AUTHOR ""
TITLE "Meaning Through Language Contrast: Volume 2"
SUBJECT "Pragmatics & Beyond, New Series, Volume 100"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Meaning Through Language Contrast
Pragmatics & Beyond New Series Editor Andreas H. Jucker University of Zurich, English Department Plattenstrasse 47, CH-8032 Zurich, Switzerland e-mail:
[email protected] Associate Editors
Jacob L. Mey University of Southern Denmark
Herman Parret Belgian National Science Foundation, Universities of Louvain and Antwerp
Jef Verschueren Belgian National Science Foundation, University of Antwerp Editorial Board
Shoshana Blum-Kulka
Catherine Kerbrat-Orecchioni
Hebrew University of Jerusalem
University of Lyon 2
Jean Caron
Claudia de Lemos
Université de Poitiers
University of Campinas, Brazil
Robyn Carston
Marina Sbisà
University College London
University of Trieste
Bruce Fraser
Emanuel Schegloff
Boston University
University of California at Los Angeles
Thorstein Fretheim
Deborah Schiffrin
University of Trondheim
Georgetown University
John Heritage
Paul O. Takahara
University of California at Los Angeles
Kansai Gaidai University
Susan Herring
Sandra Thompson
University of Texas at Arlington
University of California at Santa Barbara
Masako K. Hiraga
Teun A. Van Dijk
St.Paul’s (Rikkyo) University
Pompeu Fabra, Barcelona
David Holdcroft
Richard J. Watts
University of Leeds
University of Berne
Sachiko Ide Japan Women’s University
Volume 100 Meaning Through Language Contrast: Volume 2 Edited by K.M. Jaszczolt and Ken Turner
Meaning Through Language Contrast Volume 2 Edited by
K.M. Jaszczolt University of Cambridge
Ken Turner University of Brighton
John Benjamins Publishing Company Amsterdam/Philadelphia
8
TM
The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.
Library of Congress Cataloging-in-Publication Data Meaning through language contrast / edited by K.M. Jaszczolt and Ken Turner. p. cm. (Pragmatics & Beyond, New Series, issn 0922-842X ; new ser. 99-100) Includes bibliographical references and indexes. 1. Contrastive linguistics. 2. Semantics. 3. Pragmatics. 4. Grammar, Comparative and general. I. Jaszczolt, Katarzyna. II. Turner, Ken, 1956- III. Series. P134 M35 2002 401’.43--dc21
2002074427
isbn 90 272 5119 3 (Eur.) / 1 58811 206 3 (US) (Hb; alk. paper, vol. 1, 99) isbn 90 272 5120 7 (Eur.) / 1 58811 207 1 (US) (Hb; alk. paper, vol. 2, 100) isbn 90 272 5349 8 (Eur.) / 1 58811 332 9 (US) (Hb; alk. paper, set volumes 1-2) © 2003 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa
Table of contents
Grammaticalization Distal aspects in Bantu languages Steve Nicolle From temporal to conditional: Italian qualora vs English whenever Jacqueline Visconti
3 23
Then – adverbial pro-form or inference particle? A comparative study of English, Ewe, Hungarian, and Norwegian Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó
51
The polysemy of the Swedish verb komma ‘come’: A view from translation corpora Åke Viberg
75
Metaphor in contrast Studying metaphors using a multilingual corpus Kay Wikberg
109
Cross-language metaphors: Conceptual or pragmatic variation? Andreas Musolff
125
A contrastive cognitive perspective on Malay and English figurative language Jonathan Charteris-Black
141
Metaphorical expressions in English and Spanish stock market journalistic texts Anna Espunya and Patrick Zabalbeascoa
159
Table of contents
Cross-cultural pragmatics and speech acts Directions of regulation in speech act theory Susumu Kubo
183
On Japanese ne and Chinese ba Mutsuko Endo Hudson and Wen-ying Lu
197
‘I am asking for a pen’: Framing of requests in black South African English Luanga A. Kasanga
213
Cultural scripts for French and Romanian thanking behaviour Tine Van Hecke
237
Sociocultural variation in native and interlanguage complaints Ronald Geluykens and Bettina Kraft
251
A cross-cultural study of requests: The case of British and Japanese undergraduates Saeko Fukushima
263
Questions as indirect requests in Russian and Czech Michael Betsch
277
The language of love in Melanesia: A study of positive emotions Les Bruce
291
Everyday rituals in Polish and English Ewa Jakubowska
331
A question of time? Question types and speech act shifts from a historical-contrastive perspective: Some examples from Old Spanish and Middle English Verena Jung and Angela Schrott The contrasts between contrasters: What discussion groups can tell us about discourse pragmatics Piibi-Kai Kivik and Krista Vogelberg
345
373
The semantics/pragmatics boundary: Theory and applications Cross-linguistic implementations of specificity Klaus von Heusinger
405
The semantics– pragmatics interface: The case of grounding Esam N. Khalil
423
Table of contents
On translating ‘what is said’: Tertium comparationis in contrastive semantics and pragmatics K. M. Jaszczolt
441
Translation equivalents as empirical data for semantic/pragmatic theory Bergljot Behrens and Cathrine Fabricius-Hansen
463
Language index
477
Name index
481
Subject index
483
Contents of Volume 1
489
TSL[v.20020404] Prn:8/04/2002; 15:18
P I
Grammaticalization
F: PBNSP01.tex / p.1 (1)
Distal aspects in Bantu languages Steve Nicolle SIL International and Digo Language and Literacy Project, Kenya
.
Introduction
This chapter is comparative at two levels. The first concerns whole language families, or perhaps we should say, groups of languages of a similar typological type. The second level of comparison concerns two individual Bantu languages: Digo (or Chidigo) which is spoken along the Kenyan and Tanzanian coast between Mombasa and Tanga, and Fuliiru (or Kifuliiru) spoken in the eastern part of the Democratic Republic of Congo.1 The focus of this study is grammaticalisation involving verbs of movement, direction and position. A distinction can be drawn between those languages in which grammaticalisation involving such verbs rapidly results in the loss of the semantic component of physical movement etc. and those languages in which such semantic change does not necessarily occur. In many Bantu languages, including Digo and Fuliiru, grammaticalisation involving verbs of movement, direction and position does not necessarily entail the loss of their lexical semantic content, which is typically the case during grammaticalisation in other language families. Between Digo and Fuliiru there are differences both in the range of lexical sources each language utilises and also in the extent to which these source constructions undergo formal grammaticalisation. Before describing these differences in detail, I shall briefly discuss the features of grammaticalisation relevant to this study. Grammaticalisation is the process whereby a construction involving one of the lexical word classes, such as nouns, verbs and adjectives, develops into a gram, that is, a morpheme of one of the functional, or grammatical, classes, such as prepositions and TAM (tense-aspect-modality) markers. For example, verbs of movement are common sources of future tense markers. Semantically, grammaticalisation results in semantic bleaching (or generalisation) and functional dependence on associated lexical material; in relevance theoretic terms grammaticalisation involves a shift from conceptual to procedural encoding (Nicolle 1998a). Following
Steve Nicolle
such semantic change, an expression undergoes formal grammaticalisation typically involving increasingly restricted morphosyntactic distribution to the extent of affixation, together with corresponding phonetic reduction. The semantic and morphosyntactic changes involved in grammaticalisation can result in variation. At the semantic level, whilst the new functional usage of a gram becomes established, there may be a period during which both the lexical and the functional semantic content of a grammaticalised expression are accessible. This is known as semantic retention, and is distinguished by two characteristics. First, the lexical component of meaning derived from the source expression is only sometimes recovered during utterance interpretation (hence there is variation). Second, when this lexical meaning is recovered it is not cancelable (hence it is semantic in origin rather than pragmatic). This is discussed in detail in Nicolle (1998a, 1998b), in particular in relation to the English be going to construction which, whilst functioning as a grammaticalised future tense marker, also retains elements of its lexical source construction, but only in some of its uses. In contrast, the meaning components of physical movement or location encoded in the Bantu distal aspects described below are always recovered. They are therefore not the result of semantic retention. Secondly, at the morphosyntactic level, the various stages of grammaticalisation may be attested contemporaneously, as in the following examples from Swahili, another Bantu language. The source is the lexical verb kwisha (‘finish’) plus the perfective aspect marker -me-, and the resulting morpheme is a new perfective aspect: (1) a.
A-me-kwisha ku-soma S/he-pfv-finish inf-read S/he has finished reading OR S/he has already read. b. A-me-kwisha-soma S/he-pfv-finish-read S/he has already read. c. A-mesha-soma S/he-pfv-read S/he has already read.
The construction in (1a) is ambiguous between a reading in which the first verb, kwisha, conveys lexical meaning and a reading in which the construction me+kwisha functions as an aspectual marker with kwisha behaving morphosyntactically as an auxiliary. In (1b) the infinitive marker has been lost and the preceding auxiliary construction becomes affixed to the main verb.2 Finally in (1c), the formal process of grammaticalisation is complete and the construction me+kwisha (formerly aspect plus verb) has been reanalysed as a single TAM marker with
Distal aspects in Bantu languages
corresponding phonetic reduction.3 Further reduction of -mesha- to -sha- is also possible. Such variation is discussed in Section 4 in relation to Digo. When a verb of movement, direction or position or an expression involving such a verb is the lexical source for a gram such as a tense marker, semantic bleaching typically results in the loss of the semantic component of physical movement, direction or position. In the data discussed below, grammaticalisation involving such verbs has not resulted in the loss of the semantic component of physical movement, direction or position in the resulting TAM markers. There are two main approaches that could be adopted with respect to this situation. First, we could argue that the expressions in question are not grammaticalised. Against this position, I will demonstrate (in Section 4) that all of the expressions have undergone or are undergoing formal grammaticalisation. It is generally agreed that formal grammaticalisation follows semantic (or functional) grammaticalisation (Givón 1991; Nicolle 1998a) or occurs in parallel with it (Bybee et al. 1994: 106); a case in which formal change precedes semantic change has yet to be established. The approach adopted here (which I will explore in Section 5) is that basic members of the semantic domain of movement, direction and position have the potential to form the basis of functional categories in the same way as, for example, temporal distance and duration. Languages differ as to whether they realise such ‘distal’ categories, and also over which members of the semantic domain of movement, direction and position are treated as sufficiently basic to undergo such grammaticalisation. It appears that Bantu languages in general tend to favour the grammaticalisation of distal categories, and that Fuliiru has developed a greater variety of source expressions than Digo. In the following section, I will describe a number of grams derived from lexical verbs of motion in Digo which express what I term ‘distal aspects’, and in Section 3 I will do the same for Fuliiru. Section 4 looks at morphosyntactic evidence of formal grammaticalisation in the various distal aspects described, and I conclude with a discussion of the implications of the study.
. Distal aspects in Digo Distal aspects in Digo involve the basic categories of ‘go’ (four grams) and ‘come’ (one gram). Amongst the forms predicated on ‘go’ there is no difference in the type of going in terms of distance traveled, speed, manner etc. Rather, the forms are distinguished in part by the time at which the movement is conceived of as occurring, relative to some temporal reference point, and in part by morphosyntactic distribution patterns.
Steve Nicolle
. -yaThere are two semantic components to -ya-: it indicates that the fulfillment of the action described by the verb is subsequent to some temporal reference point, and also that there is physical movement away from an initial location which occurs during and after that temporal reference point. Like the pure future tense in Digo, -ya- can be displaced. In (2) the action of repairing a car is still future and the subject of the action is moving towards the location of this future action. (2) A-ka-lamuka a-ya-tengeza gari. He-hod-wake he-ya-repair car He woke (today) and is going (i.e. is on his way) to repair a car. In (3), taken from a first person narrative, the action of taking the sick man to be treated is described as being subsequent to the time at which the doctor will have met him; the subject (the sick man) is described as moving towards the place where the treatment would occur, which the subjunctive clause amphirike (‘to convey him’) also expresses. (3) Hu-ri-pho-fika hipho Kenyatta ná -m-kuta Dr. Maneno We-pst-rel-arrive there Kenyatta I.pst-him-met Dr. Maneno When we arrived at Kenyatta (Hospital) I met Dr. Maneno a-na-hu-godz-era ili a-m-hal-e yuya mkpwongo na he-prog-us-wait-app so.that he-him-take-sub that sick.person and waiting for us so as to receive the sick man and a-m-phirik-e kura ambako a-ya-lagul-wa. he-him-convey-sub there where he-ya-treat-pas take him to where he would (go and) be treated. The origin of -ya- is uncertain, but I believe it derives from the current lexical verb phiya ‘go’. The process of phonological reduction from phiya to -ya would then be parallel to the development of kwisha into -sha in Swahili detailed above.
. -chaA less frequently occurring distal aspect is -cha-. This occurs with the second of two verbs when the first is marked by the -ka- hodiernal past tense marker. It indicates that the action of the second verb occurs at a distance and that the subject of the verb has arrived at the location of the action. This is in contrast to the -ya- distal aspect in (2), which indicates that the subject of the verb is still progressing towards the location of the action. The origin of this form is uncertain, but it is possible
Distal aspects in Bantu languages
that it derives from the verb kpwedza ‘come’ which is derived historically from a non-finite prefix and the verb stem -dza.4 (4) A-ka-lamuka a-cha-tengeza gari. He-hod-wake he-cha-repair car He woke today and has gone to repair a car.
. Kpwenda Constructions involving the form kpwenda in Digo illustrate the phenomenon of a linguistic cycle, in which grammatical forms with similar functions develop consecutively from similar sources. The distal aspect -ya- is not the only gram to have developed from a verb meaning ‘go’. Digo has a future tense marker -nda- derived from the archaic verb *kwenda, which also meant ‘go’. This verb is still attested in neighbouring languages, in some of which (including Giryama and Pokomo) it is the source of a future morpheme also (see Nurse & Hinnebusch 1993: 376). This is what I call a ‘pure future’; the semantic component of movement has been entirely lost and the only information encoded is along the lines of: ‘The associated event occurs subsequent to some temporal reference point’. Kpwenda5 occurs again, either as a subsequent development from the same archaic Digo form from which -nda- developed or as a borrowing from a neighbouring language such as Swahili or one of the Mijikenda group, but this time functioning as a distal aspect. Kpwenda indicates that movement either precedes or is an integral part of the action described by the main verb. It differs morphosyntactically from -ya- and -cha- in that it occurs in the infinitive form (which I have used as the citation form) or in conjunction with a tense marker. The morphosyntactic status of kpwenda is discussed in Section 4. Example (5) illustrates the infinitive form and in example (6) kpwenda combines with the consecutive tense marker -chi- as follows: -chi- + -enda- = -chenda-. (5) Na-ye yuya mzee wa-kat-ika mairo kaya And-he that old_man he.pst-cut-caus running home kpwenda-iha ayae phamwenga mchewe... inf.go-call his.fellows together his.wife And the old man ran off home to go and call his colleagues and his wife . . . (6) A-ri-pho-ona hivyo, a-piga mbiru a-chi-iha They-pst-rel-see thus they.pst-hit horn they-con-call atu osi people all When people saw that, they blew a horn and called together all the people
Steve Nicolle
hipho laloni na a-chenda-mu-endza hiko weru-ni. there place and they-con.go-him-search there bush-loc. in that area and they went and looked for him in the bush.
. -ka-V-e A frequently occurring construction consists of -ka-V-e (TAM marker, verb and subjunctive suffix). This indicates that an action occurs subsequent to some temporal reference point and at a distance, but movement only begins subsequent to that temporal reference point. In the declarative mood the action described using -ka-V-e follows some other action, but when used in the imperative there is no such restriction on its use. This follows from the characterisation of this form as expressing movement subsequent to some temporal reference point; in the declarative this is established by some other event whereas in the imperative this is not necessarily so. Example (7) illustrates the declarative usage, and example (8) the imperative form. (7) Ná-lamuka chiti ligundzu sana kaya Vyongbwani, I.pst-wake.up early morning very home Vyongwani ku-dzi-tayarisha inf-refl-prepare I woke up very early in the morning at home in Vyongwani, to get ready n-ka-fundish-e hiko Golini skuli ya msingi. I-ka-teach-sub there Golini school of base. to go and teach in Golini primary school. (8) Mwanangu na-ku-voy-era uzima. Ka-som-e. My.child I.prog-you-pray-app health ka-study-sub. My child I pray you have health. Go away and study. Neither -ka- alone nor the subjunctive alone conveys this distal meaning. In (9), the story of an encounter with a lion taken from the same first person narrative as (7), the underlined occurrences of verbs with -ka- TAM markers are sequential and no change of location is possible. In (10) the subjunctive alone (indicated by the -e suffix) is neutral with respect to movement. (9) Lakini kpwa kpweli kala wa-mendza, mana kala But for truly if it.pst-like, for cop_pst a-ka-ni-ona, iye it-pfv-me-see, it But for sure if it wanted, because it saw me
Distal aspects in Bantu languages
mwandzo kala a-na-weza ku-ni-urukira a-ka-ni-olaga ama first cop.pst it-prog-can inf-me-jump.on it-seq-me-kill or first, it could have jumped on me and killed me or a-ka-ni-gbwarura a-ka-ni-richa phapho au a-ka-ni-rya. it-seq-me-maul it-seq-me-leave there or it-seq-me-eat. mauled me, and left me there or eaten me. (10) A-chi-ambira hara anjina a-kal-e mbere ili a-shuh-e. He-con-tell those others they-cop-sub ahead so.that he-fart-sub. He told the others to go ahead so that he could fart. TAM markers with the phonological form -ka- are common throughout most Bantu languages, with a wide variety of functions, including hodiernal (or near) and far past tenses, sequential and consecutive tenses, and perfect and perfective aspects. In addition, -ka- and related forms can function as future tenses and, as in Digo, distal aspects (Botne 1999). The distal form is concentrated in the southcentral Bantu area, but it also occurs in the eastern Bantu area, where Digo is located (where it is probably more widespread than Botne’s study suggests). In Digo, distal -ka- only occurs in the subjunctive form, which is in line with Botne’s observation (1999: 482) that distal -ka- is most likely to occur in the subjunctive or imperative, and that its occurrence in infinitives or indicatives almost guarantees its occurrence in imperatives (and to a lesser extent subjunctives). Botne concludes from this that distal -ka- may have originated in imperative constructions in most of the languages which exhibit it, although it may have a separate origin in subjunctive constructions in eastern Africa (1999: 504). The source constructions were most probably derived either from a motion verb or from a locative clitic which followed verbs of motion which, once deleted, left the clitic with the sense “go and V” (1999: 488–489). Over time -ka- would have become affixed to the main verb. Whatever its precise origin, the -ka-V-e distal aspect in Digo has probably been in existence for a long time. This suggests that there is nothing inherently unstable about distal aspects compared with other TAM markers, at least as far as Bantu languages are concerned. Cross-linguistically, one finds that grammaticalisation of constructions expressing physical movement typically leads to the development of pure tense markers, and it could be argued that the distal aspects in -ya-, -chaand kpwenda are simply intermediate stages in such a grammaticalisation chain. According to this view, -ya- and kpwenda are transparently innovations and therefore probably of fairly recent provenance, and are likely to develop into pure future tense markers, which are in any case transparently innovations in all the North-East Coast Bantu languages which have future tenses (Nurse & Hinnebusch 1993: 384). Turning to -cha-, its limited distribution and its relatively infrequent occurrence suggest that this might be a developing gram which failed to run its course and will soon fall into disuse (similar to shall in English). However the longevity of -ka-V-e
Steve Nicolle
provides evidence that in Bantu languages, distal aspects do not inevitably develop into pure future tenses or die prematurely.
. Kpwedza Another member of the Digo system of distal aspects indicates movement towards rather than away from some location, and not surprisingly derives from the lexical verb kpwedza meaning ‘come’. In the following example, kpwedza is in the infinitive, but, like the construction with kpwenda at the beginning of the same example, it can also occur with additional tense marking. Examples of this usage are given in Section 4 in relation to variation in the morphosyntactic frames in which kpwedza occurs. (11) N-á-kpwenda-sagala kpwakpwe siku hahu halafu n-chi-uya I-pst-go-stay his.place days three then I-con-return kaya home I went and stayed at his place for three days and then I returned home kpwedza-sema safari vyo-kala. inf.come-say journey how-cop.pst. to come and report how the journey had been.
. Distal aspects in Fuliiru Fuliiru also exhibits distal aspects derived from verbs of movement, direction and position. However, Fuliiru shows greater variation in the range of sources of distal aspects than Digo, which as we have seen has grammaticalised verbs whose basic meaning is either ‘come’ or ‘go’ (the origin of -cha- is uncertain). Van Otterloo (2000), from which the examples in Sections 3.1 and 3.2 come, classifies Fuliiru aspects both semantically and morphosyntactically. Semantically, there are three classes of verbal aspect in Fuliiru: relative time, relative place, and expectation (modality). Aspects of relative place correspond to what I have called distal aspects. The classification of Fuliiru aspectual constructions according to morphosyntactic criteria is reflected in the division between Section 3.1 and 3.2. Section 3.3, the data for which were supplied by Karen Van Otterloo (p.c.), does not describe a distal aspect, but is included as it illustrates how the same morphosyntactic form, which in Digo and many other Bantu languages indicates a distal aspect, is used in Fuliiru with a different interpretation.
Distal aspects in Bantu languages
The data presented here is all from translated material (the New Testament) which was checked for naturalness and intelligibility with a number of native speakers.
. Forms ending in -i The constructions of this morphosyntactic type are characterised by distal aspect markers ending in -i. According to Van Otterloo (p.c.) the -i suffix is a reflex of the older non-finite prefix which is no longer attested in contemporary Fuliiru. Originally, this -i suffix would have been a prefix on a main verb following an auxiliary, so distal aspect markers ending in -i are the result of reanalysis analogous to the reanalysis in English of going [to V] as gonna [V]. The non-finite marker in contemporary Fuliiru is the class 15 prefix uku-, but the distal aspects ending in -i developed much earlier, at a time when non-finite verb forms were indicated by the class 5 prefix i-, suggesting that these forms are relatively long-lived.6 Like kpwenda and kpwedza in Digo, these distal aspects may be preceded by a subject prefix plus an optional negative prefix and a tense marker (except in the subjunctive mood), or by a non-finite prefix. There is no additional subject prefix or non-finite prefix before the main verb. In contrast to the standard Digo orthography in which distal aspects are affixed to the following verb, in the Fuliiru orthography the distal aspect is usually written as a separate word.7 Although this suggests an auxiliary plus main verb analysis, Van Otterloo (2000) prefers to call these two word constructions ‘aspects’, although more recently the term ‘adverbial auxiliary’ has been used for aspectual constructions of time, place and expectation (Van Otterloo 2001). In this section I will describe four distal aspects, but this does not necessarily constitute an exhaustive list. The form in -gendi, meaning ‘go’, indicates that the event described by the main verb is preceded by the subject moving away from its current location. In (12), which is tensed, the subject is described as moving away from the location at which she had got up in order to prepare food. In (13), which is in the subjunctive mood, the subjects propose moving away from the location of the speech event in order to perform an action at a different location. (12) Kwokwo uyo maawe a-na-vyuka, a-na-gendi mú-kolera ibyo kulya. Thus that mother she-con-rise she-con-go him-serve food So the woman got up and went and served him food. (Matthew 8: 15) (13) Ko’-loziizi tu-gendi ki-randura? You-want we-go.sub it-uproot Shall we go and uproot it?
