A Babylonian Prayer to An¿na, Pp. 321–36 in Dumu-é-dub-ba-a. Studies in Honor of Åke W. Sjöberg, eds. H. Behrens, D. Loding, and M. T. Roth. Philadelphia: University Museum. 1990 A New Babylonian Descent to the Netherworld, Pp. 289–300 in Lingering Over Words (Harvard Semitic Studies 37). Atlanta, GA: Scholars. 2001 Babylonian Linguistics. Pp. 217–32 in Languages and Cultures in Contact. At the Crossroads of Civilizations in the SyroMesopotamian Realm (Orientalia Lovaniensia Analecta 96), eds. K. Van Lerberghe and G. Voet. Leuven: Peeters. Livingstone, A. 1989 Court Poetry and Literary Miscellanea. SAA 3. Helsinki: Helsinki University Press. Maul, S. M. 1988 Herzberuhigungsklagen. Die sumerischakkadischen Ersahunga-Gebete. Wiesbaden: Harrassowitz. 1997 Küchensumerisch oder hohe Kunst der Exegese? Überlegungen zur Bewertung akkadischer Interlinearübersetzungen von EmesalTexten. Pp. 253–67 in Ana sadî Labnani l¿ allik: Beiträge zu altorientalischen und mitelmeerischen Kulturen. Festschrift für Wolfgang Röllig, eds. B. Pongratz-Leisten, H. Kühne, P. Xella. AOAT 247. Kevelaer/ NeukirchenVluyn: Verlag Butzon & Bercker/ Neukirchener Verlag.
82
NATHAN WASSERMAN AND URI GABBAY
2005 Bilingual (Sumero-Akkadian) Hymns from the Seleucid-Arsacid Period. Pp. 11–116 in Cuneiform Texts in the Metropolitan Museum of Art II: Literary and Scholastic Texts of the First-millennium B.C., eds. I. Spar and W. G. Lambert, New York: Metropolitan Museum of Art. Mayer, W. R. 1987 Akkadische Lexikographie: CAD N. OrNS 56: 195–210. Metzler, K. A. 2002 Tempora in altbabylonischen literarischen Texten (AOAT 279). Münster: Ugarit-Verlag. Michalowski, P. 1989 Lamentation Over the Destruction of Sumer and Ur. Mesopotamian Civilizations 1. Winona Lake: Eisenbrauns. 2003 A Man Called Enmebaragesi. Pp. 195–208 in Literatur, Politik und Recht in Mesopotamien. Festschrift für Claus Wilcke. Orientalia Biblica et Christiana 14, eds. W. Sallaberger, K. Volk and A. Zgoll. Wiesbaden: Harrassowitz. Robson, E. 2001 The Tablet House: A Scribal School in Old Babylonian Nippur. RA 95: 39–67. 2003 Bird and Fish in the OB Sumerian literary catalogues. NABU 2003: 76–77 §68. Römer, W. H. P. 1967b Studien zu altbabylonischen hymnischepischen Texten (3). Ein Lied mit Bezug auf einen †ubartum-Feldzug Hammurapis (CT 15, 1–2)? WO 4: 12–28. 2004 Die Klage über die Zerstöring von Ur. AOAT 309. Münster: Ugarit-Verlag. Sallaberger, W. 1999 “Wenn Du mein Bruder bist . . .” Inteaktion und Textgestaltung in altbabylonischen alltagsbriefen. Cuneiform Monographs 16. Groningen: STYX. 2004 Review of Wasserman 2003a. BiOr 61: 579–86. Seminara, S. 2002 The Babylonian Science of the Translation and the Ideological Adjustment of the Sumerian Text to the ‘Target Culture’. Pp. 245–55 in Ideologies as Intelectural Phenomena. Melammu Symposia 3, eds. A. Panaino and G. Pettinato. Milano: Universita di Bologna & IsIAO. Sjöberg, Å. W. 1991 Was there a Sumerian version of the laws of Hammurabi? AuOr 9: 219–25. Sladek, W.R. 1974 Inanna’s Descent to the Netherworld. Unpublished PhD dissertation. Baltimore: The John Hopkins University.
Short
Steinkeller, P. 2003 An Ur III Manuscript of the Sumerian King List. Pp. 267–92 in: Literatur, Politik und Recht in Mesopotamien: Festschrift für Claus Wilcke. Orientalia Biblica et Christiana 14, eds. W. Sallaberger, K. Volk and A. Zgoll. Wiesbaden: Harrassowitz. Streck, M. P. 2003a Die Klage ‘Istar Bagdad’. Pp. 301–12 in Literatur, Politik und Recht in Mesopotamien. Festschrift für Claus Wilcke. Orientalia Biblica et Christiana 14, eds. W. Sallaberger, K. Volk and A. Zgoll. Wiesbaden: Harrassowitz. Tanret, M. 2000 Per aspera ad astra: L’apprentissage du cunéiforme à Sippar-Amnanum pendant la période paléobabylonienne tadrive. MHE III/ I, 2, Ghent: University of Ghent. Tinney, S. 1996 The Nippur Lament: Royal Rhetoric and Divine Legitimating in the Reign of IsmeDagan of Isin (1953–1935 B.C.). Philadelphia: The University of Pennsylvania Museum. 1999 On the Curricular Setting of Sumerian Literature. Iraq 61: 159–72. van de Mieroop, M. 1992 Society and Enterprise in Old Babylonian Ur. Berlin: D. Reimer. van Dijk, J. 1989 Ein spätbabylonischer Katalog einer Sammlung sumerischer Briefe. OrNS 58: 441–52. Veldhuis, N. 2000a Sumerian Proverbs in their Curricular Context. JAOS 120: 383–99. 2000b Kassite Exercises: Literary and Lexical Extracts. JCS 52: 67–94. 2004 Religion, Literature, and Scholarship: The Sumerian Composition Nanse and the Birds, with a catalogue of Sumerian bird names Cuneiform Monographs 22. Leiden: Brill. Volk, K. 1989 Die BalaG-Komposition Úru àm-ma-ir-ra-bi: Rekonstruktion und Bearbeitung der Tafeln 18 (19uˆ.), 19, 20 und 21 der späten, kanonischen Version, FAOS 18. Stuttgart: Steiner Verlag. 1992 Review of K. Oberhuber, Sumerisches Lexikon zu “George Reisner, Sumerisch-babylonische Hymnen nach Thontafeln griechischer Zeit (Berlin 1896)” (SBH) und verwandten Texten. Innsbruck 1990. BiOr 49: 764–72. 1993 Ad CT 58 Nr. 31. NABU 1993: 65–67 §80. 2000 Edubaåa und Edubaåa-Literatur: Rätsel und Lösungen. ZA 90: 1–30.
THE BALA˝ ÚRU ÀM-MA-IR-RA-BI
von Soden, W. and Oelsner, J. 1991 Ein spät-altbabylonisches parum-Preislied für Istar. OrNS 60: 339–43. Waetzoldt, H. 1990 Zu einer Schülertafel aus Mari. NABU 1990: 75–76 §97. Wasserman, N. 2003a Style and Form in Old-Babylonian Literary Texts. Cuneiform Monographs 27. Leiden: Brill. 2003b A Forgotten Old-Babylonian Lament over a City’s Destruction: UET 6/2,403 and its Possible Literary Context. Pp. 126–32 in Hayim and Miriam Tadmor Volume (Eretz-Israel 27), eds. I. Eph’al, A. Ben-Tor and P. Machinist. Jeru-
83
salem: The Israel Exploration Society (in Hebrew). Westenholz, A. 1974/ 77 Old-Akkadian School Texts. Some Goals of Sargonic Scribal Education. AfO 25: 95–110. Wilcke, C. 2000 Wer las und schrieb in Babylonien und Assyrien: Überlegungen zur Literarität im Alten Zweistromland. Bayerische Akademie der Wissenschaften, Philosophisch-Historische Klasse, Sitzungsberichte. Jahrgang 2000 6. München: Verlag der bayerischen Akademie der Wissenschaften / Verlag C. H. Beck.
UET 6/2, 403
NATHAN WASSERMAN AND URI GABBAY
cm
84
INTERNALLY-HEADED RELATIVE CLAUSES IN AKKADIAN: IDENTIFYING WEAK QUANTIFICATION IN THE CONSTRUCT STATE J. Cale Johnson (Cuneiform Digital Library Initiative, UCLA)
In a pair of recently published articles, Deutscher has argued that the early history of the relative clause in Akkadian is one fraught with dysfunction (2001, 2002). Old Akkadian relative clauses are described as “demonstrably dysfunctional” (2001: 405), “dysfunctional and maladaptive” (2002: 86, 95), and “an unstable way-station in the process of the emergence of a new relative clause structure in the language” (2002: 86–87). I suspect that Deutscher’s choice of pejorative terminology is meant more as an entertaining rhetorical device rather than carefully considered statement of linguistic theory, but, if nothing else, it has inspired me to attempt a redescription of the relative construction in Akkadian along lines similar to those enunciated by Deutscher. For a number of reasons that I adduce throughout the paper, I hypothesize that the relative clause in Old Akkadian generally as well as the construct relative clause in Old Babylonian is what is known as an internally headed relative clause (hereafter IHRC). To be more speci˜c, I will argue that the
“unstable way-station” in Old Akkadian is perceived as “dysfunctional” because both the relative formed via the placement of the head of the relative in the construct state as well as those formed by means of the determinative-relative pronoun are IHRCs, but that, by the time of the Old Babylonian period, the relative clause formed with the determinative-relative pronoun had been reinterpreted as an externally headed relative clause (EHRC). In this paper, I investigate the construct relative in Old Babylonian drawing on the material in the Code of Hammurapi. IHRCs have only achieved a reasonably coherent description in the past couple decades and largely in non-Western languages: the term itself derives from the particular form of IHRCs such as the following contrastive examples in (1) and (2) from Quechua (Cole 1987: 277; Basilico 1996: 499). Note that the languages that make use of IHRCs sometimes also make use of the kind of externally headed relative clause (EHRC) with which the reader will be more familiar.
This paper is based on a presentation of the same title delivered at the annual meeting of the American Oriental Society, March 18, 2005 in Philadelphia. I would like to thank
those who commented on the presentation at the AOS and/or on earlier draft versions.
85
JCS 57 (2005)
86
J. CALE JOHNSON
(1) Externally Headed Relative Clause [DP [CP nuna ranti-shaq-n] bestya] alli [DP [CP man buy-Perf-3] horse.Nom] good ‘The horse that the man bought was a good horse’
bestya-m horse-Evid
ka-rqo-n be-Past-3
(2) Internally Headed Relative Clause [CP nuna bestya-ta ranti-shaq-n] alli [CP man horse-Acc buy-Perf-3] good ‘The horse that the man bought was a good horse’
bestya-m horse-Evid
ka-rqo-n be-Past-3
The terminology derives from the placement of the head of the relative clause in the two constructions: in the externally headed relative in (1), the head of the relative clause (bestya, “horse”) has moved outside of the relative clause (to the right in Quechua) and has been assigned case by the main clause (nominative case as the subject of ka “to be”), whereas, in the internally headed relative in (2), the head of the relative clause remains inside the relative clause and bears the case appropriate to its role within the relative clause rather than the main clause. If the location of the head noun and the assignment of case were the only factors determining whether or not a relative clause is head internal, then we would be forced to argue that the Akkadian relative clause is clearly an externally headed relative, but, as it happens, (3) [DP [CP Mary [owiza wa] kage] [DP [CP Mary [quilt a] make] ‘I bought the quilt that Mary made’
ki] the]
Determiners follow the nouns they modify in Lakhota and, as can be seen in (3), the inde˜nite determiner wa modi˜es owiza, the head of the IHRC; at the same time, the de˜nite determiner ki modi˜es [CP Mary [owiza wa] kage], namely the entire IHRC as a whole. (4) *[DP [CP Mary [owiza ki] *[DP [CP Mary [quilt the]
kage] make]
ki] the]
This kind of restriction on the occurrence of a de˜nite nominal phrase in a particular morphosyntactic environment is known as a de˜niteness eˆect and has been investigated in excruciating detail in the years since Milsark’s initial description (Milsark 1974; see Chung and McCloskey
there are several additional features of IHRCs involving de˜niteness and quanti˜cation that may well provide su¯cient justi˜cation for a reanalysis of the Akkadian forms. Based in large part on a seminal study of IHRCs in Lakhota (Williamson 1987), a growing number of researchers have noted that the head of an IHRC in a variety of languages must be inde˜nite and that IHRCs also display a quanti˜cational phenomenon known as the de˜niteness eˆect (Munro 1976; Tellier 1989; Watanabe 1992; Basilico 1996; 2003; Shimoyama 1999; Nishigauchi 2003; Kim 2004). One of the examples from Lakhota cited by Williamson is in (3) below. Note that I have updated the syntactic terminology slightly and omitted the diacritics—see Williamson 1987, 171 for the original form. he Dem
ophewathu I.buy
If the inde˜nite determiner wa is replaced by the de˜nite determiner ki, as in the following example, the IHRC is no longer grammatical (Williamson 1987: 171):
he Dem
ophewathu I.buy
2002 for an extensive set of references and a history of investigation; see Chung and Ladusaw 2004 and Hallman 2004 for the most recent work on the syntax of inde˜nites and the de˜niteness eˆect). The behavior of the de˜nite and inde˜nite articles in IHRCs clearly resembles the behavior
INTERNALLY-HEADED RELATIVE CLAUSES IN AKKADIAN
of the de˜nite article in the construct state in that the de˜nite article cannot appear on the head noun, while it may or may not appear on a modi˜er of the head noun. This resemblance should be kept in mind throughout the balance of the paper. Within the broad expanse of Semitic languages, only the Semitic languages dating to the ˜rst millennium B.C.E. and later exhibit a de˜niteness eˆect in the strict sense of the term due to the fact that morphological de˜niteness seems to be limited to the later Semitic languages. Furthermore, in these later Semitic languages such as Aramaic or Biblical Hebrew, the only clear example of a de˜niteness eˆect is the restriction on the occurrence of the de˜nite article on any noun that is in the construct state. (5) Biblical Hebrew bet ham-melek house.Const Def-king ‘the house of the king’ Thus in (5) the noun in construct, bet “house,” which is also the head of the genitive construction cannot bear the de˜nite article *han- and the de˜niteness of the entire phrase is determined by the presence (or absence) of the article on the noun that follows and is not in construct, hammelek, “the king.” The older Semitic languages, however, generally lack the de˜nite article entirely and seem to make do with a system based on speci˜city rather than de˜niteness. Unlike the de˜nition of de˜niteness, which is relatively uncontroversial, the precise de˜nition of speci˜city continues to show variation and uncertainty in the literature. In English, for example, the combination of an intensional predicate such as “to look for” with an inde˜nite nominal phrase like “a magazine” in (6) tends to have two rather diˆerent interpretations as exempli˜ed in (7). (6)
I am looking for a magazine
(7a) ‘I am looking for any old magazine’ (non-speci˜c) (7b) ‘I am looking for a particular magazine’ (speci˜c)
87
In (7a), the inde˜nite noun phrase “a magazine” in (6) is interpreted as non-speci˜c—any magazine will do as long as it is a magazine. In the interpretation of (6) in (7b), however, there is a particular magazine that the speaker is searching for, but for whatever reason the speaker is unable or unwilling to oˆer a better description: the inde˜nite noun phrase interpreted as in (7b) is speci˜c, whereas the interpretation in (7a) is non-speci˜c (for a more rigorous de˜nition, see Hallman 2004, 709). Although there is no commonly accepted de˜nition of the term, the idea of an inde˜nite noun that is referentially speci˜c has produced an extensive discussion in the literature (Donnellan 1966; Ioup 1977; Pesetsky 1987; Enç 1991; Diesing 1992; Ihsane and Puskás 2001). Without delving too deeply into what is a quite involved matter, speci˜city can be associated with entities that can be referenced in subsequent discourse through the use of a pronominal element (“it is on the couch” in reference to the speci˜c magazine in [7b]), whereas pronouns cannot be used to refer to nonspeci˜c entities (“one is on the couch” in reference to the non-speci˜c magazine in [7a]). Given these considerations, I would like to suggest that early Akkadian made use of a speci˜city based system that is preserved to some degree in the way that it forms relative clauses. When the head of a relative clause is in the construct state, the nonspeci˜city of the head of the relative functions, in my view, much like the inde˜nite determiner in the Lakhota example in (3) above. If the speaker wishes to indicate that the head of the relative clause is speci˜c, however, the noun in the construct state must be focalized and moved out of the relative and replaced by a resumptive pronoun in the form of the determinative-relative pronoun: this amounts, in other words, to the formation of a cleft-sentence in which the determinative relative pronoun inside the relative clause acts as a pronoun referring to the former, now external head of the relative that is in focus. That is to say, I would argue that the use of focus in conjunction with pronominal resumption to code speci˜city resulted in the formation of a discontinuous nominal phrase in which the external head of the relative
88
J. CALE JOHNSON
formed an entirely separate nominal phrase that was resumed by the determinative-relative pronoun (see Szabolcsi 1994; Aboh 2004; Ntelitheos 2004 for the role of focus in forming discontinuous nominal phrases in Hungarian, the West African language Gungbe, and Greek respectively). Deutscher’s Hypothesis Deutscher’s proposal is a typological argument at its essence: relative clauses in which the “pronoun” inside the relative clause takes the same case as the “head” outside of the relative clause
are typologically rare. In comparison to, say, the classical languages of Europe such as Latin and Greek, which for better or worse provided an early model for typologists attempting to categorize the diˆerent kinds of relative clause found around the world, the situation in Old Akkadian does seem perverse. In (8) below, for example, as Ehrenkranz and Hirschland describe it, “[t]he relative word usually agrees in number and gender with its antecedent and takes the case appropriate to its function in its own clause” (Ehrenkranz and Hirschland 1972: 23).
(8) Latin relative clause magister [quem canis teacher.MascSgNom [Rel.MascSgAcc dog.Nom ‘The teacher [whom the dog bit] sees the boy’
momordit] he.bit]
puerum boy.Acc
videt he.sees
(9) Old Akkadian relative clause (HSS 10, 5, obv. 4–7; Michalowski 1993: 34; Kienast and Volk 1995: 182–84; Deutscher 2001: 406) {†E su a-na †E.BA / a-si-tu / a-na †E.NUMUN / li-sa-mi3-id-ma} ûm [s¿ ana †E.BA asÿt-u] ana †E.NUMUN barley.Nom [Rel.Nom to/for ration(s) 1Sg.left-Sub] for seed-grain li¶amid-ma Prec.Cause.lean.3cs-Conj ‘As for the barley [that I left for rations], let him assign it as seed-grain . . .’ In the Old Akkadian example in (9), in contrast with the example from Latin in (8), the case of the “relativizer” s¿ is identical with and determined by the “head” of the relative clause, namely ûm.1 On the basis of a perusal of several typologies of relativization from ˜fteen to twenty years ago (Lehmann 1984; Nichols 1984; Givón 1990), Deutscher quite reasonably concludes that “the Old Akkadian construction can safely be assumed to be rare” (Deutscher 2001: 408). 1. An anonymous referee notes that the external head of the construction is written logographically and is not, therefore, a particularly good example of the phenomenon in question, but the nominative determinative-relative pronoun as the direct object of the verb within the relative clause makes it clear that the case of the pronoun is not determined by the verb within the relative. The nominal case of the relative construction in this example is driven by pragmatic factors (casus pendens as topic), but I ˜nd it an interesting example nonetheless and retain it here, see Deutscher 2001: 406 for additional examples.
One [Body] Line Short
Deutscher notes several descriptive features of relative clauses in Akkadian that might at least suggest the possibility of an IHRC analysis such as (1) the appositional character of early Akkadian relatives, (2) the role of the construct state in marking the head of genitive constructions, and (3) the association between “emphasis” (see below for the role of focus, another term for “emphasis,” in my own proposal) and the relative headed by the determinative-relative pronoun (Deutscher 2001: 410–11). Indeed, Deutscher’s descriptive work captures the phenomenon I describe below quite well, but neglects the possibility that there is a reasonable syntactic model that would ˜t such a description. The development of su from the head of a RC [relative clause] to a ‘relative particle’ must have its origins in appositional patterns. The whole genitive/relative complex headed by su could be added as an apposition
INTERNALLY-HEADED RELATIVE CLAUSES IN AKKADIAN
after a noun, probably for the purpose of emphasis. (Deutscher 2001: 414)
It should be kept in mind, however, that although appositional and internally headed relatives do share a certain super˜cial similarity, they are fundamentally diˆerent in a number of ways. Where I would disagree with Deutscher is the role he gives to clause boundary reanalysis. Deutscher suggests that relative clauses headed by the determinative-relative pronoun originated through appositional structures such as in (10) below (Deutscher 2001: 412). (10) dÿn-um suHEAD [idÿn-u]RC legal.case-Nom Dem.Nom [he.judged-Sub] ‘the legal case, (namely) the one that he judged’ (11) dÿn-umHEAD [s¿ idÿn-u]RC legal.case-Nom [Rel.Nom he.judged-Sub] ‘the legal case that he judged’ Whereas Deutscher sees a progression from the appositional structure in (10) to the restrictive relativization in (11), I would like to adopt an even stronger version of Deutscher’s suggestion that relative clauses headed by a noun in the construct state provided a model for relatives formed using the determinative-relative pronoun: I would suggest that, whereas a non-speci˜c head noun would occur in the construct state at the beginning of the relative clause as in (12), a speci˜c head noun would be focalized and moved out of the relative clause entirely, while the slot it had previously occupied in the construct state within the relative was ˜lled with a kind of resumptive pronoun2 bearing all of the nominal features (including case) of the focalized head noun and thereby yielding the form in (13).3
2. One of the lesser known features of weak quanti˜cation environments that exhibit a de˜niteness eˆect (as is the case with IHRCs) is that de˜nite nominal phrases can occur in such an environment when aˆected by focus; the historical scenario envisaged here would seem to provide the focus necessary to allow the resumptive pronoun to occur as the head of an IHRC.
89
(12) [dÿnHEAD idÿn-u]RC [legal.case he.judged-Sub] ‘A/the legal case he judged’ (13) [Focus dÿn-um] [CP s¿ [Focus legal.case-Nom] [CP Rel.Nom idÿn-u] he.judged-Sub] ‘It is [Focus a legal case] (that) he judged’ Once the head noun was in focus and separated from the rest of the relative clause (now headed by the determinative-relative pronoun), these two elements would form a discontinuous nominal phrase analogous to a cleft-sentence at the clause level (“It is a legal case that he judged”). In other words, I would suggest that (10) above is not representative of any stage in the history of the Akkadian relative, and that (11) was derived from an underlying form such as (12) through focalization of the head noun and resumptive pronominalization. The fundamental diˆerence between Deutscher’s proposal and my own is the location of the head of the relative clause with respect to the relative clause as a whole. Deutscher’s proposal presumes that the head of a relative clause must be outside of the relative clause itself, hence the characterization of dÿnum as a head of the relative and s¿ as a relativizer in (11). My own proposal argues that relative clauses in the older phases of Akkadian are head internal. This hypothesis not only yields a more parsimonious explanation of relativization in Akkadian in that all relative clauses in Old Akkadian (whether headed by a noun in the construct state as in [12] or the determinative-relative pronoun as in [13]) make use of the same basic construction: where the head of the relative has been moved out of the relative clause itself, the construct state position 3. See Simpson and Wu 2002, for a convincing demonstration of the role of focus constructions in the development of relative clause constructions; Ouhalla 2004 describes how a similar process plays out in the later Semitic languages, but makes no reference to the earlier Semitic languages such as Old Babylonian. See Schachter 1973 for an earlier proposal along similar lines.
90
J. CALE JOHNSON
within the relative is still present and is occupied by a resumptive pronoun in the form of the determinative-relative pronoun.4 The (Weak) Quanti˜cational Test Even if the history of relativization in Akkadian can be redescribed in reasonable terms as a development of a basic IHRC template into two distinct constructions, one of which—the construction involving the determinative-relative pronoun—eventually forms something quite like our own EHRC, in the absence of a morphosyntactic test of some kind, my reconstruction of relativization in Akkadian remains somewhat speculative. One of the essential properties of IHRCs, however, is weak quanti˜cation. The opposition between weak and strong quanti˜ers is a more re˜ned version of the de˜niteness eˆect that I alluded to at the beginning of the paper. In English, weak quanti˜cation is clearly evident in existential sentences, where weak quanti˜ers can occur, but strong quanti˜ers cannot (Milsark 1974; Williamson 1987: 175; Hallman 2004: 709). (14) There is/are {a ˜reman / three ˜remen / many ˜remen / ˜remen} available. (15) *There is/are {the ˜reman / every ˜reman / most ˜remen / Sheila} available
The weak quanti˜ers in (14) include noun phrases headed by an inde˜nite article, a cardinal number, “several,” as well as the bare plural, whereas the strong quanti˜ers in (15) cannot occur in an existential sentence. Williamson’s description of the opposition between weak and strong quanti˜ers is as follows: “Cardinality expressions” (allowed in existentials) include the inde˜nite determiners a and some, the quanti˜ers many and few, and the cardinal numbers. In contrast, “quanti˜ed expressions” (ungrammatical in existentials) include the de˜nite determiner the, demonstratives, proper names, de˜nite pronouns, and the quanti˜ers all, every, most, and so forth. (Williamson 1987: 175)
In terms of meaning, the strong quantifers are capable of picking a particular referent out of some larger set of possible referents, whereas the weak quanti˜ers are not. Thus “most ˜remen” necessarily refers to some but not all of the contextually salient set of ˜remen, but the weak quanti˜ers in (14) refer to some cardinal number of ˜remen or to the ˜remen who are available in a particular context. This contrast between strong and weak quanti˜ers is particularly clear in languages that make use of both IHRCs and EHRCs such as Korean. Take, for example, the contrast between the two following examples (Kim 2004: 39).
(16) Externally headed relative clause in Korean (Kim 2004: 39) John-un [[ei tomangka-n]-un sey-myeng-uy totwuki]-ul capassta PN-Top [[ei run.away-Impf ]-Rel three-Cl-Gen thief ]-Acc caught ‘John caught three thieves (out of many more) who were running away’ (17) Internally headed relative clause in Korean (Kim 2004: 39) John-un [[sey-myeng-uy totwuk-i tomangka-n]-un kes]-ul capassta PN-Top [[three-Cl-Gen thief-Nom run.away-Impf ]-Rel Thing]-Acc caught ‘(Only) three thieves were running away and John caught all of them’ 4
4. The historical derivation of an externally headed relative from an internally headed relative would also ˜t nicely in Kayne’s antisymmetric theory of relative clause formation (Kayne 1994: 86–115), particularly once head-initial IHRCs are acknowledged (see Basilico 1996 for the plausibility of headinitial IHRCs). If Basilico is correct in arguing that IHRCs must be headed by speci˜c nominal phrase, then the semantic characterization of the materials in this paper would have to be
reconsidered: one possibility is that the speci˜city contrast that I have discussed herein is actually a de˜niteness contrast. Nonetheless, my characterization of the syntactic contrast on the basis of weak quanti˜cation would presumably survive such a recon˜guration of the semantics of the construction. See also Ouhalla 2004 for an application of Kayne’s antisymmetric theory of relativization to the later Semitic languages.
INTERNALLY-HEADED RELATIVE CLAUSES IN AKKADIAN The head of the EHRC in (16), namely seymyeng-uy totwuk-i, “three thieves,” has moved from its original position within the relative clause to the right and left behind a gap in its original position—this is indicated by ei in the example. But in (17), an IHRC, the same phrase sey-myeng-uy totwuk-i, “three thieves,” remains in situ within the relative clause and also displays a rather speci˜c meaning in certain contexts. Whereas (16) can be used to refer to three thieves when there are three or more thieves in the available context, the IHRC in (17) cannot pick a group of three out of a larger group of thieves: the IHRC can only refer to three thieves—no more and no less—and the three thieves that it refers to must include all the thieves who are contextually available. (18) CH §119, reverse, iii 74 – iv 4 74–75. sum-ma a-wi-lam / e-åi4-il-tum 76. is-ba-su2-ma 77. GEME2-su2 sa DUMU.ME† 77. ul-du-sum 78. a-na KU3.BABBAR it-ta-din col. iv 1. KU3.BABBAR DAM.GAR3 is-qu2-lu 2. be-el GEME2 i-sa-qal-ma 3. GEME2-su2 i-pa-tar2
91
Although various diagnostics for weak quanti˜cation in Akkadian could be adduced at this point (see the brief discussion of the absolute state with cardinal numbers at the conclusion of this paper), I would like to limit myself to a small set of relative clauses found in the Code of Hammurapi rather than dealing with the Old Akkadian material.5 In doing so, I hope to clarify the end of the historical process that begins with the IHRC in Old Akkadian and lay the groundwork for a future study of the Old Akkadian material. In the Code of Hammurapi, there are a dozen or so relative clauses that are headed by a noun in the construct state (Ravn 1941: 36–40). Take, for example, the following law, dealing with the sale of a female slave: CH §119, rev., col. 3, line 74 through col. 4, line 4 (Ravn 1941: 38).
If a claim has seized a man, and He sells a female slave of his who has given birth to children for him, The silver that the trader paid, The owner of the female slave will (re)pay and he will set his female slave free
Note in particular the construct relative clause in the ˜rst line of column 4 in (18). (19) kasap tamkar-um isqul-u silver.Const trade.agent-Nom he.weigh.Perf-Rel ‘the (amount of) silver that the trade-agent paid (for the female slave)’ 5
This statute identi˜es a particular amount of silver in any particular transaction that meets the other requirements of the statute and states that precisely the same amount of silver that the slave-owner received is to be paid back to the purchaser so as to free the mother of the slaveowner’s children. The use of the construct relative
5. Rebecca Hasselbach’s recent study of Sargonic Akkadian is not yet available to me, so any eˆort on my part to deal with the Old Akkadian materials in detail would, undoubtedly, be premature.
in this circumstance is particularly interesting because the statute does not state a particular cardinal amount like “three shekels of silver,” but rather the amount of silver that changed hands in any particular event necessarily amounts to a cardinal number and that same cardinal number is the amount that the slave-owner is to pay. The amount that the slave-owner pays to redeem his slave cannot be more than or less than the precise amount that he was originally paid, thus any kind of strong quanti˜cational reading is excluded. This was presumably meant to protect the former
92
J. CALE JOHNSON
owner of the slave from extortion on the part of the purchaser: if the statute had simply stipulated that the seller must repurchase the female slave and then set her free, it would undoubtedly have led to some form of extortion on the part of the buyer, who could presumably demand any price for the female slave. The use of the construct relative determines a particular amount of silver without specifying, for example, that the same particular pieces of silver that the owner received in the ˜rst place must be used to redeem the slave, (20) CH §232, rev., xix 82–92 82–83. sum-ma NIG2.GUR11/ uh2-ta-al-li-iq 84–85. mi-im-ma / sa u2-hal-li-qu2 86–87. i-ri-ab / u3 as-sum E2 i-pu-su 88–89. la u2-dan-ni-nu-ma / im-qu2-tu 90–91. i-na NIG2.GUR11/ ra-ma-ni-su 92. E2 im-qu2-tu i-ip-pe2-es
In lines 87–89 in (20), a construct relative (bÿt ÿpusu) serves as the direct object of two conjoined verbs: la udannin¿ma imqutu. The verbs that are conjoined by *-ma are in the subjunctive since they are governed by assum, but clearly this is not the case for bÿt ÿpusu; some other motivation is needed for the occurrence of a construct relative in this passage. As in the previous example, I would like to suggest that the house that the building contractor is required to build as a replacement for the poorly built one that has collapsed must be equivalent to, but not identical with, the one that collapsed. Without delving too far into the semantics of identity, I think it can be said that if the replacement were truly the “same” as the house that collapsed, then the new house would collapse just as easily as the old house had. What is clearly at stake in the statute is the number of square feet, or in the Mesopotamian idiom, the number of sar that the house should cover. In this respect, the use of the construct relative to code “identity of quantity” rather than “identity of substance” is perfectly reasonable and analogous to the use of the construct relative in (18) above.
which would presumably be the meaning of the relative clause if it made use of the determinativerelative pronoun. In the terminology used in Grosu and Landman, the construct relative in (19) requires “identity of quantity” but not “identity of substance” (Grosu and Landman 1998: 132). The other example that I would like to discuss in this section is also among the most complicated in the Code: CH §232, rev., col. 19, lines 82–92 in (20) below.
If (a builder) destroyed property, Whatever he destroyed, He will replace, and because he did not strengthen a house that he built, and it collapsed, Out of his own property, He will build a house (equivalent to) the one that fell. The syntactic phenomenon that is of particular interest in (20), however, is the way in which reference is made to the house that has collapsed later on in the statute. Elsewhere in the Code, when a nominal phrase that has occurred earlier in the protasis is repeated in the apodosis so as to anaphorically refer back to the same individual, it is typically repeated with the addition of the appropriate form of the distal demonstrative s¿ immediately after it. This does not, however, seem to be that case with the construct relative in line 92. (21) summa [awÿlum] . . . , [awÿlum s¿] iddâk ‘If a man . . . , that man will be killed’ (22) assum . . . [(bÿt) ÿpusu] . . . imqutu, [bÿt imqutu] ippes ‘Because . . . the house collapsed, he will build (a house equivalent to) the house that collapsed’ The apparent impossibility of using a demonstrative pronoun for anaphoric reference in the apodosis in this example strongly suggests that bÿt imqutu is non-speci˜c. One of the standard diagnostics for speci˜c noun phrases is the possibility
INTERNALLY-HEADED RELATIVE CLAUSES IN AKKADIAN
of referring back to a speci˜c, inde˜nite noun with an anaphoric element such as a demonstrative as in (21): the impossibility of doing so in (22) provides a substantial piece of evidence that relatives that use a lexical noun in the construct state to indicate the head of the relative clause are nonspeci˜c. The Diˆerence a Strong Quanti˜er Makes in Old Babylonian The last piece of evidence that I would like to oˆer in support of an interpretation of the construct state in Old Babylonian as a form of weak quanti˜cation is the contrastive distributional be-
93
havior of certain strong quanti˜ers in Old Babylonian. Like weak quanti˜cation above, there are few if any previous analyses of strong quanti˜cation in Old Babylonian. On the basis of both standard translational eqivalencies and, as we will see in a moment, diˆerences in morphosyntactic distribution, two lexemes that we might expect to exhibit properties associated with strong quanti˜cation are kalu “all” and kilallan “both.” In semantic terms, these two lexemes are relative close to the prototypical examples of strong quanti˜ers in English such as “each” or “every,” and they also make use of a distinctive morphosyntactic pattern as in (23) below.6
(23) Resumptive strong quanti˜cation in Old Babylonian (ARM 1, 76, obv., line 17: [DUMU].ME† ka-lu-su-nu su-ma-am i-su-u2) mar¿ kalu-sunu sum-am isû son.Nom.Pl all-3Pl.Poss name-Acc they.have ‘All the sons are famous, lit. have a name’ Huehnergard mentions this construction and refers to kalu as a quanti˜er (Huehnergard 2000: 92), but I do not know of any other discussion of the construction in the secondary literature. CAD (sub kalu) lists primarily Old Assyrian examples with a limited number of examples from Mari, which I take as examples of Old Babylonian practice, but further clari˜cation of the dialectical situation is needed. The syntax of the construction seems to be fairly clear: kalu and the noun that it quanti˜es over are not in a bound relationship such as the construct state but rather in an appositional structure. The diˆerence, however, between the standard appositional construction and the construction in (23) is that there is also a kind of resumptive pronoun in the form of a possessive pronoun that agrees with the quanti˜ed noun in person, number and gender and that possesses the quanti˜er kalu. 6 The construction in (23) can, perhaps, be associated with strong quanti˜cation in the Old Babylonian period and, presumably, earlier in the history of Akkadian. In simple, distributional terms, it should be fairly clear that the strong quanti˜cational construction in (23) is a kind of
analytic genitive construction that avoids the use of the construct state. If the construct state codes weak quanti˜cation at some point in the history of Akkadian morphosyntax as detailed above, then it would make a certain amount of sense for strong quanti˜ers to be in complementary distribution with the construct state as well. Since most of the examples in the CAD of kalu in the construct state—where it is actually functioning as a strong quanti˜er—derive from ˜rst millennium sources and the ˜rst millennium also witnesses the development of the de˜nite article in the languages that surrounded and in˜ltrated previously Akkadianspeaking regions, a certain amount of contamination from the quanti˜cational systems of other Semitic languages such as Aramaic would not be surprising in the later texts. 6. Mark Baker has noted the problem of using rough equivalency of lexical meaning as a way of identifying strong quanti˜ers in Mohawk (Baker 1995), and it remains unclear whether or not kalu would qualify as a strong quanti˜er under the narrower de˜nition adopted by Baker. The contrast between kalu and mala, “as much as there is,” may be informative in this regard, but I have not yet been able to ascertain which of the two is proportional and, consequently, the strong quanti˜er of the two.
94
J. CALE JOHNSON
(24) Externally headed relative clause in Korean (Kim 2004: 39) John-un [[ei tomangka-n]-un sey-myeng-uy totwuki]-ul capassta PN-Top [[ei run.away-Impf ]-Rel three-Cl-Gen thief ]-Acc caught ‘John caught three thieves (out of many more) who were running away’ (25) Internally headed relative clause in Korean (Kim 2004: 39) John-un [[sey-myeng-uy totwuk-i tomangka-n]-un kes]-ul capassta PN-Top [[three-Cl-Gen thief-Nom run.away-Impf ]-Rel Thing]-Acc caught ‘(Only) three thieves were running away and John caught all of them’ Partitives, Speci˜city, and the Construct State It may be helpful at this point to return to the contrast between an EHRC and an IHRC in Korean: examples (16) and (17) above, repeated here as (24) and (25). In both (24) and (25), note that the head of the relative clause is itself a genitive construction, seymyeng-uy totwuk. (26)
sey-myeng-uy totwuk three-Human-Gen thief
(27a) ‘three of the thieves’ (in an EHRC as in [24]) (27b) ‘the three thieves’ (in an IHRC as in [25]) In the absence of any other context, this expression means simply “three thieves,” but as noted above, it takes on two very diˆerent interpretations depending on whether it is the external or internal head of a relative clause: as an external head, it has a partitive meaning as in (27a), “three of the thieves,” but as an internal head, it can only refer to an indivisible group of three thieves as approximated by the translation in (27b): “the three thieves.” The role of partitivity in diˆerentiating strong and weak quanti˜ers has a rather involved history extending back to Milsark (1974) that I will pass over here, but one particularly clear example of the opposition between the two interpretations in (27) and its relation to speci˜city has emerged in Enç’s study of partitives in Turkish (Enç 1991). Enç points out that, unlike inde˜nite nouns in English which can be either speci˜c or nonspeci˜c, Turkish draws a regular contrast between speci˜c and non-speci˜c noun phrases in certain morphosyntactic positions. Inde˜nites in the direct
object position, for example, are always unambiguously speci˜c or non-speci˜c in Turkish as a function of case-marking: If the [noun phrase] bears the accusative case morpheme -(y)i [or one of its allomorphs], it is obligatorily interpreted as speci˜c . . . . If the [noun phrase] does not carry case morphology, it is obligatorily interpreted as nonspeci˜c (Enç 1991: 4).
This is roughly the same mechanism that, as I argued above, exists in Old Babylonian. As an example of this opposition, Enç oˆers a pair of contrasting examples (Enç 1991: 4–5, ex. 12 and 13). (28) Ali bir piyano-yu kiralamak istiyor PN one piano-Acc to.rent wants ‘Ali wants to rent a certain piano’ (29) Ali bir piyano kiralamak istiyor PN one piano to.rent wants ‘Ali wants to rent a (non-speci˜c) piano’ Where the Turkish data takes on a far greater interest, however, is when Enç constrains the discourse environment in which speci˜c and nonspeci˜c noun phrases are used. Enç stipulates a discursive interaction in which the ˜rst utterance is (30) below, while the second utterance is one of the pair of alternatives in (31) and (32). (30) odam-a birkaç çocuk girdi my.room-Dat several children entered ‘several children entered my room’ (31) iki kız-ı tanıyordum two girl-Acc I.knew ‘I knew two girls’ (32) iki kız tanıyordum two girl I.knew ‘I knew two girls’
INTERNALLY-HEADED RELATIVE CLAUSES IN AKKADIAN
Whereas in (31), the example in which “two girls” bears the accusative case and is consequently speci˜c, iki kız-ı can be used to refer to two members of the group of children in (30), the example in (32), where the expression iki kız is nonspeci˜c, cannot be used to refer to two members of the group. This opposition obviously bears a great deal of similarity to the discourse behavior of the two types of Korean relative clause repeated above in (24) and (25). Moreover, only the example involving a speci˜c noun phrase in (31) can be paraphrased with a partitive construction in Turkish. (33) kız-lar-dan iki-sin-i tanıyordum girl-Pl-Abl two-Agr-Acc I.knew ‘I knew two of the girls’ Given the similarities between the IHRC in Korean and the absence of case-marking as an in-
95
dication of non-speci˜city in Turkish, the obvious question is: in what ways can the construct relative IHRCs in Old Babylonian be seen as analogous to the constructions in Korean and Turkish. What was labelled as “identity of quantity” above in my discussion of the construct relative in Old Babylonian is quite similar to Enç’s discourse analytical constraints on the coreference of non-speci˜c nouns in Turkish: when a noun phrase in Turkish is nonspeci˜c, it can share an “identity of quantity” with a previous noun phrase, but not an “identity of substance,” just as construct relatives in Old Babylonian can share an “identity of quantity,” but not an “identity of substance” with an expression earlier in the discourse. Korean does not exhibit the same kind of interclausal coreference restriction, but only an EHRC, which is presumably speci˜c, can be further modi˜ed by a partitive expression as exempli˜ed in (34), whereas an IHRC such as in (25) cannot form a partitive construction.