(Matthew 13: 28)
Steve Nicolle
The form -yiji (‘come’) functions similarly to -gendi except that the subject is described as moving towards rather than away from some location. In the example below, the speaker is responding to a request to go to a certain location to heal a sick person, and it is this location rather than the location of the speech event which is taken as the locus of movement. (14) Yesu a-na-mú-shuvya: “N-ga-yiji mú-kiza.” Jesus he-con-him-say I-fut-come him-heal Jesus said: “I will come and heal him.”
(Matthew 8: 7)
The distal aspect formed from -hikiri (‘arrive’) is similar to -yiji (‘come’) in that it indicates movement towards a location but (as in -gendi) the location is the deictic centre (the location of the speech event or some other previously established location). The example in (15) consists of a negative subjunctive form of -hikiri followed by the bare stem of the main verb. In (16) -hikiri is tensed and the main verb, which is in the same form as (15), is written as part of the same word; this is merely orthographic convention and there is no other morphosyntactic difference between these two instances of -hikiri. (15) Ikyanya u-ga-laalik-wa ku’-buhya u-ta-hikiri bwatala When you-fut-invite-pas loc-wedding you-neg-arrive.sub sit ku’-kitumbi kye-mbere. loc-chair pos-front When you are invited to a wedding, do not arrive-and-sit on the front seat. (Luke 14: 8) (16) A-na-hikiri’-gwa mu’-magulu ga’-Yesu, iri a-na-mú-tangira He-con-arrive-fall loc-feet of-Jesus while he-prog-him-give kongwa thanks And he arrived-and-fell at Jesus’ feet, while giving him thanks. (Luke 17: 16) The final distal aspect of this morphosyntactic type to be discussed here, -sigali (‘remain’), differs from the previous three in that it indicates lack of movement rather than movement in a particular direction. Prior to the utterance in (17) the speaker, Jesus, has stated that he will go away (that is, he will go to heaven) and so the hearers will search for him in vain, the implication being that this is because they will remain on earth. In Fuliiru, the use of -sigali indicates that when the subjects die this will not involve any change of location, either literally (they will not follow Jesus to heaven) or figuratively (they will die as they lived: ‘in’ their sins).
Distal aspects in Bantu languages
(17) Na’-kundu mu-ga-n-jakula mu-ga-sigali fwira And-although you-fut-me-search you-fut-remain die mu’-byaha biinyu. loc-sins your And even though you will search for me, you will remain-and-die in your sins. (John 8: 21)
. Forms ending in -a and followed by -gaIn constructions of the second morphosyntactic type, the main verb is a bound form which is preceded by a subject prefix and the TAM marker -ga-8 as follows: SP-ga-V. This verbal complex is preceded by a distal aspect marker ending in -a. The distal aspect may occur in one of three basic morphosyntactic frames: 1. preceded by a subject prefix (which must be the same as that preceding the main verb) followed by a TAM marker; 2. preceded by another aspect construction ending in -i (again the subject prefix must be the same as that preceding the main verb); or 3. preceded by a locative prefix. In the following examples, (18) illustrates the distal aspect -lenga (‘pass’) preceded by the iterative aspect marker -kizi which behaves morphosyntactically in the same way as -gendi, -yiji etc. (described in Section 3.1 above). Example (19) illustrates -genda (‘go’) preceded by a locative prefix. (18) Yikyo kimenyeeso, Abayahudi bingi That sign Jews many ba-na-kizi lenga ba-ga-ki-soma. they-con-itv pass they-ga-it-read And many Jews passed by and read that sign.
(John 19: 20)
(19) Uyimule uyu mukazi; a-li mu’-genda a-ga-banda ulubi Remove this woman; she-cop loc-go she-ga-make noise i’nyuma liitu. behind us Get rid of this woman; she is going along making a racket behind us. (Matthew 15: 23) Constructions with this morphosyntactic form also occur with -yija (‘come’) and other distal and non-distal aspects. These constructions are similar in meaning to the corresponding forms ending in -i (Section 3.1) but indicate that the action is ongoing, hence they often co-occur with the iterative aspect marker -kizi as in example (18) or the locative prefix mu- as in example (19).9
Steve Nicolle
. Non-distal -ka-V-e Fuliiru exhibits a form with the same morphological form, -ka-V-e (-ka- TAM marker, verb stem, -e subjunctive suffix), as Digo but with a different function. I will briefly describe the meaning of -ka-V-e in Fuliiru before offering a few tentative suggestions as to its origins and its relation to the distal form. The -ka-V-e construction in Fuliiru indicates that there is a period of time, often including a mentioned event, that intervenes before the occurrence of the event described by the -ka-V-e construction. In (20) the act of telling is marked by -ka-V-e and the intervening event is the arrival of the speaker. In (21) the event associated with -ka-V-e is the naming of a child and the intervening event is the child’s birth. (20) U-bwatal-e ku’-kitumbi kye’-nyuma, gira mango You-sit-sub loc-seat pos-back so.that when ka-ku-laalika a-yija rel-you-invite he-come Sit in the back seat, so that when the one who invited you comes, a-ka-ku-bwir-e kwokuno: Mwira wani, u-yiji he-ka-you-say-sub thus friend my you-come bwatala hano ha’-mbere. sit here pos-front he will speak to you thus: My friend, come and sit here at the front. (Luke 14: 10) (21) A-ga-buta umwana wo’-butabana. No’-yo’-mwana, She-fut-bear child pos-male And-rel-child u-ka-m-yinik-e iziina lya’-Yesus you-ka-him-name-sub name pos-Jesus She will bear a son. And that child, (you should) name him Jesus. (Matthew 1: 21) The -ka-V-e construction contrasts with two other future or prospective forms: the simple future with -ga- (agagenda ‘s/he will go at some future time’) and the far future formed by -ye-V-e (ayegende ‘s/he will go at some time in the distant future’) which indicates that the associated event is sure to occur but at a time which is uncertain. The -ka-V-e construction and the -ye-V-e construction each consist of two formatives but constitute a single morpheme. It is possible that the -ka-V-e construction in Fuliiru is an independent innovation, but given the wide geographical spread of this form there is a high likelihood that it is derived from the same form or forms that gave rise to the future and distal -ka- constructions in a number of other Bantu languages. As we saw in Section
Distal aspects in Bantu languages
2.4, cross-linguistically all distal -ka- forms occur in the subjunctive or imperative moods, and conversely all future -ka- constructions occur in the indicative (Botne 1999: 496). As the examples above show, the Fuliiru -ka-V-e construction can occur in both the subjunctive (imperative) and the indicative, so an origin in either a distal or a future -ka- construction is possible. Nurse and Muzale (1999: 529) note that a number of Great Lakes languages (the same sub-group as Fuliiru, following Nurse’s 1999, classification) form a far future in -(a)ka-V-e, which might suggest an origin in a Proto-Bantu future -kaconstruction. However, these languages are only situated on the eastern (Kenyan) side of Lake Victoria (whereas Fuliiru is spoken in the Democratic Republic of Congo), so the far future form in -(a)ka-V-e may be a more recent localised innovation, as Nurse and Muzale suggest (ibid.). It is possible, however, that Nurse and Muzale are incorrect, and that the far future form in -(a)ka-V-e is a reflex of a Proto-Bantu future -ka- construction which has been replaced elsewhere within the Great Lakes sub-group by more recent innovations. In this case, we could imagine a situation in which a far future of the form -ka-V-e was replaced in Fuliiru by a newer innovation, and the morphological form previously associated with far future was reanalysed. Since events envisaged as occurring in the far future are typically viewed as following some earlier future event or events, it is conceivable that a far future could come to be reanalysed as a gram with a meaning along the lines of: ‘The associated event occurs after another action has been completed.’ If we assume that Nurse and Muzale (1999: 529) are correct in viewing the Great Lakes far future form in -(a)ka-V-e as a more recent localised innovation, then it is equally possible that the -ka-V-e construction in Fuliiru originates from a distal aspect as from a far future. The semantic mechanism of change that would need to operate in such a case is generalisation, which can be defined as “the loss of specific features of meaning with the consequent expansion of appropriate contexts of use for a gram” (Bybee et al. 1994: 289). The meaning of -ka-V-e in Digo can be paraphrased as: ‘The associated event occurs after another action has been completed by the subject of the verb, and that action involves moving to another location.’ If we subtract two components of meaning: 1. ‘by the subject of the verb,’ and 2. ‘that action involves moving to another location,’ we arrive at the same paraphrase of the meaning of the Fuliiru form suggested above: ‘The associated event occurs after another action has been completed.’ The implications of this hypothesis are explored in Section 5.
Steve Nicolle
. Grammatical considerations So far I have focused on the semantics of distal aspects in Digo and Fuliiru, and whilst I have described the morphosyntactic characteristics of these forms, I have not yet evaluated the classification of these constructions as aspects using morphosyntactic criteria. Are we dealing here with morphological aspect markers, auxiliary or serial verb constructions, or something else? The basic pattern in finite verb phrases in many Bantu languages, including Digo and Fuliiru, is of a verbal complex consisting of the following slots: Subject Prefix – TAM Prefix – (Object Prefix) – Verb Stem – (Suffixes) – Final Vowel Note also that in Fuliiru (but not in Digo) the subject prefix and TAM prefix can be replaced by a construction consisting of a copula form followed by a separate verbal complex prefixed with a locative as in example (19) above. If we take this pattern as representative of prototypical or fully grammaticalised TAM marking, then the following distal aspects in Digo are unambiguously TAM markers: -ya- (Section 2.1) -cha- (Section 2.2) -ka-V-e– (Section 2.4) The following constructions behave morphosyntactically like TAM markers in that the following (main) verb does not have its own subject prefix or infinitive prefix, but they behave like verbs in that they may themselves be preceded by a TAM marker: In Digo:
kpwenda (Section 2.3) kpwedza (Section 2.5) In Fuliiru: -gendi, -yiji, -hikiri, -sigali (Section 3.1) I consider these to be aspects, that is fully grammaticalised members of the TAM systems of Digo and Fuliiru, since the absence of a repeated subject prefix and the restricted number of such forms in both languages make it unlikely that these are serial verb constructions. The verb-like characteristics of these constructions are a feature of newly grammaticalised constructions. Such constructions typically exhibit variation, as in the case of mekwisha versus mesha in Swahili (see Section 1). Variation is also a feature of the use of kpwedza in Digo. Examples (22) and (23) are translations of the same passage from the New Testament (Matthew 10: 34). (22) was the original version which had been translated by a native speaker of Digo and checked
Distal aspects in Bantu languages
for acceptability by another native speaker, before being presented in context to a group of about ten people, none of whom had been involved in the previous translation process. These native speakers accepted the grammaticality and intelligibility of (22) but felt, since this was representing the spoken language, that (23) would be more natural. (22) M-si-fikiriy-e kukala ná-kpwedza kureha amani dunia-ni. You-neg-think-sub that I.pst-come inf-bring peace world-loc Do not think that I came to bring peace to the world. S-edz-ere ku-reha amani, bali ná-reha upanga. I.neg-come-pst inf-bring peace rather I.pst-bring sword I did not come to bring peace, rather I brought a sword. (23) M-si-on-e kukala ná-kpwedza-reha amani dunia-ni. You-neg-see-sub that I.pst-come-bring peace world-loc Do not think that I came to bring peace to the world. S-edz-ere-reha amani, ela ná-kpedza-reha kondo. I.neg-come-pst-bring peace but I.pst-come-bring conflict I did not come to bring peace, but I came to bring conflict. In (22) kpwedza functions morphosyntactically like an auxiliary verb; it is preceded by a TAM marker (in the second occurrence, sedzere, the past tense marker is indicated by the suffix -ere rather than by a prefix as is usually the case), and the main verb is preceded by an infinitive prefix. In the revised version, (23), the infinitive prefix has been removed from the first two occurrences of -reha (‘bring’) and kpwedza functions as a TAM prefix on the main verb. The meaning of these constructions is, however, identical, suggesting that kpwedza is a newly grammaticalised construction. In addition, the tensed occurrence of -reha in the final clause of (22), náreha (‘I brought’), has been rephrased as nákpedzareha (‘I came to bring’). The Fuliiru forms -genda, -yija and -lenga behave morphosyntactically more like verbs than TAM markers in that the subject prefix is repeated on the main verb and is followed by the TAM prefix -ga-. These considerations favour a serial verb analysis. However, the fact that only the -ga- TAM marker can occur on the main verb indicates that these constructions are morphosyntactically restricted, which in turn suggests that formal grammaticalisation may have begun. The forms ending in -a (followed by a main verb marked by -ga-) are drawn from the same limited sub-set of verbs as the forms ending in -i and are similar in meaning, which suggests that they represent a more recent innovation which is following a similar grammaticalisation path as the forms ending in -i. In addition, the forms ending in -a do not take verbal extensions (with the exception of the emphatic extension -ag, which always occurs at the end of a word10 ) and cannot be modified by an adverb phrase (which would split the ‘aspect’ from the following verb), suggest-
Steve Nicolle
ing that these are best characterised as auxiliaries or aspect markers rather than as independent verbs.
. Discussion This study has assumed that distal aspects are genuinely grammaticalised elements of at least two Bantu languages. In Digo, at least one distal aspect, -ka-V-e, has probably been in existence for a long time. This suggests that distal aspects are not merely intermediate stages along a grammaticalisation chain from lexical verbs of movement, direction or location to non-distal (‘pure’) TAM markers. Rather, it is possible for a grammaticalisation chain to culminate in a distal aspect. The existence of the same form in Fuliiru with a non-distal meaning may be evidence of a further stage of grammaticalisation (specifically through the mechanism of generalisation) operating on a distal aspect. If this is the case, it need not imply that distal aspects are less grammaticalised than pure aspects (which indicate relative time). Schwenter (1994) presents evidence that the present perfect in the Alicante dialect of Spanish is developing into a hodiernal past/perfective with concurrent loss of the semantic component of current relevance. He goes on to argue that the resulting hodiernal form will eventually lose the semantic component that restricts associated events to a single day and will become a general perfective marker. This does not mean that the present perfect in Spanish is not a real functional category, or that the hodiernal past is less grammaticalised (functionally or formally) than the perfective. What it indicates is that it is possible for one functional category to develop into another through generalisation (see Nicolle 1998a: 22). At the outset, I stated that this chapter would be contrastive at two levels. Within the Bantu language family I have contrasted distal aspects in two individual languages: Digo and Fuliiru. I have demonstrated how Digo differentiates between a number of forms derived from the basic semantic domain of movement and direction. These distal aspects all differ semantically and vary morphosyntactically from fully grammaticalised TAM prefixes to newly grammaticalised forms exhibiting formal variation. Fuliiru also distinguishes a number of distal aspects, but these exhibit fewer characteristics of formal grammaticalisation and are derived from a wider range of lexical source constructions than in Digo. Whilst at least two of the distal aspects in Digo, as well as the pure future tense marker, are derived from verbs meaning ‘go’, in Fuliiru only -gendi and -genda are derived from the lexical verb ‘go’ and the difference in meaning between these is unclear. Conversely, Fuliiru has developed distal aspects from lexical verbs that have not undergone grammaticalisation in Digo, such as ‘arrive’, ‘remain’ and ‘pass’.
Distal aspects in Bantu languages
There is no morphosyntactic reason why Digo and Fuliiru should differ in these ways, so the variation is more likely to be semantic in origin. In Section 1 I suggested that basic members of the class of verbs covering the semantic domain of movement, direction and location could undergo grammaticalisation to become distal aspects. The verbs ‘go’ and ‘come’ are the most basic members of this domain, and they frequently contribute to the source constructions of both distal aspects and pure tenses. The system of distal aspects in Digo appears to be based on these two basic verbs, whereas Fuliiru extends the set to include ‘arrive’, ‘remain’ and ‘pass’, also arguably basic members of the domain. If the variation between Bantu languages concerning which verbs are acceptable sources for distal aspects is dependent on which verbs are considered to be ‘basic’ members of the domain, then we would expect more languages to have distal aspects based only on ‘come’ and ‘go’ (the most basic members of the set) than on other verbs. It should also be possible to construct an implicational scale, such that the existence of distal aspects derived from less basic verbs implies the existence of distal aspects based on more basic verbs such as ‘go’ and ‘come’. A further hypothesis is that languages which draw on a larger set of verbal sources will allow more of these to co-occur in a verb phrase. Thus, Fuliiru, with a large set of distal aspects (and other aspects or adverbal auxiliaries sharing the same morphosyntactic characteristics) allows up to three aspects to precede the main verb, whereas Digo, with a small set allows only one. (TAM information can only occur on the first verbal auxiliary in a phrase.) Correspondingly, where there is large set of aspects or adverbial auxiliaries, these will exhibit fewer features of formal grammaticalisation (phonological reduction, affixation) than members of small sets. Hence, in Fuliiru distal aspects and other adverbial auxiliary constructions are written as separate words whereas in Digo they are written as single words. At a more general level, the Bantu language family, represented by Digo and Fuliiru, can be contrasted with other language families. This raises the general question of why distal aspects, unlike, say, future tenses, occur in Bantu languages but not in many other language families. Another way of putting the same question is to ask why it is that lexical sources which invariably undergo semantic bleaching during grammaticalization in many languages are not necessarily semantically bleached during grammaticalization in (at least some) Bantu languages. For example, a movement verb which constitutes the source lexeme for a future tense should (according to grammaticalisation theory) undergo semantic bleaching with the result that the original lexical semantic component of movement becomes no longer recoverable. Even if there is a period during which semantic retention is operative, the semantic component of movement will only occur in addition to the grammaticalised future tense function (see Nicolle 1998b). What marks Bantu languages out from many other languages is that supposedly lexical components of meaning are
Steve Nicolle
retained in fully grammaticalised items as an integral part of the meaning of such grams. The reason for such macro-variation may lie in the morphosyntactic characteristics of language families rather than in conceptual differences concerning which semantic domains are amenable to grammaticalisation. It could be argued that at a functional level English has distal aspects derived from go and come occurring in the syntactic frame ‘go/come and V’. These often occur in imperatives (‘Go and buy the paper.’ ‘Come and eat.’) but may also occur in other moods (‘She went and saw the doctor this morning.’ ‘Can you come and give me a hand later?’). A limited range of other verbs may also occur in this syntactic frame, including stay (but not remain), run (but not walk) and, in some dialects, try (but not attempt: ‘I’ll try/*attempt and finish before five.’). It is possible that such constructions exhibit some degree of formal grammaticalisation, however the morphosyntactic structure of English seems to be relatively resistant to grammatical innovations. In Bantu languages however, formal grammaticalisation of TAM markers occurs relatively rapidly, which may favour the retention of meaning types which in other languages are subject to semantic bleaching. A relevant study would be to compare the kinds of categories grammaticalised in agglutinating languages (such as Bantu) compared to inflecting and isolating languages. It would also be necessary to look in some detail at word formation in agglutinating languages, since the ease of affixation in Bantu languages is a feature not only of their agglutinating character but also of the relative position within the verb phrase of TAM morphemes and lexical verbs. The fact that auxiliary verbs, taken in the widest sense of the phrase (see Heine 1993), and TAM markers both usually precede the main verb stem, with very little intervening material, means that it is easy for an auxiliary verb to become an affix in most Bantu languages. If genetically unrelated agglutinating languages with similar structures to Bantu exhibit a higher concentration of ‘non-standard’ aspectual categories such as direction and expectation than languages of other typological types, this would provide some support for the above hypothesis.
Glossary app caus con cop fut inf
Applied/Applicative suffix Causative suffix Consecutive tense Copula Pure Future tense Infinitive prefix
pas pfv pos prog pst refl
Passive suffix Perfective aspect Possessive marker Progressive aspect Past tense Reflexive marker
Distal aspects in Bantu languages
itv loc neg
Iterative aspect Locative marker Negative marker
rel seq sub
Relative marker Sequential tense Subjunctive suffix
Notes . Digo is classified as E.73 (Guthrie 1967–1971), North-East Coast (Nurse 1999). Initial research into Digo was conducted in Kenya by Martien de Groot and Andy Clark under Research Permit of the Office of the President No. OP.13/001/17 C 180/20. Fuliiru is classified as D/J.63, Great Lakes. I am grateful to Roger and Karen Van Otterloo for the Fuliiru data. . There are two pieces of evidence in favour of this analysis: First, verbs in Swahili do not occur in isolation except as imperatives. Second, stress in Swahili occurs on the penultimate syllable of polysyllabic words, and in (1b) only the penultimate syllable of the main verb is stressed. . The perfective marker -me- originally also derived from a lexical verb meaning ‘finish’. The reoccurrence of similar grammaticalisation processes or ‘chains’ (Heine 1993: 53) within a single language known as a ‘linguistic cycle’ (Hodge 1970; see also Ono 1992). Sometimes a linguistic cycle may result in the loss of the earlier grammaticalised morpheme, but typically the old and the new co-exist and are semantically distinct. . I am grateful to Barb Heins of SIL Mozambique for this suggestion. . In Digo /kw/ is realised as a double plosive, written /kpw/. Although the verb stem in kpwenda is strictly speaking -enda and the initial kpw- is the non-finite prefix, in some tenses, the whole form kpwenda occurs. The same applies to the verb kpwedza ‘come’ (see examples (22) and (23) below). . Class 5 prefixes with the form i- still function as non-finite markers in some of the northwestern Bantu languages in Guthrie’s zones A and C (Hadermann 1999: 434–435). . Example (16) is written as a single word because the main verb stem, gwa (‘fall’), is monosyllabic, and phonologically all monosyllabic words are cliticised to the preceding or following word. The apostrophe indicates a ‘semantic’ word break. . When -ga- is the only TAM marker in a clause it functions as a future (or relative future) tense marker, as in examples (14), (15) and (17). However, this is not the function of -ga- in distal aspect constructions, where its primary function appears to be grammatical. For this reason it is glossed simply as -GA- in the examples. . Note the similar use of a locative particle in the German ‘progressive’ construction with the preposition am (‘at’): sie ist am lesen ‘she is in the process of writing’. . Van Otterloo (p.c.) cites the fact that the emphatic extension can be suffixed to an aspect marker as an argument for the fact that a word break exists at that point because the emphatic extension must directly precede the final vowel.