(34) Externally headed relative clause in partitive construction (Korean)7 John-un [[ei tomangka-n]-un sey-myeng-uy totwuki]-dul joong doo-myeng-ul capassta PN-Top [[ei run.away-Impf ]-Rel three-Cl-Gen thief ]-Pl among 2-Human-Acc caught ‘John caught two of the three thieves who were running away’ 7
As demonstrated above for Old Babylonian as well, a non-speci˜c noun phrase cannot refer to a speci˜c noun phrase if the “identity of substance” interpretation is the one that is intended. Where the prior noun phrase is a non-speci˜c IHRC in Old Babylonian, a following noun phrase that refers to it must, likewise, be a non-speci˜c IHRC as exempli˜ed in (21) and (22), repeated below as (35) and (36). (35) summa [awÿlum] . . . , [awÿlum s¿] iddâk ‘If a man . . . , that man will be killed’ (36) assum . . . [(bÿt) ÿpusu] . . . imqutu, [bÿt imqutu] ippes ‘Because . . . the house collapsed, he will build (a house equivalent to) the house that collapsed’
Thus, as noted above, the use of an IHRC in the form of an Old Babylonian construct relative, necessitates the use of another IHRC in the second half of (36) to “refer” or perhaps better to “equate” the newly rebuilt house that is required of the negligent builder to the ˘awed house that had previously collapsed. If an “identity of substance” had been found necessary—if, for example, a verb that did not involve creation or building such as nadanu(m) ‘to give’ had been used instead—then we might reasonably expect that an externally headed relative using the determinative-relative pronoun would have been used so as to allow for a subsequent act of reference to be achieved using an anaphoric device such as the demonstrative pronoun in (35). Implications
7. Thanks to Grace Park for the modi˜ed example in (34).
The role of cardinality in weak quanti˜cational environments in general and the construct state
96
J. CALE JOHNSON
IHRCs in Old Babylonian in particular cannot be overestimated. One of the hallmarks of the research tradition that has studied weak quanti˜cation, extending from Milsark’s foundational work (1974) down to several recent treatments (Hallman 2004; Chung and Ladusaw 2004), has been an association between phrases headed by a cardinal number (the archtype in some sense of other weak quanti˜ers such as the inde˜nite determiner in English and an analogous use of “one” as a kind of inde˜nite determiner in a variety of Semitic languages) and a variety of morphosyntactic phenomena such as the de˜niteness eˆect in existential sentences (in English) and IHRCs (in Old Babylonian and Korean). Likewise, within the constraints of Old Babylonian morphosyntax, one also ˜nds a rough similarity between the morphological form of cardinal numbers, generally known as the absolute state and the morphosyntactic device that I argue codes weak quanti˜cation in Old Babylonian, namely the construct state. These two morphosyntactic devices cannot be equated: certain exponents of grammatical number as well as indicators of grammatical gender to the degree that they are coded through portmanteau morphemes are retained in the construct state but lost in the absolute state.8 But, as in the Turkish example dealt with in the preceding section, they share one crucial feature, namely the absence of nominal case-marking and the speci˜city that can presumably be associated with such case-marking. Given the extremely small amount of material that the proposal described herein is based on, further corroboration is needed both within the known corpora of Old Babylonian letters and the Old Akkadian materials when they become available. What is of more importance, perhaps, than the particular empirical components of this proposal is my eˆort to make sense of relativization in Akkadian through appeal to ongoing work
8. There are additional phonological constraints aˆecting the preservation of certain vowels in the construct state—including residual case marking in a few instances—that I pass over here.
in syntactic theory rather than typological abnormality. The model that I have sketched out in this paper provides an empirically testable scenario, namely that (1) several factors would seem to suggest that the construct state exhibits weak quanti˜cation and that weak quanti˜cation is used to indicate the head of an IHRC in the form of the construct relative in the Old Babylonian dialect of Akkadian, (2) if the construct state in such IHRCs is a weak quanti˜cation environment, then the only way in which a demonstrative element such as the determinative-relative pronoun could occur in such an environment would be if the underlying head of the relative were in focus as if it were a kind of cleft sentence, and (3) a focus construction resulting in a discontinuous nominal phrase would also result in a full copy of all the nominal features (including case) in the form of the determinative-relative pronoun that acts as a resumptive pronoun when the noun in focus is moved out of the relative clause. This scenario would yield the kind of relative regularly found in Old Akkadian, in which the external head of the relative and the determinative-relative pronoun within the relative clause itself exhibit the same case-marking. In the Old Babylonian dialect, however, the resulting form of such a relative clause could easily have been reinterpreted as a kind of externally headed relative with the corresponding loss of case-marking distinctions on the determinative-relative pronoun, while the original construct relative should still retain the features that we might expect to ˜nd in an IHRC. In this paper, I have oˆered several examples of construct relatives in the Code of Hammurapi that seem to exhibit just such a feature, namely weak quanti˜cation of the noun that occurs in the construct state. References Aboh, Enoch O. 2004 “Topic and Focus within D,” Linguistics in the Netherlands 21: 1–12. Baker, Mark C. 1995 On the Absence of Certain Quanti˜ers in Mohawk. Pp. 21–58 in Quanti˜cation in Natural Languages, ed. E. Bach, E. Jelinek,
INTERNALLY-HEADED RELATIVE CLAUSES IN AKKADIAN
A. Kratzer, and B. Partee, ed. Dordrecht: Kluwer Academic Publishers. Basilico, David 1996 Head Position and Internally Headed Relatives Clauses. Language 72: 498–532. 2003 The Topic of Small Clauses. Linguistic Inquiry 34: 1–35. Chung, Sandra, and William A. Ladusaw 2004 Restriction and Saturation. Linguistic Inquiry Monographs 42. Cambridge: MIT Press. Chung, Sandra, and James McClosky 2002 Existentials at the Interface—NSF Project No. BCS-0131767. http://ohlone.ucsc.edu/~jim/ PDFFiles/nsf.pdf. Cole, Peter 1987 The Structure of Internally Headed Relative Clauses. Natural Language and Linguistic Theory 5: 277–302. Deutscher, Guy 2001 The Rise and Fall of a Rogue Relative Construction. Studies in Language 25: 405–22. 2002 The Akkadian Relative Clauses in Crosslinguistic Perspective. ZA 92: 86–105. Diesing, Molly 1992 Inde˜nities. Linguistic Inquiry Monographs 20. Cambridge: MIT Press. Donnellan, Keith 1966 Reference and De˜nite Descriptions. Philosophical Review 75: 281–304. Ehrenkranz, Jean F., and Edward C. Hirschland 1972 Latin Relative Clauses. Pp. 23–29 in The Chicago Which Hunt: Papers from the Relative Clause Festival, April 13, 1972, ed. P. Peranteau, J. Levi and G. Phares. Chicago: Chicago Linguistics Society. Enç, Mürvet 1991 The Semantics of Speci˜city. Linguistic Inquiry 22: 1–25. Givón, Talmy 1990 Syntax: A Functional-typological Introduction. 2 vols. Amsterdam: John Benjamins. Grosu, Alexander, and Fred Landman 1998 Strange Relatives of a Third Kind. Natural Language Semantics 6: 125–70. Hallman, Peter 2004 NP-interpretation and the Structure of Predicates. Language 80: 707–75. Huehnergard, John 2000 A Grammar of Akkadian, revised edition. Harvard Semitic Studies 45. Atlanta: Scholars Press. Ihsane, Tabea, and Genoveva Puskás 2001 Speci˜c is not De˜nite. Generative Grammar in Geneva 2: 39–45.
97
Ioup, Georgette 1977 Speci˜city and the Interpretation of Quanti˜ers. Linguistics and Philosophy 1: 233–45. Kayne, Richard 1994 The Antisymmetry of Syntax. Cambridge: MIT Press. Kienast, Burkhart, and Konrad Volk 1995 Die sumerischen und akkadischen Briefe des III. Jahrtausends aus der Zeit vor der III. Dynastie von Ur. Freiburger altorientalische Studien 19. Stuttgart: Franz Steiner. Kim, Min-Joo 2004 Event-structure and the Internally-headed Relative Clause Construction in Korean and Japanese. PhD dissertation, University of Massachusetts Amherst. Lehmann, Christian 1984 Der Relativsatz: Typologie seiner Strukturten, Theorie seiner Funktionen, Kompendium seiner Grammatik. Tübingen: Narr. Michalowski, Piotr 1993 Letters from Early Mesopotamia. SBL Writings from the Ancient World 3. Atlanta, GA: Scholars Press. Milsark, Gary 1974 Existential Sentences in English. PhD dissertation, MIT. Munro, Pamela 1976 Mojave Syntax. New York: Garland. Nichols, J. 1984 Another Typology of Relatives. Proceedings of the Annual Meeting of the Berkeley Linguistics Society 10: 524–41. Nishigauchi, Taisuke 2003 Head-Internal Relative Clauses in Japanese and the Interpretation of Inde˜nite NPs. Manuscript, Dobe Shoin Graduate School. Ntelitheos, Dimitrios 2004 Syntax of Elliptical and Discontinuous Nominals. MA thesis, UCLA. Ouhalla, Jamal 2004 Semitic Relatives. Linguistic Inquiry 35: 288–300. Pesetsky, David 1987 Wh-in-situ: Movement and Unselective Binding. Pp. 98–129 in The Representation of Inde˜niteness, ed. E. Reuland and A. ter Meulen. Cambridge: MIT Press. Ravn, O. E. 1941 The So-called Relative Clause in Accadian, or The Accadian Particle sa. Copenhagen: NYT Nordisk Forlag. Schachter, Paul 1973 Focus and Relativization. Language 49: 19–46.
98
J. CALE JOHNSON
Simpson, Andrew, and Zoe Wu 2002 Agreement, Shells, and Focus. Language 78: 287–313. Shimoyama, Junko 1999 Internally Headed Relative Clauses in Japanese and E-type Anaphora. Journal of East Asian Linguistics 8: 147–82. Szabolcsi, Anna 1994 The Noun Phrase. Pp. 179–274 in The Syntactic Structure of Hungarian, ed. F. Kiefer and K. Kiss. Syntax and Semantics 27. New York: Academic Press.
Tellier, C. 1989 Head-Internal Relatives and Parasitic Gaps in Mooré. Current Approaches to African Linguistics 6: 298–318. Watanabe, Akira 1992 Subjacency and S-structure Movement of Wh-in-situ. Journal of East Asian Linguistics 1: 255–91. Williamson, Janis 1987 An Inde˜niteness Restriction on Relative Clauses in Lakhota. Pp. 168–233 in The Representation of Inde˜niteness, ed. E. Reuland and A. ter Meulen. Cambridge: MIT Press.
Esarhaddon, Egypt, and Shubria: Politics and Propaganda Israel Ephçal (The Hebrew University of Jerusalem)
In the winter of 673 B.C., at the end of Esarhaddon’s seventh year, the Assyrian army set out to conquer Egypt. The campaign ended in what was probably one of Assyria’s worst defeats.1 A
few months later, Esarhaddon went to war against Shubria, a small kingdom at the foot of the Taurus Mountains, east of the upper Tigris and west of Lake Van.2 According to Esarhaddon’s “Letter to the God” the reason for the attack on Shubria was that its ruler had oˆered asylum to Assyrian political fugitives and had refused to hand them over
The contents of this paper were delivered at a symposium held on November 2003 at the Israel Academy of Sciences and Humanities on the occasion of the eightieth birthday of Prof. Hayim Tadmor. References to Esarhaddon’s inscriptions follow R. Borger, Die Inschriften Asarhaddons, Königs von Assyrien, AfO Beih. 9 (Graz, 1956); to those of Ashurbanipal, M. Streck, Assurbanipal und die letzten assyrischen Könige bis zum Untergange Ninivehs (Leipzig: J. C. Heinrichs, 1916); and to the chronicles, A. K. Grayson, Assyrian and Babylonian Chronicles (Locust Valley: J. J. Augustin, 1975), unless otherwise speci˜ed. 1. The only certain evidence of the failure of the ˜rst Egyptian campaign appears in the Babylonian Chronicle, which reads: “Year 7 (of Esarhaddon), Adar 5. Assyrian army was defeated in Egypt” (Chron. 1 iv 16). In The Esarhaddon Chronicle for that year we read: “Year 7, Adar 8. Assyrian army [. . .] to uru†á-amîlê mes” (Chron. 14:20). The attempt to harmonize the two texts (for the view that uru†á-amîlê mes is simply a derivative of Silê see G. Fecht, “Zu den Namen ägyptischer Fürsten und Städte in den Annalen des Assurbanipal und der Chronik des Asarhaddon,” Mitteilungen des deutschen archäologischen Instituts, Abteilung Kairo, 16 [1958] 116–19) encounters epigraphic and phonetic problems (see A. Spalinger, “Esarhaddon and Egypt: An Analysis of the First Invasion of Egypt,” OrNS 43 [1974] 300–301). Interestingly, the Esarhaddon Chronicle does not mention any defeat of the Assyrian army in Egypt. Another approach associates the destination of the Assyrian campaign, according to the Esarhaddon Chronicle, with the city of uru†á-amîlê mes in the region of Bit-Amukani in southern Babylonia (M. Dietrich, Die Aramäer Südbabyloniens in der Sargonidenzeit (700–648), AOAT 7 (Neukirchen-Vluyn: Butzon and Bercker Kevelaer, 1970), 19, 39, 56 n. 1). This city was the seat of the governor of Bit-Amukani under Esarhaddon. This view requires the Assyrian army to have been active at one and the same time in two places some one thousand kilometers apart, but this is also problematic. The diˆerence
between the texts should be evaluated on the basis of a comprehensive examination of the textual relationship between the two chronicles, and the results are far from conclusive; see J. A. Brinkman, “The Babylonian Chronicle Revisited,” in Lingering over Words: Studies in Ancient Near Eastern Literature in Honor of W. L. Moran, eds. T. Abusch, J. Huehnergard and P. Steinkeller (Atlanta: Scholars Press, 1990), esp. 91–94. The ˜rst Egyptian campaign was not the only failed venture of Esarhaddon’s army. The chronicles and the queries to the Sun God document a campaign that took place in Esarhaddon’s sixth year (675/4 B.C.) against Mugallu, ruler of Melid (see Chron. 1 iv 10; Chron. 14:15; SAA IV 3–5, 9. The chronicles record only the beginning of the campaign but not its outcome). Judging from the silence of the royal inscriptions about this campaign and the reference to Mugallu in a letter to Esarhaddon dated 671 B.C. (see SAA X 351), the campaign does not appear to have been successful. Nevertheless, the Melid campaign may have been limited in scope (the letters imply that it was commanded by the chief eunuch), so that its failure left less of an impression; whereas the failure of the Egyptian campaign could not have been insigni˜cant, given the geographical conditions involved. As noted, we have non-literary documentation of the Melid campaign, but as of the present, no non-literary documentation of Esarhaddon’s ˜rst Egyptian campaign has come to light. 2. On Shubria as a border state on the northern boundary of the Assyrian Empire and its history in the ninth to seventh centuries B.C., see B. J. Parker, The Mechanics of Empire: The Northern Frontier of Assyria as a Case Study in Imperial Dynamics (Helsinki: Helsinki University Press, 2001), 230–46.
99
JCS 57 (2005)
100
ISRAEL EPHçAL
to the Assyrian king despite repeated demands.3 The “Letter to the God” reports that when the king of Shubria realized that Esarhaddon, not content with diplomatic action, was marching on his country with his army, he rapidly declared his total and immediate surrender. However, Esarhaddon rejected these touching gestures and cruelly devastated Shubria:4 he laid siege to the capital Uppume and occupied it, deported many of the kingdom’s people, divided it into provinces, and changed the names of many cities. Shubria was conquered in the month of Tebeth (or Adar, see below), in year 8 of Esarhaddon’s reign (672 B.C.). Fourteen to sixteen months later, in Nisan of Esarhaddon’s tenth regnal year, he launched a new attack on Egypt.5 This time he 3. Gbr. II i 16, 21–22, 29–30. According to Gbr. I 3, the fugitives in question were highly placed members of Assyrian governing circles who were opposed to Esarhaddon, not mere deserters ˘eeing to Shubria to escape corvée labor or military service (for the latter see Parker, Mechanics 233, 242–43). Leichty, who identi˜es these fugitives as Esarhaddon’s rivals—his brothers, headed by Ardi-mulissi—with whom he had been engaged in a struggle since his accession to the throne, suggests that the brothers and their supporters ˘ed to Urartu in 680 B.C. (after Esarhaddon had prevailed over his opponents and ascended the throne); they were later forced to escape from there to Shubria (together with their supporters from Urartu), for fear that the authorities of Urartu would hand them over to Esarhaddon. This also explains the need to write the gods a detailed account of the Shubrian campaign and of Esarhaddon’s activities, about two months after the end of the campaign, in order to arrange his two sons’ inheritance of the kingdom (E. Leichty, “Esarhaddon’s ‘Letter to the Gods’,” in Ah, Assyria . . . : Studies in Assyrian History and Ancient Near Eastern Historiography Presented to Hayim Tadmor, eds. M. Cogan and I. Ephçal, (Jerusalem: Magnes Press, 1991), 52–57. 4. As Esarhaddon wrote the king of Shubria: “When did you hear a mighty king command twice? (Whereas) I, an almighty king, have addressed you the third time and you have not listened to the words of my lips . . . you were not afraid and ignored my letter” (Gbr. II i 29–31). This passage implies that before Esarhaddon marched on Shubria (and before his ˜rst Egyptian campaign) the two kings negotiated at length, but the king of Shubria refused to hand over the political fugitives or to expel them. Now, after the failure of the Egyptian campaign, Esarhaddon needed a military achievement, and he therefore rejected the Shubrian king’s admission of guilt and his declaration that he would accept whatever burden Esarhaddon imposed upon him. The time for negotiating was over; it was time for a violent punitive expedition. 5. Taken into account in this time calculation is the fact that Esarhaddon’s eighth regnal year was a leap year, with an additional month of Second Adar.
successfully defeated the army of Taharqa (of the 25th, Nubian, Dynasty), occupied the Nile Delta, and captured the capital Memphis. This military victory marked the beginning of Assyrian rule in Egypt, which lasted some twenty years. Esarhaddon’s inscriptions report that, after the initial occupation, he sacked the country, deported many inhabitants (among them the members of the royal family, including the crown prince Ushanahuru, and specialists in valued professions, for example, physicians, diviners, dream interpreters, snakecharmers, and various metal-workers), appointed governors and senior o¯cials, imposed annual taxes and changed city names. In this article, I propose a political interpretation of the accounts of Esarhaddon’s second Egyptian campaign and his war against Shubria, with special attention to the accompanying propaganda and literary activity.6 The operative course of these wars will not be discussed, nor will Esarhaddon’s policy toward Egypt after its occupation.7 Any discussion of Esarhaddon’s Egyptian campaigns must take into consideration that, before the advent of the internal-combustion engine as an aid to moving armies and supplying water in the Middle East, military movements in the Sinai Desert were limited because of the need for a 6. To the relief of Esarhaddon’s scribes, the decision (taken at the beginning of his reign) not to list military campaigns in his historical inscriptions by ordinal number or by the regnal year (palû) in which they were undertaken—as was the practice of previous Assyrian kings—meant that they did not have to cope with the problem of reporting the failed campaign and could simply ignore it. Compare, e.g., how Sennacherib’s scribes dealt with his failure to capture Jerusalem, H. Tadmor, “Sennacherib’s Campaign to Judah: Historical and Historiographical Considerations,” Zion 50 (1985) 74–78 (Hebrew); or how Sargon’s scribes tried to create a far from realistic picture of consecutive victories, year after year, H. Tadmor, “The Campaigns of Sargon II: A Chronological-historical Study,” JCS 12 (1958) 22–40, 77–100. An exception to this rule is the New Account Edition, composed after the conquest of Egypt in Esarhaddon’s second Egyptian campaign, which, unlike Esarhaddon’s other inscriptions, refers to military campaigns by number. In this edition we ˜nd references to “the second campaign” (against Sidon; see Nin. D = Tadmor, S 2, and the observations of Borger, 38, 49); and to “the tenth campaign” (against Egypt; see Frt. F obv. 6). 7. For Esarhaddon’s military and administrative measures during his Egyptian war and their description in his inscriptions see Spalinger, OrNS 43 (1974) 295–326.
ESARHADDON, EGYPT, AND SHUBRIA: POLITICS AND PROPAGANDA
regular water supply. The di¯culties of crossing the desert were particularly acute when the ends of the east–west roads in the desert were controlled by rival political or military entities. An army that crossed the Sinai Desert in order to ˜ght a war upon reaching inhabited land on the other side was thus taking a considerable risk. If the battle was lost, the attacker could not even minimize his defeat by breaking oˆ contact and retreating. With the desert at his rear, he would be annihilated.8 This was presumably the fate of Esarhaddon’s ˜rst Egyptian campaign. Some idea of the possible military and political outcomes of such a failure may be obtained by studying Nebuchadnezzar’s defeat in his Egyptian campaign of 601 B.C. After this debacle, Nebuchadnezzar spent two full years or more in the eˆort to rebuild his army, and serious political disturbances shook the Babylonian Empire for many years.9 The collapse of Esarhaddon’s ˜rst Egyptian campaign could have had similar results. As we shall see, however, Esarhaddon, unlike Nebuchadnezzar, reacted quickly and ˜rmly, and was thus able to recover and move on to conquer Lower Egypt only two and a half years after his defeat. First, however, Esarhaddon had to demonstrate that the failure of the ˜rst campaign at the entrance to Egypt did not mean that Assyria’s power had been spent. Esarhaddon’s energetic and merciless treatment of Shubria may well have been intended to remind everyone—both external enemies and, in particular, his own subjects and potential rivals in the Empire10—that 8. For the logistical and strategical signi˜cance of the Sinai Desert as a barrier between the Fertile Crescent and Egypt up until World War I see I. Ephçal, The Ancient Arabs (Jerusalem and Leiden: Magness Press and Brill, 1982), 137–42; E. Galili, The Battle of Raphia 217 B.C.E.: Tactics, Strategy and Logistic in the Hellenistic World ( Jerusalem: Bialik Institute, 1999), 194–212 (Hebrew). 9. See Ephçal, “Nebuchadnezzar the Warrior: Remarks on his Military Achievements,” IEJ 53 (2003) 180–83. 10. In Oppenheim’s view, the real addressees of the “Letter to the God” were the residents of the historical city of Ashur; while the city enjoyed royal privileges, opposition to Esarhaddon was particularly strong there. Accordingly, the extended account of the harsh treatment meted out to the king of Shubria was meant to demonstrate to the hostile elements in
101
Assyria was still a power to be reckoned with, and that severe punishment awaited anyone who underestimated her ruler and rebelled.11 The crushing of Shubria, only a small kingdom, was not an exceptional challenge for Assyria’s military forces, even after their defeat in Egypt. Indeed, the Assyrian army marched on Shubria without delay, after a rapid reorganization. Only about ten to twelve months elapsed from the defeat in Egypt to the conquest of Shubria. The Assyrian documents at our disposal of the type known as “Letters to the God” are very few in number. Hence, it would seem that such documents were not written after every military campaign, but only rarely, to mark military events that were considered particularly important. Such is the case with Sargon’s “Letter to Assur,” written after his defeat of the army of Urartu, a powerful rival. Shubria, as already intimated, could not be considered a rival of the same order of magnitude. More probably, just as the conquest of Shubria was a political reaction to Esarhaddon’s military debacle in Egypt, his “Letter to the God” was a suitable reaction to that same defeat in the form of literary propaganda. 12 This literary work has
Ashur that Esarhaddon was determined to react ˜rmly to any violation of the oath of loyalty to him (A. L. Oppenheim, “NeoAssyrian and Neo-Babylonian Empires,” in Propaganda and Communication in World History, Vol. I: The Symbolic Instrument in Early Times, eds. H. D. Lasswell, D. Lerner and H. Speier [Honolulu: University Press of Hawaii, 1979], esp. 125–37). 11. In a query to the Sun God (SAA IV 18), Esarhaddon asked whether Rusa II, king of Urartu, would attack Shubria with his Cimmerian allies and annex the forts on its border to the area under their control (while the date of the query is not known, it clearly preceded Esarhaddon’s occupation of Shubria, which eliminated Shubria as a political entity and split it into two provinces). Possibly, Esarhaddon, by conquering Shubria, anticipated the action of the king of Urartu and thereby warned him not to take advantage of Assyria’s weakness after its Egyptian debacle. 12. Our reference to this source as a “Letter to the God” accords with the conventional scholarly attitudes to Esarhaddon’s inscriptions. For criticism of its de˜nition as a “Letter to the God” see L. D. Levine, “Observations on ‘Sargon’s Letter to the Gods’,” Eretz-Israel 27 (2003) 117, note 6. This criticism has implications as to the historiographic purpose of the work, with which I am inclined to agree.