Steve Nicolle
References Botne, R. (1999). Future and distal -ka-’s: Proto-Bantu or nascent form(s)? In J.-M. Hombert & L. M. Hyman (Eds.), Bantu Historical Linguistics: Theoretical and Empirical Perspectives (pp. 473–515). Stanford, CA: CSLI. Bybee, J. L., Perkins, R. D., & Pagliuca, W. (1994). The Evolution of Grammar: Tense, Aspect and Modality in the Languages of the World. Chicago: University of Chicago Press. Givón, T. (1991). Serial verbs and the mental reality of “event”: Grammatical versus cognitive packaging. In E. C. Traugott & B. Heine (Eds.), Approaches to Grammaticalization, Vol. 1 (pp. 81–127). Amsterdam: John Benjamins. Guthrie, M. (1967–1971). Comparative Bantu: An Introduction to the Comparative Linguistics and Prehistory of the Bantu Languages. Vols. I–IV. Farnborough: Gregg International. Hadermann, P. (1999). Les formes nomino-verbales de classes 5 et 15 dans les langues bantoues du Nord-Ouest. In J.-M. Hombert & L. M. Hyman (Eds.), Bantu Historical Linguistics: Theoretical and Empirical Perspectives (pp. 431–471). Stanford, CA: CSLI. Heine, B. (1993). Auxiliaries: Cognitive Forces and Grammaticalization. Oxford: Oxford University Press. Hodge, C. (1970). The linguistic cycle. Language Sciences, 13, 1–7. Nicolle, S. (1998a). A relevance theory perspective on grammaticalization. Cognitive Linguistics, 9, 1–35. Nicolle, S. (1998b). Be going to and will: A monosemous account. English Language and Linguistics, 2, 223–243. Nurse, D. (1999). Towards a historical classification of East African Bantu languages. In J.-M. Hombert & L. M. Hyman (Eds.), Bantu Historical Linguistics: Theoretical and Empirical Perspectives (pp. 1–41). Stanford, CA: CSLI. Nurse, D., & Hinnebusch, T. (1993). Swahili and Sabaki: A Linguistic History. Berkeley etc.: University of California Press. Nurse, D., & Muzale, H. R. T. (1999). Tense and aspect in Great Lakes Bantu languages. In J.-M. Hombert & L. M. Hyman (Eds.), Bantu Historical Linguistics: Theoretical and Empirical Perspectives (pp. 517–544). Stanford, CA: CSLI. Ono, T. (1992). The grammaticization of the Japanese verbs oku and shimau. Cognitive Linguistics, 3, 376–390. Schwenter, S. (1994). The grammaticalization of an anterior in progress: Evidence from a peninsular Spanish dialect. Studies in Language, 18, 71–111. Van Otterloo, R. (2000). Kifuliiru aspects “on stage”. Paper presented at the SIL Bantu Initiative Workshop, Ruiru, Kenya, February 2000. Van Otterloo, R. (2001). Bantu adverbial auxiliaries. Paper presented at the SIL Bantu Initiative Tense Aspect Mood Meetings, Limuru, Kenya, February 2001.
From temporal to conditional* Italian qualora vs English whenever Jacqueline Visconti University of Birmingham, UK
.
Introduction
Grammaticalization, the process whereby “lexical material in highly constrained pragmatic and morphosyntactic contexts is assigned functional category status” (Traugott & Dasher 2002: 81), plays a crucial role in the evolution of ‘complex conditional connectives’, such as Italian qualora or English supposing (that) (Visconti 2000), which ‘originate’ from lexical material of various kinds – verbs, nouns, adjectives, prepositions. The cross-linguistic diachronic investigation of these connectives, for reasons that will be highlighted in this paper, sheds a new light on the interaction between lexicon and structure in the development of clause linkage markers. The first part of this paper focuses on Italian qualora (§2). The analysis identifies the basic semantic structure underlying the evolution of this connective, highlighting the role of mood and aspect in the process. The second part focuses on English whenever, one of its ‘equivalents’ as suggested by bilingual dictionaries (§3). The data are related to two important issues in semantic change theory: (i) unidirectionality, i.e. the tendency of meanings to shift from one domain to the other (e.g. from TEMPORAL to CONDITIONAL) but not vice versa; (ii) subjectivization, “the development of a grammatically identifiable expression of speaker belief or speaker attitude towards what is said” (Traugott 1995: 32). The question is addressed whether the evolution of qualora constitutes a case of grammaticalization (vs other phenomena, such as lexicalization) (§4). Finally, the nature of the tertium comparationis (Krzeszowski 1990; Fisiak 1990; Jaszczolt 1995) and the importance of the diachronic dimension in contrastive analysis are discussed (§5).
Jacqueline Visconti
. Qualora: From temporal to conditional According to dictionaries and grammars of Present Day Italian, qualora, which is defined as “nel caso che”, “nell’eventualità che” ‘in the event that’, “se” ‘if ’, “se eventualmente” (GDLI), “se mai”, “se per caso” ‘if per chance’ (Zingarelli), has conditional value and selects the subjunctive mood, as in the examples: (1) Qualora avvenissero dei mutamenti, vi prego di informarmene tempestivamente [Zingarelli]. ‘Should any changes occur, please inform me immediately’. (2) Qualora restituisca il mal tolto, rinuncerò a denunciarlo [Devoto-Oli]. ‘If he gives back what he has stolen, I won’t report him to the police’. Most dictionaries refer also to a “temporal” value of the connective, which is indicated as “desueto” ‘obsolete’, “antico” ‘archaic’ (DISC; Treccani), or “letterario” ‘literary’ (Garzanti; Zingarelli). In fact, the latest example provided by dictionaries of such a temporal value dates back to the XIX century: (3) Naturalmente l’animale odia il suo simile, e qualora ciò è richiesto all’interesse proprio, l’offende. (G. Leopardi, Pensieri [GDLI]) ‘By nature animals hate their kin, and whenever it is their interest to do so, they will offend them’. To trace the different fates of the temporal and conditional values, I examined qualora in texts from previous centuries. The data are taken from two corpora: (i) Tesoro della Lingua Italiana delle Origini (henceforth TLIO), elaborated by the Opera del Vocabolario Italiano, which comprises a data-base of nearly 1400 texts from the period prior to 1375;1 (ii) Letteratura Italiana Zanichelli 3.0. (henceforth LIZ), which contains a data-base of 770 literary texts from 1250 to approximately 1950.2 In TLIO qualora occurs 77 times. The first element that strikes us in the corpus is the prevalence of the indicative mood, vs the subjunctive found in Present Day Italian: only 9/77 subjunctive (12%); vs 68 indicative (88%). Of the indicative examples, 49 (64%) are present indicative, as in: (4) Qualora vol, la femena se mostra sempl’e plana e mena relegione como fose nonana; mai s’ela se vé l’asio, ben fai volta sotana: per l’un no lassa l’autro cortese né vilana. (Proverbiaquedicuntur, XIII in.- venez. [TLIO]) ‘When she wants, the female shows herself simple, plain and religious, but given the chance, she lifts up her skirt: she won’t leave one [man] for another, whether she’s a gentlewoman or a peasant’. eleven (14%) are imperfect, as in:
From temporal to conditional
(5) Omè, dove lascio io i cari amici? Dove le feste e il sommo diletto? Ove i cavalli, omai fatti mendici del lor signore? Ove quel ben perfetto ch’amor mi dava, qualora i pudici occhi d’Emilia vedeva e l’aspetto? (Boccaccio, Teseida, 1339–1341? [TLIO]) ‘Oh my, where do I leave my dear friends? Where the parties and the great pleasure? Where the horses, who have lost their master? Where that perfect good that love gave me, whenever I looked at Emilia’s honest eyes and face?’ five (6%) future, as in: (6) Ma noi non negheremo però che i savi non conoscano il male, e pur lo fanno; ma diremo che essi per quello non perdono il senno, con ciò sia cosa che, qualora essi vorranno, con la ragione ch’elli hanno, la volontà raffrenare, elli nell’usato senno si rimarranno. (Boccaccio, Filocolo, 1336–1338 [TLIO]) ‘But we won’t deny however that the wise know evil, and indeed practise it; but we will say that they don’t thereby lose their wisdom, since whenever they want, with the reason that they have, to curb their will, they will stay with their usual wisdom’. and only three (4%) perfect, as in: (7) Perdonami, o figliuolo di Maria, Per lo prezioso sangue che succhiasti Da Lei, qualora entrasti Nel verginal suo seno immaculato. (Petrarca, Disperse e attribuite, 1374 [TLIO]) ‘Forgive me, o Mary’s son, for the precious blood that you sucked from Her, when you entered her immaculate virginal bosom’. The second surprising fact is that, in nearly all examples, as in the four quoted above, qualora has a temporal, rather than a conditional value: (4), (5), (6), (7), (8): (8) Elgli à due campane in Gallia, che suonano spessamente per lo vento; sonsi aveduti quelli della contrada, che qualora elle suonano per abbattenza ad uno punto, nascie nella contrada uno huovo, che non sanno d’onde si vengnia, il quale poi ricolgono le lor servigiali, e nasciene uno animale c’ à ffatte le menbra come huomo. (Fr. da Barberino, Regg., 1318–1320 - tosc. [TLIO]) ‘There are two bells in Gaul, that often toll because of the wind; the inhabitants have noticed that when(ever) they ring per chance at one point, an egg is born of unknown origin, which their servants collect and from which an animal is born that has limbs like a man’.
Jacqueline Visconti
Thus, the value of qualora would appear to have undergone a transformation from essentially temporal to conditional. In the first part of this paper, I shall try to account for this semantic shift.
. The data: Moods and tenses To capture the trend from temporal to conditional in a more systematic way, I searched LIZ for all occurrences of qualora, subdividing the results according to centuries. The results are summarized in Tables 1 and 2. Despite the non-representativeness of some of the data (in particular for the XV century, which yielded nine occurrences from the same text), the chart highlights a few interesting trends:3 i.
the ratio between subjunctive and indicative is inverted in the XVIII century, eventually leading to a total disappearence of the indicative two centuries later (as confirmed by search of synchronic corpora in Visconti 2000); ii. the overwhelming majority of the examples throughout the centuries are of present indicative, followed by the present and imperfect subjunctive and by the imperfect and future indicative; iii. the occurrences of perfective tenses, such as simple past and present perfect, are extremely rare. Table 1. Occurrence of qualora by mood, tense and century (LIZ) Mood and tense I n d i c a t i v e S u b j.
XIII
XIV
XV
XVI
XVII
XVIII
Present
2
30
2
16
31
20
Imperfect
–
10
1
1
3
Simple Past
–
4
–
1
Future
–
5
4
Pres. Perfect
–
–
Total ind.
2
Present
XIX
XX
Total
6
–
107
3
1
–
19
–
3
1
–
9
5
–
2
–
–
16
–
–
–
1
1
–
2
49
7
23
34
29
9
–
153
–
–
–
6
3
26
6
–
41
Imperfect
–
8
2
1
3
8
8
1
31
Pluperfect
–
–
–
–
–
–
1
2
3
Total subj.
–
8
2
7
6
34
15
3
75
Total examples
2
57
9
30
40
63
24
3
228
From temporal to conditional
Note that, though not in LIZ, the subjunctive is present, although marginal, right from the start: (9) Qualor que l’un de questi non fose [SUBJ.] a Venezia. . . vogo que ser Marco Çen . . . sia en logo de quello commesario. (Testi veneziani, 1282 [GDLI]) ‘Should any of these not be in Venice . . . I want sir Marco Çen . . . to be commissary at his place.’ (10) A cui la donna rispose: “Signor mio dolce, il quando potrebbe essere qualora più ci piacesse [SUBJ.], per ciò che io non ho marito a cui mi convenga render ragione delle notti; ma io non so pensare il dove”. (Boccaccio, Decameron, c. 1370 [TLIO]) ‘To whom the woman replied: “My sweet Lord, the when could be whenever we like most, as I don’t have a husband to whom I have to explain where I spend my nights, but I cannot think of a where”.’ Hence, the questions arise: (i) what triggers the change in mood selection? (ii) what is the role of the subjunctive in the semantic shift from temporal to conditional? (iii) is aspect (in particular the opposition between imperfective and perfective) a relevant factor in this process? Table 2. Percentage occurrence of qualora by mood, tense and century (LIZ) Mood and tense
XIII
XVIII
XIX
XX
Present
100
77.5
32
25
–
47
Imperfect
3
7.5
5
4
–
8
–
3
–
5
4
–
4
9
44.5
17
–
3
–
–
7
–
–
–
–
–
2
4
–
1
100
86
78
77
854
46
38
0
67
Present
–
–
–
20
7.5
41
25
–
18
Imperfect
–
14
22
3
7.5
13
33
33
14
Pluperfect
–
–
–
–
–
–
4
67
1
Total subj.
0
14
22
23
15
54
62
100
33
Total p. century
100
100
100
100
100
100
100
100
100
I n d i c a t i v e
S u b j.
XV
XVI
XVII
52.5
22
53
–
17.5
11
Simple Past
–
7
Future
–
Pres. Perfect. Total ind.
XIV
Total
Jacqueline Visconti
To address these questions, the precise mechanism underlying the evolution of qualora needs to be pinned down. I shall turn to this question in the following sections.
. Generic and habitual contexts The distribution of tenses noted in the previous section is not casual. The overwhelming majority of the examples across the centuries consists of ‘generic’ sentences, in which two sets of eventualities are set in an ideal correspondence without reference to a specific time or entity in a defined context: (3) Naturalmente l’animale odia il suo simile, e qualora ciò è richiesto all’interesse proprio, l’offende. (G. Leopardi, Pensieri, XIX cent. [LIZ (GDLI)]) ‘By nature animals hate their kin, and whenever it is their interest to do so, they will offend them’. followed by ‘habitual’ sentences,5 in which two sets of eventualities are said to cooccur/have co-occurred with some kind of regularity: (11) [. . . ] qualora non era con la sua Ambrosia, ogne contentezza e piacere in summa tristicia se convertia. (Masuccio Salern., Novellino [LIZ]) ‘[. . . ] whenever he was away from his Ambrosia, all happiness and pleasure turned into deepest sadness’. Cases of semelfactive interpretation, such as (7) above, or (12) and (13) below, are rare: (12) Tamas perdon mi chiede d’avermi a torto offesa/Me lo scordai qualora sposa d’Alì fui resa. (C. Goldoni, Ircana in Ispaan, XVIII cent. [LIZ]) ‘Tamas asks my pardon for having wrongly offended me/I forgot about it when I was given as Ali’s bride’. (13) Vinsi pugnando, e non usata forza/Provai nel braccio mio, qualora, o bella,/Il tuo nome invocai. (C. Goldoni, Belisario, XVII cent. [LIZ]) ‘I won fighting, and unusual strength/I felt in my arm, when, oh beautiful one,/I invoked your name.’ vs cases of generic or habitual interpretation, even with the future (14) or the simple past (15): (14) [. . . ] delle quali cose tutte avendo fatta copiosa conserva la buona madre di famiglia, qualora averrà che per alcuno impedimento non sian portate vivande di piazza a bastanza per la tavola o per la famiglia, o qualora da
From temporal to conditional
qualche forestiero saran sopragiunti, potrà in un punto arricchire la mensa in modo che non lassi desiderar la copia delle vivande comprate. (T. Tasso, Il Padre di famiglia, XVII cent. [LIZ]) ‘[. . . ] having made copious provision of all of which the good housewife, should it happen that for any reason not sufficient fresh food is brought from the piazza for the family table, or should they be joined by a visitor, will readily be able to enrich the table so that a lack of fresh produce is not felt.’ (15) So che, a pugnar qualora/Partisti armato o vincitor tornasti,/Gli ultimi e i primi baci erano i suoi. (P. Metastasio, Siroe, XVIII cent. [LIZ]) ‘I know that, to fight whenever/you left armed or returned the victor,/the last and first kisses were hers.’
. Qualora: A universal quantifier? To try to account for the data noted in §§2.1–2.2, let us look more closely at the meaning of the connective. Visconti (2000) argues that the distinctive feature of the semantics of qualora is the introduction of a particular quantificational structure over instants or intervals (ibid.: 132–135). The existence of such a component and its nature are related to the analysis of the two lexical units which form the connective, quale and ora. Among the values of quale (Lat. quale(m), of Indoeuropean origin) the dictionaries list that of indefinite adjective, with the meaning of “qualsiasi”, “qualunque”, ‘any’, ‘whichever’ (GDLI; Zingarelli; Treccani), as in the example: (16) Petre rotonde veioce, che venner dal fossato; da quale parte volgome, rompomece il costato. (Iacopone, XIII sec. [GDLI]) ‘Fast and round stones, coming from the ditch; whichever way I turn, they break my ribs’. If quale, like qualsiasi or qualunque (“any”, or “whichever”), assumes within qualora the function of denoting one arbitrarily chosen element of a class, thus evoking a potentially indefinite set of elements, the elements of the set are identified by the second component: ora, which, among its various meanings, is a “momento particolare del tempo in cui si determina un evento”, “circostanza, frangente”, and, archaic, “volta” = ‘time’ (GDLI). Hence, it is argued: “it makes sense to hypothesize that the lexical semantics of qualora contains a component expressing a generic universal quantification over the set of instants in which p (the denotation of an element quod libet of the class of moments in which the state of affairs evoked in p
Jacqueline Visconti
is realized), to assert that in those instants, i.e. “at any time”, or “every time” p, q” (Visconti 2000: 135) [my translation]. In this definition the denotation of an element quod libet of the class of moments in which the state of affairs in p is realized is equated with the expression of a generic universal quantification over the same set of instants, following Longobardi’s (1988) definition of qualsiasi and qualunque as “quantificatori universali [. . . ] intrinsicamente generici” ‘intrinsically generic universal quantifiers’ (ibid.: 645–651).6 This description does indeed account for a certain type of contexts in which Present Day Italian qualora is frequently found – i.e. normative, ‘atemporal’ propositions, of ideal validity, hypothetical precepts, which establish that every time a certain state of affairs occurs a certain effect takes place, such as (17) and (18): (17) Tali disposizioni indicano quale legge debba essere applicata dal giudice italiano qualora egli sia chiamato a risolvere controversie che non coinvolgono soltanto cittadini italiani o che riguardano fatti che non si sono svolti soltanto nel territorio italiano. (L. Bobbio, Corso di diritto [IRC]) ‘Such provisions indicate which law should be applied by Italian judges should they be called to resolve controversies which do not involve only Italian citizens or which concern acts which were not only enacted on Italian territory.’ (18) Il deputato eletto attraverso l’organizzazione del partito diventa un mandatario, se non degli elettori, del partito che lo penalizza revocandogli la fiducia qualora egli si sottragga alla disciplina, la quale quindi diventa un surrogato funzionale del mandato imperativo da parte degli elettori. (N. Bobbio, Stato, governo, società [IRC]) ‘Deputies elected through the party organisation become mandated, if not by the electorate, by the party which penalises them by withdrawing confidence should they shirk discipline, which thus becomes a functional surrogate of the necessary mandate from the electorate.’ However, the characterization of qualora as a universal quantifier does not account for examples such as (2) above, in which the connective seems to operate an existential, rather than universal, quantification:7 (2) Qualora restituisca il mal tolto, rinuncerò a denunciarlo [Devoto-Oli]. ‘If he gives back what he has stolen, I won’t report him to the police’. Although I have so far spoken of “universal quantification”, following Visconti (2000: 133–135), I believe a more precise characterization of this function to be the key to the diachronic puzzle outlined in the preceding sections.
From temporal to conditional
. Arbitrariness and intensionality In §2.3 I have argued that quale within qualora has the function of denoting one element a piacere of a class, thus evoking a potentially indefinite set of elements. According to this definition, qualora is not a universal quantifier such as “every”, but a free-choice operator of the “any”-type, which selects an element quod libet of the class of moments in which the state of affairs evoked in p is realized.8 Crucial to this definition are (i) the arbitrariness of the object on which qualora operates; (ii) the consequent intrinsic intensionality of qualora. Consider examples (19) and (20): (19) Che il timore sia, come ho detto altrove, più naturale all’uomo della speranza, e che l’uomo inclini più a quello che a questa, veggasi che qualora gli uomini ignorano le cagioni degli effetti o naturali o artifiziali, ordinariamente ne temono. (G. Leopardi, Zibaldone, XIX cent. [LIZ]) ‘That fear is, as I have said elsewhere, more natural for men than hope, and that men are more inclined to the former than the latter, can be seen in that whenever men ignore the causes of effects either natural or artificial, ordinarily they fear them.’ (20) E di questo albero e di questa fontana vi dirò mirabile cosa: che qualora l’amiraglio vuole far pruova della virginità d’alcuna giovane, egli nell’ora che le guance cominciano all’Aurora a divenire vermiglie, prende la giovane, la quale elli vuol vedere se è pulcella o no, e menala sotto questo albero. . . (Boccaccio, Filocolo, 1336–1338 [TLIO]) ‘And of this tree and of this fountain I will tell you an admirable thing: that whenever the admiral wishes to prove the virginity of some young lady, he, in the moment that Aurora’s cheeks start to blush, takes the young lady, the one he wants to see if she is a virgin or not, and leads her under this tree. . . ’ In (19), qualora selects an arbitrary member of the set of times in which the eventuality of human beings not knowing the causes of natural or artificial effects is true, to assert a correspondence between the selected member and a member of the set of times in which the eventuality of human beings fearing those effects is true. Similarly, in (20), qualora selects an arbitrary member of the class of times in which the eventuality of the admiral being willing to prove the virginity of a certain young lady is true, to assert a correspondence between that member and a member of the set of eventualities of the admiral taking the young lady under a certain tree. In both cases, qualora can be paraphrased by “at any time”. Indeed, both the free-choice meaning of qualora and its intensional nature are mirrored in the current debate on the nature of any.9 Of particular interest to the present study is the account offered by Tovena and Jayez (1999). Opposing current
Jacqueline Visconti
views of any as either a universal or an existential quantifier, the authors point at the (more abstract) logical notion of “arbitrariness” – in the sense of Fine (1985)10 – as the key notion in the analysis of both free-choice and negative polarity any. The two cases, state the authors, “are parallel in that the type of the situation, as described by the sentence, is not reducible to a finite conjunction of individual subtypes” (ibid.: 55). An object, following Fine, “is arbitrary with respect to a set P of properties if it has all and only the properties in P” (Tovena & Jayez 1999: 46) and hence it can be replaced by any object with the same properties. This definition provides us with a useful means of capturing a key feature – arbitrariness – in the semantics of qualora. The temporal entity over which quale operates is arbitrary because it can be replaced by any other one in which p is true: the eventuality over which qualora operates is arbitrary because it can be replaced by any other eventuality of the set. This defines the intrinsic intensionality of qualora. I shall now turn to another important element in the evolution of qualora: aspect.
. Aspect As noted in §2.2, the majority of the examples of qualora from the XIII to the XX centuries consists of generic sentences, in which two sets of eventualities are set in an ideal correspondence without reference to a specific time or entity in a defined context, such as (3), and habitual sentences, in which two sets of eventualities are said to co-occur/have co-occurred with some kind of regularity, such as (11). The characterization which I have so far outlined for qualora, and in particular its intensional nature, explains this interesting finding. What the two types of examples have in common is namely their intensional character: they do not refer to a particular, specific eventuality, nor to a finite list of eventualities belonging to the real world, but to the infinite set of possible worlds in which p is true. In this respect, although I have so far distinguished generic and habitual sentences (the subject of the former being interpreted in a classlike way, whereas the subject of the latter is an individual object), these two categories can be subsumed into one, which I will name ‘intensional’. This feature is related to the aspectual configuration of the examples, viz. (habitual) imperfective vs perfective. Further support for this idea is found in a recent study on the interaction between aspect and intensionality. Starting from the assumption that imperfective habitual sentences contain a quasi-universal quantification over events (cf. Bonomi 1995; Delfitto & Bertinetto 1995), Lenci and Bertinetto (2000) argue that the habitual imperfective aspect is typically intensional, whereas the perfective aspect, as well as introducing a different quantification (i.e. existential), differs from the former in
From temporal to conditional
that it is typically extensional. Following Dahl (1975) and others, they reserve the term “habitual” only for nomic generalisations about events: The most salient feature of nomic sentences is that they express lawlike generalisations, which cannot be reduced to quantifications over specific and limited sets of objects and which show a sort of intensional behaviour. (ibid.: 254–255)11
This is precisely the kind of contexts in which temporal qualora is most frequently found: present and imperfect indicative sentences that do not have an episodic, factual interpretation and that concern possible, non-actual cases. On the basis of the arguments outlined so far, I argue that: (i) the free-choice semantics of qualora leads to a potential ambiguity of the connective between a temporal and a conditional reading;12 (ii) the intensional character of temporal qualora favours its presence in intensional contexts (‘nomic’, or ‘habituals’); (iii) such contexts, which realize “the extrapolation of a regularity to be projected in the infinite intensional universe” (Bertinetto 1997: 212 [my translation]), ‘attract’ the subjunctive mood; (iv) the more and more frequent association of qualora with the epistemically marked subjunctive mood triggers in the speakers’ consciousness the association of qualora with the expression of a modal value.
. A model diachronic anaysis for qualora What I am arguing is that the semantic shift of qualora from temporal to conditional can be explained by reducing its semantics to a more abstract structure, the basic component of which is a free-choice operator over intervals, or eventualities. The arbitrariness of the temporal entities on which quale operates and the consequent intrinsic intensionality of qualora encourage its appearance in contexts that enhance the co-occurrence with the subjunctive. Consider how the indicative can be replaced by the subjunctive in ‘intensional’ (11), but not in ‘extensional’, semelfactive, contexts (7): (11) [. . . ] qualora non era [IND.]/fosse [SUBJ.] con la sua Ambrosia, ogne contentezza e piacere in summa tristicia se convertia. (Masuccio Salern., Novellino [LIZ]) ‘[. . . ] whenever he was away from his Ambrosia, all happiness and pleasure turned into deepest sadness’. (7) Perdonami, o figliuolo di Maria, Per lo prezioso sangue che succhiasti Da Lei, qualora entrasti [IND.]/*entrassi [SUBJ.] Nel verginal suo seno immaculato. (Petrarca, Disperse e attribuite, 1374 [TLIO]) ‘Forgive me, o Mary’s son, for the precious blood that you sucked from Her, when you entered her immaculate virginal bosom’.
Jacqueline Visconti
Thus, intensionality is the key factor in the diachronic analysis of the connective: (i) it triggers the switch in mood selection (from an ‘extensional’ to the ‘intensional’ mood by definition); (ii) it accounts for the modal value that qualora has in Present Day Italian, if we conceive modality as the “relativization of the validity of sentence meanings to a set of possible worlds” (Kiefer 1994: 2515). Note the affinity between free-choice operators, which evoke “a potentially indefinite set of events”, and the subjunctive mood, as identified by Wandruszka (1991): “The subjunctive is found after [. . . ] qualsiasi [‘any’]. The representation of the totality of a set is here communicated analytically by means of the semantic structure ‘it is irrelevant which element is concerned’,13 which may evoke the impression of a potentially indefinite set, even more so if it is a future event. The subjunctive is obligatory” (ibid.: 462–464 [my translation]). Two types of contexts are of particular interest in following the gradual enhancement of the frequency of the combination with the subjunctive in intensional contexts: i.
those in which the opposition between present indicative and subjunctive is neutralized in a morphologically ambiguous form, such as: (21) Se l’imagine di Polemone Filosofo [. . . ] veduta da quella meretrice dentro una camera, valse a spaventarla ed a raffrenarla mentre ch’era già in procinto di commettere disonestà, che dovrà fare il sembiante dell’onnipotente Iddio, giustissimo Giudice di tutte le nostre operazioni, qualora, o persuasi dalle lusinghe del senso o sollecitati dagli stimoli dell’affetto, a violare le divine leggi ci apparecchiamo con qualche scelleratezza? (G. B. Marino, Dicerie Sacre, XVII cent. [LIZ]) ‘If the image of Polemone Filosofo [. . . ] seen by that prostitute in a room, was enough to scare her and curbed her while she was already in the process of being dishonest, what will the image of the omnipotent God do, just Judge of all our acts, whenever, either persuaded by the allurements of the senses or solicited by the stimuli of desire, to violate the divine laws we prepare ourselves for some wickedness?’ (22) In che pecco qualora altrui mostr’io le cose belle? (G. B. Marino, Adone, XVII cent. [LIZ]) ‘How do I sin when I show beautiful things to others?’
ii. those in which indicative and subjunctive co-occur, such as: (23) Negli uomini, grandi sono le varietà di forze, di spirito, di corpo e di condizione, varietà tutte che pongono l’uomo sociale in istato di avvedersi della sua debolezza e de’ suoi bisogni, qualora si trova [IND.] solo, e della
From temporal to conditional
sua perfezione, qualora venga [SUBJ.] dagli altri soccorso ed aiutato. (AA.VV. Il Caffè, XIX cent. [LIZ]) ‘In men, great are the varieties of strength, of spirit, of body and of status, all varieties which place the social man in the position to notice his weakness and his needs, whenever he finds himself alone, and of his perfection, whenever he is succoured and helped by others.’ (24) La gloria, gli onori, le ricchezze, il potere, tutti diventano mali, e mali insopportabili all’uomo, tosto che accompagnino tenacemente ogni momento della sua esistenza. Chi possede questi beni, e da lungo tempo s’è abituato a possederli, ritrova i momenti più deliziosi della vita qualora gli riesca [SUBJ.] di confondersi col popolo ed essere dimenticato nella folla de’ gregari [. . . ] L’uomo in somma si consola qualora esce [IND.] da quello stato che incessantemente lo accompagna, e il villano entrando nella città prova quella gioia che sente il cittadino all’uscirne. (AA.VV. Il Caffè, XIX cent. [LIZ]) ‘Glory, honours, riches, power, all become evils man cannot tolerate, as soon as they accompany tenaciously every moment of his existence. He who has these gifts, and has been used to having them for a long time, finds life’s most delicious moments when he manages to mingle with the people and be forgotten in the gregarious crowd [. . . ] Man in short consoles himself whenever he leaves the state which is incessantly with him, and the yokel entering the town tastes the same joy that the townsman feels on leaving it.’ It is interesting to note the analogy with the use of quando with the imperfect subjunctive, as in (25): (25) Ma quando voi aveste [SUBJ.] a determinare un’altezza, come, per esempio, quanto sia alto questo palco dal pavimento che noi abbiamo sotto i piedi; essendo che da qualsivoglia punto del palco si possono tirare infinite linee [. . . ] di quali di coteste linee vi servireste voi? (Galileo, Dialogo sopra i massimi sistemi: 13, XVII cent.) ‘But when you had to measure a height, like, for example how high this ceiling is from the floor beneath our feet; given that from whichever point on the ceiling an infinite number of lines can be drawn [. . . ] which of the said lines would you make use of?’ According to historical grammars of Italian: “Solo nel caso in cui un avvenimento perfettivo è pensato in modo iterativo, la lingua (specie in antico) può usare il congiuntivo imperfetto” ‘only when a perfective event is thought of in an iterative
Jacqueline Visconti
way,14 Italian (especially Old Italian) may allow the imperfect subjunctive’ (Rohlfs 1969: 76), as in: (26) Molte volte io mi dolea, quando la mia memoria movesse [SUBJ.] la fantasia ad immaginare quale Amore mi facea. (Dante, Vita Nuova, 16 [Rohlfs]) ‘Many times I suffered, when my memory moved my imagination to imagine what Love did to me.’ “The subjunctive”, it is stated, “is justified here by the fact that we are not decribing a precise event, but simply considering a potential case” [my translation] (ibid.: 76–77).15 In this respect, the difference between quando and qualora is on the diachronic axis: the use of the subjunctive and hence the conditional value remain an option for the former in Present Day Italian (cf. Visconti 2000: 72, 76), whereas the change in mood selection provokes the insurgence of a modal component of the epistemic kind16 in the semantics of the latter, conventionalized in Present Day Italian.
. Whenever Modality seems to be the key to understanding the differences between qualora and whenever, which is suggested as a straightforward equivalent to qualora by several bilingual dictionaries (Cambridge; Cassell’s; Skey), qualified as “antico” ‘archaic’, or “antico letterario” ‘archaic literary’ by a few others (Hazon Garzanti; SansoniHarrap). Clearly, the equivalence of these two connectives in Present Day Italian/English is problematic, as shown by the oddity of the translations, in which qualora is used in purely hypothetical contexts and in counterfactual conditionals: (27) Qualora la seduta fosse rimandata, avvertimi subito. ‘*Whenever the meeting should be/were postponed, let me know at once’. (28) Qualora avessi perso l’aereo, sarei rimasto a Parigi. ‘*Whenever I had missed the plane, I would have stayed in Paris’. On the other hand, most of the examples of Old Italian qualora may be translated by whenever: (29) Omè, dove lascio io i cari amici? Dove le feste e il sommo diletto? Ove i cavalli, omai fatti mendici del lor signore? Ove quel ben perfetto ch’ amor mi dava, qualora i pudici occhi d’Emilia vedeva e l’aspetto? (Boccaccio, Teseida, 1339–1341? [TLIO])
From temporal to conditional
‘Oh my, where do I leave my dear friends? Where the parties and the great pleasure? Where the horses, who have lost their master? Where that perfect good that love gave me, whenever I looked at Emilia’s honest eyes and face?’ (30) E dicovi così che, qualora egli avvien che noi insieme ci raccogliamo, è maravigliosa cosa a vedere i capoletti intorno alla sala dove mangiamo e le tavole messe alla reale e la quantità de’ nobili e belli servidori, così femine come maschi [. . . ]. (Boccaccio, Decameron, 1370 [TLIO]) ‘[. . . ] And so I tell you that, whenever it happens that we gather all together, it is a marvel to see the drapes around the walls of the hall where we eat and the tables [. . . ]’. To trace both the roots of the original equivalence and its disappearance, I looked at some early texts. The data are from the diachronic component of the Helsinki Corpus (850–1707), the Lampeter Corpus of Early Modern English (1640–1740), the Corpus of Early English (1418–1680) and the online Oxford English Dictionary.17
. The data Tables 3 and 4 summarize the results of the scrutiny of moods and tenses selected by whenever throughout the centuries. With all the caution that such an examination induces (the data are less numerous than for qualora, most of the texts are from the Early Modern period), the tables provide evidence of two facts: i. the subjunctive is virtually non-existent in combination with whenever; ii. the clear majority of the examples are present indicative, followed by simple past, future and present perfect indicative. Such a tense and mood configuration has an interesting semantic counterpart. The examples, with a few exceptions that we shall consider later, fall within two categories: Type (i), ca. 75%, in which a systematic correlation between two series of eventualities is established,18 such as: (31) As þe popis clerkis feynen þat þei done miraclis whanne evere þei syngen, moo and more woundirful þan ever dide Crist or his apostlis. (Wyclif, c. 1380 [OED]) (32) After the Miser had Bought what he had left, for half the value, he forbid him his House, and whenever he met him, he pass’d by him as a Stranger. (Edward Ward, Labour in vain, London, 1700 [Lampeter])
Jacqueline Visconti
Type (ii), where an unspecified moment in time is evoked, in which the eventuality expressed in the main clause is located, as in: (33) I know it has been alledged, that the Test Act has been the Security of the Church; and whenever it is repealed, the Church will be in Danger. But this cannot possibly be made appear to be true. On the contrary, it will be found to be the Discredit and Ruin of it. (Anon. The rights and liberties of subjects vindicated, London, 1732 [Lampeter]) (34) These things, together with the known Endeavours of the French to procure an Interest amongst the Natives of that Country, and especially with Don Pedro and Corbet, in order to a Settlement, make it evident enough that it is the Interest of Spain the Scots should rather have it than the French, who have already been tampering with the Spaniards as well as with the Indians, and doubt not to have a large share of America whenever the King of Spain dies. (A. Foyer and A. Fletcher (attr.), A defense of the Scots settlement at Darien, Edinburgh, 1699 [Lampeter])
Table 3. Occurrence of whenever by mood, tense and century Mood and tense I n d i c a t i v e
S u b j.
XIII
XIV
XV
XVI
XVII
XVIII
XIX
Total
Present
–
1
2
1
13
13
1
31
Imperfect
–
–
–
–
–
–
–
0
Simple Past
–
–
–
–
3
6
2
11
Future
–
–
–
1
–
3
–
4
Pres. Perfect
–
–
–
–
–
2
1
3
Total ind.
0
1
2
2
16
24
4
49
Present
–
–
1
–
–
1
–
2
Imperfect
–
–
–
–
–
–
–
0
Pluperfect
–
–
–
–
–
–
–
0
Total subj.
0
0
1
0
0
1
0
2
Total examples
0
1
3
2
16
25
4
51
From temporal to conditional
Table 4. Percentage occurrence of whenever by mood, tense and century Mood and tense I n d i c a t i v e
S u b j.
XIII
XIV
XV
XVI
XVII
XVIII
XIX
Total
Present
–
100
67
50
81
52
25
60
Imperfect
–
–
–
–
–
–
–
0
Simple Past
–
–
–
–
19
24
50
22
Future
–
–
–
50
–
12
–
8
Pres. Perfect
–
–
–
–
–
8
25
6
Total ind.
96
Present
–
–
33
–
–
4
–
4
Imperfect
–
–
–
–
–
–
–
0
Pluperfect
–
–
–
–
–
–
–
0
Total subj.
Total p. century
4 100
100
100
100
100
100
100
100
Even though whenever appears to operate two different kinds of quantification in type (i) – universal – and type (ii) – existential, these two categories share one important feature: the disanchoring of the eventuality in p from a specific temporal localization in the actual world. In other words, according to the definition suggested in §2.5, both types are ‘intensional’: they do not refer to a specific eventuality or to a list of eventualities belonging to the real world, but to the infinite set of possible worlds in which p is true. An (apparent) exception to this statement is offered by particular, ‘extensional’ examples, such as (35): (35) He gave me a good supper last night when ever I came within his doors. (Sorel’s Com. Hist. Francion VIII, 1655 [OED]) paraphrased as “at the very time or moment when; as soon as” by the OED. However, the almost total disappearance of this structure, “now only in Scottish and Irish use” (OED), provides a further argument to the hypothesis of the intensional nature of whenever.
Jacqueline Visconti
. Whenever: A free-choice operator If we look more closely at the semantics of the connective, we notice that whenever is characterized by the same structure as that underlying qualora: the wh-element when+the indefiniteness marker ever, together denoting an unspecified moment in time, resulting in a free-choice indefinite marker such as: “at any time when”.19 In the case of whenever the free-choice component is etymologically related to universal quantification: ever, as pointed out by Leuschner (1996), derives from the Proto-Germanic adverb *aiw (> Old English a¯ + either in feore ‘in life’ or + byre ‘in the event’ > 5fre > Middle English efre > ModE. ever); it is argued: “There is no doubt that both a¯ and 5fre, in contrast with Modern English ever, can be translated as either ‘ever’ or ‘always”’ (ibid.: 474): (36) *aiw > a¯ (+in feore/byre) > 5fre > efre > ever. The free-choice semantics of whenever, in analogy to qualora (§2.3), would explain its frequency within intensional contexts, such as (37) and (38): (37) And therefore if no Laws must be made to promote the Making or Consumption of our own Goods, nor to hinder the Importation or Consumption of any from abroad, it must inevitably follow, that when ever a Nation falls into Luxury, and the People to Idleness, or to spend their Time in Imployments unprofitable to a Nation, such a Nation must be reduced to beggery by Trade. (J. Pollexfen, England and East India inconsistent in their manufactures, London, 1697 [Lampeter]) (38) It appears from what has been said, that our Saviour has given no Commission to exercise the Office of a Bishop in such a particular Place, but to such as submit to the Government of that Country in which they reside: and if they are not Bishops but in that one Place where first they were design’d to preside, if they have not (I say) any larger Commission, then whenever they refuse to acknowledge the Civil Government, their whole Commission is void, and they are not any longer Bishops. (H. Humphry, A letter [. . . ], Oxford, 1692 [Lampeter]) The question, however, arises: if whenever and qualora are both intensional in nature, why has whenever not undergone the same process as qualora and acquired conditional value?
From temporal to conditional
. Mood and modality The difference, I shall argue, concerns the lack of interaction between whenever and the subjunctive mood. Only 2 instances (of which example 40 is rather marked) were found out of 51 occurrences (4%): (39) Ser, on to hir loggyng, When euer it please yow, I shall be your gyde. (Generydes, c. 1440 [OED]) (40) Whenever, and which way soever it be; great is the peculiar and distinguishing Excellence of This Charity. (J. Trapp, The dignity, and benefit, of the priesthood, London, 1721 [Lampeter]) A first relevant factor is clearly the loss of importance of the subjunctive mood throughout the history of the English language, starting from Old English onwards.20 Instead, the expression of modality in English came to reside with modal verbs such as can, may, shall, etc.21 Nevertheless, only in 3 cases (6%) does whenever co-occur with modal verbs.22 The explanation I propose is the following: as whenever, like qualora, is intensional in nature, it would be expected to attract the subjunctive mood. However, given the different mood configuration in the two languages and in particular the loss of importance of the subjunctive in the history of English, the mechanism for the transition from temporal to conditional was not available in the case of whenever, the lack of contexts in which the connective interacts with the subjunctive not allowing modality to become a conventionalized feature in its semantics.
. Conditionals and grammaticalization The mechanism for the shift in mood selection (and hence in value) of Italian qualora, according to the analysis presented in this paper, is grounded compositionally in the lexical semantics of the connective, which results from the combination of quale as free-choice adjective + ora, a noun indicating time. An interesting question to be addressed, which I shall only mention here,23 is to what extent the development of qualora constitutes an instance of grammaticalization. The boundary between grammaticalization and, e.g., lexicalization, is a fascinating issue itself.24 As “hallmarks” for identifying changes that fall within the domain of grammaticalization, Tabor and Traugott (1998) suggest: (i) “morphosyntactic change”; (ii) “pragmatic/semantic change”; (iii) “gradualness in the sense that some subtypes of a new construction become possible before others” (ibid.: 235). In Traugott (1998) the “decategorialization of a V, N, or A in a con-
Jacqueline Visconti
struction and eventual reanalysis as a grammatical one, e.g. case, tense, aspect, mood marker, complementizer, discourse particle” and the “shift to specifier or operator status” (ibid.: 9) are considered as essential correlates for grammatical as well as semantic change.25 In this respect, as a case of categorial downshifting from Adjective + Noun to Preposition/Connective, qualora constitutes an instance of grammaticalization, its development differing from clear cases of lexicalization, such as compounding, as in black + board, or shifts within major categories, such as that from bottle (N) to bottle (V), and also from more borderline cases, such as the fusion of the parts of a compound noun into stem-suffix, as in child + hood, or king + dom.26 Another argument in favour of this hypothesis is the presence in the evolution of qualora of other features typical of grammaticalization: (i) the increase in structural scope, following Tabor and Traugott (1998); the development of a meaning which is both (ii) more abstract27 and (iii) more subjective: the assumption of an epistemic value in the evolution of qualora is indeed an interesting case of subjectivization, “the process whereby speakers/writers come over time to develop meanings for lexemes that encode or externalize their perspectives and attitudes [. . . ]” (Traugott & Dasher 2002: 30).28 Thus, the history of qualora provides a further case in favour of the much-debated hypothesis of the unidirectionality of language change:29 structurally, as an example of functional shift towards the role of operator; semantically, as an example of the tendency of meanings to shift from a more objective to a more subjective domain (from TEMPORAL to CONDITIONAL) and not vice versa. If we focus on the step-by-step change of qualora from an [Adjective + Noun] sequence to a connective, we notice another important feature confirming the evolution of qualora as a case of grammaticalization. The first occurrences in the Italian data-bases are the following, all from the same XIV–XV century text: (41) Vigilate adunque; imperò che non sapete in qual ora il vostro Signore debba venire. (Bibbia, XIV–XV (tosc.) [TLIO]) ‘Thus stay awake; because you do not know in which hour your Lord should come’. (42) Ma questo sappiate, che se il padre della famiglia sapesse in qual ora venisse il ladro, certamente egli vigilarebbe, e non permetteria che fosse scavata la sua casa. (ibid.) ‘Of this be aware, that if the father knew at which time the thief would come, he would surely watch and would not allow his home to be dug’.
From temporal to conditional
(43) E voi siate apparecchiati; imperò che in qual ora non pensate, il Figliuolo dell’uomo verrà. (ibid.) ‘Be prepared; because at which hour you think not, the Son of man will come’. In (41) and (42) qual ora appears within a PP in an indirect interrogative, in (43) in a PP with adverbial function modified by a relative clause. The indirect interrogative is overwhelmingly the most frequent configuration also in Latin, as shown by a search of the Latin data-bases of the Patrologia Latina and the Acta Sanctorum.30 Consider the Latin ‘equivalents’ of the above quoted biblical examples: (44) Vigilate quia nescitis qua hora Dominus vester venturus sit. (Historia Mortis et Miraculorum ex Mss. Strozziano, Caput I [Acta Sanctorum]) (45) Nam si scires pater-familias qua hora fur veniret, vigilaret vtique & non sineret perfodi domum. (Vita et Translation S. Valentini, Caput III [Acta Sanctorum])31 (46) Vigilate utique, milites Christi, quia qua hora non putatis, Filius hominis veniet. (Luc. XII, 40 [Patrologia Latina]) The first feature to be noted is the loss of case marking from the Latin ablative qua hora, the function of which is fulfilled in the Italian examples by the preposition in (qual ora). An interesting case for the reanalysis of qual ora as adverbial conjunction is (43), in which the preposition in could be omitted: ‘qual ora non pensate [IND.], il Figliuolo dell’uomo verrà’. More data are necessary before we can come to a conclusion. My prediction is that the analysis of a larger corpus will allow the identification of (43) as one of the contexts in which the syntactic reclassification of qual ora as a temporal connective is favoured.
. Conclusion This paper provides a contrastive diachronic analysis of two connectives, Italian qualora and English whenever, in which the relationship between temporal and conditional values is crystallized in different forms. Whereas in Old Italian qualora is constantly used with the indicative mood and has temporal value, in Present Day Italian it is exclusively found with the subjunctive and has conditional value (§2.1). Despite their apparent homomorphism as free-choice operators over temporal entities (§§2.4; 3.2), the same semantic shift does not occur for English whenever, which maintains its temporal value from Middle and Early Modern to Present Day English (§3). The explanation offered for such contrastive evidence focuses on
Jacqueline Visconti
the role of aspect, mood and modality in the evolution of the connectives (§§2.5; 2.6; 3.3). Sharing the assumption that “in order to contrast languages methodically, a comprehensive ‘criterion of measurement’ is required [which] will differ from one level of analysis to another” (Jaszczolt 1995: 562), I propose that the tertium comparationis (Krzeszowski 1990; Fisiak 1990) between qualora and whenever consists in their free-choice semantics. This is defined on the basis of the notions of (i) the arbitrariness of the temporal entities on which the connectives operate; (ii) the intrinsic intensionality of the connectives, which favours their appearance in nomic, or habitual contexts (§2.4; 2.5). The difference on the diachronic axis is due to the different organization of (aspect and) moods in the two languages,32 and, in particular, to the different role played by the subjunctive mood in adding a component of epistemic modalization to the meaning of qualora (whereby p is presented as nonfactual), which is lacking in whenever. Hence the oddity of the examples, in which qualora is used, as it is in Present Day Italian, in purely hypothetical contexts and in counterfactual conditionals: (47) Qualora la seduta fosse rimandata, avvertimi subito. ‘Should the meeting be postponed, let me know at once’. (48) Qualora avessi perso l’aereo, sarei rimasto a Parigi. ‘If I had missed the plane, I would have stayed in Paris’. (49) Qualora la seduta fosse rimandata, avvertimi subito. ‘*Whenever the meeting should be/were postponed, let me know at once’. (50) Qualora avessi perso l’aereo, sarei rimasto a Parigi. ‘*Whenever I had missed the plane, I would have stayed in Paris’. As the instantiation of the value of the connective as either temporal or conditional is dependent upon mood selection (i.e. indicative vs subjunctive), this paper underlines: (i) the relevance of considering a whole construction33 rather than a single expression as a source of semantic change; (ii) the importance of a systematic investigation of the interaction between lexicon and structure in the development of clause linkage markers; (iii) the role of specific semantic and pragmatic contexts (in this case, ‘generic’ and ‘habitual’) in favouring semantic change. An interesting area for further research is a more general diachronic evaluation of the role of the subjunctive mood in phenomena of semantic change, both in Italian and in other Romance languages.