102
ISRAEL EPHçAL
several unique literary features. In its style and the detailed enumeration of actions described in the text, it was an in˘uential model for the account of the second Egyptian campaign. The texts of the “Letter to the God” and the account of the actions against Shubria and Egypt are similar in their descriptions of how the god Marduk saved the Assyrian army at critical junctures of its activity, in the detailed lists of deportees grouped by profession, in the division of the occupied country into subdistricts, and in the changing of city names. The two compositions were written within the span of at most two years. The conquest of Egypt was a tremendous military, economic and political achievement that also facilitated royal propaganda. After the victory, Esarhaddon took the title of “King of kings of Egypt, Pathros and Cush”13 (in a variant of inscription Klch. D, he is referred to as “King of Egypt, defeater of the king of Meluhha”14). This
Despite the considerable propaganda and political signi˜cance ascribed at the time to the conquest of Shubria, as intimated in the “Letter to the God,” the event is not mentioned in Esarhaddon’s monuments discovered in the western part of the Empire (of which see below). The conquest of Egypt during the second campaign overshadowed the achievement in Shubria, to such an extent that the latter lost its importance. 13. See Mnm. A obv. 16; Trb. B 5; Klch. D 3; Ass. H 5–7; cf. also Mnm. C 5–6. 14. J. M. Russell, The Writing on the Wall: Studies in the Architectural Context of Late Assyrian Palace Inscriptions (Winona Lake: Eisenbrauns, 1999), 293–94. AsBbA obv. 28–29 describes Esarhaddon as “King of kings of Tilmun, Magan and Meluhha.” The juxtaposition of Magan and Meluhha with Tilmun in this title, in an inscription making no reference to Egypt and its booty (probably written before the occupation of Egypt; see Borger, 78, for the distinction between AsBbA–D and AsBbE–H), indicates a connection with the activity of Esarhaddon’s army in the Persian Gulf region (the campaign to the land of Bazu?). For references to Tilmun, Magan, and Meluhha in Mesopotamian non-literary sources from the pre-Dynastic period up to the seventh century B.C., clearly associated with the Gulf region and the Indian Ocean, see W. Heimpel, “Das Untere Meer,” ZA 77 (1987) 22–91. On the other hand, Esarhaddon’s title in Klch. D is clearly related to the conquest of Egypt (cf. also Frt. F 7[?], 15, and similarly in Ashurbanipal’s A i 52). For the signi˜cance of Magan and Meluhha (without Tilmun) as designations of Egypt and Nubia see Heimpel, “Magan,” RlA 7 (1987–1990) 196; “Meluhha,” RlA 8 (1993–1997) 55.
One [Body] Line Short
title, an expression of achievement and power, appears not only in his historical inscriptions, but also in the inscription on the great colossi in his palace at Calah.15 It does not represent political reality: Esarhaddon’s rule did not extend much beyond Memphis. South of Memphis, Taharqa was still in control; although badly weakened by his loss of the Delta region as well as by the capture and deportation of his son, the crown prince, his wives and other members of his family, he had not totally been defeated. Assyrian rule in Egypt was extended only under Ashurbanipal, who wrested Thebes from Taharqa and defeated the latter’s successor Tantamani. Chronological Framework In order better to understand the sequence of Esarhaddon’s political, military and propagandistic activities from his ˜rst Egyptian campaign until his death, careful note must be taken of the time factor aˆecting his decisions and his course of action. Hence, the available chronological data concerning his actions merits some attention. The data derive from several sources, not all equally reliable. When there is a discrepancy between the data in the Babylonian Chronicles (which are later than Esarhaddon) and those in Esarhaddon’s inscriptions, the latter are to be prefered. For the reader’s convenience, and to avoid erroneous matching of the Assyrian date to the standard (B.C.) calendar, events will be dated according to Esarhaddon’s regnal years: year 7, Adar: Defeat of the Assyrian army in Egypt. Failure of the ˜rst campaign year 8, Tebeth(?)/Adar: Conquest of Shubria year 9, Nisan–Iyyar: Administration of oath of loyalty to Ashurbanipal and Shamash-shumaukin as the future heirs to Esarhaddon’s throne year 10, Nisan: Beginning of second Egyptian campaign year 10, Tammuz/Tishri: Victory in Egypt. Conquest of Memphis 15. See Russell, The Writing on the Wall, 149–51; 293–94.
ESARHADDON, EGYPT, AND SHUBRIA: POLITICS AND PROPAGANDA
year 12, Marcheshvan 10: Death of Esarhaddon on the way to Egypt Comments Year 7: Chron. 1 iv 16. For the problematic aspect of the account in Chron. 14:20 see note 1. Year 8: Cf. Chron. 1 iv 19–21: “Year 8 of Esarhaddon. On the broken (hepi; i.e., ‘[the tablet from which the text was copied is] broken’) day of Tebeth Shubria was captured (and) sacked. In Kislev its booty entered Uruk”; Chron. 14:24: “Adar, year 8, cf. line 23), 10 (of the month). The army of Assyria . . . Shubr[ia].” The text in line 21 of the Babylonian Chronicle “Kislev, its (Shubria’s?) booty entered Uruk,” is puzzling. That the booty of Shubria was brought to the cities of Assyria, as reported in the “Letter to the God,” Gbr. II iii 21– 22, is plausible. However, it is hard to understand why it should be brought to Uruk in southern Babylonia; see Brinkman, “The Babylonian Chronicle Revisited,” 94. Given the indication that the source employed by the author of the chronicle was broken, the word “Tebeth” in line 19 may not refer at all to Uruk and to booty being brought there, but to some other event; hence no conclusion may be drawn here as to the date of Shubria’s occupation. Having examined the tablet on which the Babylonian Chronicle is written, Brinkman comments further that the reading of the year numbers 7 and 8 is not certain, and the text may possibly be read as 8 and 9, respectively (“The Babylonian Chronicle Revisited,” 95, note 128). For our purposes, this alternative reading is irrelevant for the time elapsing from Esarhaddon’s ˜rst Egyptian campaign to the occupation of Shubria. However, the proposed alternative reading of the numbers raises another di¯culty: it does not leave su¯cient time after the conquest of Shubria to prepare for the second Egyptian campaign. According to the “Letter to the God,” the decisive battle for Uppume, the capital of Shubria, took place after its wall had been damaged by ˜re on the night of 21 Kislev (Gbr. II ii 1–9). It seems plausible, therefore, that the occupation of Shubria
103
was completed not long after the fall of its capital. Ten to twelve months had therefore passed from the defeat in Egypt to the conquest of Shubria. Within this span of time, Esarhaddon had to return from Egypt to Assyria (a journey of at least two and a half months in administrative movement), to prepare for the campaign against Shubria, to march on Shubria (a further two weeks in administrative movement), and to conquer it, an undertaking that involved at least some siege warfare—also a time-consuming procedure.16 This crowded schedule attests to Esarhaddon’s determination in acting against Shubria in the context of his eˆorts to reestablish Assyria’s political prestige, which had been mortally damaged in the ˜rst Egyptian campaign. Year 9: The chronicles report nothing for Esarhaddon’s ninth year. This silence was not a result of lack of sources, but of the fact that indeed no military action was taken that year. This is clearly evident from Frt. F obv. 1–6: The end of the account of actions taken against Shubria and the description of the second Egyptian campaign constitute a continuous text, no diˆerentiation being made between the two accounts. The same sequence may be observed in column V of prism Nin. S (see Nin. S 3 col. b; Nin. S 4 col. b in Tadmor, “An Assyrian Victory Chant,” 273–76; see note 18 below). It transpires from several documents that in year 9, when Esarhaddon stayed in Assyria, he consolidated his rule and guaranteed a smooth transmission of power after his death by administering an oath to his scribes, the palace staˆ, the royal guard, and other groups in the months of Nisan and Iyyar (see SAA II 6:664–665; Streck, Asb. 2–3; and also SAA X 6, 717 ); he also organized
16. For the length of the long military campaigns and their eˆect on the imperial administration, see Ephçal, “On Warfare and Military Control in the Ancient Near Eastern Empires: A Research Outline,” in eds. H. Tadmor and M. Weinfeld, History, Historiography and Interpretation ( Jerusalem: Magnes Press, 1983), 99–101. 17. For Esarhaddon’s Median royal guard’s oath of loyalty see M. Liverani, “The Medes at Esarhaddon’s Court,” JCS 47 (1995) 57–62.
104
ISRAEL EPHçAL
his army for the major project ahead—the second campaign against Egypt; cf. Nebuchadnezzar’s limited activity after the failure of his Egyptian campaign (Ephçal, “Nebuchadnezzar the Warrior,” 183–85). Year 10, Nisan: Chron. I iv 23; cf. similarly Chron. 14:25; Frt. F obv. 10. Year 10, Tammuz/Tishri: Chron. I iv 24–25; Chron. 14:26. According to the Babylonian Chronicle, the Egyptian army was defeated in three battles on 3, 16 and 18 Tammuz, whereas the Esarhaddon Chronicle speaks of a major Egyptian defeat on 3 Tishri. The diˆerence in the names of the months implies the length of time at Esarhaddon’s disposal for the siege of Tyre during his advance on Egypt (for this siege see Frt. F obv. 12–14): According to the data in the Babylonian Chronicle, he could have assigned at most two or three weeks to the siege, whereas the Esarhaddon Chronicle leaves almost three months. The date on which Nin. S was written, even if it is a “pseudo-date” (see below), con˜rms the information of the Babylonian Chronicle dating of the occupation of Memphis. The opening section of the fragment Borger, Nin. E, col. ii = Nin. S 4 col. v refers to Esarhaddon’s victory in a battle that took place in Egypt on the third day of a month whose name has not survived. According to the Zincirli Stela, lines 37–40, the march from Ishkhupri to Memphis took ˜fteen days. Combining these two points of chronological data, one is led to the conclusion that the battle mentioned in Borger, Nin. E, col. ii = Nin. S 4 col. v must have taken place in Ishkhupri on the third of the month (of Tammuz), and the battle for Memphis, on 18 Tammuz (as reported in the Babylonian Chronicle). The time diˆerence between the third battle (18 Tammuz) and the occupation of Memphis (four days later, according to the Babylonian Chronicle) indicates that the hostilities did not develop into a real siege of Memphis, and that in fact Esarhaddon had no time to set up the siege machines listed in the Zincirli Stela (Mnm. A rev. 41–43). It is more plausible that the author of the stela text, writing that Memphis was overrun in “half a day,” was using a conventional technical
formula (known, e.g., from the account of Sennacherib’s conquest of the cities of Judah, Luckenbill, The Annals of Sennacherib, 32–33 iii 21– 23), and added the exaggerated expression in order to magnify a victory that had been achieved in a shorter time. Year 12: Chron. 1 iv 30–31; Chron. 14:28–29. Esarhaddon lived 28 to 31 months more after occupying Egypt (the discrepancy is owing to the diˆerence between the dates speci˜ed in the chronicles as to the month in which Egypt was taken). All the literary and propaganda activity relating to the conquest of Egypt took place within that relative short span of time. Documents Reporting the Conquest of Egypt The New Account Edition Shortly after the conquest of Egypt, an account was written of the campaign, describing the departure from Assyria, the march through Phoenicia, Palestine and the Sinai Desert, the ˜ghting in Egypt and its occupation, as well as the aftermath of the occupation: the looting, deportation, organization of Assyrian administration in the occupied area. The account has not survived in its entirety, but only in the form of a draft, part of which was written on a broken tablet (Borger, Frt. F) and on fragments of an octagonal prism (Nin. S).18 Of Borger, Frt. F only 36 lines survive. They provide a detailed account, in ˘orid literary terms, of Esarhaddon’s progress with his army through Palestine and the Sinai Desert to Ishkhupri, the site of the ˜rst battle at the approaches of Egypt. Preserved in the ˜ve lines before this account is the end of the account of the campaign against Shubria. 18. For Nin. S and its reconstruction see Tadmor, “An Assyrian Victory Chant and Related Matters,” in From the Upper Sea to the Lower Sea: Studies on the History of Assyria and Babylonia in Honor of A. K. Grayson, ed. G. Frame (Leuven: Peeters, 2004), 273–76. The sections of the prism are referred to according to Tadmor’s system.
ESARHADDON, EGYPT, AND SHUBRIA: POLITICS AND PROPAGANDA
The three fragments of Nin. S describe Esarhaddon’s activities in Shubria and Egypt. In one fragment, in col. v of the prism, there is an account of the war against Shubria (Borger, Die Inschriften Asarhaddons, 115 = Nin. S 3, col. b). Another fragment, also containing part of the same column, is the beginning of the account of the second Egyptian campaign (Borger, Nin. E, col. ii = Nin. S 4, col. b). The surviving passage begins on the day of the ˜rst battle of the campaign on Egyptian soil.19 The third fragment, which contained parts of columns vi–vii of the prism, preserves lists of Egyptian deportees grouped by profession, lists of Assyrian governors appointed in Egyptian cities whose names had been changed, as well as details of the annual tax to the king of Assyria and funds assigned for sacri˜ces to the gods imposed on Egypt (Borger, Frt. J = Nin. S 5, cols. a–b).20 It follows from this reconstruction of Nin. S that the account of the Egyptian campaign and the actions taken thereafter was extremely detailed, amounting to no less than two hundred lines (the average number of lines in an octagonal Assyrian prism is 95).21 The date the text was inscribed is at the end of the prism, phrased in solemn terms: “The month of itu dMA{ (a logogram denoting the Elamite name for the month Tammuz22), the lim[mu] of . . . , in the year in which [Memphis was captured and] its booty [ca]rried oˆ.” Reconstructing the name of the eponym [Kanunai] (the eponym of Esarhaddon’s tenth year, 671/70 B.C.), one necessarily infers that
19. The description of Esarhaddon’s setting out for battle in lines 1–19 is remarkably similar to the ˘owery literary account of Sennacherib going forth to the battle of Khalule (see D. D. Luckenbill, The Annals of Sennacherib, OIP II [Chicago: University of Chicago, 1924], 44–45: 63–81). 20. For a transcription and translation with notes concerning this fragment see also H. Onasch, Die assyrischen Eroberungen Ägyptens, Ägypten und Altes Testament 27 (Wiesbaden: Harrasowitz, 1994), 31–37. 21. The Zincirli and Nahr el-Kelb inscriptions contain passages that did not survive in the fragments of the New Account Edition as preserved (see Mnm. A rev. 40–46; Mnm. C 7–23). On the assumption that these passages are based on that account, they may also be used in its reconstruction. 22. See E. Reiner, “Inscription from a Royal Elamite Tomb,” AfO 24 (1973) 100.
105
the activities in Egypt, as listed in the prism, including the organization and execution of the deportation, appointment of governors and organized plunder, as described in the inscription, as well as the changing of Egyptian city names, must have been carried out within eight days after the capture of Memphis (the date of which, 22 Tammuz, is speci˜ed in the Babylonian Chronicle). Within that short time it was also necessary, of course, to write the prism in Egypt, which was later brought to Nineveh, where its fragments were discovered. Since all these activities could not have been completed in a few days, we may conjecture that the date speci˜ed for the writing of the prism was not true but an ideologically motivated “pseudo-date,” like the formula “accession year” of Esarhaddon speci˜ed as the date on which his Babylon Inscription (Bab. A–G) was composed, in order to create the impression that he had rebuilt the ruined city of Babylon immediately after ascending the throne.23 Another solution to the di¯culty is to consider the words “in the year in which [Memphis was captured and] its booty [ca]rried oˆ” as a chronologically inde˜nite phrase meaning “when,” as it occurs occasionally in Neo-Assyrian economic documents, and to reconstruct the missing name in the date formula as [Shulmu-beli-lashme], the eponym of Esarhaddon’s eleventh year (670/69 B.C.).24 Also worthy of mention here is a small, partly damaged tablet, K 8692, with an account, in annalistic style, of a tremendous quantity of booty plundered and deportees “whose skin is as black as pitch,” the members of the family of a ruler whose name was not preserved in the text.25 No parallel has been found in the inscriptions of Esarhaddon and Ashurbanipal for the content of this tablet.
23. See M. Cogan, “Omens and Ideology in the Babylonian Inscription of Esarhaddon,” in History, Historiography and Interpretation, 85–87. 24. See H. Tadmor and E. Weissert in Tadmor, “An Assyrian Victory Chant,” 276. 25. See W. G. Lambert, “Booty from Egypt?,” JSS 33 (1982), 61–70. For the question of the relationship of this text to Esarhaddon’s second Egyptian campaign see also Onasch, Die assyrischen Eroberungen Ägyptens, 25–28.
106
ISRAEL EPHçAL
Monuments After Egypt had been conquered and the account composed, monuments of Esarhaddon were erected in public places, with Assyrian inscriptions based on the account; the number of the addressees of these obviously propagandistic products was evidently larger than that of the potential readers of the tablets and the prisms. Some of these monuments have been discovered in the west of the Empire: Two stelae found at Til Barsip (today Tell Ahmar), one at Zincirli (the capital of Samåal), an inscribed relief of Esarhaddon carved on a rock near the mouth of Nahr el-Kelb (ten kilometers north of Beirut), and a fragment of a stela found at Qaqun, Israel (six kilometers northwest of Tul Karm).26 Esarhaddon’s stelae from Zincirli and Til Barsip are the largest of the Assyrian royal stelae: The Zincirli Stela is 322 centimeters high (together with its base, the monument reached a height of 346 cm). The Til Barsip monuments are 380 and 330 centimeters high, respectively. All three are similar in layout: They picture Esarhaddon standing; above him are divine symbols, before him are two smaller human ˜gures (in the Zincirli Stela they reach his knees; in the Til Barsip stelae, his waist) in supplicatory postures. One, bearded and wearing a conical hat, is standing;
26. Zincirli Stela (Borger, Mnm. A): For a transcription and translation of the passage relating to the conquest of Egypt and the subsequent actions (rev. 37–50), as well as notes, see also Onasch, Die assyrischen Eroberungen Ägyptens, 24–29. For a drawing of the inscription, with transcription and translation, see E. Schrader, “Inschrift Asarhaddon’s, Königs von Assyrien,” in F. von Luschan, Ausgrabungen in Sendschirli. I. Einleitung und Inschriften (Berlin: W. Spemann, 1893), 30–43; see this volume also for the technical details of the stela (location, height, weight, etc.). Til Barsip, Stela A (Borger, Mnm. A): For a drawing of the stela and technical details, as well as a transcription and translation of the inscription, see F. Thureau-Dangin and M. Dunand, Til-Barsib (Paris: Paul Geuthner, 1936), 151–55. For Stela B, see pp. 155–56. Nahr el-Kelb Inscription (Borger, Mnm. C): For a photograph of the monument and drawing of the inscription, with a transcription and translation, see F. H. Weissbach, Die Denkmäler und Inschriften an der Mündung des Nahr El-Kelb (Berlin and Leipzig: W. de Gruyter, 1922). Qaqun inscription: See note 38.