From temporal to conditional
Notes * I would like to thank Elizabeth C. Traugott for her support and for many fascinating discussions during my stay at Stanford. Thanks also to Angela Ferrari, Martin Maiden, Lucia Tovena and Arnold Zwicky for many interesting remarks, and to the British Academy and the Arts and Humanities Research Board for funding the project. . Thanks to Pietro Beltrami for his help with this data-base. . Data for diachronic research, which are obviously found in written texts, are not necessarily representative of changes shared by the whole speech community, as they are usually produced by the upper strata of the society and often belong to text-types which require particular rhetorical strategies, such as literary, religious, pedagogical, legal, etc. (cf. Traugott & Dasher 2002: 45–47). However, this limitation does not affect substantially the present account, since most of the connectives analysed belong to high and formal varieties of the language. . For a more in-depth account cf. Visconti (in preparation). . Note the apparent anomaly in the trend marked by the Seicento, which could be related to puristic influences, in analogy with the phenomena identified in the history of the Italian language by D’Achille (1990). Bembo, for instance, uses qualora with the indicative only, as in: ‘Ma qualora esse la lettera del mezzo lasciano [IND.] adietro [. . . ]’ (Prose 1525); similarly, in the Vocabolario degli Accademici della Crusca (1612), which refers back to examples from Petrarch and Boccaccio, qualora is defined as a “temporal adverb”, with the value of ogni volta che ‘every time that’, alcuna volta ‘any time’, quando ‘when”’, as in: ‘Qualora egli avviene che insieme ci raccogliamo’ (Boccaccio 1370). . In this paper I follow the traditional distinction between “generic” sentences, the subject of which is interpreted in a classlike way (“kind-referring NPs” in Krifka et al. 1995: 2), and “habitual” sentences, expressing regularities, the subject of which is an individual object (“characterizing sentences” in Krifka et al. 1995: 2–3) (on the terminology cf. also Lenci & Bertinetto 2000, n. 4). The two phenomena share an interesting property: with the former we abstract from particular objects, with the latter we abstract from particular events (cf. Krifka et al. 1995: 4). We shall see the relevance of this property to our analysis in §2.5. . His definition, however, runs into the following problem (Tovena, p.c.): if qualsiasi realized a universal quantification, then the expression “in qualsiasi momento” ‘at any time’ could be paraphrased by “continuamente” ‘at all times’ (for all the x that are moments), which is not the case. Rather, qualsiasi and qualunque seem to express the same concept of free-choice that captures more appropriately the semantic value of qualora, as we shall see in §2.4. . For qualora, it was pointed out to me (Tovena, p.c.), one finds the same alternance between universal and existential quantification which is so intriguing in ‘any’ (cf. §2.4 and references therein). . On the difference and relationship between free-choice elements and universal quantifiers cf. Haspelmath (1993: 87–91, 153–154); Haspelmath (1995). . For a synthesis cf. Hand (1999: 180ff.); Haspelmath (1993: 87–91); Tovena and Jayez (1999: 40–43).
Jacqueline Visconti . Cf. Tovena and Jayez (1999: 46, 55). . Cf. also Jayez (1999): “As pointed out by Giannakidou, habitual sentences license nonveridical determiners such as any, in contrast with the progressive form. If non-veridicality is a form of negative sensitivity to actual events (see Giannakidou 1998), this is as expected. Habitual sentences do not refer to actual events but to series of events with global properties” (ibid.: §5.1). . It will be interesting to reflect on the ontological status of what I call so far ‘ambiguity’ (‘underspecification’?) (cf. Visconti in preparation). . Cf. the notion of arbitrariness as the “irrelevance of individual choice with respect to a judgement” (Tovena & Jayez 1999: 55). . Note the cross-linguistic importance of iterativity in this process: “The overlap between iterative temporal and ‘real’ conditional clauses is quite clear: an antecedent which has on more than one occasion been fulfilled and has on each occasion led to a given outcome gives rise to a (factual) statement ‘whenever x, then y’ (= ‘if x, then always y’). When and if are primarily distinguished by the degree of certainty they convey. Such an epistemic notion is just not relevant to generic statements: hence in this case the very significant degree of overlap between if and when (ever)” (Harris 1986: 430). Cf. also the conditional reading of a when-clause such as: ‘When Bill came home, John left’, where when is treated as whenever, i.e. the “frame of the set of occasions” (Traugott 1985: 295). . Cf. the use of Middle English when with the “modally marked form” (Visser 1966: 879), as in ‘Whan an unaged prynce is made knyght or be crowned king’, in which “the contrast between the certainty and the uncertainty of a future action is expressed [. . . ] (The prince will become a knight but may become a king)” (ibid.: 879). . Cf. Tekavcic (1972: 520–524): “to the subjunctive belongs the sphere of subjectivity: first of all the domain of will [. . . ] and finally the rather intellectual domain of doubt, uncertainty” [my translation] (ibid.: 521); Rohlfs (1969: 59–60, 76–77): “the subjunctive is the mood of the uncertain, of the doubt, possibility, supposition [. . . ]” [my translation] (ibid.: 59). Cf. also Wandruszka (1991: 15ff.); Visconti (2000: 76–78, 86) and references therein. . Thanks to Geoff Barnbrook, Mike Fraser, Sharon Krossa, Paul Taylor, and Glen Worthey for their help with the corpora. The periods for English are approximately as follows: Old English (450–1150); Middle English (1150–1500); Early Modern English (1500–1770); Modern English (1770–1970); Present Day English (1970–). . Cf Quirk et al. (1986): “Whenever is primarily used to introduce a frequency adverbial clause [indefinite frequency] indicating that the situation is repeated: It may also imply that the two clauses overlap in time if at least one of the clauses is durative” (ibid.: 1083; 529), as in: ‘They come here when(ever) they feel like it.’; ‘It rains whenever we’re camping’. . On the derivation of free-choice determiners from a wh-determiner plus an indefiniteness marker and on the hypothesis that universal distributive determiners derive from free-choice cf. Haspelmath (1995: 371–374); Leuschner (1996: 472). . Cf. Denison (1998: 160ff.); Visser (1966: 786–941 [789]) and references herein. . Cf. Denison (1998: 164); Visser (1966: 1632–1674, 1705–1734, 1742–1791).
From temporal to conditional . The examples are: (a) For whereas the former would look like the putting themselves into a condition of giving check to their Prince, whenever a ‘Capricio’ should take them, and they should fancy themselves agrieved; all that can be aimed at, or possibly compassed by the latter, is to have Justice equally administred according to the known Laws, which is no less his Majesties Interest than his Duty, to make wise and careful Provision for. (R. Ferguson, The late proceedings and votes of the parliament of Scotland, Glasgow, 1689 [Lampeter]) (b) The Pleasure we receive from external Objects, must be suited to the Appetite we intend to gratify. And that which results from the Actions of our Lives, whenever we would have it intense or durable, must be supposed adapted to our reasonable Faculties, and to spring from a Consciousness of having done well. (Anon., The rights and liberties of subjects vindicated, London, 1732 [Lampeter]) (c)
Whenever an Englishman would cry ‘All right!’ an American cries ‘Go ahead!’ (Dickens, Amer. Notes, 1842 [OED])
. For a more in-depth discussion cf. Visconti (in preparation). . Cf. Ramat and Hopper (1998); Traugott (1998). . Cf. the criteria in Traugott and Dasher (2002: 84–85, 87). . On the ambiguous status (lexical or grammatical?) of the category of Preposition cf. Traugott (1998: 1). . Cf. Tabor and Traugott (1998: 229) and references herein. . Cf. also Traugott (1995). . For a synthesis cf. Traugott (2000). . The Acta Sanctorum contain materials on the lives of the saints (Vitae, Passiones, Miracula, Translationes, Gloriae posthumae, inscriptions, etc.) from the beginning of the Christian era to the end of the sixteenth century. The Patrologia Latina Database comprises the works of the Church Fathers from Tertullian in 200 AD to the death of Pope Innocent III in 1216. . One also finds a few cases of quali hora (much rarer: in the Patrologia Latina the ratio is 545 hits of qua hora vs 1 hit of quali hora): ‘Si videres quoque in Quadragesima quali hora damus operam ut dormiamus, utique fatereris satis nos mature surgere posse ad nocturnos (Udalricus Cluniacensis, Antiquiores consuetudines cluniacensii monasterii, Liber I, Caput XLI, MLXXXVI)’. The precise relation of the more marked form quali hora vs qua hora (qualis deriving from *quo (QVI)+ -alis) to the Italian marker qualora needs to be investigated. . Cf. Bertinetto (1997: 183–203); Bybee et al. (1994). . ‘Construction’ is used here to refer to the sequence [connective+verb]. (For an overview of the notion of construction in cognitive and functional linguistics cf. Thompson 2001; on the role of constructions in grammaticalization cf. Traugott forthcoming.)
Jacqueline Visconti
References Bertinetto, P. M. (1997). Il dominio tempo-aspettuale. Torino: Rosenberg and Sellier. Bonomi, A. (1995). Aspect and quantification. In P. M. Bertinetto et al. (Eds.), Temporal Reference, Aspect and Actionality, Vol. 1 (pp. 93–110). Torino: Rosenberg and Sellier. Bybee, J., Perkins, R., & Pagliuca, W. (1994). The Evolution of Grammar. Tense, Aspect and Modality in the Languages of the World. Chicago/London: University of Chicago Press. D’Achille, P. (1990). Sintassi del parlato e tradizione scritta della lingua italiana. Roma: Bonacci. Dahl, Ö. (1975). On generics. In E. L. Keenan (Ed.), Formal Semantics of Natural Language (pp. 99–111). Cambridge: Cambridge University Press. Delfitto, D., & Bertinetto, P. M. (1995). A case study in the interaction of aspect and actionality: The imperfect in Italian. In Temporal Reference, Vol. 1 (pp. 125–142). Denison, D. (1998). Syntax. In S. Romaine (Ed.), The Cambridge History of the English Language, Vol. 4 (pp. 1776–1997). Cambridge: Cambridge University Press. Fabricius-Hansen, C., & S5bø, K. J. (1983). Über das Chamäleon wenn und seine Umwelt. Linguistische Berichte, 83, 1–35. Fine, K. (1985). Reasoning with Arbitrary Objects. Oxford: Basil Blackwell. Fisiak, J. (Ed.). (1990). Further Insights into Contrastive Analysis. Amsterdam: John Benjamins. Harris, M. B. (1986). The Historical Development of Conditional Sentences in Romance. Romance Philology, 39/3, 405–436. Hand, M. (1999). Semantics vs pragmatics: Any in game-theoretical semantics. In K. Turner (Ed.), The Semantics-Pragmatics Interface from Different Points of View (pp. 179–198). Oxford: Elsevier. Haspelmath, M. (1993). A Typological Study of Indefinite Pronouns. Ph.D., Free University of Berlin. Haspelmath, M. (1995). Diachronic Sources of All and Every. In E. Bach et al. (Eds.), Quantification in Natural Languages (pp. 363–382). Dordrecht: Kluwer. Jaszczolt, K. (1995). Contrastive analysis. In J. Verschueren, J. Östman, & J. Blommaert (Eds.), Handbook of Pragmatics (pp. 561–565). Benjamins. Jayez, J. (1999). Imperfectivity and Progressivity: The French Imparfait. SALT. Kiefer, F. (1994). Modality. In R. E. Asher & J. M. Y. Simpson (Eds.), The Encyclopedia of Language and Linguistics (pp. 2515–2520). Oxford: Pergamon Press. Krifka, M., Pellettier, F. J., Carlson, G. N., ter Meulen, A., Chierchia, G., Link, G. (1995). Genericity. An Introduction. In G. N. Carlson & F. J. Pellettier (Eds.), The Generic Book (pp. 1–124). Chicago: University of Chicago Press. Krzeszowski, T. P. (1990). Contrasting Languages. Mouton. Lenci, A., & Bertinetto, P. M. (2000). Aspect, Adverbs, and Events. Habituality vs Perfectivity. In Higginbotham et al. (Eds.), Speaking of Events (pp. 245–287). Oxford: Oxford University Press. Leuschner, T. (1996). Ever and Universal Quantifiers of Time: Observations from some Germanic Languages. In K. Jaszczolt & K. Turner (Eds.), Contrastive Semantics and Pragmatics (pp. 469–484). Oxford: Elsevier.
From temporal to conditional
Longobardi, G. (1988). I quantificatori. In L. Renzi (Ed.), Grande grammatica italiana di consultazione, Vol. 1 (pp. 645–696). Bologna: Il Mulino. Marques, R. R. (2000). Semantic and Pragmatic Constraints on Mood Selection. Paper presented at the Second International Conference in Contrastive Semantics and Pragmatics (Cambridge, 10–13 September 2000). Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J. (1985). A Comprehensive Grammar of the English Language. London: Longman. Ramat, A. G., & Hopper, P. (Eds.). (1998). The Limits of Grammaticalization. Amsterdam/ Philadelphia: John Benjamins. Rohlfs, G. (1966–1969 [1954]). Grammatica storica della lingua italiana e dei suoi dialetti, 3 voll. Torino: Einaudi. Tabor, W., & Traugott, E. C. (1998). Structural scope expansion and grammaticalization. In The Limits of Grammaticalization (pp. 229–272). Tekavcic, P. (1972). Grammatica storica dell’italiano, 3 voll. Bologna: Il Mulino. Tovena, L. M., & Jayez, J. (1999). Any: from scalarity to arbitrariness. In F. Corblin et al. (Eds.), Empirical Issues in Formal Syntax and Semantics, Vol. II (pp. 39–57). The Hague: Holland Academic Graphic. Thompson, S. A. (2001). Constructions and Conversation. Plenary paper presented at the International Cognitive Linguistics Conference (UCSB, 22 July 2001). Traugott, E. C. (1985). Conditional Markers. In J. Haiman (Ed.), Iconicity in Syntax, Proceedings of a Symposium (Stanford, California, 24–26 June 1983), (pp. 289–307). Amsterdam: John Benjamins. Traugott, E. C. (1995). Subjectification in grammaticalization. In D. Stein & S. Wright (Eds.), Subjectivity and Subjectivisation (pp. 31–54). Cambridge: Cambridge University Press. Traugott, E. C. (1998). Lexicalization and grammaticalization. In A. A. Cruse, F. Hundsnurscher, M. Job, & P. R. Lutzeier (Eds.), Lexikologie-Lexicology (pp. 1–14). Berlin: Walter de Gruyter. Traugott, E. C. (2000). From Etymology to Historical Pragmatics. Plenary paper presented at the Studies in English Historical Linguistics Conference (UCLA, May 27, 2000). Traugott, E. C. (forthcoming). Constructions in Grammaticalization. In R. D. Janda & B. D. Joseph (Eds.), A Handbook of Historical Linguistics. Oxford: Blackwell. Traugott, E. C., & Dasher, R. B. (2002). Regularity in Semantic Change. Cambridge: Cambridge University Press. Visconti, J. (2000). I connettivi condizionali complessi in italiano e inglese. Alessandria: Edizioni dell’Orso. Visconti, J. (in preparation). Conditionals and Grammaticalization. Visser, F. Th. (1966[–1969]). A Historical Syntax of the English Language, Vol. II. Leiden: E. J. Brill. Wandruszka, U. (1991). Frasi subordinate al congiuntivo. In Grande grammatica, Vol. 2 (pp. 415–481).
Jacqueline Visconti
Dictionaries Battaglia, S. (1961). Grande dizionario della lingua italiana [GDLI]. Torino: UTET. Devoto, G., & Oli, G. C. (1979). Vocabolario della lingua italiana. Firenze: Le Monnier. Il grande dizionario Garzanti della lingua italiana. (1987). Milano: Garzanti. Sabatini, F. (1997). Dizionario Italiano Sabatini-Coletti [DISC]. Giunti: Firenze. Vocabolario della Lingua Italiana. (1986–1994). Roma: Istituto della Enciclopedia Italiana Treccani. 5 voll. Zingarelli, N. (1996). Vocabolario della Lingua Italiana. Bologna: Zanichelli. Simpson, J. A., & Weiner, E. S. C. (1989). The Oxford English Dictionary [OED]. Oxford: Clarendon Press. Cassell’s Italian Dictionary. (1967 [1958]). New York: Macmillan. Reynolds, B. (1985). Cambridge Italian Dictionary. Cambridge: Cambridge University Press/Milano, Signorelli. Hazon (1990). Grande dizionario inglese-italiano italiano-inglese. Milan: Garzanti. Skey, M. (1995). Dizionarion inglese-italiano italiano-inglese. Oxford University Press, Oxford/SEI, Torino. Sansoni-Harrap (1970–1976). Standard Italian and English Dictionary. London, Harrap & Florence/Rome, Sansoni.
Then – adverbial pro-form or inference particle? A comparative study of English, Ewe, Hungarian, and Norwegian Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó University of Trondheim, NTNU / Eötvös Loránd University, Budapest
.
Introduction
The topic of this paper is the process of grammaticalization (e.g. Hopper & Traugott 1993; Heine, Claudi, & Hünnemeyer 1991; Heine 1993) that we claim to have affected the truth-conditional adverbial anaphor then in English and the corresponding lexical items in two other European languages – a close genetic relative: Norwegian, and a genetically unrelated language: Hungarian. Using Relevance theory as our theoretical framework of description (Sperber & Wilson 1995) we compare the adverbial markers in the three European languages mentioned to the corresponding conditional anaphor ekema in Ewe, a language belonging to the Gbe sub-branch of the Kwa branch of the Niger-Congo phylum, dialects of which are spoken in the region from Lower Volta in Ghana through Togo and Benin to Western Nigeria. We are going to account for why the Ewe anaphor ekema has not undergone a grammaticalization process similar to the one we postulate for English then, Hungarian akkor and Norwegian da, and we point out why we think certain data from Ewe give support to the proposed analyses of then, akkor and da.
. Grammaticalization of pronouns Grammaticalization is traditionally viewed as “the process by which full lexical items become grammatical morphemes” (Croft 1990: 230). The term “grammat-
Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó
ical morpheme” is meant to encompass both bound morphemes and so-called function words like connectives, pre- and postpositions, determiners, articles, and pronouns, including demonstrative ones. If a pronoun is a function word and a function word or a bound morpheme is what a lexical item ends up as when it undergoes all stages in a grammaticalization process, then we should not expect something that originates as a pronoun to grammaticalize. But there are attested cases of pronouns which have gone through a diachronic development of this sort. Hopper and Traugott (1993: 167) mention the two different clause linkage markers that in a sentence like (1) as examples of “the increased grammaticalization of already grammatical items in specific contexts (in this case, the context of clause combining)”. (1) That there will be an earthquake within the decade that will destroy the whole town is likely. The grammaticalized complementizer function of the predecessor of present-day that existed alongside the segmentally identical pronominal even in Old English. Another good example of pronominal grammaticalization is seen in the Norwegian extraposition construction in (2) compared with (3) where the comma indicates that the ‘that’-complement is an added afterthought. (2) Det er en fordel at Jon er så fysisk sterk. ‘It is an advantage that Jon is so physically strong.’ (3) Deti er en fordel, [at Jon er så fysisk sterk]i . ‘Thati is an advantage, [that Jon is so physically strong]i .’ The subject det is unstressed in (2) but stressed in (3) where there is a prosodic break before the ‘that’-complement. A demonstrative pronoun is exactly what one would expect to find when the complement is added as an afterthought. A stressed pronominal is used felicitously if its intended referent, in this case a higher-order entity, is at least activated for the conversing parties (see Gundel, Hedberg, & Zacharski 1993), while an unstressed pronoun requires a referent with the highest possible cognitive status, what Gundel et al. (1993) refer to as ‘in focus’. Gundel, Borthen, and Fretheim (1999) showed that the cognitive status of a higher-order entity is influenced by the number of times the entity has been processed in a segment of discourse. Norwegian det must be pronounced with stress the first time it occurs in B’s turn in (4) but it will normally be unstressed when it reappears in B’s final sentence, as indicated by the respective glosses ‘that’ and ‘it’. (4) A: [Jon er veldig glad i å svømme]i . ‘Jon is very fond of swimming.’
Then – adverbial pro-form or inference particle?
B: deti er ingen ulempe. Deti er en fordel. ‘Thati ’s no drawback. Iti ’s an advantage.’ An utterance like (3) with an afterthought ‘that’-clause is likely to be produced only if the speaker feels uncertain that the cognitive status of the higher-order entity referred to is high enough for a pronominal anaphor. In that type of situation the cognitive status of the subject pronoun would presumably be at most activated, not in focus, and an unstressed pronoun det is therefore unnatural in (3). Why, then, is it in order to let the pronoun det in (2) be unstressed? Det in that sentence, we shall argue, has lost its ability to refer through a process of grammaticalization and has ended up as an expletive form. What that development implies is that the higherorder entity referred to by means of the description in the extraposed clause does not have to meet the activation constraint on the felicitous use of the anaphor det. Compare the two dialogues (5) and (6) where the form det is an unstressed subject pronoun that has lost its referential status in the former but a demonstrative pronoun requiring an activated referent in the latter dialogue. (5) A: Jon har et veldig hardt program foran seg. ‘Jon has a very tough program ahead of him.’ B: Det er en fordel at han er så fysisk sterk. ‘It’s an advantage that he’s so physically strong.’ (6) A: Jon har et veldig hardt program foran seg. ‘Jon has a very tough program ahead of him.’ B: #det er en fordel, at han er så fysisk sterk. #‘That’s an advantage, that he’s so physically strong.’ No activated higher-order entity matching the description in the afterthought clause is available in (6), hence the cross-hatch sign in front of speaker’s B’s utterance, symbolizing discourse unacceptability. In (5) the assumption about Jon’s physical strength is presented as knowledge that B believes A to possess (‘be an advantage’ is a factive predicate), but the extraposition construction with its referentially empty subject pronoun is felicitous even if B happens to lack that knowledge, or to have a different opinion of Jon’s physical condition. In (6), however, the fact that Jon is physically strong is presented not just as retrievable information but as information believed to be currently in speaker and hearer’s working memory. Exactly the same is going on in (7) as in (6), although the two occurrences of det are both unstressed in (7). The copying of the pronoun in the position before the ‘that’-clause encodes the same cognitive status for its intended referent as the stressed pronoun in (6) does. (7) A: Jon har et veldig hardt program foran seg. ‘Jon has a very tough program ahead of him.’
Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó
B: #Deti er en fordel, deti , at han er så fysisk sterk. it is an advantage it that he is so physically strong #‘That’s an advantage, that he’s so physically strong.’ Compare (7) with (7 ) where A’s statement evokes the pragmatic implication that is verbalized in the right-detached complement of B’s response (‘Jon is physically strong’). (7 ) A: Jon løftet til og med den steinen der borte. ‘Jon lifted even that stone over there.’ B: Deti er en fordel, deti , at han er så fysisk sterk. ‘That’s an advantage, that he’s so physically strong.’ An unstressed pronoun without a dislocated partner encodes a higher cognitive status than the corresponding form in (7 ) with a dislocated copy of the pronoun. The regular unstressed pronoun, exemplified by the second occurrence of det in (4), demands that the discourse entity referred to is in focus for the speech participants; the dislocated alternative only demands that the referent is activated. If det were a referring expression in the utterance of (5) B, it would refer to the state of affairs described in the extraposed clause. However, as the unacceptability of (6) B and (7) B shows, det as a referring pronoun refers anaphorically, not cataphorically. It is not the right-detached afterthought complement which determines whether the discourse conditions for use of the higher-order entity pronoun are met. Rather, the difference between (7) and (7 ) demonstrates that the acceptability of B’s use of det depends on the assumptions evoked by A’s utterance and B’s decoding and inferential processing of its meaning. We can therefore conclude that det is neither an anaphoric nor a cataphoric pronoun in (5) B. The word must have lost its referential status in a process of grammaticalization which led to the emergence of what linguists know as the canonical extraposition construction in which the extraposed clause is not added as an afterthought and the subject is an expletive form. The present paper argues that the adverbial discourse anaphor then is yet another pronominal that has undergone a grammaticalization process resulting in loss of its truth-conditional meaning. But apart from the fact that the source meanings of Norwegian det and English then both include a truth-conditional component which has disappeared in the respective grammaticalization processes, these two grammaticalized products have very little in common. Norwegian det could assume the role of formal, expletive subject because the formal subject category exists in the grammar of Norwegian independently of the extraposition construction. Adverbial then gets to be used not only in contexts where it is understood to constrain the truth of the proposition expressed by the sentence it occurs in but also in contexts where it makes a different communicative contribution. The placement
Then – adverbial pro-form or inference particle?
of then in a parenthetical position at the end of the utterance makes it mutually manifest to speaker and hearer that the word is to be processed as a non-truthconditional marker. This is a syntactic slot reserved for modality markers which may constrain the speaker’s communicated attitude to the proposition expressed, like English parentheticals of the type . . . , I think, . . . , I suppose, . . . , I wonder, . . . , I take it, or which may instruct the hearer to construe the proposition of the utterance as metarepresenting (cf. Noh 1998, 2000; Wilson 1999) a thought which the speaker attributes to someone else, as in reports rounded off by a parenthetical . . . , he said, or . . . , from what I understand. The historical process that takes us from a temporal/conditional anaphor like English then to an inference particle is a feature familiar from a number of languages. And frequently the non-truth-conditional nature of the erstwhile anaphor is signalled formally by the word’s being lifted out of the sentence structure proper so that it comes to occupy an appositional slot instead. This is for example equally true of French alors, English then, and the corresponding words in the Scandinavian languages. The widespread tendency to use one and the same word as temporal/conditional anaphor and as inference particle suggests that there is an affinity between the two functions, and the nature of this affinity is what will be explored in the present paper, with reference to the three languages English (Section 3), Hungarian (Section 4), and Norwegian (Section 5). Finally we examine the corresponding function words in Ewe (Section 6) where the anaphora and the inference particle are not cognates.
. English Consider (8a–c). (8) a. Charlie is coming soon. I’ll confront him with a couple of facts. b. Charlie is coming soon. Then I’ll confront him with a couple of facts. c. Charlie is coming soon. I’ll confront him with a couple of facts, then. We understand all three versions (8a–c) to mean that the speaker is going to confront Charlie with some information at the time when Charlie arrives, but (8b) is the only one of them which contains an item that we are able to identify as a temporal modifier, then, referring to the time of Charlie’s arrival. The parenthetical then in (8c) seems to be more related to the anaphoric then used as a conditional marker than to the temporal marker appearing in (8b), yet it is not quite the same as the conditional marker. While it is possible to process then in (8b) as a conditional marker and get a relevant result, even on that interpretation there is a meaning difference between (8b) and (8c).
Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó
Adopting a relevance-theoretic perspective (Sperber & Wilson 1995) we would claim that (8a–c) are truth-conditionally identical. What is linguistically encoded in the second sentence in all three versions must be inferentially enriched, or saturated in the same way. A temporal anchoring of the future event described by the speaker is necessary both in (8a) and in (8c). In (8b) the temporal anchoring is made mutually manifest by virtue of the sentence-initial then. The crucial difference between (8b) and the other two is that there is an adjunct then there which encodes a constraint on the proposition expressed, though its conceptual content is established through the hearer’s inferential search for the intended antecedent of this anaphor. The relationship between then as a temporal modifier and as a conditional protasis is not the focus of this paper. An initial then, as in (8b), is semantically different from a right-detached then, as in (8c), but that difference is not one between a temporal and a conditional meaning; rather, the former is a truth-conditional adjunct and the latter is a non-truth-conditional inference marker representing the conversational interactants’ shared set of contextual assumptions bearing directly on the relevance of the utterance. We consider this to be a case of polysemy. The existence of a potential truth-functional difference between the respective final sentences in the three alternatives a., b. and c. above may be easier to appreciate when the frame sentence contains a modal auxiliary like may. Cf. the triplet (9a–c). (9) a. Charlie may be coming soon. I’ll confront him with a couple of facts. b. Charlie may be coming soon. Then I’ll confront him with a couple of facts. c. Charlie may be coming soon. I’ll confront him with a couple of facts, then. A stressed, fronted then, as in (9b), could in principle be interpreted either conditionally or temporally. On the conditional interpretation of then in (9b), the truth of the consequent ‘I’ll confront Charlie with a couple of facts’ is contingent upon the truth of the assumption that Charlie is coming. The speaker’s epistemic modal may in the preceding utterance shows that she is not sure he is coming. A parenthetical then on the other hand seems to have no impact on the explicature (Sperber & Wilson 1995; Carston 1998) of the speaker’s utterance, which is therefore the same in (9a) and (9c). (9c), like (9a), causes us to infer that the speaker does not doubt that Charlie is coming, in spite of the modal reservation expressed in the first declarative. The modal auxiliary may in that first sentence could simply be an echo of what the speaker has interpreted as a reservation expressed by the individual who is the source of the speaker’s information about Charlie. Now look at the three alternative follow-up interrogatives in (10a–c).
Then – adverbial pro-form or inference particle?
(10) I would even be willing to share the expenses. a. then will you be satisfied? b. Will you be satisfied then? c. (#) Will you be satisfied, then? The accented occurrences of then in (a) and (b) are anaphoric representatives of the condition specified in the first utterance, but the unaccented parenthetical then in version c. does not seem to represent the condition expressed in the first utterance. In (10a–b) the proposition embedded under the interrogative operator includes the condition linguistically represented by then, but then in (10c) is not included in the proposition expressed. (10a) and (10b) invite the implicature that the speaker realizes that the hearer was dissatisfied with the situation alluded to. His dissatisfaction has caused her to propose to do something in order to appease him. Our experience is that native speakers of English tend to frown upon an utterance like (10c) in the context of the preceding information. Their reaction is understandable, because the hearer has presumably not (yet) said anything that warrants the conclusion that he is satisfied with the speaker’s offer to share expenses. Parenthetical then in (10c) is a grammaticalized then which has lost the truth-functional meaning of temporal and conditional then. It belongs in a context where the speaker is heard to react to some stimulus coming from the interlocutor. A parenthetical then in English is often utterance-final but it can also occupy other positions where an apposition may be inserted. One frequent place for it is between the subject nominal and its complement, to indicate that what follows then is a conclusion about the topic specified in the subject phrase. The text presented in (11) is a passage from a paper by Diana Archangeli (Archangeli 1997) – and we are now only interested in the use of then in the fourth line, not in the author’s phonological message. (11) The linguist notes that the negative prefix takes the form im-, which ends with a labial nasal (m), whenever it precedes an adjective which begins with a labial stop (p, b), otherwise it takes the form in-: imbalance, impolite, but inoperative, intangible, infallible, inviolable. The prefix, then, is analyzed as having an input form, in-, which relates to two different output forms, in- and im-, depending on the context in which the prefix is placed. (Archangeli 1997: 3) What exactly is the function of the appositive then placed between the topical subject NP the prefix and the rest of the declarative? Then may be said to represent the set of assumptions mentioned in the first period of the quotation, which in the author’s opinion permits one to draw the conclusion, or extract the generalization expressed in the second and last period of the quotation. Steve Nicolle has
Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó
reminded us of the ‘framing’ effect of parenthetical then in (11). The word serves to refocus the prefix per se rather than its set of allomorphs. If then were left out here, the careful reader would presumably still infer that the statement about the prefix in- is to be understood as a generalization based on the foregoing observations about its possible realizations, but there would then be no linguistic signal that causes the reader to automatically read the last period that way. By presenting the word then as an apposition and deliberately failing to integrate it in the clause structure the author evidently takes it for granted that the reader is capable of working out the logical relations between the linguistic structures in the first and the second period of (11), and the appositional then is presumably added simply in order to speed up the reader’s processing and thereby contribute to the relevance of this passage in her text. It has a procedural, rather than conceptual meaning in the sense of Blakemore (1987) and Wilson and Sperber (1993). If then had been syntactically integrated in the clause as a regular adverbial adjunct – what we have actually done in (11 ) and (11 ) by tampering slightly with Archangeli’s text – then the logical relations between the meanings of the two periods would be asserted rather than presupposed. (11 ) The linguist notes that the negative prefix takes the form im-, which ends with a labial nasal (m), whenever it precedes an adjective which begins with a labial stop (p, b), otherwise it takes the form in-: imbalance, impolite, but inoperative, intangible, infallible, inviolable. The prefix is then analyzed as having an input form, in-, which relates to two different output forms, in- and im-, depending on the context in which the prefix is placed. (11 ) The linguist notes that the negative prefix takes the form im-, which ends with a labial nasal (m), whenever it precedes an adjective which begins with a labial stop (p, b), otherwise it takes the form in-: imbalance, impolite, but inoperative, intangible, infallible, inviolable. Then the prefix is analyzed as having an input form, in-, which relates to two different output forms, in- and im-, depending on the context in which the prefix is placed. If Archangeli’s text had been as shown in (11 ) or (11 ), the word then would have been an anaphor, a conditional pro-form whose antecedent would have to be sought in the preceding period. Then is truth-conditional in (11 ) and (11 ), as it was in (10a–b) and in (8b) and (9b). It is a non-truth-conditional inference marker in its parenthetical positions where its communicative role is not to contribute to the explicature of the utterance but to direct the hearer’s attention to a relation between one or more premises and a conclusion which would presumably
Then – adverbial pro-form or inference particle?
be inferable even without the particle but which its presence makes more manifest to the hearer. We believe the parenthetical placement of then to be the linguistically overt sign of a grammaticalization process which has resulted in the loss of the referring property of the original conditional anaphor. The systematic syntactic difference between the parenthetical and non-parenthetical positions offers procedural information about how the addressee is supposed to process the word in a concrete utterance. We propose to describe the relationship between the conditional anaphor then and the inference particle then in the following way: the anaphor represents one inferentially determinable proposition, or a very small set of determinable propositions, which serves as its antecedent, while the inference particle which we claim to be the result of grammaticalization metarepresents that mutually manifest set of propositions which constitute the context without which the utterance would be irrelevant. The loss of the truth-conditional semantic component which is a general property of anaphora is seen to have some structural consequences. There is, as noted above, a strong affinity between parenthetical elements and non-truth-conditional meaning. Parentheticals typically contribute to so-called higher-level explicatures (Wilson & Sperber 1993) like expressions of the speaker’s propositional attitude, rather than to the “ground-floor” explicature, which is the explicitly communicated assumption inferentially developed from a logical form encoded by the grammatical form of the utterance (Sperber & Wilson 1995). Parenthetical expressions like if you like, I’m sure and extra-clausal tags like don’t you think? all constrain the speaker’s ostensively communicated propositional attitude, like the inference particle then. Their extra-clausal syntactic position and the lack of stress that goes with that position are indicative of a grammaticalization process involving the loss of truth-conditional meaning.
. Hungarian There exists a non-truth-conditional use of the Hungarian marker akkor (‘then’) in addition to its truth-conditional use as a discourse anaphor (either temporal or conditional). If there is contextual evidence which enables the hearer to infer that the speaker’s utterance is meant to be an assertion, then akkor normally affects truth conditions by virtue of the specific condition that the antecedent of this anaphor imposes on the propositional form; however, if the utterance is intended to be a question, then it is less certain that akkor is meant to stand for one specific conditional protasis, rather it may metarepresent whatever set of propositions constitutes what the speaker considers to be a mutually manifest context of utterance, like the parenthetical then in English.
Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó
Even in assertions there seems to be a choice between an anaphoric and a non-truth-conditional use of akkor. Unlike the situation in English there is no appositional or right-detached slot for akkor that encodes its non-truth-conditional meaning but its syntactic placement in the sentence does give the addressee some guidance in this matter. A sentence-initial akkor, as in (12) B1 uttered in response to A’s epistemically modified observation (cf. the English example (9) above), is a fairly safe choice if the speaker’s intention is to assert the truth of a conditional where akkor is intended to represent the protasis, and normally a lot safer than placing the word in the medial or final position illustrated by (12) B2 and B3 , respectively. (12) A: Lehet, hogy Karcsi is itt lesz a hétvégén. ‘Maybe (that) Charlie will be here, too, at the weekend.’ B1 : Akkor tudok neki mesélni egy-két dolgot. then I.can to.him tell one-two things ‘Then I can tell him a few things.’ B2 : Tudok akkor neki mesélni egy-két dolgot. I.can then to.him tell one-two things ‘I can then tell him a few things.’ B3 : Tudok neki mesélni egy-két dolgot akkor. I.can to.him tell one-two things then ‘I can tell him a few things then.’ While the sentence-initial position of akkor in B1 offers the procedural information that the truth of the apodosis is contingent upon the truth of the proposition expressed in the complement of A’s sentence, the postverbal position of akkor in (12) B2 and its final position in (12) B3 are both consistent with the interpretation that speaker B has taken the truth of the assumption ‘Karcsi will be here in the weekend’ for granted. Akkor is to be understood as an inference marker (see Vaskó 1999) pointing to the speaker’s pragmatically inferred assumption that this gives him an opportunity to tell Karcsi something he ought to be made aware of. Use of a global intonation contour associated with questions, one that involves a rising tune culminating in a F0 maximum in the penultimate syllable and then a low-pitched final syllable (cf. Fónagy 1998), overrides the information derivable from the syntactic position of akkor, not only in (13) B2 and B3 but even when it is imposed on the sentence structure employed in B1 where the position of akkor would otherwise favor a truth-conditional interpretation. (13) A: Teljesen sötét van a házban. ‘It is totally dark in the house.’
Then – adverbial pro-form or inference particle?
B1 : Akkor nincsenek itthon./? then they.are.not at.home ‘Then they are not at home./?’ B2 : Nincsenek akkor itthon./? they.are.not then at.home ‘They are then not at home./?’ B3 : Nincsenek itthon akkor./? they.are.not at.home then ‘They are not at home then./?’ Some speakers of English find it hard to accept a then-modified negative declarative like ‘They haven’t gone out, then?’ or negative interrogative like ‘Haven’t they gone out, then?’ as requests for confirmation of the assumption that the people referred to have gone out. This kind of functional link between the inference marker then and the conditional anaphor then is what we would expect to find if the two thens have a common source. On the other hand, if then (optionally preceded by adversative but) is placed in a left-dislocated position before a negative interrogative sentence, then it is apparently easier to associate the linguistic form with a challenge to the interlocutor’s explicit or implicit belief that the people referred to had not left the house: ‘(But) then, haven’t they gone out?’. Steve Nicolle has informed us, however, that he accepts both interrogatives ‘Then, haven’t they gone out?’ and ‘Haven’t they gone out, then?’ as requests for confirmation of the positive proposition. Hungarian negatives modified by akkor seem to have much the same pragmatic potential as the corresponding English structures. The answer with akkor in (14) is infelicitous if B intends to convey that the people referred to have presumably switched off all lights because they have gone to bed. (14) A: Teljesen sötét van a házban. ‘It is totally dark in the house.’ B: #Akkor nem feküdtek le? then not they.lay down #‘Then they have not gone to bed?’ =‘Haven’t they gone to bed (since there’s no light on)?’ In some languages there appears to be a general rule against the use of a negative interrogative in a request for an affirmation of the positive counterpart proposition. Could Hungarian be one of them? B’s reaction is still not quite felicitous even if we remove akkor in (14), but an overt sign of reduced speaker commitment due to the addition of modal markers – the expanded verb form feküdhettek with the suffix -het expressing possibility, as indicated in (14 ) B1 , or esetleg (‘perhaps’), as in (14 ) B2 – improves the situation considerably.
Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó
(14 ) A: Teljesen sötét van a házban. ‘It is totally dark in the house.’ B1 : Nem feküdhettek le? ‘Couldn’t they have gone to bed?’ B2 : Nem feküdtek esetleg le? ‘Haven’t they perhaps gone to bed?’ While akkor led to a bad result in (14) where B is describing a situation based on the inference he is drawing in processing A’s information and seeing the dark house, it works without a hitch in (15) and (16) below where the speaker’s purpose is to prompt some extra-linguistic behavior. These are acts of encouraging the hearer not to behave in accordance with the negative proposition but in accordance with its positive counterpart. (15) A: Az elöadás kilenckor kezdödik. the performance at.nine begins ‘The performance begins at nine.’ B: Akkor miért nem megyünk? Then why not we.go ‘Then why don’t we go?’ (16) A: Elment a busz. has.left the bus ‘The bus has left.’ B: Nem kellene akkor taxival menni? not should.we then by.taxi go ‘Then, shouldn’t we take a taxi?’ Like English conditional then the nongrammaticalized conditional anaphor akkor was originally confined to utterances in the epistemic domain, to conclusions based on a set of premises. The use of akkor in an utterance like (14) B is presumably a lot harder to dissociate from its prototypical use in deductions than the use of akkor in admonitions and hortatives like (15) B and (16) B, which is a secondary, derived use anyway. Nevertheless, the sentence structure in (14 ) B where akkor is placed between the verb form feküdhettek and the verbal particle le meaning ‘down’ does not sound bad at all if the utterance is produced with a pitch excursion to high tone for accent on akkor, which is indicative of its anaphoric function. (14 ) A: Teljesen sötét van a házban. ‘It is totally dark in the house.’ B: Nem feküdhettek akkor le? ‘Then, couldn’t they have gone to bed?’
Then – adverbial pro-form or inference particle?
Akkor in (14 ) B stands for one specific condition, the premise provided by speaker A. The accent on akkor in this utterance shows that it should be classified as an adverbial anaphor, not as a non-truth-conditional particle metarepresenting the context at large. There is no evidence that the inference particle akkor is ever used to communicate the existence of a conflict between the speaker’s beliefs and the beliefs that the speaker attributes to the hearer.
. Norwegian The grammaticalization of Norwegian da, meaning either temporal or conditional ‘then’, has gone further than in English, and further than in Hungarian. Notice first of all that an unaccented right-detached occurrence of this Norwegian word for ‘then’ can freely modify a negative declarative used in a request for confirmation of the corresponding positive proposition. Consider the following Norwegian versions of the Hungarian dialogues (14) and (16). (17) A: Det er helt mørkt i huset. ‘It’s totally dark in the house.’ B: Har de ikke gått til sengs, da? have they not gone to bed then ‘(But) then, haven’t they gone to bed?’ (18) A: Bussen har gått. ‘The bus has left.’ B: Skal vi ikke ta en taxi, da? shall we not take a taxi then ‘Then, shouldn’t we take a taxi?’ Da in (17) B and (18) B metarepresents the set of contextual assumptions that would make the respective negative propositions true but it also allows one to draw a conclusion about the speaker’s own epistemic attitude to the proposition in (17) and the speaker’s deontic attitude to the proposition in (18). The answer to the question whether we understand the speaker’s attitude to the proposition expressed to be positive or negative depends on the context of utterance. Only a dissociative attitude to the negative propositions expressed makes B’s response in (17) and in (18) optimally relevant in the sense of Sperber and Wilson (1995), so the inference marker da will here be associated with a communicated dissociative attitude. In English it is impossible for the inference particle then to cooccur with a positive-polarity item in a negative sentence, but in Norwegian this is a perfect combination.
Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó
(19) De har ikke gått til sengs allerede, da? they have not gone to bed already then ‘(But) don’t you think they have gone to bed already (*then)?’ Due to the cooccurrence of the marker da and the positive-polarity item allerede (‘already’) in (19) the negative proposition expressed there can again only be understood to metarepresent an assumption which is not shared by the speaker of (19), i.e. which is not the speaker’s explicature. (19) has no explicature, because even if the sentence form is declarative, the inference particle da reveals that the utterance is an interpretation of someone’s, for example the interlocutor’s, thought. When there is a conflict between the metarepresented proposition expressed by the utterance and the attitude to it indicated by the inference particle, then the speaker’s act is invariably heard as a request to the addressee to either agree or disagree with the speaker’s communicated opinion. On the other hand, when the context of utterance permits us to infer that there is no conflict between the proposition expressed and the epistemic or deontic attitude of the speaker, then the speech act can just as well be an assertion, a prediction, or a declaration of intention. Cf. (20) where da does not metarepresent the proposition expressed by A but rather what B takes to be a context shared by A and B. (20) A: Jeg har tenkt å gå hjem nå. ‘I’m going home now.’ B: Vi ses igjen i morgen klokka ti, da. ‘We’ll meet again tomorrow at ten o’clock, then.’ In contradistinction to (20) B, (20 ) B expresses a material implication whose antecedent proposition is found in A’s utterance (‘#If you’re going home now, then we’ll meet again tomorrow at ten’). (20 ) A: Jeg har tenkt å gå hjem nå. ‘I’m going home now.’ B: #Da ses vi igjen i morgen klokka ti. #‘Then we’ll meet again tomorrow at ten o’clock.’ One particular use of da as inference particle will now be shown to have parted company with the anaphor da altogether, at least in some dialects of Norwegian. The anaphor and the inference particle can cooccur in conducive questions. Da is a stressed intraclausal anaphor in (21); in (22) this anaphor is seen to cooccur with the segmentally identical utterance-final inference particle, which is produced with a Low boundary tone (L%) in (22) B1 and a High boundary tone (H%) in the alternative response (22) B2 .
Then – adverbial pro-form or inference particle?