the other, a beardless Negroid, wearing a uraeus headdress, is kneeling. In the Til Barsip stelae these two ˜gures are portrayed in conventional submissive poses, but the Zincirli one shows Esarhaddon holding ropes attached to rings in their lips, and the Negroid ˜gure’s hands and legs are in chains.27 The kneeling ˜gure is generally identi˜ed as Ushanahuru, Taharqa’s son, whom Esarhaddon deported, as noted in both Zincirli (line 43) and Nahr el-Kelb (line 12) inscriptions. As to the second, standing ˜gure, Thureau-Dangin suggested, on the basis of the reference to Abdimilkutti at the base of the Til Barsip stela, that it represents Abdi-milkutti, king of Sidon. However, in view of the explicit reports that Abdi-milkutti unsuccessfully ˘ed from the Assyrian king only to be decapitated and have his head brought to Assyria (Nin. A ii 71–74; iii 32–34; see also Chron. 1 iv 67; Chron. 14:14), it is di¯cult to identify him as the supplicating ˜gure in the relief. Others, therefore, have identi˜ed him as Baal, king of Tyre, who was besieged by the king of Assyria during the Egyptian campaign, since he had been loyal to Taharqa and cooperated with him (see Frt. F obv. 12–14). However, nowhere in Esarhaddon’s inscriptions is there any mention of Baal’s capture or surrender. He is known to have continued to rule Tyre during Ashurbanipal’s reign as well (see C i 24).28
27. For captives led by ropes tied to rings piercing their lips see also P. E. Botta and M. E. Flandin, Monument de Niniveh (Paris: Imprimerie nationale, 1849), pl. 83. A variant of this depiction may be seen in the rock relief of Annubanini, king of the Lulubi, in which enemies are led with rings through their noses; see J. Börker-Klähn, Altvorderasiatische Bildstelen und Vergleichbare Felsreliefs, Baghdader Forschungen 4 (Mainz am Rhein: P. von Zabern, 1982), no. 31 (the date of the relief is a scholarly bone of contention). For a verbal description of the scenes under discussion, compare the account of Manasseh in 2 Chr 33:11. For a rope passed through the pierced cheeks of an enemy captive (Uaitaç, son of Birdada, king of the Arabs), who was then exhibited in public, see Ashurbanipal A ix 106–107; and cf. Job 40:25–26. 28. Miglus takes the view that there is no need to seek a direct link between the supplicating ˜gures in the Zincirli and Til Barsip stelae with speci˜c military events in the western regions of the Empire during Esarhaddon’s reign (and hence no need to identify the ˜gures with known persons), and that the stelae were not intended mainly to intimidate the subjects
ESARHADDON, EGYPT, AND SHUBRIA: POLITICS AND PROPAGANDA
Carved on each face, on either side of the Zincirli and Til Barsip stelae, are two standing human ˜gures. The right-most ˜gure is depicted in Assyrian style, that on the left, in Babylonian style. These are presumably Esarhaddon’s sons, the heirs apparent: Ashurbanipal would rule Assyria and the empire, and Shamash-shuma-ukin would be in charge of Babylonia.29 Assyrian o¯cials and subjects swore loyalty to them in Nisan and Iyyar of Esarhaddon’s ninth year, ten months before the launching of the second Egyptian campaign.30 In the Zincirli Stela, the inscription is carved on both sides, front and back. The inscription on one of the Til Barsip stelae (Stela A in ThureauDangin and Dunand, Til Barsib, 151–55) is carved on only the obverse, and it seems to be incomplete;31 the other Til Barsip stela, Stela B (ThureauDangin and Dunand, Til Barsib, 155–56), bears
of those regions and deter them from rebelling, but to demonstrate impressively Esarhaddon’s power to the internal opposition to his rule, thus bolstering the succession of his sons Ashurbanipal and Shamash-shuma-ukin to the throne (P. A. Miglus, “Die Stelen des Königs Asarhaddon von Assyrien: Siegesdenkmäler oder ein politisches Vermächtnis?,” BaM 31 [2000] 195–209). However, this view is hard to reconcile with the fact that the stelae were prepared a relatively long time after the succession had been arranged. The oath of loyalty to Ashurbanipal and his brother took place in Iyyar 672 B.C., whereas the preparation of the stelae could not have begun before the end of 671 (after the conquest of Egypt, and after the consolidation of the royal reports of the organization of the administration in Egypt). The absence of an inscription on the Til Barsip Stela B may indicate that the stelae were made not long before Esarhaddon’s death (March 669) and were therefore never completed. Moreover, Ashurbanipal and Shamash-shuma-ukin are not mentioned at all in the texts of the stelae. 29. Ashurbanipal stands on the right, as be˜ts his senior status, and Shamash-shuma-ukin, on the left; compare SAA X 185: 5–12. 30. See above, Year 9. The inclusion of Esarhaddon’s sons in these stelae should not be seen as the main reason for their existence. As already noted, they are not even mentioned in the inscriptions. 31. Only 19 of the approximately 34 lines originally on the obverse of the stela are legible. The line incised in the lefthand margin seems to be the end of the text (compare this line with the inscription in the Zincirli Stela, rev., 50, 52). Still missing are the benedictory and maledictory formulas aimed at persons dealing with the inscription in the future.
107
no inscription, surely because Esarhaddon died before a text could be carved.32 The height of the niche carved in the rock in the Nahr el-Kelb estuary is 189 centimeters. Shown in this niche is the standing ˜gure of Esarhaddon topped by divine symbols, as in the Zincirli and Til Barsip stelae. Carved on the relief is an inscription of some forty lines. There are no supplicatory ˜gures, nor any portrayals of
32. The Zincirli excavations uncovered another Assyrian royal stela (see von Luschan, Ausgrabungen in Sendschirli, 28– 29). However, since it was badly weatherworn and damaged by ˜re, the excavators left it in place at the site. Comparison with the two Til Barsip stelae suggested to Porter that this was one of a pair with Esarhaddon’s stela (Borger, Mnm. A), and she therefore conjectured that it bore an Aramaic or Phoenician inscription, and that the two stelae were addressed to audiences of two distinct types, speaking diˆerent languages—one to readers of Assyrian and the other to readers of the local language (B. N. Porter, “Language, Audience and Impact in Imperial Assyria,” IOS 15 [1995] 51–72, notes 3 and 9). This conjecture is, I believe, wrong: Esarhaddon’s stela (Mnm. A) is made of dolerite, is 3.22 m high, and stood at the city gate; whereas the damaged stela was of limestone, is only 190 cm high, and stood in the northeast palace, some 320 m from the other stela (see von Luschan, Ausgrabungen in Sendschirli, 10). It is hardly possible that one of the stelae was displaced after the occupation of Samåal (Esarhaddon’s stela weighs about six tons, the other about half of that ˜gure). Moreover, all bilingual inscriptions discovered to date are incised on the same or adjacent stones. There is no reason to believe that stelae originally distant from one another (the distance between the inscriptions at Til Barsip was also more than six hundred meters) were written in two languages and intended for two diˆerent audiences. Porter, pointing out the diˆerences in the visual depictions on the Zincirli and Til Barsip stelae, attributed these diˆerences to the diˆerent intended audiences, based on the political history of each of the centers where the stelae were placed (at Til Barsip and Samåal); see Porter, “Language, Audience and Impact,” 51–72; idem, “ ‘For the Astonishment of All Enemies’: Assyrian Propaganda and Its Audiences in the Reigns of Ashurnasirpal II and Esarhaddon,” Bulletin of the Canadian Society for Mesopotamian Studies 35 (2000) 7–18; idem, “Assyrian Propaganda for the West: Esarhaddon’s Stelae for Til Barsip and Samåal,” in Essays on Syria in the Iron Age, ed. G. Bunnens (Leuven: Peeters, 2000), 143–76. To my mind, the available data are insu¯cient for a de˜nition of the political positions of the people of Samåal and Til Barsip, to the extent that it might explain the visual diˆerences between the reliefs. As will be inferred below from an examination of the Nahr el-Kelb and Qaqun inscriptions, the inscribed texts do not support Porter’s conjecture. More probably, the stelae were visually distinct because they were made by diˆerent artists, each with his own taste, but following the same general pattern.
108
ISRAEL EPHçAL
Esarhaddon’s sons, but the general and chronological context of the monument is clear from the inscription, which is concerned with Esarhaddon’s occupation of Memphis and describes the booty taken and the make-up of the deportees exiled from Egypt. As against the almost completely identical layouts of the monuments from Zincirli and Til Barsip, and the similarity of the ˜gure of the king (and of most of the divine symbols) in the Nahr elKelb Relief to the standard depictions of Assyrian kings,33 the inscriptions on these monuments are diˆerent: The Til Barsip Inscription (Stela A) resembles a summary inscription, presenting Esarhaddon as having acted against the Arabs,34 Hilakku, Ellipi, Barnaki, the Manneans and the Scythians, the Cimmerians and Khubushna, and the name of Abdi-milkutti of Sidon, is inscribed at the base. These actions took place over the span of a few years before 673 B.C., before the ˜rst Egyptian campaign. The inscription makes no mention of the conquest of Shubria or Egypt, but, inasmuch as it was never completed, it is quite possible that these actions should also have been included. The Zincirli Inscription is quite diˆerent. Written in the ˜rst person, it lists Esarhaddon’s titles, praises him as the elect of the gods, and describes his divinely assisted power, as well as his piousness towards the gods and his acts on their behalf. The text is imbued with a sense of piety, centering mainly not on the king but on the gods. Of all the military events during his reign, the only ones explicitly mentioned are those that took place in Egypt during the second campaign. Their description occupies only fourteen of the ninety-
33. Thus, for example, there is a marked similarity between Esarhaddon’s monument at Nahr el-Kelb and the ˜ve other rock-cut Assyrian monuments at the same site, which are certainly not of Esarhaddon, though the exact identity of the monarchs to whom they were dedicated is a matter of scholarly debate (see Börker-Klähn, Altvorderasiatische Bildstelen und Vergleichbare Felsreliefs, 211–12). 34. Various episodes associated with the Arabs are related in most of the surviving text of the inscription (lines 7–19; perhaps some of the no longer extant lines preceding line 7 originally referred to the Arabs as well).
two lines in the inscription (obv. 37–50). It begins with the defeat of Taharqa’s army and the pursuit of the ˘eeing army from Ishkhupri to Memphis, a ˜fteen-day journey.35 The inscription goes on to say that Memphis was captured during a half-day siege using tunnels and breaches, destroyed and put to the torch. Taharqa’s wife and concubines, the crown prince Ushanahuru, and Taharqa’s other sons and daughters were deported to Assyria. His possessions, including innumerable horses, cattle and sheep, were seized. The “root” of Nubia was uprooted from Egypt; kings, governors, o¯cers and other o¯cials were appointed over the country, and taxes were imposed, as well as oˆerings to be made regularly to Ashur and the great gods. The purpose and meaning of the monument are clari˜ed by lines 50–53: Esarhaddon commanded the erection of the monument and the inscription and the mighty deeds both of the god Ashur and of himself (with the god’s help), so that enemies would see them forever. The inscription ends with a request to the future ruler (rubû arkû), who would see Esarhaddon’s name carved on the monument, that, upon having the inscription read to him, he would anoint it with oil, oˆer a sacri˜ce, and glorify the name of Ashur (rev. 56–57). Presumably, therefore, the message of the stelae, addressed to those entering the gates of Zincirli and Til Barsip, was primarily visual, as conveyed by the size of the reliefs and the scenes portrayed in them; the inscriptions themselves were of secondary importance.36
35. The three lines preceding the passage in question (rev. 35–37) state that “after Ashur and the great gods, my lords, had commanded me, I advanced con˜dently and safely (over) distant roads, wearying mountains, tremendous sands, a place of thirst.” This may be a general allusion to the di¯culty of an advance through the Sinai Desert, but it may also refer to other di¯cult regions through which Esarhaddon’s armies had marched in previous years, such as the Bazu campaign; cf. Nin. A iv 53–60. 36. In other words, one might say that the relief was aimed at Esarhaddon’s contemporaries, whereas the text was addressed primarily to future generations. The secondary nature of the propaganda message of the text in the Zincirli Stela may be inferred from its peculiar placing: the reverse side (including most of the text) was up against the wall of the gate, and the inscription was not readily readable; see the data in von Luschan, Ausgrabungen in Sendschirli, 11–12.
ESARHADDON, EGYPT, AND SHUBRIA: POLITICS AND PROPAGANDA
The historical account in the Nahr el-Kelb Inscription opens with Esarhaddon’s entry into occupied Memphis (lines 7–8), without any account of the march through the desert or of battles. There ensues a list, more detailed than that in the Zincirli Inscription, of the booty taken from Taharqa (lines 9–22) and of the deportees from Egypt, including the members of Taharqa’s family and valued professionals (lines 23–27).37 Lines 30–34 are severely damaged; they preserve the names of rulers who probably cooperated with Taharqa: line 30: “. . . son of Binzuqi . . .”; line 31: “. . . Ashkelon”; line 32: “of Taharqa and their fortresses . . .”; line 33: “. . . Tyre”; line 34: “. . . 22 kings.” Lines 35– 40 have not survived. As mentioned, a fragment of another inscription was found at Qaqun. It comprises twenty lines, each preserving only a few signs.38 The Zincirli and Nahr el-Kelb inscriptions report Esarhaddon’s conquest of Egypt and his actions after the ˜rst battle there, while the Qaqun fragment describes an earlier stage of the military campaign, beginning with the departure from the city of Ashur, the attack on Baal, king of Tyre, who was an ally of Taharqa, and the advance to southern Palestine and farther to the Sinai Desert, with its fearful di¯culties (compare Frt. F obv. 10–rev. 8). Since it is clear that the account of the ˜rst stage of the campaign alone could not possibly ˜ll out a historical inscription, it may be assumed that the account in the Qaqun Inscription was longer and more detailed than in the other monuments, including the occupation of Egypt and the subsequent activities. This enhances the similarity of the Qaqun Inscription to the expanded edition of the
37. The list of deported professionals probably continued in the now illegible lines 28–29; cf. Frt. J i 9–17. 38. The inscription is to be published by Elnathan Weissert, to whom I am indebted for providing me with a reconstructed text. One piece of information preserved in this inscription is unique among known Esarhaddon inscriptions: It reports that the Assyrian army crossed the Sinai with the assistance of the people of Mibsam (for this nomadic group cf. Gen. 25:13). This reference to the people of Mibsam is unique for any of the known Assyrian sources (in Frt. F rev. 1–2 the owners of the camels on which water skins for the Assyrian army were carried while it advanced in Northern Sinai were designated with the common noun “Arabs”).
109
account in Frt. F and in Nin. S, rather than to the large-scale monuments. The diˆerence between the Qaqun Inscription and the monuments is also evident in the size of their cuneiform signs.39 This, coupled with the known direct proportion between the size of a monument and the size of the script, indicates that the Qaqun Inscription is not one of the series of large-scale monuments. The propaganda message of the Qaqun Inscription lay not in its size, but in the fact that it was erected in a region where not many people could read Assyrian. To summarize the accounts of the Egyptian conquest, the intensity with which accounts were written and with which monuments were erected for propaganda purposes is particularly evident in the west of the Empire. As noted, this propaganda eˆort, which took a variety of forms, was undertaken within a period of at most two years. Change of Toponyms Toponyms with the structure Kar/Dur RN/DN were very common in Mesopotamia, from the beginning of the Old Babylonian period on. The new element in the Neo-Assyrian period was not, therefore, just that certain places were given names of this structure, but that such toponyms were used as the new names of occupied cities as a palpable expression of Assyrian hegemony, as distinct from cities that were newly established as a result of royal initiatives.40 Extant Neo-Assyrian royal inscriptions mention ˜fty cities whose names were changed after occupation by the Assyrian army, from Ashurnasirpal II to Esarhaddon. Forty-one of them are included in Pongratz-Leisten’s study; to these one should add
39. The height of the lines in the various inscriptions is as follows: Zincirli Stela—ca. 43 mm; Til Barsip Stela—ca. 37 mm; Nahr el-Kelb Inscription—ca. 31 mm; Qaqun inscription—ca. 17 mm. 40. B. Pongratz-Leisten, “Toponyme als Ausdruck assyrischen Herrschaftsanspruchs,” in Ana sadî Labnâni lû allik, Beiträge zu altorientalischen und mittelmeerischen Kulturen (Festschrift für W. Röllig), ed. B. Pongratz-Leisten, H. Kühne and P. Xella, AOAT 247 (Neukirchen-Vluyn: Butzen and Bercker, 1997), 325–43.
110
ISRAEL EPHçAL
nine Egyptian cities whose names were changed after their capture by Esarhaddon in 671 B.C.41 Esarhaddon’s changing of names of occupied cities that had been administrative centers in Egypt is also mentioned in Ashurbanipal’s inscriptions.42 Two-thirds of the new toponyms (thirty-three, to be precise) were given under Esarhaddon. A closer examination of the circumstances reveals that thirty-two of these names were given in less than two years between the conquest of Shubria and the composition of the report of the conquest of Egypt. Toponyms that were changed before Esarhaddon’s reign are quite diˆerent from those changed during his time. Prior to his reign, the majority had the structure Kar/Dur RN/DN.43 From his time on, almost all (thirty-one of the presently surviving thirty-two) had the format of slogans, highly reminiscent of the names of many gates in the large cities of Assyria.44 These
41. For the list of omitted cities see Borger, Frt. J ii 2u–11u (= Bu 91–5–9, 218); the list is also included in Onasch, Die assyrischen Eroberungen Ägyptens, 31. The beginning is apparently broken. From the existing traces of lines 8u, 10u in Frt. J ii it appears that other toponyms were listed. To estimate the number of toponyms in the entire list, it should be compared with the list of administrative centers in Egypt in Ashurbanipal’s inscriptions, which include (the Egyptian names of) 21 cities besides the names of twenty “kings” that Esarhaddon appointed over them after his conquest of Egypt (A i 90–109; C ii 87–92; see Onasch, Die assyrischen Eroberungen Ägyptens, 36. H. Verreth, “The Egyptian Eastern Border Region in Assyrian Sources,” JAOS 119 [1999] 234–47] believes that the list in Ashurbanipal’s inscriptions is composite, and that some cities were listed in it twice). Two of them are mentioned in Ashurbanipal’s “Harran Tablets” with both their Egyptian and Assyrian names ({athariba/Athribis = Limmer-issakku-Assur; see Asb., Harran obv. 65; Sais = Kar-Bel-matati; see Asarh., Smlt. obv. 25; Asb., Harran obv. 61; see also Onasch, Die assyrischen Eroberungen Ägyptens, 35). There are grounds for the assumption that the names of the other cities were also changed. This would bring the total number of cities whose names Esarhaddon changed up to at least 43 (Kar-Esarhaddon, built on the ruins of Sidon + 21 cities in Shubria + 21 cities in Egypt). 42. See A. C. Piepkorn, Historical Prism Inscriptions of Ashurbanipal, AS 5 (Chicago: University of Chicago, 1933), Prism E i 14–15. 43. See Pongratz-Leisten, “Toponyme als Ausdruck,” 331–35. 44. The only exception to the rule is the toponym DurEsarhaddon given to one of the conquered Shubrian cities,
slogans express the sovereignty of the Assyrian king, without specifying his name, and his connection with the god Ashur (e.g., “Do not ˘out the king’s word,” “May the representative (issâkku) of Ashur45 shine”), the size and power of Ashur, the national god (e.g., “Who shall compare with Ashur,” “Ashur has extended his yoke,” “Long live he who delights Ashur”), as well as “patriotic” slogans (e.g., “I have restored his land to Ashur,” “May [the land] increase and not diminish”). Such an impressive format for the names of cities was an innovation in the Assyrian Empire and, to a large degree, in the entire ancient Near East.46 Even in Esarhaddon’s fourth regnal year, one still ˜nds the old format, in which the king’s name was incorporated in the changed toponym (as in KarEsarhaddon, the city built on the ruins of Sidon in 676 B.C.). It would appear, therefore, that the marked change in this practice began at the time of the Shubria campaign and continued with the conquest of Egypt as a deliberate, universally applied, act of propaganda. Among the theophoric toponyms conforming to the pattern Kar/Dur DN before Esarhaddon’s day, one ˜nds the names of the gods Ashur, Nergal, Nabu, Sin, Adad, Ishtar and Enlil;47 in the newer
which conforms to the old format. For the names of gates in this “slogan” format see, e.g., A. R. George, Babylonian Topographical Texts, OLA 40 (Leuven: Department Orientalistiek and Peeters, 1992), 67, 177. 45. For this title of Esarhaddon see Borger, “Die Inschriften Asarhaddons (AfO Beiheft 9),” AfO 18 (1957–1958) 113, 10a, 9. 46. The only earlier example of this format is the toponym Ana-Assur-utêr-asbat (“I restored to Ashur and captured”) given to the city “that the people of the land of {atti call Pitru” and mentioned in Shalmaneser III’s inscriptions (RIMA 3 23:85– 86; 51–52:40–43; 64–65:38–40). The name implies that the site of the city was in territory disputed by the Assyrians and the Arameans. The historical signi˜cance of the Assyrian toponym may be inferred from the proximity of the reference to AnaAssur-utêr-ahbat to that of “(the city of) Mutkinu on the other side of the River Euphrates, which was settled by Tiglathpileser (I, 1114–1076 B.C.) (and) in the days of Ashur-rabi (II, 1012–972), king of Assyria, the king of the Arameans took it by force.” These cities were restored to Assyria by Shalmaneser III (RIMA 3 19:35–38). 47. See Pongratz-Leisten, “Toponyme als Ausdruck assyrischen Herrschaftsanspruchs,” 331–35.
ESARHADDON, EGYPT, AND SHUBRIA: POLITICS AND PROPAGANDA
toponyms, however, Ashur’s name takes pride of place, far ahead of all the other gods. His name is the theophoric element in fourteen of the seventeen theophoric toponyms. The others are Belmatati,48 Marduk and Banitu = Ishtar, each in one toponym. As we know, Esarhaddon’s policy toward Babylonia, from the earliest days of his reign, was conciliatory, quite the opposite of that of his father Sennacherib. Esarhaddon’s overtures to the Babylonians, with an eye to reinforcing his position as king of Babylonia (in parallel with his position as king of Assyria) and forging the two nations into a single unit, are evident in many of his inscriptions; they were expressed by honoring the god Marduk and enhancing his position.49 Thus, Marduk was granted a position of honor in the descriptions of Esarhaddon’s campaigns against Shubria and Egypt. In the “Letter to the God,” the miraculous victory at the battle of Uppume, capital of Shubria, is attributed to Marduk, “King of the Gods,” who had diverted the wind when the besieged Shubrians tried to set ˜re to the siege ramp built by Esarhaddon’s warriors before the city, so that the city wall itself caught ˜re and
48. In the Assyrian royal inscriptions Bel-matati is a designation of the god Ashur (see K. L. Tallqvist, Akkadische Götterepitheta [Helsinki: Societas orientalis Fennica, 1938], 48). 49. See Porter, Images, Power, and Politics: Figurative Aspects of Esarhaddon’s Babylonian Policy (Philadelphia: American Philosophical Society, 1993), passim, esp. 141–42.