(21) A: Skriv navnet ditt her, er du snill. ‘Write your name here, please.’ B: Får jeg pengene mine da? ‘Do I get my money then?’ (22) A: Skriv navnet ditt her, er du snill. ‘Write your name here, please.’ B1 : Får jeg pengene mine da, da-L%? get I the.money mine then then ‘So do I get my money then?’ B2 : Får jeg pengene mine da, da-H%? get I the.money mine then then ‘But do I really get my money then?’ In all of B’s responses in (21)–(22), the accented anaphor da metarepresents the condition expressed in A’s imperative (‘If I write my name there, do I then get my money?’). Both in (22) B1 and (22) B2 the utterance-final inference particle metarepresents that set of contextual assumptions that makes it true to say that B gets his money if he signs the designated document. The communicative difference between B1 with da-L% and B2 with da-H% is that the former, like (21) B, is a question whether the speaker’s interpretation of the intent of A’s request is correct, but when the boundary tone is H% as in (22) B2 , the speaker seems to be worried that the proposition in the host sentence might not be true (see Fretheim 2000), that the practical consequences are that he won’t have the money he feels he is entitled to and that his signature on that paper presumably won’t make a difference. Da has the same encoded meaning as inference particle in the two prosodically distinct utterances (22) B1 and (22) B2 but the boundary tone suggests that the speaker is hopeful in B1 and pessimistic in B2 . It would be false to conclude that the two opposite boundary tones aligned with da encode contrary epistemic attitudes to the proposition expressed in the host sentence. The skepticism associated with da-H% in (22) B2 would be almost as strongly present if the speaker had placed a focal phrase-accent (in the form of a fundamental frequency (F0) maximum) on the clause-initial finite verb får (‘get’), for polarity focus, as well as one on the clausefinal conditional anaphor da. The latter intonational phrasing is not compatible with the boundary tone H% in the inference particle (for reasons that need not concern us), so it is the entire combination of prosodic and syntactic form in (23) B that has pragmatic implications similar to (22) B2 . (23) A: Skriv navnet ditt her, er du snill. ‘Write your name here, please.’ B: får jeg pengene mine da, da-L%? ‘(But) can I be absolutely sure that I do get my money then?’
Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó
The polarity focus pattern exemplified by (23) B works equally well with declarative syntax. (23 ) B expresses the same as (23) B. (23 ) A: Skriv navnet ditt her, er du snill. ‘Write your name here, please.’ B: Jeg får pengene mine da, da-L%? I get the.money mine then then ‘(But) can I be absolutely sure that I do get my money then?’ To sum up what data like (23)–(23) permits us to conclude, the accented da metarepresents the condition specified by A, the unaccented right-detached da metarepresents the set of assumptions making up an accessible context that would make the proposition (including the conditional adjunct) true, and the global intonation pattern reveals the speaker’s skeptical attitude, just as the boundary tone H% on da did so after the interrogative in (22) B2 . Segmental attrition is a well-known linguistic accompaniment of grammaticalization and a reliable indicator that grammaticalization has taken place. It is interesting to note that in some dialects of Norwegian (South-East Norwegian, including Oslo), the particle da can be produced without the onset consonant [d] if and only if the linguistic signal is otherwise one that is compatible with a communicated dissociative attitude to the proposition expressed. That is to say, the eroded segmental form a encodes a lack of compatibility between the speaker’s dissociative attitude to the proposition expressed and the attitude which the speaker attributes to the hearer, or, occasionally, to some other contextually identifiable person who is held to believe the proposition to be true. Thus the shorter form a would be acceptable in (17) B, (18) B, (19), (22) B2 , (23) B and (23 ) B, but grammatically illformed in (20) B and (22) B1 , two utterances whose linguistic form does not suggest any discrepancy between the state of affairs that the speaker believes to be either true or commendable (i.e. worth making true) and the state of affairs represented by the proposition expressed. Moreover, the consonant-free form a is quite impossible when da is a truth-conditional adverbial adjunct. While there are a number of linguistic environments and inferred contexts in which da cannot be replaced by a, there is no situation where a has to replace da in South-East Norwegian speech. The full form da is always an acceptable alternative. Still, the emergence of a changed phonological form is the ultimate proof that the link between the truth-conditional anaphor and the inference particle is cut off. Observe that the dissociative attitude use of the inference particle da in Norwegian is on the one hand the only use that permits loss of the onset consonant [d] and on the other hand the only use that lacks a direct counterpart in the use of English then and Hungarian akkor as particles. The choice between da and a is only available when the meaning of the particle is maximally distant from the
Then – adverbial pro-form or inference particle?
meaning of the historical origin, the anaphor, so there is clearly an iconic basis for the phonological attrition process: likeness in segmental form suggests some likeness in function, less likeness in segmental form suggests less similarity in function. Our conclusion is that Norwegian da has gone further along the grammaticalization path than English then and Hungarian akkor by opening for a use of the inference particle da which implies the speaker’s doubt that the proposition represents a true state of affairs, and which leads to the loss of phonological material that is otherwise obligatorily present in the anaphor as well as in the inference particle.
. Ewe Ewe has a set of lexicalized adverbial phrases which are conceptually underspecified so that they may be said to function like anaphora to be pragmatically enriched by association with an antecedent. One of these is the consequence marker, or conditional anaphor ekema (‘then’, or ‘so’, ‘consequently’), which is lexically distinct from the temporal anaphor γemaγi (‘then’). Historically ekema is composed of two Ewe demonstratives, first eke which today is used with the meaning of proximal ‘this’ only in certain dialects of Ewe (the standard word corresponding to English ‘this’ being esia) and then ema which is the word for ‘that’ in present-day Standard Ewe. Ekema has a fixed syntactic position in the beginning of the clause. There are also a number of monomorphemic clause-final particles in Ewe, many of which have a modal meaning which justifies the label ‘question particle’. This set of non-truth-conditional markers includes an inference particle, 2e, which represents exactly those contextual assumptions prompted by the interlocutor’s verbal behavior which warrant the speaker’s inference expressed by the propositional form of the utterance. Thus from a Standard Average European point of view one might feel that the utterance-final particle 2e has a function which closely resembles that of the English non-truth-conditional then and similar particles in languages like Hungarian and Norwegian. The particle 2e has a wide range of uses, some of which are appropriately glossed as ‘then’ and some not. Two of its uses are illustrated in (24) and (25). (24) A: Nuka wf ge nele etsf? what do fut 2sg.be tomorrow ‘What will you be doing tomorrow?’ B: ]dime mayi agble. morning.inside 1sg.fut.go farm ‘I will go to the farm in the morning.’
Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó
A: Emegbe 2e? afterwards q ‘Afterwards, then?’ (= ‘And afterwards?’) (25) Mele dfa me hafi wò va 2e. 1sg.be work.def inside before 3sg come part ‘(But) I was actually working here before he came.’ (i.e. before he started in this job) (25) might have been uttered by someone who finds it unfair that one of his workmates is getting promoted but he himself is not, even though this other person has been with the company for a shorter period than the speaker. By using 2e here the speaker is alluding to the pertinent set of contextual assumptions considered to be mutually manifest to the interactants. The utterance serves the function of a reminder, and at the same time it is an implicit protest. This use, which is incidentally matched by a similar use of the Norwegian particle da, shows that we miss something by calling 2e a “question” particle tout court, but a discussion of the multifunctionality of this interpretive inference particle in Ewe falls outside the scope of this paper. Boateng has carried out a test among native speakers of Ewe in which she asked them to offer what they considered to be the best translation into Ewe of a number of English sentences given as responses by the second speaker in a conversational dyad. 8 respondents participated in the test, whose result is presented in (26)–(28). (As in previous examples, B1 , B2 etc. are meant to be alternative reactions to speaker A’s information. Occurrences of ekema and 2e in the informants’ responses are given in italics.) (26) A: The theatre performance begins at nine o’clock. B1 : Then there’s no chance that I can be there. Ewe: Ekema mfnukpfkpf a2eke meli be manf then chance none neg.be compl 1sg.fut.be afima o. there neg B2 : Then I can’t watch it. Ewe: Ekema nyemate]u akpfe o. Then 1sg.neg.fut.can fut.see.3sg neg B3 : I can’t watch it, then. Ewe: Ekema nyemate]u akpfe o. then 1sg.neg.fut.can fut.see.3sg neg Nyemate]u akpfe o 2e 1sg.neg.fut.can fut.see.3sg neg q
(8/8)
(8/8)
(4/8) (2/8)
Then – adverbial pro-form or inference particle?
Ekema nyemate]u akpfe o 2e? then 1sg.neg.fut.can fut.see.3sg neg q B4 : So I can’t watch it, then. Ewe: Ekema nyemate]u akpfe o 2e? then 1sg.neg.fut.can fut.see.3sg neg q Ekema nyemate]u akpfe o. then 1sg.neg.fut.can fut.see.3sg neg B5 : Then why don’t we go? Ewe: Ekema nukata míate]u ayi o? then why 1pl.neg.fut.can fut.go neg (27) A: The last bus has left. B1 : Then what do we do? Ewe: Ekema nuka mia wf? then what 1pl do B2 : What about a taxi, then? Ewe: Ekema taxi 2e? then taxi q Alo mitsf taxia? or 1pl.take taxi.q Ekema miatsf taxi oa? then 1pl.(neg.)fut.take taxi neg.q B3 : Then we have to get a taxi. Ewe: Ekema elebe miakpf taxi. then 3sg.must 1pl.fut.see taxi Ekema miatsf taxi. then 1pl.fut.take taxi Ekema elebe midi taxi 2e? then 3sg.must 1pl.look-for taxi q B4 : So we’ll have to find a taxi, then. Ewe: Ekema ehiaã be miadi taxi 2e? then 3sg.need compl 1pl.fut.look-for taxi q Ekema ehiaã be miadi taxi. then 3sg.need compl 1pl.fut.look-for taxi Ekema miadi taxi. then 1pl.fut.look-for taxi Ekema elebe miadi taxi, alo? then 3sg.must 1pl.fut.look-for taxi or (28) A: B1 :
I learn so much from coming here. Then why don’t you come more often?
(2/8)
(4/8) (4/8)
(8/8)
(8/8)
(6/8) (1/8) (1/8)
(6/8) (1/8) (1/8)
(3/8) (2/8) (2/8) (1/8)
Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó
Ewe: Ekema nukata womava zige2ee o? then why 2sg.neg.fut.come many-times neg Ekema elebe nava afisia enuenu. then 3sg.must 2sg.come here often B2 : Why don’t you come more often, then? Ewe: Ekema nukata mava afisia edziedzi o? then why neg.3sg.come here often neg B3 : Why, then, don’t you come more often? Ewe: Ekema nukata mava zige2ee o? then why neg.3sg.come many-times neg Nukata mevana kabakaba o? why neg.3sg.come.hab quickly neg
(7/8) (1/8)
(8/8)
(6/8) (2/8)
Ekema turned out to be the preferred choice whenever speaker B’s English stimulus looked like a conclusion based directly on what B considers to be a premise provided by the interlocutor A. All eight respondents decided to translate then as ekema in (26) B1 , B2 and B5 . In (26) B3 the marker then in the source language was placed at the end, separated graphically from the rest by a comma to indicate its parenthetical position. This time six of the eight respondents used the sentenceinitial consequence marker ekema, but two of those six people let ekema cooccur with sentence-final 2e, and the remaining two respondents used 2e at the expense of ekema. Ekema was used by all eight respondents in (26) B4 as a translation of English so, but four of them again let initial ekema interact with a sentence-final 2e. The combination of ekema and 2e was also preferred by a majority of six in (27) B2 : ‘What about a taxi, then?’. The two markers appear to be mutually supportive in that Ewe sentence. Notice also the negative polarity of the last Ewe interrogative under (27) B2 , Ekema miatsf taxi oa?, where the negative particle o (here with the interrogative suffix a added to it) makes it superfluous to mark negation in the verb form itself. One respondent chose to translate ‘What about a taxi, then?’ by giving the interrogative a negative form indicated by the particle o and providing the verb form with the future tense marker: miatsf. This reveals that in Ewe, as in so many other languages, negative interrogatives may be used for the purpose of requesting confirmation of the corresponding positive proposition. The marker was ekema, though, and not 2e. We received no data where 2e was attached to a negative interrogative performed with the expectation of an affirmative answer, and this is indeed an impossible utterance type in Ewe. (29a), where the interrogative mood is indicated by the suffix -a, could be interpreted to mean ‘Let’s take a taxi’ in an appropriate context (and ‘We’re not going to take a taxi, then?’ in a different one), but (29b) can only mean ‘We are not going to take a taxi, then?’. The utterance-final particle does not permit the hortatory interpretation licensed by ekema.
Then – adverbial pro-form or inference particle?
(29) a.
Miatsf taxi oa? 1pl.fut.take taxi neg.q ‘Shouldn’t we take a taxi?’/‘Let’s take a taxi.’ b. Miatsf taxi o 2e? 1pl.fut.take taxi neg q ‘We’re not going to take a taxi, then?’/‘Is it true that we’re not taking a taxi?’
Thus we can conclude that 2e is not an interpretive particle with the same use potential as Norwegian da which can signal that the speaker’s own propositional attitude conflicts with the attitude attributed to the hearer and made explicit by the metarepresented proposition expressed. The utterance-final particle 2e never occurred in the respondents’ translations of (28) B1 –B3 . Ekema reigned alone in (28). This, we believe, is due to the fact that the ‘why’-questions in (28) express a higher-level explicature set in the domain of deontic rather than epistemic modality. The particle 2e can apparently only be used to ask conducive questions in the epistemic domain. In contrast the conditional anaphor ekema can open an Ewe sentence when the illocutionary act performed is an assertion but the word can also be prefaced to an act of advising, recommending or requesting. The appropriate Ewe gloss for then in (28) is therefore ekema, not 2e, regardless of the position of then in the English source sentence. It follows from what we have just said that speakers of Ewe must use ekema to capture the meaning of then in a dialogue like (30). (30) A: I’m cold. B: Put on your coat then. Ewe: Ekema do awua? (with a tone of frustration) Ekema do awu. *Do awu 2e. The same applies to the Ewe translations of English imperatives used to perform a command, or an impatient admonition: (31)
Then stop it!/Stop it then! Ewe: Ekema dzudzf. *Dzudzf 2e.
(32)
Hurry up then! Ewe: Ekema wf kaba. *Wf kaba 2e.
The fact that 2e as a conducive question particle cannot be used in communicative acts where the speaker intends to influence the hearer’s behavior is seen to be one
Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó
major difference between the communicative functions of ekema and 2e. Yet we have witnessed that there are certain situations in which either one of the two particles is applicable, and they may also cooccur in acts where the speaker is asking the hearer to confirm the speaker’s inference (2e) that a propositional form which includes an explicit condition (ekema) holds true. The functional overlap between the two markers ekema and 2e is reminiscent of the borderline cases in Hungarian where it is mainly accessible context-dependent information that prompts the addressee to assign either a truth-conditional or a non-truth-conditional interpretation to a given occurrence of akkor. At least with some speakers of Ewe, ekema is seen to encroach on the territory of 2e as an interpretive marker of epistemic attitude, just like our truth-conditional European anaphora do; conversely, there is no sign that 2e ever encroaches on that part of the meaning range of ekema which is to do with marking of a propositional attitude of the deontic type. Nor has 2e adopted any semantic properties that would ever place it in the set of truth-conditional markers, even if the context might appear to favor a material implication interpretation. Notwithstanding those pragmatic similarities between ekema and 2e which Boateng’s empirical investigation revealed, there is absolutely no sign that Ewe speakers would ever allow themselves to let the marker ekema appear in any other position than the sentence-initial one. Ekema has not acquired any structural properties pertaining to 2e. There are no signs of a process of grammaticalization of ekema that might eventually lead to inclusion of that form in the set of utterancefinal particles to which the marker 2e belongs. In some dialects of Ewe, ekema regularly reduces to ke but the syntactic position of this monosyllabic form is just as fixed as that of the trisyllabic form. Conversely, 2e has not assumed any syntactic privileges that might cause one to question the rigidity of the grammatical distinction between an adverbial marker like ekema and an inference particle like 2e. The two items belong to completely separate sets of markers despite their functional overlap. We believe, nevertheless, that the functional overlap we have observed does reveal something about the cognitive mechanisms that link then, akkor and da as adverbial pro-forms to then, akkor and da used as inference markers in requests for confirmation and more generally as particles metarepresenting a manifest context.
. Conclusion We have demonstrated that in certain situations the two Ewe function words ekema and 2e contribute to the relevance of a given utterance in much the same way. When they cooccur, their pragmatic roles are presumably not completely separated. Their pragmatic similarities derive from the fact that they both represent
Then – adverbial pro-form or inference particle?
contextual premises without which the speaker would not have had an occasion to produce her verbal stimulus. We believe the partial functional overlap of ekema and 2e in Ewe to furnish us with an argument in support of the analysis of then, akkor and da proposed in this paper. What this overlap indicates is that the parallel developments in English, Hungarian and Norwegian from anaphoric adverbial adjunct to non-truth-conditional inference particle are highly natural processes of grammaticalization prompted by the language users’ recognition of a common cognitive basis. Both types of marker represent contextual assumptions which the hearer has to bring to bear in the processing of the modally modified sentence in order to arrive at a relevant interpretation of it. One major difference that we have pointed out is that the anaphor cues activation of one particular contextually retrievable condition which is sufficient for the truth of the proposition expressed, while the particle represents the pertinent context at large, the whole set of assumptions that would secure the truth of the proposition expressed. Ewe, with its distinct lexical sets of truth-conditional adverbial adjuncts on the one hand and non-truth-conditional particles on the other, has a grammar that simply does not give the conditional anaphor ekema a chance to grammaticalize so that it assumes not only part of the function but even the look of a non-truthconditional inference particle. The latter category is already occupied by a function word, 2e, whose encoded meaning and communicative role corresponds closely to what we have characterized as the end product of the grammaticalization process envisaged for the three European languages that we examined, a process which we have argued to have gone further in Norwegian than in English and Hungarian. Unlike the English parenthetical particle then and its counterpart akkor in Hungarian, the Norwegian particle da can be used to metarepresent an implicit set of context-bound truth conditions some of which the speaker believes to be not satisfied. Norwegian uses da freely in interpretations of thoughts that the speaker disagrees with or finds a reason to question. And colloquial South-East Norwegian has developed a new particle a (or ’a, if you like) which functions just like da, except that it encodes the speaker’s dissociative attitude to the proposition expressed and therefore requires that the utterance contains linguistic elements, syntactic or prosodic, which are all consistent with the communication of that kind of attitude.
Acknowledgement We are grateful to Steve Nicolle for comments that led to some last-minute changes.
Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó
References Archangeli, D. (1997). Optimality Theory: An introduction to linguistics in the 1990s. In D. Archangeli & D. T. Langendoen (Eds.), Optimality Theory: An Overview (pp. 1–33). Oxford: Blackwell. Blakemore, D. (1987). Semantic Constraints on Relevance. Oxford: Blackwell. Carston, R. (1998). Pragmatics and the Explicit-Implicit Distinction. Ph.D. thesis, University College London. Croft, W. (1990). Typology and Universals. Cambridge: Cambridge University Press. Fónagy, I. (1998). Intonation in Hungarian. In D. Hirst & A. Di Cristo (Eds.), Intonation Systems: A Survey of Twenty Languages (pp. 328–344). Cambridge: Cambridge University Press. Fretheim, T. (2000). Procedural encoding of propositional attitude in Norwegian conditional clauses. In G. Andersen & T. Fretheim (Eds.), Pragmatic Markers and Propositional Attitudes. Pragmatics & Beyond New Series, no. 79 (pp. 53–84). Amsterdam: John Benjamins. Gundel, J. K., Hedberg, N., & Zacharski, R. (1993). Cognitive status and the form of referring expressions in discourse. Language, 69(2), 274–307. Gundel, J., Borthen, K., & Fretheim, T. (1999). The role of context in pronominal reference to higher order entities in English and Norwegian. In P. Bouquet, L. Serafini, P. Brézillon, M. Benerecetti, & F. Castellani (Eds.), Modeling and Using Context. Berlin: Springer-Verlag. Heine, B. (1993). Auxiliaries: Cognitive Forces and Grammaticalization. Oxford: Oxford University Press. Heine, B., Claudi, U., & Hünnemeyer, F. (1991). Grammaticalization: A Conceptual Framework. Chicago: University of Chicago Press. Hopper, P. J., & Traugott, E. C. (1993). Grammaticalization. Cambridge: Cambridge University Press. Noh, E.-J. (1998). A relevance-theoretic account of metarepresentative uses in conditionals. In V. Rouchota & A. H. Jucker (Eds.), Current Issues in Relevance Theory. Pragmatics & Beyond New Series, no. 58 (pp. 271–304). Amsterdam: John Benjamins. Noh, E.-J. (2000). The Semantics and Pragmatics of Metarepresentation in English: A Relevance-theoretic Account. Pragmatics & Beyond New Series, no. 69. Sperber, D., & Wilson, D. (1995). Relevance: Communication and Cognition (2nd edition). Oxford: Blackwell. Vaskó, I. (1999). Pragmatiske partikler i norsk og ungarsk. Ph.D. thesis, Eötvös Loránd University, Budapest. Wilson, D. (1999). Metarepresentation in linguistic communication. UCL Working Papers in Linguistics, 11, 127–164. Wilson, D., & Sperber, D. (1993). Linguistic form and relevance. Lingua, 90, 1–25.
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.1 (75)
The polysemy of the Swedish verb komma ‘come’ A view from translation corpora Åke Viberg Uppsala University, Sweden
.
Introduction
Polysemy has been a major area of focus in recent lexical research (Ravin & Leacock 2000), especially within computational linguistics and cognitive semantics. Within computational linguistics, interest has focused on identification/disambiguation of word senses in actual texts (Ide & Véronis 1998; Kilgariff & Palmer 2000). Polysemy is also a central area of recent work within the relational approach used in electronic lexical databases such as WordNet (Fellbaum 1998) and EuroWordNet (Vossen 1998). Synonymous words are grouped into synsets (e.g. pull, draw). Polysemous words appear in different synsets, each of which represent one of their meanings (e.g. run, operate vs. run, function, work). There is, however, no way of showing how various meanings are related within this approach. Basically, the various senses of a word are listed. The relatedness of various senses of a word and the cognitive processes accounting for sense extensions have been the major area of interest within cognitive semantics (Langacker 1988; Lakoff 1987; Newman 1996) and within the generative lexicon (Pustejovsky 1995). This paper is one in a series of papers dealing with the polysemy of Swedish verbs from a crosslinguistic perspective (Viberg 1984, 1999, 2002). In particular, the present study of the Swedish verb komma ‘come’ is a follow-up of an earlier study of the Swedish verb gå ‘go’ (Viberg 1996, 2000). The primary aim of the paper is to identify the major meanings of komma from a crosslinguistic perspective showing what properties represent universal tendencies (to the extent these are known) and what properties appear to be more language-specific. This means the paper will be relatively data-oriented. However, some indication will also be given as to how the meanings are related (the problem of semantic representation). Another aim is to isolate the formal markers that can be used as disambiguation cues
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.2 (76)
Åke Viberg
in a text (the problem of semantic interpretation) and to account for the choice of translations. The analysis is based on translation corpora. The major source is the complete set of occurrences of Swedish komma and English come in the English Swedish Parallel Corpus (ESPC, see Altenberg & Aijmer 2000). In addition, data will be presented from a small pilot corpus which is being prepared by the author and consists of extracts from novels in Swedish with translations into English, German, French and Finnish and originals in the same languages translated into Swedish. As an introduction, a survey will be given of the major meanings of komma. In Table 1, examples from the pilot corpus are given divided into three major groups of meanings: Motion in physical space, Extended lexical meanings, and Grammatical meanings. (The capitals at the end of the examples serve to identify the author and text.) Examples (1)–(3) represent the uses of komma as a motion verb. In example (1) rendered in English as the newly engaged couple came on a visit, the closest equivalent of komma is used in all the translations: come in English, kommen in German, venir in French and tulla in Finnish. The examples in (2)–(3) show some of the variation that appears with respect to uses describing motion in physical space. In Swedish, there is a great number of verbal particles with directional meanings which can be productively combined with komma. In example (2), komma is combined with the particle över ‘over, across’. Characteristically, French must use one of its many directional verbs traverser ‘cross, move across’ as a translation in this case. The other languages express the direction with a particle (or prefix in German). In example (3), a particle lacking any direct equivalent even in German or English is used. The particle fram in this frequent use means something like ‘to reach an intended destination’. The translations are all verbs of arrival: arrive in English, arriver (in nominalized form) in French and saapua ‘arrive’ in Finnish. German uses ankommen which is a prefixed form of kommen functioning as a verb of arrival. This example also reflects another characteristic of Swedish komma, which will be discussed in greater detail in a later section, namely that the deictic component is very general. The combination of komma and fram is generally used as a verb of arrival without much concern about the spatial location of speaker and listener. The wide extension of the deictic element is also reflected in the English translation of example (2), where komma över is translated by get across. The verb get in combination with a particle or other spatial complement is a relatively frequent translation of komma in situations where the deictic component is relatively far extended from its prototype. The verb komma has many extended, non-spatial meanings. In many cases, the interpretation of such cases is guided by subtle formal cues such as the presence of a non-spatial complement like the following ones taken from the English originals in the ESPC corpus: He had come face to face with his destiny. (RL), The human species is coming of age here. (CSA), and the science faculties were coming
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.3 (77)
The polysemy of the Swedish verb komma ‘come’
Table 1. Examples of translation patterns of the major meanings of komma Swedish
English
German
French
Finnish
Einige Wochen später kamen die Frischverlobten zu Besuch. Es war nicht leicht, mit Mia hinüberzukommen.