111
burnt to the ground.50 The New Account Edition of the march to Egypt through the Sinai Desert states that “Marduk, the great master, [came] to my aid . . . He revived my hosts.”51 In light of this observation, it is remarkable that the tendency to glorify Marduk is not apparent in the changed names of the cities of Shubria and Egypt; as noted, the god Ashur takes pride of place in that context, whereas Marduk’s name is featured in only one toponym (Bit-Marduk). It would seem that the literary account of the military actions in Shubria and Egypt, on the one hand, and the changing of toponyms, on the other, re˘ect two distinct ideological systems, one aimed at the Babylonians in order to win their sympathy, the other, exclusively Assyrian. Hence one should regard the changing of toponyms and the composition of the accounts as actions undertaken in two diˆerent spheres.
50. Gbr. II ii 3–9. In this passage Marduk is referred to as “King of the gods” (line 5), whereas in other passages in the same work “King of the gods” designates Ashur (see Gbr. II i 1, 18). 51. Frt. F rev. 9–10. Labat (followed by Porter) wrongly attributed to Marduk another such act of deliverance—bringing down rain for the thirsty Assyrian soldiers on their march to Egypt (see R. Labat, “Rapports sur les conférences: Assyrien,” Annuaire 1973–1974. École pratique des Hautes Études, IV e section: Sciences historiques et philologiques, (Paris, 1974), 65; Porter, Images, Power, and Politics, 141): The fragmentary inscription 79-7-8, 196, mentioned by Labat (= Borger, Frt. G), is not referring to a miraculous provision of water for the Assyrian army (by Marduk) but to transport of water on camels (as in Frt. F).
CUNEIFORM TABLETS IN THE BRISTOL PUBLIC LIBRARY Michael Jursa (Universität Wien) and Elizabeth E. Payne (Yale University)
from the antiquities dealer Edgar J. Banks.1 During the summer of 2002, the tablets were brought to the Yale Babylonian Collection for conservation and translation, at which time they came to the attention of the authors. The eight tablets published here comprise six Ur III administrative documents and one royal inscription of Sîn-kasid (edited by E. E. Payne), as well as one NeoBabylonian document (edited by M. Jursa). M. Jursa’s research on this article was conducted within the framework of an ongoing project titled “Economic History of Babylonia in the First Millennium B.C.” funded by the Fonds zur Förderung der Wissenschaftlichen Forschung (Vienna). Unpublished texts in the British Museum, the Harvard Semitic Museum, and the Yale Babylonian Collection are cited with permission of the Trustees of the British Museum, P. Steinkeller (Harvard University) and B. Foster and U. Kasten (Yale University) respectively.
The following tablets comprise the cuneiform holdings at the Bristol Public Library in Bristol, Connecticut. The tablets were purchased in 1921
The authors would like to thank the Bristol Public Library, especially the Library’s Director Francine Petosa, for permission to publish these texts, and Mr. Jay Manewitz for his assistance and for bringing the correspondence between the Library and Banks to our attention. We would also like to thank Benjamin R. Foster for allowing us to work on the tablets. In preparation of the Ur III material, E. E. Payne would also like to thank William W. Hallo, Marcel Sigrist, and Walther Sallaberger for their assistance with di¯cult passages. Any mistakes that remain are, of course, our own. Abbreviations are those of W. von Soden, Akkadisches Handwörterbuch (AHw) and The Sumerian Dictionary of the University Museum of the University of Pennsylvania (PSD), with the following exceptions and additions: Bongenaar, Ebabbar = A. C. V. M. Bongenaar, The Neo-Babylonian Ebabbar Temple at Sippar: Its Administration and Its Prosopography, Uitgaven van het Nederlands Historisch-Archaeologisch Instituut te Istanbul 80 (Leiden: Nederlands HistorischArchaeologisch Instituut te Istanbul, 1997), CTMMA = Cuneiform Tablets in the Metropolitan Museum of Art, Ebeling, NBU = E. Ebeling, Neubabylonische Brief aus Uruk (Berlin: im Verlage des Herausgebers, 1930–34), FLP = tablet in the Free Library of Philadelphia, Frame, Fs. Dion III = “A NeoBabylonian Tablet with an Aramaic Docket and the Surety Phrase p¿t sep(i) . . . nasû,” in The World of the Aramaeans III. Studies in Language and Literatur in Honour of Paul-Eugène Dion, eds. P. M. Michèle Daviau, J. W. Wevers and M. Weigl (She¯eld: She¯eld Academic Press, 2001), 100–133, HSM = tablet in the Harvard Semitic Museum, Kümmel, Familie = H. M. Kümmel, Familie, Beruf und Amt im spätbabylonischen Uruk. Prosopographische Untersuchungen zu Berufsgruppen des 6. Jahrhunderts v. Chr. in Uruk, Abhandlungen der Deutschen Orient-Gesellschaft, 20 (Berlin: Gebr. Mann Verlag, 1979), NBC = tablet in the Nies Babylonian Collection (at Yale), NCBT = tablet in the Newell Collection of Babylonian Tablets (at Yale), PTS = tablet in the Princeton Theological Seminary, Sack, CD = R. H. Sack, Cuneiform Documents from the Chaldean and Persian Periods (Selinsgrove: Susquehanna University Press, London and Toronto: Associated University
Presses, 1994), Sigrist, SAT II = M. Sigrist, Texts from the Yale Babylonian Collections, Part I, Sumerian Archival Texts II (Bethesda: CDL Press, 2000), YBC = tablet in the Yale Babylonian Collection. 1. Three letters in the ˜les of the Bristol Public Library, dated between 2 June and 20 June 1921, recount the purchase of these tablets. According to the ˜rst of these letters, Banks oˆered the library a lot containing twenty tablets: the eight tablets published here, plus nine additional Ur III tablets, one from Girsu, one dated, by Banks, to ca. 2250 B.C., and one Neo-Babylonian mathematical text. Due to a con˘ict regarding the price, the Library purchased only these eight tablets and the ˜nal destination of the remaining twelve tablets is not known. Banks continued to do business with the library on at least one other occasion. In a letter dated 2 June 1924, Banks oˆered three “very rare cone-shaped symbols of Istar.” The library did not purchase these objects.
113
JCS 57 (2005)
114
MICHAEL JURSA AND ELIZABETH E. PAYNE ªgiri3º ip-qú-sa sár-ªraº-ab-du 10. iti ezen ªdºme-ki-gál mu dsu-den.zu lugal úriki-ma-ke4 UE ma-da za-ab-sa-ªliºki mu-hul LeE 1 gu4 “One barley-fattened ox was expended from Puzur-Enlil (for a ceremony) at midnight on the seventh day, before Aridubuk, the man of †asrum, entered the city. Nanna-kam, the messenger, was the maskim; (conducted) under the authority of Ipqusa, the sarrabdu. Month 12, †S7. (Total) 1 ox.” Commentary
Tablet 1
1. Ur III Administrative Text from Drehem (41x37x17mm); (†u-Sîn 7/ XII/7) Expenditure of an ox. obv. 1 gu4 niga a-ªri-duº-bu-uk ªlú saº-as-ruki ªdub-sagº iri-a ku4-ra-ni d ªnanna-kamº sukkal maskim 5. ªáº-gi6-ba-a rev. u4 7-kam ki puzur4-den-líl-ta ba-zi
line 2: In addition to this text, Aridubuk of †asrum (RGTC 2, 178–79) is known from three other sources: Sigrist, AUCT I, 414: 3 (AS4/X/8); Fish, Rylands, 455: 6 (†S9/XII/14) (see below); de Genouillac, Bab. 8 (1924), Pl. VII, no. 30:7 (†S2/ XI/24). An Aridubuk without gentilic is also known from SET 297: iii 15 (n.d.). line 3: The meaning of dub-sag in this line is unclear. Alster and Vanstiphout oˆer the general meaning of “˜rst anything” from the literal meaning “˜rst tablet”2 and this general meaning can be found in both administrative and literary sources.3 The same, basic meaning for dub-sag continues to be found in bilingual sources of the ˜rst millennium where dub-sag = mahrû, mahru, muttu and qudmu. The phrase under consideration here (dub-sag iri-a ku4-ra-ni) can be found in four additional texts: Fish, Rylands 455: obv. 7 (†S9/XII/14; a text also involving Aridubuk of †asrum), JCS 10, 28, no. 5: rev. 6 (IS1/II/25), FAOS 16, 1172: obv. 6 (†S6/ VII/4), Sigrist, SAT II, 316: obv. 6 (†43/I/9).
2. B. Alster and H. Vanstiphout, ASJ 9 (1987) 34. 3. To cite one example from each genre, see, bala dub-sag = “˜rst term of duty” (PSD B, 66; Ur III) and, more clearly, Ersemma No. 97 line 40u and following, (M. E. Cohen, Sumerian Hymnology: The Ersemma [Cincinnati: Hebrew Union College, 1981], 76, 81; V. Scheil, RA 8 [1911] 164, OB), where gal5-lá dub-sag amas-e ku4-r[a-e . . .] “The ˜rst demon to enter the sheepfold . . .” is followed by gal5-lá min-kam-ma amas-e ku4-r[a-e . . .] (l. 41u), gal5-lá es5-kam-ma amas-e ku4-ra-e (l. 42u) and so forth through seven demons.
CUNEIFORM TABLETS IN THE BRISTOL PUBLIC LIBRARY
In treating this line, Sollberger translates: “as he entered the City for the ˜rst time.”4 I would propose, however, that dub-sag iri-a ku4-ra-ni might better be understood, in this context, as “before he entered the city.” This suggestion is supported by the later Akkadian equivalents of dub-sag and by the fact that Aridubuk is known to have entered the city on at least two occasions (Bristol 1; Fish, Rylands 455). The phrase, however, merits further discussion.
115
2. Ur III Administrative Text from Umma (49x48x17mm); (Ibbi-Sîn 1/ XII/-) Delivery of reed bundles. obv. 1.0.0 sa-gi gu-kilib-ba 13 sa-ta ki a-du-mu-ta mu-DU sà é-te-na-ka rev. 5. kisib inim-dsára iti ddumu-zi mu di-bí-den.zu lugal Seal: inim-dsára, dub-sar, dumu lugal-iti-da
4. E. Sollberger, JCS 10 (1956) 19 and notes to line 20.
“3600 reed bundles, with 13 bundles per bale, from Adumu. Delivery in Etena. Sealed by Inimsara. Month 12, IS 1. Seal: Inim-sara, the scribe, son of Lugal-itida.”
Tablet 2 Tablet 3
116
MICHAEL JURSA AND ELIZABETH E. PAYNE
3. Ur III Administrative Text from Drehem (36x30x16mm); (†ulgi 44/ VIII/11) Withdrawal of animals. obv. 6 udu niga níg-kù du6-kù-ga sà é dªen-lílº-lá 2 udu-niga 1 ªsila4?º 5. dªen-lílº 2 udu niga 1 ªsila4º rev. dnin-ªlíl-lẠd ªnanseº-usum-gal ªmaskimº ªitiº u4 11 ba-zal 10. ªzi-gaº ki na-ªlu5º ªiti suº-es-sa mu si-mu-ru-ªumkiº ù lu-lu-bu-fiumflki ªaº-rá 10 lá-1 kam-as ªba-hulº “6 barley-fattened sheep . . . for the Holy Hill, in the temple of Enlil, 2 barley-fattened sheep and 1 lamb (for) Enlil, 2 barley-fattened sheep and 1 lamb (for) Ninlil (were) withdrawn from Nalu. Nanse-usumgal (was) the administrator. The 11th day of the month having passed. Month 8, †44.” Commentary line 2: No parallels for this line are known to me and it seems likely that níg-dab5 du6-kù-ga is intended. The writing of kù for ku (= dab5) would then be a scribal error in anticipation of the kù following almost immediately thereafter.
Tablet 4
4. Ur III Messenger Text from Umma (32x27x13mm); (Ibbi-Sîn 2/XI/1) The messenger text published here belongs to Group L, as identi˜ed by R. McNeil, “The “Messenger Texts” of the Third Ur Dynasty” (Ph.D. Diss., University of Pennsylvania, 1970), 168–70; see also, F. D’Agostino and F. Pomponio, Umma Messenger Texts in the British Museum, Part I. NISABA I (Rome: Dipartimento di Scienze One Line Short
CUNEIFORM TABLETS IN THE BRISTOL PUBLIC LIBRARY
117
dell’Antichità dell’Università degli Studi di Messina, 2002), 16. It is one of the latest known examples of this type. obv. Ilú-dba-ba6 I su-ªéº-a I lú-dnin-ªsuburº I puzur4-ªéº-a I 5. ur-dab-ba-ú I ur-dba-ba6 I lú-diskur Id en-líl-lá-ì-sa6 sukkal-me-és 10. Isu-dIDIM I lú- den-líl-lá LoE Iur-dnin-gír-suki I lú-dingir-ra rev. Isu-dnin-subur 15. 3 sìla kas 2 sìla ninda 3 gín ì 2 gín ªnagaº 5 gín sum-ta su-nígin 3 bán 7 sìla kas-sig5 2 sìla [kas]gin su-nígin 5 sìla ninda zì-sig15 [(x)] ª2 bánº 1 sìla ninda su-nígin !- sìla 2 9 gín ì ªsu-nígin !- º sìla 3 6 gín naga su-nígin 1 sìla 5 gín sum 20. u4 1-kam iti pa4-ú-e mu en dinanna más-ªeº ì-pàd “Lu-Baba, †u-Ea, Lu-Ninsubur, Puzur-Ea, UrAbba’u, Ur-Baba, Lu-Iskur, and Enlil-isa, the messengers; and †u-IDIM, Lu-Enlil, Ur-Ningirsu, Lu-dingir, and †u-Ninsubur (received) 3 sila beer, 2 sila bread, 3 gin oil, 2 gin naga-herb, and 5 gin onions each. (For a) total (of): 3 ban 7 sila beer of good quality and 2 sila beer of normal quality; 5 sila bread of “ground ˘our” and 2 ban 1 sila bread; !- sila 2 9 gin oil; !- sila 3 6 gin naga-herb; 1 sila 5 gin onions. Day 1, month 11, IS2.”
Tablet 5
Commentary line 5: A parallel to this name was kindly brought to my attention by M. Sigrist. In YBC 16034 (unpubl.) obv. 6u we ˜nd the name Ilú-dabba-ú.
5. Ur III Administrative Document (82x58x24 mm); (Amar-Sîn 9/-/-) This document is a summary report of wool delivered to the city of Umma. The wool recorded
118
MICHAEL JURSA AND ELIZABETH E. PAYNE
was used to weave textiles of various types and qualities. For the textile industry during the Ur III period, see H. Waetzoldt, Untersuchungen zur neusumerischen Textilindustrie, Studi Economici e Tecnologici 1 (Rome: Centro per le Antichità e la Storia della’Arte del Vicino Oriente, 1972). obv. i. 17!- ma-na 2 síg túg nì-lám 3-kam-ús 2 gú 12!- ma-na 2 síg túg-guz-za 3-kam-ús 45!- ma-na 2 síg túg nì-lám 4-kam-ús 25!- ma-na 2 síg túg nì-lám gin 5. 3 gú 30 lá-1 ma-na síg túg-guz-za 4-kam-ús 3 gú 40@- ma-na 3 síg túg-guz-za-gin 1 gú 15 ma-na síg túg us-bar su-nígin 12 gú 5@- ma-na 3 sí[g] [u]r-ru na-[gada] 10. [x+] ª1º!- ma-ªnaº 2 [síg túg] [nì]-lám [x-kam-ús] (rest of column 1 broken) ii. 15 m[a-na síg túg] nì-lám ª4º-[kam-ús] 12!- ma-na 3 s[íg túg nì]-lám gin 1 gú 53 ma-na 10 gín síg túg-guz-za 4-kam-ús 1 gú 46@- ma-na 3 síg túg-guz-za gin 5. 31@- ma-na 3 síg túg us-bar su-nígin 6 gú 43 + [x m]a-na [síg] lú-dªutuº [na-gada] síg udu ªé?º-[x x] 13 ªmaº-[na síg x] (rest of column 2 broken) (beginning of column 3 broken) rev. iii. ª40 ma-na 10 gº[ín síg túg]-guz-za 4-ka[m-ús] 2@- ma-na 3 síg túg nì-lám [x x] 36!- ma-na 2 síg túg-guz-za gin 18 ma-na síg túg us-bar 5. su-nígin 1 gú 45%- ma-n[a 6 síg] ur-ru na-[gada] 4 ma-na s[íg túg nì]-lám [x x] 17!- m[a-na 2 síg túg]-guz-z[a x] [(x+)] ª4@- ? maº-[na 3 síg túg x] (beginning of column 4 broken) iv. [su-nígin x+] ª4º!- m[a-na 2 síg] d lú- utu [na-gada] síg udu ªénsiº-[ka] 27!- ma-na 2 síg udu bar-su!(=kù)-ga
Two Lines Short
5u. ur-dma-mi na-gada su-nígin 33 gú 1!- ma-na 2 síg-kur-ra hi-a síg-kur-ra mu-DU sà ummaki mu en dnanna kar-zi-daki ba-h[un] (column i) “17!- mina 2 wool for a nilam-garment of the third quality; 2 talents, 12!- mina 2 wool for a guzza-garment of the third quality; 45!- mina 2 wool for a nilam-garment of the fourth quality; 25! - mina 2 wool for a nilam-garment of normal quality; 3 talents 29 mina wool for a guzzagarment of the fourth quality; 3 talents 40@- 3 mina wool for a guzza-garment of normal quality; 1 talent 15 mina wool for an us-bar-garment. Total: 12 talent 5@- mina 3 wool (from) Urru, the shepherd. [x +] 1!- mina 2 wool for a nilam-garment [of the x quality] . . . (rest of column broken). (column ii) “15 mina wool for a nilam-garment of the fourth quality; 12!- mina 3 wool for a nilamgarment of normal quality; 1 talent 53 mina 10 shekels wool for a guzza-garment of the fourth quality; 1 talent 46@ - mina 3 wool for a guzzagarment of the normal quality; 31@- mina 3 wool for an us-bar-garment. Total: 6 talent 43 [+x] mina wool (from) Lu-Utu, the shepherd, wool of the sheep of the . . . house. 13 mina wool . . . (rest of column broken). (column iii; beginning broken) “. . . 40 mina 10 gin wool for a guzza-garment of the fourth quality; 2@- mina 3 wool for a nilam-garment of the [x quality]; 36!- mina 2 wool for a guzza-garment of normal quality; 18 mina wool for an us-bargarment. Total: 1 talent 45% - mina 6 wool (from) Urru, the shepherd. 4 mina wool for a nilamgarment [of the x quality]; 17!- mina 2 wool for a guzza-garment [of the x quality]; [(x+)] 4@- (or 3 %- ?)6 m[ina wool for an x garment of the x quality]. (column iv; beginning broken) “. . . [Total: x+] 4!- 2 m[ina wool] (from) Lu-Utu, the shepherd; wool of the sheep of the governor; 27!- mina 2 wool from the bar-su-ga-sheep; Ur-Mami, the shepherd; Total: 33 talents 1!- mina 2 wool of the mountain (sheep); the wool of the mountain (sheep) delivery in Umma. AS 9.”
CUNEIFORM TABLETS IN THE BRISTOL PUBLIC LIBRARY
Commentary line ii 7: Compare SET 275, 17–18 (Umma, AS 6/—/—) where Lu-Utu (na-gada) is among the shepherds collectively referred to as the “shepherds of the mountain sheep” (SET 275, 24: ki sipa udu-kur-ra-ke4-ne-ta) and MVN 16, 965: 16–17 (Umma, AS 6/—/—) where Lu-Utu is described as a shepherd of the mountain sheep (na-gada / udu-kur-ra). Such characterizations are in keeping with the summary line of this tablet (i.e., delivery of wool of the mountain (sheep)), but Lu-Utu’s association with the wool of the governor in iv 2u–3u is unexpected. line iv 5u: See MCS 8/4 (1959) 93 (unnumbered), rev. col. i, line 8 for ur-dma-mi na-gada mentioned in a similar text (Umma, †S 1/—/—).
119
6. Ur III Administrative Text from Drehem (29x27x14mm); (Amar-Sîn 5a/I/5) Receipt of animals. obv. 1 dara4? nit[a2] 1 gukkal 1 más-gal 1 u8 5. 1 sila4 1 más-ga rev. ba-ús u4 5-kam ki a-hu-we-er-ta d sul-gi-iri-mu 10. su ba-ti iti mas-dù-kú mu ús-sa en dnanna ba-hun LeE 6 udu “†ulgi-iri-mu received 1 male ibex, 1 fat-tailed sheep, 1 full-grown goat, 1 ewe, 1 lamb, and 1 suckling goat, (all) dead, from Ahu-Wer. Day 5, month 1, AS 5. (Total) 6 sheep.” 7. Building Inscription of Sîn-kasid (56.8x45.1x20.0mm) (RIME 4.4.1.2; not copied) obv. dEN.ZU-kà-si-id nita kala-ga lugal unuki-ga lugal am-na-nu-um rev. é-gal nam-lugal-la-ka-ni mu-dù “Sîn-kasid, mighty man, king of Uruk, king of the Amnanum, built his royal palace.”
Tablet 6
120
MICHAEL JURSA AND ELIZABETH E. PAYNE 15 sá Idu.gur-tinit se-pir-ri u lúumbisag Idutu-mu-ùru a-sú sá I sil-la-a a Isu-ªtu?º-ú u.e. tin.tirki i[t]i.ªdu6º ud.17.kam mu.sag.nam.sagsic.lugal Id 20 u.gur-lugal-ùru lugal tin.tirki l.e. e-lat !- ma.ªnaº 2 ªsámº zú.lum.ma xxx “One and a half minas of silver, owed to PanÿBel-adaggal, steward of the courtier Bel-sarruusur, for work on the Pal(a)gu canal. He (PBA) has been paid at the house of the courtier Bel-sarruusur by Nergal-dan, son of Mukÿn-zeri, from the silver which is at the disposal of Nergal-dan, son of Mukÿn-zeri, and (of) the palace scribe. “Witnesses: Nabû-mÿtu-uballit, the courtier; Bel-dÿn¿åa-epus, the courtier; Nabû-bel-ilÿ, son of Arad-Nabû; Iqÿsaya, son of Nergal-uballit, the alphabet scribe; and the scribe: †amas-sumu-usur, son of Sillaya of the †utû family. “Babylon, 17th day of Tasrÿtu, accession year of Nergal-sarru-usur, king of Babylon. “This is in addition to half a mina (of silver), the price of the dates . . . (one word illegible).” Tablet 8
8. Neo-Babylonian receipt (Neriglissar 0/VII/17) 1 1!- ma.na 2 kù.babbar I d sá igi- ªenº-[a]-dag-gal lúgal é sá Iden-lugal-ùru lúsag lugal a-na dul-lu ídpa-la-gu 5 ina muh-hi é Iden-lugal-ùru lú sag lugal ina ªkùº.babbar sá ina igi Id u.gur-ªda-aº-nu a-sú sá Iªduº-[nu]mu[n] u ªlúºumbisag é.gal l.e. ina suII Idu.gur-ªdaº-a-nu 10 a-sú Idu-numun ma-hi-ir rev. lúmu-ªkinº-nu Idag-ús-tinit lú sag lugal Iden-di-nu-a!-e-pu-us lú sag lugal Idag-en-dingirmes a-sú sá Iìr-dag Iba-sá-a a-sú
The interpretation of this text must focus on the two protagonists—Bel-sarru-usur and Nergaldan, who are known from other sources,5 and on the o¯ce of the “palace scribe,” tupsar ekalli. A list of attestations of named holders of this title is given in Table 1. Royal o¯cials designated as “palace scribe” (tupsar ekalli) are mentioned in several archives from both northern and from southern Babylonia. In the Eanna archive, not surprisingly, they appear most often as agents of the crown.6 The attestations 5. Among the other men mentioned in the text, only the witness in line 11, the sa resi Nabû-mÿtu-uballit, can be identi˜ed with certainty in another text: YOS 19, 288, Nabonidus 3. 6. YOS 3, 132: a palace scribe sends men to measure (royal) ˜elds and assess the harvest; it is explicitly stated that the ˜elds of Eanna not be included. NCBT 53: sheep delivered to a palace scribe. YOS 17, 207: a certain Ubar-Nabû, (royal) o¯cial (sa-resi) and messenger of a palace scribe, receives silver. Similarly, in Sack, CD 16, unnamed messengers of the sakin mati
CUNEIFORM TABLETS IN THE BRISTOL PUBLIC LIBRARY
121
Table 1 †amas-ibni Arad-Nabû Arad-Nabû †amas-sarru-usur †amas -sarru-usur Madan-sarru-usur Madan-sarru-usur
Nabopolassar 7/III/14 Nebuchadnezzar 15/IV/6 Nebuchadnezzar 18/II/13 Amÿl-Marduk 0/VII/15 Amÿl-Marduk 0/VII/20 Nabonidus 14 fiNabonidus 15–Cyrus 7fl10
NBC 4514 (Eanna archive) Sack, CD 167 (Eanna archive) JCS 54, 117 YBC 3926 (Eanna archive) PTS 2046 (Uruk; dossier of †umu-ukÿ(n)8 BIN 2, 109 (Uruk; dossier of †umu-ukÿ(n) BM 59098 (Ebabbar archive)9 CT 22, 242 (Ebabbar archive)
7 8 9 10
of named palace scribes listed above are unusual; more often these men remain unnamed, perhaps because their names, like those of high o¯cials, are either well-known or deemed irrelevant since the o¯cial title is su¯cient for the purpose of identi˜cation. They have high social standing,11
and the palace scribe Arad-Nabû are being issued silver too. A palace scribe is also mentioned in TCL 9, 76:33 (collated) in a damaged context. 7. Sack’s reading of the name as Musezib-Bel is wrong. Beaulieu, in his discussion of Arad-Nabû in YBC 3926, notes the sepÿru sa mar sarri of that name known from Nebuchadnezzar’s so-called Hofkalender (“Ea-dayyan, Governor of the Sealand, and Other Dignitaries of the Neo-Babylonian Empire,” JCS 54 [2002], 11857 ), but adds that this probably is another individual. The name is very common. An additional reference to Arad-Nabû (without title) can probably be found in YBC 4146, see below. 8. The text was identi˜ed by B. Jankovic3. 9. The text reads as follows (transliteration of C. Waerzeggers): 4 gur 4 pi se.bar es-ru-ú, sá Iddi.kud-lugal-ùru, lúdub.sar é.gal, a-na é.babbar.ra it-[ta-din], [. . .] ªx xº [. . .] (rest of obv. and beginning of rev. lost) rev. 1u: ªmuº.14.kam Idag-i, lugal tin.tirki (followed by the drawing of a symbol, resembling those discussed by Jursa in Zawadzki and Jursa, “†amas-tirri-kusur, a smith manufacturing weapons in the Ebabbar temple at Sippar,” WZKM 91 [2001] 36023 and 36226). Bongenaar, Ebabbar, 106 suggests (on the basis of CT 22, 242) that the palace scribe Madan-sarru-usur be identi˜ed with a courtier of that name attested more often in the Ebabbar archive. This is possible, but not absolutely certain, given the small degree of variation in the onomasticon of royal o¯cials. A tithe paid by the courtier Madan-sarru-usur is attested in CT 55, 349. 10. The date is based on the identi˜cation of the addressee Marduk-sumu-iddin with the sangû Sippar of that name: Bongenaar, Ebabbar, 29.106. 11. NBC 4514: a palace scribe appears as witness in a contract dealing with temple aˆairs. He heads the witness list and thus takes precedence over the men named after him. These are the qÿpu of Eanna, the sakin temi of Uruk, and the satammu of Eanna, the highest-ranking o¯cials based at Uruk at the time. The text is cited here with the permission of L. Pearce who will publish it in a forthcoming YOS volume.
and most often act through agents, only important tasks warranting their personal presence.12 They are occasionally involved in legal matters,13 and Eanna o¯cials turn to them for pertinent information.14 One letter from the Ebabbar archive shows a tupsar ekalli inspecting the staˆ of the royal resident (qÿpu) in Ebabbar.15 The same man is known to have paid a barley tithe to the temple, similarly to other royal o¯cials.16 Sometimes these o¯cials are either responsible for, or owners of, large estates. On one occasion, the palace scribe †amas-sarru-usur was owed substantial amounts of barley, apparently the yields or rents of ˜elds in {urrabatu, by the well-known entrepreneur †umu-ukÿn.17 The business relationship
12. JCS 54, 117 (YBC 3926): the palace scribe Arad-Nabû personally delivers large amounts of silver from the palace to the temple. This o¯cial is probably to be identi˜ed with the Arad-Nabû mentioned in YBC 4146. This is a contract dealing with the ˜nancing of the manufacture of bricks intended as a contribution of Eanna to Nebuchadnezzar’s palace building project in Babylon. Arad-Nabû is the only protagonist in this text who is not identi˜ed with a patronymic: he must have been well known. 13. YOS 3, 109: a palace scribe has talked to the sakin temi regarding the aˆairs of Nabû-belsunu and has brought them to a conclusion. TCL 9, 107: a palace scribe requests the handingover of a certain man’s property. Contrary to Ebeling, NBU, p. 265, and Kümmel, Familie, 121, Nabû-sumu-ukÿn, son of Nabû-ahhe-sullim, in TCL 9, 107: 9f. is not to be identi˜ed with the palace scribe mentioned in line 11 of that letter. The passage reads: “(regarding) Nabû-sumu-ukÿn, son of Nabûahhe-sullim: the palace scribe has sent (him) to you (il-tap!rak!-ka), saying: . . .” 14. TCL 9, 93: a messenger is sent by the fisatammufl to a palace scribe to make inquiries. 15. CT 22, 242. 16. BM 59098 (above note 9). 17. BIN 2, 109. See F. Joannès, Textes économiques de la Babylonie récente, Études assyriologiques cahier no. 5 (Paris, Éditions Recherche sur les Civilizations, 1982), 130–41. On the
122
MICHAEL JURSA AND ELIZABETH E. PAYNE
of these two parties is clari˜ed by an undated administrative text dealing with the division of barley between †umu-ukÿn and an unnamed palace scribe.18 The barley is the yield of ˜elds cultivated by plough teams under the supervision of †umu-ukÿn. From this one can deduce that the land in question was the palace scribe’s estate, or was at least managed by him. An “estate of the palace scribe,” bÿt tupsar ekalli, is attested in the Ebabbar archive. It is mentioned in three tithe lists19 and has to be added to the list of estates of important state o¯cials subject to paying tithes to the Ebabbar.20 Similarly, evidence from the Nûr-Sîn-archive proves the existence of a bÿt tupsar ekalli on land belonging to Ezida.21 Finally, there is a “canal of the palace scribe,” nar-tupsar-ekalli, which is attested in a Murasû text.22 This is reminiscent of other canals named after high o¯cials: nar-sîn-magir, nar-masenni. Cumulatively it seems that the title more likely refers to a high royal o¯cial possibly equal in rank
early phase of †umu-ukÿn’s career. PTS 2046 concerns the same aˆair. †umu-ukÿn will be re-studied by B. Jankovic3 in a forthcoming contribution based on unpublished material. 18. NBC 4569. 19. BM 42607 (M. Jursa, NABU 1998/70), 62436 and 101262. 20. See M. Jursa, Der Tempelzehnt in Babylonien vom siebenten bis zum dritten Jahrhundert v. Chr., Alter Orient und Altes Testament 254 (Münster: Ugarit Verlag, 1998), 93–97: Bÿtmar-sarri, Bÿt-mukÿl-appati, Bÿt-Nabû-naåid (in a text from the accession year of Nabonidus), Bÿt-rab-mungi, Bÿt-rab-qannate, Bÿt-sar-babili, and M. Jursa, NABU 1998/70: Bÿt-rab-kari, Bÿtrab-urate, Bÿt-mar-ekalli. Dar. 198 mentions a sprawling uru Bÿt-tupsar-ekalli, parts of which near the Nar-sarri and the villages Gil¿su and Bÿr-ili were cultivated by ploughmen of the Ebabbar: M. Jursa, Die Landwirtschaft in Sippar in neubabylonischer Zeit, Archiv für Orientforschung, Beiheft 25 (Wien: Institut für Orientalistik, 1995), 212a. Barley from the Bÿt-tupsar-ekalli is also mentioned in BM 64077, Darius 2/ VIII/5. 21. C. Wunsch, Die Urkunden des babylonischen Geschäftsmannes Iddin-Marduk, Cuneiform Monographs 3 (Groningen: Styx, 1993), No. 214 (M. Jursa, Review of Wunsch, IddinMarduk, AfO 42/43 [1995/96], 261a). The unpublished text HSM 1890.4.5 shows a slave of Iddin-Marduk//N¿r-Sîn cultivating a date garden in the ˜eld(s) of the palace scribe (ina a.sà lúumbisag é.gal); the dates he has to pay as rent are referred to as níg.ga lugal man sú: this is certainly a mistake for níg.ga dag . . . “property of Nabû, the lord of the universe.” 22. RGTC 8, 387.
with men like the treasurer (rab kasiri) or the sakin temi, and not to ordinary scribes working for the palace. Such men are referred to as tupsarr¿ sa ekalli in the letter TCL 9, 98 from the Eanna archive:23 this seems to be a simple designation for palace clerks and not a formal title like tupsar ekalli.24 The background for the appearance of the palace scribe in the present text cannot be ascertained with certainty. I would assume that the business relationship between the agricultural entrepreneur †umu-ukÿn (who was well connected with the royal administration) and the palace scribe †amas-sarru-usur discussed earlier may have been similar to that of the anonymous palace scribe and Nergal-dan. At the least it can be demonstrated that Nergal-dan was a businessman involved in similar activities as †umu-ukÿn, albeit on a much smaller scale. Nergal-dan, son of Mukÿn-zeri of the Sîn-leqeunninnÿ family, is one of the main protagonists of the dossier to which this text belongs. Other family members attested are his sons †amas-ahu-iddin and †amas-reåûsunu, his brother Bel-leåi and his nephews Nadin and Gimillu, sons of Bel-leåi.25 They are mentioned in Eanna texts and private contracts—the remnant of their archive—which were in all likelihood found intermixed with the Eanna temple archive during the large-scale illicit excavations at Uruk at the beginning of the past century. These men had manifold connections to 23. The lúumbisagme sá é.gal request the Eanna pay barley owed to the royal administration. 24. Yet another type of scribe is represented by the “scribes of the kings:” these were notaries responsible (i.a.?) for drafting sale contracts for real estate. See H. D. Baker and C. Wunsch, “Neo-Babylonian Notaries and Their Use of Seals,” in Seals and Seal Impressions. Proceedings of the XVL e Rencontre Assyriologique Internationale, Part II, eds. W. W. Hallo and I. J. Winter (Bethesda: CDL, 2001), 200. 25. See Frame, Fs. Dion III, 103–33. In addition to the texts mentioned by Frame, Nergal-dan is a protagonist in the following tablets: SAKF 135+145 and 160 (see below), Sack, CD 98 (= Jursa, NABU 1998/124), YBC 3869, YOS 19, 31, 67 and 312, possibly also VS 20, 68. For †amas-ahu-iddin see also Sack, CD 76, GC 1, 343, YOS 6, 91, 102, 104, 132, 162 and 197, YOS 19, 68, 73, 56, 312 and CTMMA 3, 103. The alleged family name . . . a I mu-dutu given in the edition of CTMMA 3, 103 is incorrect. The passage must in fact be read sá! mu.an.na! 1! gín “for 1 shekel per year.” The text is closely related to YOS 19, 68.
CUNEIFORM TABLETS IN THE BRISTOL PUBLIC LIBRARY
Eanna. Nergal-dan was a member of the inner circle of the temple community by virtue of the fact that he owned a prebend of butcher’s service before Urkaåÿtu jointly with his brother Bel-leåi,26 Bel-leåi’s son Gimillu was a prebendary baker,27 and both Bel-leåi and his other son Nadin had business contacts with the temple.28 Nergal-dan’s son †amas-ahu-iddin acted for the temple on several occasions and is once attested as “messenger of the bel-piqitti of Eanna.”29 Most of the private texts concern rather smallscale transactions, probably with an entrepreneurial background. The majority of the documents are debt notes for dates, barley and/or silver.30 According to these texts, some of the family’s activities were aimed at the commercial exploitation of the agricultural production in the hinterland of Uruk. Nergal-dan rented a ˜eld there,31 and he, Gimillu and Nadin dealt in kasia (cuscuta?) selling it to individuals32 and maybe also to the temple.33 †amas-ahu-iddin was also 26. NBDMich. 62. See P.-A. Beaulieu, The Pantheon of Uruk During the Neo-Babylonian Period, Cuneiform Monographs 23 (Leiden/Boston: Brill/Styx, 2003), 261. Cf. also PTS 2061, published by Frame, Fs. Dion III, 101–33, a contract from the Eanna archive involving Nergal-dan and one of the highest o¯cials of Eanna. 27. AnOr 8, 22. 28. GC 1, 95: Bel-leåi buys wool from the temple; BIN 1, 104: Nadin owes the temple 200 kor of dates as a compensation for the same amount of kasia which he must have failed to deliver. He also appears as a witness in BM 114468. 29. YOS 6, 119. Cf. GC 1, 343, YOS 6, 162, YOS 19, 56 and the small dossier YOS 6, 102. 132. 197 and Sack, CD 76 (lastly M. Jursa, NABU 1996/99) which shows †amas-ahu-iddin collecting outstanding debts owed to the temple. 30. Nergal-dan and his relatives appear both as creditors (e.g., YOS 19, 31), and as debtors (e.g., YOS 6, 164). Note SpTU 2, 56 (Amÿl-Marduk 0/VII/28), in which Nergal-dan’s nephew Nadin appears as one of six men owing more than two minas of silver. The text, which probably has a business partnership as background, is unusual in that it was excavated during the German campaigns of 1970/71 but has clear prosopographic links to the Eanna material found at the beginning of the twentieth century. 31. Sack, CD 98 = NABU 1998/124. 32. YOS 6, 105. As concerns Nergal-dan, there is an unpublished tablet fragment in the Florentine Museo Archeologico numbered 94028, dated Nabonidus 3/V/22, in which Nergaldan is owed one “measure” (1et mi-is-ha-ta) of kasia. The debtor’s name is lost. 33. The background of BIN 1, 104 (above note) could be a sale.
123
owed bricks34 and wooden beams,35 which may suggest an involvement in construction work. He certainly sold textiles36 and maybe also had them manufactured.37 The two branches of the family shared business interests, which, however, led to problems. Nergal-dan and his nephew Nadin had entered into a formal partnership by the ˜fth year of Nabonidus at the latest.38 It was probably in connection with this partnership that Nergaldan had to stand surety for the payment of a debt owed by his nephew one year later.39 The partnership was still functioning the following year, according to an unpublished text,40 but must have run into di¯culties a little later, probably after the death of Nergal-dan. A series of surety agreements from the eighth and tenth year of Nabonidus documents attempts by †amas-ahu-iddin to collect debts owed to him by his cousins Nadin and Gimillu.41 Among the tablets from the archive, those most closely connected to the present text are two texts now housed in the Museo Archeologico in Florence and published (inadequately) about ˜fty years ago by Oberhuber in SAKF. In October 2003, I collated these tablets with the kind permission of the curator, M. Cristina Guidotti, and joined SAKF 135 to 145. In the following editions, new readings resulting from collation are not marked explicitly.
34. YOS 6, 104 and YOS 19, 73. 35. BIN 2, 123. 36. YOS 6, 91. 37. This is suggested by YOS 19, 68 and CTMMA 3, 103, wherein †amas-ahu-iddin hires a fuller for a certain period of time. 38. YOS 19, 67, Nabonidus 5/X/5. In this text they receive silver for a business venture from a certain †amas-bar. 39. YOS 6, 48, Nabonidus 6/III/27. 40. PTS 3056 (Nabonidus 7/VIII/2; non vidi), according to Frame, Fs. Dion III, 103. This text includes a reference to another son of Nergal-dan, †amas-reåûsunu. 41. JCS 9, 25 (Nabonidus 8/IV/17), YOS 6, 119 (Nabonidus 8/VI/4), TCL 12, 96 (Nabonidus 10/VIII/20), YOS 6, 153 (Nabonidus 10/IX/22). See J. W. Snyder, “Babylonian Suretyship Litigation: a Case History,” JCS 9 (1955) 25–28, and Frame, Fs. Dion III, 104. YOS 19, 312, also from the tenth year of Nabonidus, concerns the settlement of accounts between †amasahu-iddin and the son of another debtor of his father.
124
MICHAEL JURSA AND ELIZABETH E. PAYNE
SAKF 135+145 [the new museum numbers, not given by Oberhuber, are 93970 and 93980] 1 ina 10 ma.na kù.babbar sá Iden-lugal-ùru lúsag lugal a-na Ire-mut a-sú sá Ire-fimut fl?-ªdutuº Id u.gur-da-an [a-sú sá Idu-numun] 5 [u Id]ªagº-[e]n-ªsú-nu a-súº s[á Inumun?si.s]á? a-na kaskalII id-din-nu ina lìb-bi sá 1 ma.na kù.babbar se.bar a-ki-i ki.lam sá ªiti.xº [s]e.bar ina tin.tirki 10 ªiº-na-ás-si rev. i-di u gi-mir-ri ina ka-ªreº-sú-nu ú-tar-uå lú mu-kin-nu Idag-sesme-mu a-sú sá IGI† xx [x x] 15 a Iden [ ] lú umbisag Ida[g?-x x a-sú sá] [Il]i-si-ru tin.[tirki] iti.sig4 ud.1.kam mu.2.kam lú-damar.utu lugal tin.tirki 20 sá 15 gín kù.babbar mi-res-t[ú] l.e. [u]l-tu [ka-r]e-sú-nu [I]d u.gur-[d]a-an i-na-ás-si “Of the 10 minas of silver which the courtier Bel-sarru-usur has given to Rem¿t, son of Rem¿t?†amas, Nergal-dan, [son of Mukÿn-zeri], and Nabûbelsunu, [son of Zeru?]-lÿsir, for a business venture, thereof he (B†U) shall receive in Babylon barley for 1 mina of silver, according to the price of month [. . .]. (The debtors) shall repay the wages (for porters) and the transport costs from their common account. (Witnesses). Babylon, ˜rst day of Simanu, year 2 of Amÿl-Marduk, king of Babylon. Nergal-dan shall take trade goods for 15 shekels of silver from their common account.”
SAKF 160 [museum number 93995] 1 3 ma.na 55 gín kù.babbar sá Iden-lugal-ùru lú sag lugal ina muh-hi Idu.gur-da-an a-sú sá Idu-numun u Idag-en-sú-nu a-sú sá Ili-si-ru a-na mi-res-ti 5 i-nam-di-nu sal-sú sá mi-res-ti-sú Id ag-en-sú-nu ú-za-aå-az-ma ina iti.ab i-na-ás-si 1 ma.na 18!- gín 2 kù.babbar u ur5.ra-sú ina muh-hi ma-né-e 12 gín kù.babbar 10 2 gín kù.babbar ina ka-re-e-sú-nu lo.e. nu-up-tu4 i-nu-up-pu-uå (witnesses; l. 14: Isu-la-a) 19 urubàd-sá-garimé-da-ku-ri 20 iti.apin ud.26.kam mu.sag.nam.lugal.la Id u.gur-lugal-ùru lugal tin.tirki “3 minas, 55 shekels of silver, owed to the courtier Bel-sarru-usur by Nergal-dan, son of Mukÿn-zeri, and Nabû-belsunu, son of Lÿsiru: they shall use (it) for trade goods. Nabû-belsunu shall take a third of his trade goods as his share in the month of Tebetu, (that is, goods worth) 1 mina, 18!- 2 shekels of silver, and its interest, 12 shekels per mina. From their common account, they shall pay 2 shekels of silver as an additional payment. (Witnesses) D¿ru-sa-Bÿt-Dak¿ri, 26th day of Arahsamnu, accession year of Nergal-sarru-usur, king of Babylon.” According to the two Florentine documents, Bel-sarru-usur had supplied several Urukean entrepreneurs, among them Nergal-dan, with silver to be used for trade ventures.42 The objective, or 42. Note that according to the unpublished tablet YBC 3869 Nergal-dan and Nabû-belsunu entered into a business partnership with another royal o¯cial in the ˜rst year of Neriglissar.
CUNEIFORM TABLETS IN THE BRISTOL PUBLIC LIBRARY
at least one of the objectives, of the business partnership is mentioned explicitly only in SAKF 135+: to supply barley to the capital, Babylon. The new text published above adds a new aspect in that it demonstrates that Nergal-dan and Belsarru-usur cooperated not only in trade, but apparently also in agricultural production: this is the most plausible interpretation for their joint
125
responsibility for the work on the Pal(a)gu canal.43
43. The word palgu, “canal,” is apparently used here as a designation of a speci˜c watercourse. The same canal may be meant in FLP 1531 (Eanna archive, Nebuchadnezzar 16/VI/ 13), where bricks are to be transported from the Girû-district (ul-tu garimgi-ru-ú) to the opening (ká) of the pa-al-gu. This text was also written in Babylon.
A NOTE ON THE FIRST REGNAL YEAR OF PHILIP III (ARRHIDAEUS) Edward M. Anson (University of Arkansas at Little Rock)
Chronicle’s eight regnal years for Philip III is to assume that the cuneiform text began Philip’s reign in 324.4 He argues that Alexander made his brother “King of Babylon” in 324. This contention was eˆectively challenged by Tom Boiy, who argued convincingly through a comparison of the various surviving king lists and chronicles from Babylonia, that the listing of both a seventh and an eighth year was most probably the result of posthumous dating.5 Conclusive evidence for the veracity of Boiy’s position is found in a Babylonian astronomical text. BM 34093 states, “Year 2 of [king] Philip, [month I]; . . .” (obv. 1) “The 27th, moonrise to sunrise: 17o. The 28th, around 3o before sunset, solar eclipse . . . gusty west and north wind; it set eclipsed. . . .” (rev. 23).6 This solar eclipse occurred on September 26, 322 B.C. 7 The dating to the second year of Philip III’s reign is con˜rmed by
In 1992, A. B. Bosworth put forth the argument that Philip III was created King of Babylonia in 324/323 by his half brother Alexander III (Great), making the ˜rst regnal year of his Babylonian reign 324/323 B.C.1 Bosworth based his position on the fact that the Babylonian Chronicle gives Philip III eight regnal years (BM 34660 obv. 19), while Diodorus (19.11.5), likely copying his source Hieronymus of Cardia,2 states that Philip ruled six years, four months, placing his death either in October, or in early November 317.3 According to Bosworth, the best way to accommodate the
1. “Philip III Arrhidaeus and the Chronology of the Successors,” Chiron 22 (1992) 75–79. 2. Given Diodorus’ di¯culties, while abridging his sources, of estimating time intervals, exact time references are in all likelihood copied directly from his source (see E. M. Anson, “Diodorus and the Date of Triparadeisus,” AJP 107 [1986] 209–11), which for his Books 18–20 is generally believed to be Hieronymus of Cardia (M. J. Fontana, Le lotte per la successione di Alessandro Magno dal 323 al 315 [Palermo: Presso l’Academia, 1960], 151–237; J. Hornblower, Hieronymus of Cardia [Oxford: Oxford University Press, 1981], 2–3; K. Sacks, Diodorus Siculus and the First Century [Princeton: Princeton University Press, 1990], 21, 40, 41, see 99, 158; E. M. Anson, Eumenes of Cardia: A Greek Among Macedonians [Leiden: Brill, 2004], 11–25). 3. The precise date would depend on whether Philip III’s reign was calculated from Alexander’s death, or from the formal settlement at the end of the strife that broke out in Babylon. The period of con˘ict and the conclusion of the settlement had taken approximately one week (Curt. 10.10.9). Whether he died in October or November would depend on whether the actual length of his reign was precise, rounded up, or rounded down. Justin lists the length of Philip’s reign as six years (14.5.10); Porphyry as seven (FGrH 260 F-3).
4. Bosworth, “Chronology,” 75–79. 5. “Dating Methods During the Early Hellenistic Period,” JCS 52 (2000) 118 and n. 15; cf. E. M. Anson, “The Dating of Perdiccas’ Death and the Assembly at Triparadeisus,” GRBS 43 (2003) 377–80. 6. A. J. Sachs and H. Hunger, Astronomical Diaries and Related Texts from Babylonia, Vol. 1: Diaries from 652 B.C. to 262 B.C. (Vienna: Verlag der Österreichischen Akademie der Wissenschaften, 1988), 220–27. 7. F. R. Stephenson, Historical Eclipses and Earth’s Rotation (Cambridge: Cambridge University Press, 1997), 79, 131–32, 141–42, 540; M. Kudlek and E. H. Mickler, Solar and Lunar Eclipses of the Ancient Near East from 3000 B.C. to 0 with Maps (Neukirchen-Vluyn: Butzon & Bercker Kevelaer, 1971), 30. Kudlek and Mickler use dates equating a year 0 to 1 B.C. Therefore, -321 = 322 B.C.
127
JCS 57 (2005)
128
EDWARD M. ANSON
the listing on both the upper and left edges of this tablet as “from month 1 to month 6 of year 2 of king Philip.”8 Since the Babylonian year started in April,9 and Alexander died on June 11, 323 B.C.,10 Philip III’s ˜rst regnal year would have been 323/
322 and his second 322/321. This astronomical tablet, consequently, con˜rms that, as Boiy asserted on the basis of other data, Philip III’s ˜rst regnal year was 323/322, not, as Bosworth proposed, 324/323.
8. Sachs and Hunger, Diaries, 228–29. 9. R. A. Parker and W. H. Dubberstein, Babylonian Chronology, 626 B.C.–A.D. 75 (Providence, RI: Brown University Press, 1956), 36.
10. See now L. Depuydt, “The Time of Death of Alexander the Great: 11 June 323 B.C. (-322), ca. 4:00–5:00 PM,” WO 28 (1997) 117–35.
A NEW SCHEME FROM URUK FOR THE RETROGRADE ARC OF MARS J. M. Steele (University of Durham)
The cuneiform tablet W.20030/111 was found in the excavation of the site of the Bit Res sanctuary in Uruk during the eighteenth campaign (1959–1960) of the Deutsche Orient-Gesellschaft.1 It was one of a group recovered from a paved area in front of the gate in the southeastern wall, probably left there by grave robbers.2 W.20030/ 111 is a small ˘ake (5 x 4!- cm) 2 from the reverse of what must have been a quite sizable tablet. The bottom edge is preserved, but probably only a third of the height and a sixth of the width of the original tablet now remains. A copy of the tablet by J. van Dijk was published as number 86 in J. van Dijk and W. R. Mayer, Texte aus dem Res-Heiligtum in Uruk-Warka, Baghdader Mitteilungen Beiheft 2 (Berlin: Gebr. Mann Verlag). In the introduction to that volume, the authors quote the following description of the tablet by A. Sachs (p. 19):
The tablet preserves parts of three columns of a numerical table separated by double vertical rulings. For each entry there is an index number running from 1 to 30 followed by a three place sexagesimal number that, in the preserved part of the text, increases each line by 0,0,36. A single line colophon ˜lls the remaining space at the bottom of the reverse. I give below a transcription of the tablet from van Dijk’s copy. The context implies that, despite being the reverse of the tablet, the columns run from left to right, as in many tabular astronomical texts. The colophon at the end of the text reads: “[. . .] son of Anu-bel-sunu, son of Nidinti-Ani, lamentation priest? [of Anu? . . .].” Anu-bel-sunu is well known as owner or scribe of many astronomical and other texts,3 and father of the scribe Anu-abauter, a tupsar En¿ma Anu Enlil and also known as an astronomical scribe. It is quite probable, therefore, that Anu-aba-uter’s name should be restored at the beginning of the colophon. Given that Anu-aba-uter was a highly competent astronomical scribe it seems unlikely that the text contains simply a run of numbers increasing by 0,0,36 for at least thirty-four columns, tallied into groups
Tabelle mit Zahlen in Gruppen von jeweils 30 (vielleicht Tagen oder Graden) mit der festen Diˆerenz 0,0,36; Bedeutung unbekannt. Wenn die erste Kolumne der Tabelle mit 1 0,0,0 begann, haben wir auf unserem Fragment Kolumne XXXII–XXXIV.
1. My work on this tablet was made possible through the award of a Royal Society University Research Fellowship. I thank John Britton for his suggestion to include ˜g. 1 and Norbert Roughton for making available his database of computed planetary phenomena. 2. J. van Dijk, “Die Tontafeln aus dem res-Heiligtum,” UVB 18 (1962) 43–61; J. van Dijk, “Die Tontafelfunde der Kampagne 1959/60,” AfO 20 (1963) 217–18.
3. J. M. Steele, “A 3405: An Unusual Astronomical Text from Uruk,” Archive for History of Exact Sciences 55 (2000) 103–35; L. E. Pearce and L. T. Doty, “The Activities of Anubelsunu, Seleucid Scribe,” in Assyriologica et Semitica, Festschrift für Joachim Oelsner anläßich seines 65. Geburstages am 18. Februar 1997, eds. J. Marzahn and H. Neumann (Münster: Ugarit-Verlag, 2000), 331–41.
129
JCS 57 (2005)
130
J. M. STEELE
Rev.
I [9,30] [9,30,36] [9,31,12] [9,31,48] [9,32,24] [9,33] [9,33,36] [9,34,12] [9,34,48] [9,35,2]4
II [21] [22] [23] [24] [2]ª5º [2]6 [2]7 28 29 30
[9,48] [9,4]8,[3]6 [9],49,13 9,49,48 9,50,24 9,51 9,5[1],36 9,52,12 [9],52,48 9,53,24
III
1u 2u 3u 4u 5u 6u 7u 8u 9u 10u
[21] [22] [23] [24] [25] [26] [27] [28] [29] [30]
21 22 23 24 25 26 27 28 29 30
[10,6] [10,6,36] 10,[7,12] 10,[7,48] 10,[8,24] 10,:9 10,:9,[36] 10,10,[12] 10,10,[48] 10,11,[24]
11u
[. . .] A sá Id1–EN-sú-nu A sú INi-din-tu4-d1 lú-GALA? d? d [1 . . .]
Critical Apparatus Rev. II, 3u [9],49,13 is a mistake for 9,49,12. Rev. III, 6u A two wedge separation mark (transcribed :) separates the 10 from the 9.
of thirty. As Sachs suggested, these groups of thirty should probably be interpreted as either days in ideal months (= tithis or 1/30 of a synodic month in mathematical astronomy) or degrees within zodiacal signs. Either meaning would imply an astronomical context for the tablet. Within the corpus of mathematical astronomical texts we ˜nd three types of text in which numbers are listed for consecutive days or degrees within zodiacal signs: (i) daily motion tables (e.g., ACT Nos. 310, 650–655) 4; (ii) lists of the “pushes” used to determine two of the four phenomena of Mercury in Systems A1 and A2 for every degree (e.g., ACT Nos. 800a–800e); and (iii) lists of the length of Mars’ retrograde arc from ˜rst station to acronychal rising and from acronychal rising to second station (e.g., ACT Nos. 803 and 804). For W.20030/111 interpretation (i) can be ruled out immediately as none of the planets move as slowly as 0;0,36 degrees per day. Similarly, we can dismiss interpretation (ii) as the pushes used in Mercury’s System A1 for determining the diˆer4. P. J. Huber, “Zur täglichen Bewegung des Jupiter nach babylonischen Texten,” ZA 52 (1957) 265–303.
ence in longitudes and dates between ˜rst and last visibilities are always greater than the values recorded on W.20030/111, and for System A 2 either greater or change by amounts much bigger than 0;0,36 per degree.5 Thus we are left with either interpretation (iii), where the table relates to Mars’ retrograde arc, or we must assume that the table is of a kind previously unattested. Alone of the outer planets, the Babylonian System A scheme for Mars treats the two synodic phenomena on the retrograde arc, acronychal rising (Q) and second station (Y), as “satellite” phases of the ˜rst station (F), rather than as independent phenomena to be computed by the usual rules of System A. 6 Neugebauer, ACT,
5. O. Neugabauer, Astronomical Cuneiform Texts (London: Lund Humphries, 1955), 294–99. 6. O. Neugebauer, ACT, 303–6; O. Neugebauer, A History of Ancient Mathematical Astronomy (Berlin: Springer, 1975), 400–401 and 458–60; A. Aaboe, “A Late-Babylonian Procedure Texts for Mars, and Some Remarks on Retrograde Arcs,” in From Deferent to Equant: A Volume of Studies in the History of Science in the Ancient and Medieval Near East in Honor of E. S. Kennedy, eds. D. A. King and G. Saliba (New York: New York Academy of Sciences, 1987), 1–14.
A NEW SCHEME FROM URUK FOR THE RETROGRADE ARC OF MARS
131
Table 1. Schemes R, S, T and U for determining the retrograde arc from F to Q for Mars (where k is the number of degrees within a zodiacal sign of F). Aries Taurus Gemini Cancer Leo Virgo Libra Scorpio Sagittarius Capricorn Aquarius Pisces
R
S
T
U
6;24 6;48 6;48 7;12 7;12 6;48 6;48 6;24 6;24 6;0 6;0 6;24
6;24 + k x 0;0,48 6;48 6;48 + k x 0;0,48 7;12 7;12 – k x 0;0,48 6;48 6;48 – k x 0;0,48 6;24 6;24 – k x 0;0,48 6;0 6;0 + k x 0;0,48 6;24
6;30 7;0 7;0 7;30 7;30 7;0 7;0 6;30 6;30 6;0 6;0 6;30
6;30 + k x 0;0,30 6;45 + k x 0;0,30 7;0 + k x 0;0,30 7;15 + k x 0;0,30 7;30 – k x 0;0,30 7;15 – k x 0;0,30 7;0 – k x 0;0,30 6;45 – k x 0;0,30 6;30 – k x 0;0,30 6;15 – k x 0;0,30 6;0 + k x 0;0,30 6;15 + k x 0;0,30
305–6 identi˜ed four schemes used for determining the retrograde arc between F and Q and named them R, S, T and U. In R the retrograde arc varies between 6 and 7;12 degrees in steps of two zodiacal signs by 0;24 degrees. S keeps the same maximum and minimum but oddly the retrograde arc stays constant for one zodiacal sign and then increases or decreases linearly by 0;24 over a second zodiacal sign, whilst T is identical to R except the maximum retrograde arc is increased to 7;30 degrees in steps of 0;30. Finally, in U the retrograde arc varies linearly from 6 to 7;30 degrees. These schemes are summarized in Table 1. In each the retrograde arc from F to Q is determined by entering the table with the longitude of F. Preserved evidence for the second part of the retrograde arc from Q to Y is much scarcer; however, for scheme S Neugebauer found that this arc was equal to 3/2 the arc from F to Q, and hence the total retrograde arc would be 5/2 that from F to Q, and postulated that the same ratio would apply for the other schemes. Con˜rmation of this suggestion has since been found not in Babylonia but in the Pañcasiddhantika of Varahamihira, an Indian work that incorporates many elements of Babylonian planetary theory. In chapter XVII, verses 29–33, Mars’ total retrograde arcs in each zodiacal sign are presented, and the arcs agree
with what would be expected by applying the 5/2 scaling factor to scheme R.7 Applying the scaling factor 3/2 to schemes R and S we ˜nd that the length of the retrograde arc from F to Y should vary between 9 and 10;48 degrees, and applying it to T and U we would ˜nd a variation between 9 and 11;15 degrees. The preserved numbers on W.20030/111 are therefore of the right magnitude to be the length of the arc between F to Y. By comparison with other tables of Mars’ retrograde arc, ACT Nos. 803 and 804, both of which also come from Uruk, we would expect columns I and II to contain information for adjacent zodiacal signs. Since the numbers in both these columns vary, we cannot have here a version of schemes R, S or T for F to Y, but rather a scheme of similar type to U. However, from U we would expect that the line-by-line variation would be 0;0,45 degrees, whereas we ˜nd here 0;0,36. However, 0;0,36 is just what we would expect from a linearly varying scheme with maximum and minimum 9 and 10;48 degrees as in schemes R and S. Thus we appear to have here evidence for a new scheme—I propose to call it scheme Q—in which
7. O. Neugebauer and D. Pingree, The Pañcasiddhantika of Varahamihira, vol. II, Det Kongelige Danske Videnskabernes Selskab Historisk-Filoso˜ske Skrifter 6/1 (Copenhagen: Munksgaard, 1970–71), 170.
132
J. M. STEELE
Table 2. Retrograde arcs according to scheme Q (where k is the number of degrees within a zodiacal sign of F). Aries Taurus Gemini Cancer Leo Virgo Libra Scorpio Sagittarius Capricorn Aquarius Pisces
F to Q
Q to Y
F to Y
6;24 + k x 0;0,24 6;36 + k x 0;0,24 6;48 + k x 0;0,24 7;0 + k x 0;0,24 7;12 – k x 0;0,24 7;0 – k x 0;0,24 6;48 – k x 0;0,24 6;36 – k x 0;0,24 6;24 – k x 0;0,24 6;12 – k x 0;0,24 6;0 + k x 0;0,24 6;12 + k x 0;0,24
9;36 + k x 0;0,36 9;54 + k x 0;0,36 10;12 + k x 0;0,36 10;30 + k x 0;0,36 10;48 – k x 0;0,36 10;30 – k x 0;0,36 10;12 – k x 0;0,36 9;54 – k x 0;0,36 9;36 – k x 0;0,36 9;18 – k x 0;0,36 9;0 + k x 0;0,36 9;18 + k x 0;0,36
16;0 + k x 0;1 16;30 + k x 0;1 17;0 + k x 0;1 17;30 + k x 0;1 18;0 – k x 0;1 17;30 – k x 0;1 17;0 – k x 0;1 16;30 – k x 0;1 16;0 – k x 0;1 15;30 – k x 0;1 15;0 + k x 0;1 15;30 + k x 0;1
the retrograde arc between Q to Y varies linearly between a maximum of 10;48 degrees when F is at 1 degree in Leo and a minimum of 9 degrees when F is at 1 degree in Aquarius. The retrograde arcs given according to this new scheme Q are presented in Table 2. The preserved columns of W.20030/111 can now be seen to refer to Pisces (I), Aries (II) and Taurus (III). Successive lines give for each degree within the signs of the zodiac the associated length of the retrograde arc from acronychal rising to second station. In ˜g. 1 the total length of the retrograde arc given by scheme Q is compared to the true length of the retrograde arc given by modern computation for the period 400 B.C. to 100 B.C.8 Interestingly, although the magnitude of variation is much too
8. The modern values for the length of the synodic arc were taken from a database of all phenomena of the planets prepared and kindly made available by N. A Roughton of Regis University, Denver. For details of this database, see N. A. Roughton, “A Study of Babylonian Normal-Star Almanacs and
small, the phase of the variation is almost perfectly aligned with nature.9 The discovery of this new scheme for determining the longitudes of the retrograde phases of Mars illustrates how little we know about the treatment of Mars’ retrograde phases in Babylonian astronomy. For example, the reverse of the procedure text ACT Nos. 804 contains a linear zigzag function that Neugebauer suggested might be a scheme for the retrograde arc from Q to Y, but has a mean value about two degrees higher than that of scheme Q. Furthermore, we have almost no information on how the duration of the retrograde phases was calculated. Nevertheless, this new scheme con˜rms Neugebauer’s hypothesis that the total retrograde arc of Mars was always taken to be 5/2 that from ˜rst station to acronychal rising.
Observational Texts,” in Under One Sky: Astronomy and Mathematics in the Ancient Near East, eds. J. M. Steele and A. Imhausen (Münster: Ugarit-Verlag, 2002), 367–78. 9. See already Aaboe, Fs Kennedy, 11.
A NEW SCHEME FROM URUK FOR THE RETROGRADE ARC OF MARS
133
20
Retrograde Arc (degrees)
18
16
14
12
10
8 0
60
120
180
240
300
360
Longitude of First Station (degrees)
Fig. 1. The length of the retrograde arc of Mars from ˜rst station to second station for the period 400 B.C. to 100 B.C. (circles) and as given by scheme Q (solid line).