Quelques semaines plus tard, les fiancés vinrent en visite. Elle eut du mal à traverser en portant Mia.
Vastakihlautuneet tulivat vierailulle muutaman viikon kuluttua.
A. Motion in physical space (1) Några veckor senare kom de nyförlovade på visit. IB
A few weeks later, the newly engaged couple came on a visit.
(2) Det var inte lätt att komma över med Mia. KE
The stony riverbed made it difficult to get across with Mia. By the time we finally arrived, I was practically dead.
(3) När vi kom fram var jag nästan död. IB
Ei ollut helppo päästä yli Mian kanssa.
Als wir A notre arrivée Kun saavuimme ankamen, war j’étais presque perille, olin ich beinahe tot. mort. melkein kuollut.
B. Extended lexical meanings Mental (4) Sen kom han ihåg Edward Kennedy. KE
Then he remembered Edward Kennedy. He’ll realise he’s got the wrong day.’
(5) Han kommer nog på att han tagit fel på dag. KE Verbal communication
Dann fiel ihm Edward Kennedy ein. Ihm fällt bestimmt ein, daß er sich im Tag geirrt hat.
(6) Påståendet kom oväntat. KE
That statement Diese was Behauptung unexpected. kam unerwartet. C. Grammatical meanings
Puis il se souvint d’Edward Kennedy. Il va sûrement s’apercevoir qu’il s’est trompé de jour.
Sitten hän muisti Edward Kennedyn. Kyllä hän huomaa erehtyneensä päivästä.
Cette Väite oli yllättävä. affirmation tomba sur elle à l’improviste.
Future: Pure prediction (7) Ho kommer å She’ll survive överleve oss us all. alle, MFred*
Die wird uns noch alle überleben.
* Standard Swedish: Hon kommer att överleva oss alla.
Elle nous enterrera tous.
Äiti elää varmaan pitempään kuin kukaan meistä.
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.4 (78)
Åke Viberg
Table 1. (continued) Swedish
English
German
French
Finnish
Maja-Lisa stopped complaining, but she never really liked their new home.
Maja-Lisa hörte auf zu jammern, aber sie konnte sich in ihrem neuen Heim nicht einleben.
Maja-Lisa cessa de se plaindre, mais elle n’était pas heureuse.
Maja-Lisa lakkasi valittamasta, mutta ei koskaan oppinut viihtymään uudessa kodissaan.
Mia gave her a look which made her seem adult.
Mia warf ihr einen Blick zu, der sie erwachsen aussehen ließ.
Le regard que lui lança Mia était celui d’une adulte.
Mia loi häneen katseen joka sai tytön näyttämään aikuiselta.
Past, ‘hindsight’ future (8) Maja-Lisa slutade klaga men hon kom aldrig att trivas i det nya hemmet. MFred Causative (9) Mia gav henne en blick som kom henne att se vuxen ut. KE
to grips with Lysenko’s perversion of genetics. (MAW). Another cue is the semantic category of the subject: A premonition had come to him that this was it. (RR), The third blow came from Rook’s right foot. (JC), The chance came in an unexpected way. (RF). The examples can be classified with respect to the semantic fields to which the extended meanings belong. In Table 1, two of the most frequent fields of this type are shown. Example (4) shows the use of komma in combination with the particle ihåg, which originally meant ‘into mind’. The expression komma ihåg is actually the most frequent exponent of the mental meaning ‘remember’ in Swedish. All the other languages use mental verbs in the translations except German which uses another expression based on spatial metaphor: einfallen ‘fall into’. Example (5) shows another case of extension into the mental field; komma på means ‘come on(to)’ and has a sentential that-S-complement. Even in this case, the translations are mental verbs except in German which uses the same spatial metaphor as in the preceding example. Example (6), which represents an extension into the field Verbal communication, has the literal meaning ‘The statement came unexpectedly’. English and Finnish resort to the verb ‘be’ as translations ‘The statement was unexpected’, whereas French exploits a different spatial metaphor based on tomber ‘fall’. This verb in many languages is used in various extended meanings which highlight the unexpectedness of an event (or the lack of control on the part of its subject). As will be described in more detail below, it turns out that the verb ‘come’ can appear with extended meanings belonging in most cases to the same range of semantic fields in
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.5 (79)
The polysemy of the Swedish verb komma ‘come’
all the languages involved, whereas the more precise meanings within these fields often contrast markedly between languages. The verb meaning ‘come’ has a strong tendency to develop grammatical meanings. In spite of the fact that the verb ‘come’ in all of the languages in the pilot corpus has a number of extended meanings which can be regarded as grammatical, the grammatical meanings of Swedish komma in general lack a counterpart in these languages. In addition, some of the finer nuances of these grammatical meanings cannot be directly expressed in the other languages. The most frequent grammatical use of komma is its employment as a future marker. The translations of example (7) show the closest correspondents in the other languages: will (’ll) in English, werden in German and the inflexional future in French. In Finnish, the main verb appears in the present tense without any grammatical marker of future time. There is, however, an adverbial varmaan ‘surely’ which reflects the predictive meaning rather well (Mother lives surely longer than all of us). There are several alternatives in all the languages in the sample but none of them cover the purely predictive meaning of the Swedish future marker exactly. The use of the past form of komma as a future marker has been singled out as a special meaning. It is used to mark what turned out to be the case at some time following the focus time in a past narrative. It is presented as a fact by the author, whereas it belongs to the unrealised future for the characters involved in the evolving narrative. As in example (8), it often turns out to lack a direct expression in the translations. The future event is presented as a simple fact in the translation without the double time perspective introduced by the ‘hindsight’ future marker in Swedish. Swedish komma can also be used as a periphrastic causative as shown in example (9). In this use, the Causee appears as a direct object and the caused event as an infinitive with the infinitive marker att. As a periphrastic causative, komma is a marked alternative to få ‘get’ which is much more frequent in this use. The other languages tend to use the most unmarked periphrastic causative as a translation such as make in English, faire ‘make’ in French, which, however, is not used in example (9), and saada ‘get’ in Finnish. In the few German examples in the corpus, various periphrastic causative verbs appear such as lassen ‘let’ in this example.
. Basic uses of the verb komma The basic contrast between gå and komma in Swedish will be discussed in 2.1. In the analysis of Swedish gå ‘go’, the uses with a human subject moving in concrete space have been referred to as Human locomotion and treated separately (Viberg 2000). For komma, it might be motivated to treat human subjects together with other concrete subjects, but to simplify comparison, human subjects will be treated separately in Section 2.2. The movement of various types of non-human
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.6 (80)
Åke Viberg
subjects represent various degrees of semantic extension and will be treated in a later section (3.2).
. The contrast between ‘come’ and ‘go’ Human locomotion which serves as a basic reference point for most of the frequent motion verbs refers to motion in concrete space with a human subject. In this use, komma contrasts with Swedish gå along several parameters as shown schematically in Table 2. The most important parameter is Deixis. The other parameters which have been described in the study on ‘go’ (Viberg 2000) will be treated only briefly here since they are neutral with respect to komma. With a human subject, the Swedish verb gå always describes self-propelled motion on foot by an intentional agent. None of these restrictions hold for komma, even though such an interpretation might very well be a default even for this verb. Movement can be conceptualized as a schema consisting of a Source, a Path and a Goal. In principle, ‘come’ and ‘go’ can describe the same itinerary: Peter went from Paris to London vs. Peter came from Paris to London. The contrast is based on perspective-taking and can be understood in relation to a reference point, Origo, from which the movement is seen. As shown in Table 3 based on Di Meola (1994), Origo is situated at the Goal when ‘come’ is used and at the Source when ‘go’ is used. Basically, the verb ‘go’ is Source-oriented, which is reflected among other things in the interpretation of temporal adverbials. He went home at five o’clock means that he left some other place at five o’clock, while He came home at five o’clock means that he arrived home at that time. The Source is the temporal anchor for go, Table 2. Basic semantic parameters differentiating Swedish gå and komma
Deixis Source of energy Sensorimotor Cognitive
gå
komma
Source-oriented Self-propelled Body movement: On foot Intentional
Goal-oriented Neutral Neutral Neutral
Table 3. The deictic contrast between come and go ‘come’ ‘go’
Source Source
Path Path
Goal O Goal
O O= Origo. Prototype for ‘come’: Speaker, Coding place, Coding time
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.7 (81)
The polysemy of the Swedish verb komma ‘come’
while the Goal serves that function for come. The characterization of ‘go’ as Sourceoriented and ‘come’ as Goal-oriented represents one important point of contrast between the two verbs. It is, however, so general that it also covers the contrast between some related verbs: Peter left for Paris at five o’clock (Source-oriented) vs. Peter arrived from Paris at five o’clock (Goal-oriented). What is specific, at least for ‘come’, is that the identification of Origo is based on deixis which involves properties of the speech situation such as the roles of speaker and addressee, the place of speaking (coding place) and the time of utterance (coding time), which can be different from reference time, the time at which the described event takes place. Fillmore (1977) in a classical paper approaches the problem via a number of revised hypotheses starting with the following one: Hypothesis I: (a) “come” and “bring” indicate motion toward the location of the speaker at coding time; (b) “go” and “take” indicate motion toward a location which is distinct from the speaker’s location at coding time. In this formulation, ‘go’ is only negatively specified with respect to the deictic characteristics, but the characterization of ‘go’ as source-oriented still holds. Motion toward the location of the speaker at coding time may be regarded as a prototype, even if Fillmore does not use this term. The more complex restrictions which are gradually introduced in the following hypotheses accordingly may be regarded as a shift from the prototype in gradual steps, from speaker to addressee, from coding time to reference time, etc. The following hypothesis summarizes all the possibilities involving speaker and/or adressee: Hypothesis V: “come” and “bring” indicate motion toward the location of either the speaker or the addressee at either coding time or reference time, or towards the location of the home base of either the speaker or the hearer at reference time. In third person discourse, Origo can be seen from the point of view of the evolving narrative. This is characterized in the following way (the new part of Hypothesis VII): “come” and “bring” also indicate, in discourse in which neither speaker nor addressee figures as a character, motion toward a place taken as the subject of the narrative, toward the location of the central character at reference time, or toward the place which is the central character’s home base at reference time. This may be seen as a further extension from the prototypical case, since the speech situation is no longer referred to. In third person narrative, only perspective taking is involved and not deixis in a strict sense.
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.8 (82)
Åke Viberg
It is an open question whether the schema given above can serve as a universal prototype for the contrast between ‘go’ and ‘come’. It has long been known, for example, that there are languages such as Garo, a Tibeto-Burman language spoken in Bangladesh, which completely lack a lexical contrast between ‘come’ and ‘go’. In this language, there is a verbal root re- meaning ‘move’ which can be combined with affixes to form derived verbs with meanings close to ‘go’ (reang-) and ‘come’ (reba-) (see Burling 1970: 10–13). It thus appears as if a similar semantic contrast is present even if it is not expressed in different lexical roots. There are also languages with more extended systems. In Diuxi Mixtec spoken in Oaxaca, Mexico (Kuiper & Merrified 1975), there is a set of 10 motion verbs which contrast with respect to both deictic distinctions and whether or not a location is the homebase of the subject of the verb. A more radical challenge against the assumption that a contrast set out along the lines in Table 3 represents a language universal comes from the thought-provoking study by Wilkins and Hill (1995), who show that languages vary with respect to what constitutes the prototypes of deictic motion verbs. The question of universality will, however, not be discussed in any depth in this study.
. Human locomotion The verb komma is more frequent in Swedish than its cognate come in English. In the Swedish originals in the English Swedish Parallel Corpus (ESPC), there are 1764 occurences of komma out of a total of approximately half a million words. In the English originals, which contain practically the same number of words as the Swedish ones, there are 934 occurrences of come. (The proportions are similar when the frequency of komma in the Swedish SUC corpus are compared to the frequency of come in the British National Corpus, BNC. See Table 1 in Viberg 2000.) As will be shown below, komma has a semantically wider extension than come. This is reflected in the extent to which the two cognates are used as translations of one another. Whereas come is translated with komma in 56% of all its 934 occurences, komma is translated by come only in 37% of all its 1764 occurrences. One important reason for this is that the range of grammatical meanings is greater for komma. However, this is not the only reason. As will be demonstrated in this section, komma has in general a wider semantic range even as a concrete motion verb. The verbs meaning ‘come’ are most similar across the languages included in this study when they refer to Human locomotion, i.e. when the subject is human and the verb refers to motion in physical space. In Table 4, the major translations of English come into Swedish are shown to the left. In the columns to the right of this, English and French translations of Swedish komma are shown. The primary equivalent in each case is the most frequent translation but to a highly variable extent. When it refers to Human locomotion, the Swedish verb komma is translated
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.9 (83)
The polysemy of the Swedish verb komma ‘come’
Table 4. Swedish translations of English come and English and French translations of Swedish komma when the verbs refer to Human locomotion in concrete space Semantic features
Goal-oriented
In vehicle/ Non-selfpropelled
Swedish translations of come komma dyka upp ‘turn up’ nå‘reach’ anlända ‘arrive’
307 5 2 1
English translations of komma
French translations of komma*
come get+Part. arrive turn up bring
(re)venir arriver s’approcher
97 70 14
(r)entrer (res)sortir (re)monter (re)descendre (re)passer
29 25 7 8 15
420 59 35 14 7
åka ‘ride’ fara ‘travel’ resa köra ‘drive’ flyga ‘fly’ cykla inträda
1 6 4 1 3 1 1
enter
12
återvända följa ‘follow’
1 3
return follow
17 2
go
29
Conflated direction
Divergence: come/go
gå (stiga )
Various Paraphrase Zero Total Human locomotion Proportion of all uses of ‘come’
16 9
aller
7
34
93 21 13
58 15 21
395 42%
722 41%
366 64%
* The Swedish texts translated into French represent only fiction
with come in 420 out of a total of 722 examples of this type in ESPC. This amounts to 58%, which should be compared to the percentage of come as a translation when all uses are taken into consideration (37%). The English verb come is translated with komma to an even greater extent and accounts for 78% of the translations of come when it refers to Human locomotion (compared to 56% when all uses are included). When komma is translated into French, venir is the most frequent translation but accounts for only 27% in the function Human locomotion. In the leftmost column of Table 4, the translations are grouped according to certain shared semantic features. As can be observed, a number of Goal-oriented
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.10 (84)
Åke Viberg
verbs other than the primary equivalent of Swedish komma are used as translations in English and French. In French, arriver reaches 19% which approaches the percentage of venir. In Swedish, verbs describing movement in a vehicle (such as ‘drive’, ‘ride in a car’) are used more frequently than in English. As translations of come, the frequency of these verbs is moderate. The reason this group is singled out is that there is a rather prominent group of translations of English go, which cannot be translated with Swedish gå when a vehicle is involved (see Viberg 2000). When komma is translated into French, the directional verbs (entrer, sortir, monter, descendre, passer) together account for as much as 23% of the translations. This is related to the fact that Swedish is a satellite-framed language (Talmy 1985), where komma like other motion verbs can be combined with a large number of directional verbal particles. As can be seen in the section Divergence come/go, the verb meaning ‘come’ to some extent is translated with the verb meaning ‘go’ but this applies to a moderate extent. What is most remarkable is the degree to which Swedish komma is translated with other Goal-oriented verbs than the one meaning ‘come’. In concrete, spatial uses, Swedish komma in several respects is extended further from the deictic prototype than English come as described by Fillmore. There are a number of uses where Origo is defined in a general way independent of the speech situation. Such cases are discussed in depth in Di Meola’s (1994) insightful study of German kommen and gehen. One case is Endpoint Focus which in addition to the general Goal-orientedness highlights the final phase. The verb komma in combination with the languagespecific particle fram discussed in connection with example (3) in Table 1 tends to be translated with verbs meaning ‘arrive’. (Only the Finnish verb means ‘come’.) Först på kvällen kom de fram till Röbäck där kyrkan låg. KE They didn’t reach Röbäck, where the church was, until evening. Röbäck, wo die Kirche stand, erreichten sie erst am Abend. Elles n’arrivèrent à Röbäck qu’en fin de journée. Vasta illalla he tulivat Robäckin kirkonkylään. In this example, reach is used as an English translation, but arrive is more frequent: Till Birka kom Nordens förste kristne apostel Ansgar omkring år 830. AA
About 830 Ansgar, the first Christian apostle in Scandinavia, arrived at Birka.
The Swedish verb komma is actually the predominant translation of arrive. In the English original texts, there are 111 occurrences of arrive. 67 (60.3%) of these are translated by komma. The second most frequent translation is anlända, the most direct equivalent of ‘arrive’, which is used 14 times. Get can also be used as a translation of Swedish komma fram:
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.11 (85)
The polysemy of the Swedish verb komma ‘come’
Det var fortfarande ganska mörkt It was still fairly dark when he got to när han kom fram till huset. LG the house. The most interesting case in many ways is the use of komma to describe the Overcoming of an Obstacle to Movement as in the following example: Genom den hålvägen kom man väl inte fram förrän till våren. AL No one was going to be able to get through that tunneled path until spring. Si ça continue, /—/, on ne pourra pas passer par là avant le printemps. Origo is here a desired place which is hard to reach. In the English translation, get + Particle is used. The verb get has a tendency to function as a verb of success similar to ‘manage’, ‘succeed’ (see Viberg 2002). What is particularly striking is that komma in this function can be used as a Source-oriented verb as in the following two examples: Komma ut med släpvagnen genom grinden var inte alltid det lättaste. LG Många gånger kom vi inte iväg förrän vid halv ett på natten, ibland inte förrän klockan två. CE
Getting out of the gate with the trailer on wasn’t always easy. Very often we didn’t get away until half-past-twelve at night, sometimes not until two in the morning.
In the last example, the temporal adverbial refers to the time of departure which is critical for verbs of Disappearence, but even if these examples are Source-oriented and with focus on leaving the present location, Origo could still be said to represent a desired place, which is negatively specified as a place different from the present one. There are also metaphorical uses based on the idea of overcoming an obstacle: Hon hade behövt honom att prata She needed to talk to him to get med för att komma ifrån den här away from this uneasy feeling. oron. MG
. Extended lexical meanings Extensions with respect to the deictic component have already been discussed in the preceding section. Section 3 will primarily be devoted to such extensions which depend on the appearance of a non-prototypical subject or a PP complement which does not refer to concrete physical space.
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.12 (86)
Åke Viberg
. Extensions related to the verbal complement Extensions marked by the verbal complement proceed in small steps. Certain concrete spatial complements are so closely associated with a characteristic activity (e.g. ‘bed’ => ‘sleep’) that movement to the place implies the start of the activity: Han kunde utforma tjugo till trettio brev om dagen och kom sällan i säng före midnatt. KF
He could write twenty to thirty letters a day, and he seldom went to bed before midnight.
The use of komma in this Swedish example (instead of the unmarked gå till sängs ‘go to bed’) implies that some impediment was involved and means something like ‘manage to get to sleep’. When ‘come’ is used to indicate geographical origin, the spatial complement is still concrete but the dynamic motion component has turned into a rather static notion of ‘origin’, a kind of property ascribed to the subject. Both English and Swedish frequently use ‘come’ to express this meaning, but there are more specific expressions like Swedish härstamma från ‘stem from’ used as a translation in some examples: The Amundsens came from Hvaler, a cluster of islands at the mouth of the Christiania Fjord. RH
Familjen härstammade från Hvaler, en ögrupp i Oslofjordens mynning.
Spatial Source can be extended to cover Cause (one aspect of the Event Structure Metaphor). With an abstract complement of the preposition ‘from’, origin is reinterpreted as cause and the subject as a result: The pride came from an instinct. NG
Stoltheten hade sin grund i en instinktiv känsla.
There are many different types of extended uses of ‘come’ in combination with a PP or other verbal complement. To a great extent, the interpretation of such expressions can be motivated with reference to metaphor theory (Lakoff & Johnson 1980, 1999). In spite of the fact that there is a great similarity across languages at the conceptual level with respect to which metaphors are used, the lexical realization appears to be rather language-specific and to involve partly idiom-like phrases. One exemple will be given, which appears to be typical in many respects. Both Swedish komma and English come can appear in phrases denoting Existence, but the usage patterns are rather language-specific, a fact which is reflected in the translations. The English expressions come into being / come into existence are based on the metaphor Existence Is Being Located Here (Lakoff & Johnson 1999: 205).
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.13 (87)
The polysemy of the Swedish verb komma ‘come’
The verb used in the Swedish translation is based on another spatial metaphor for existence and is a compound of upp- ‘up’ and stå ‘stand’: The new science of brain/mind which Freud envisaged came into being in the Second World War, in Russia. OS
Den nya vetenskapen om hjärnan/medvetandet som Freud förutsåg uppstod under andra världskriget, i Sovjetunionen.
In Swedish, komma can be combined with till ‘to’ as a stressed particle which implies there is some general Goal such as ‘the world’ or ‘reality’. The choice of the English translation is governed by the semantic category of the entity that comes into existence: Men det skulle dröja över två århundraden, innan stadens nästa universitet kom till. LI
But more than two hundred years were to elapse before the next university of Lund was founded.
Ett nytt redskap kom till – en yxa gjord av flinta med vilken de kunde ta sig fram genom skogen. AA
A new implement was invented – an axe of flint to enable the hunter to hack his way through the forest.
The examples mentioned so far are based on the same metaphor but contrast with respect to the lexical realization. There is one use of English come which is based on a slightly different metaphor related to existence, where the existence of something (especially merchandise) can be observed in various forms and kinds. Swedish in this case uses the existential verb finnas (lexicalized passive form of finna ‘find’) used in the expression det finns which is the primary equivalent of English there is. Look, it even comes as an anaesthetising spray. JB
Se här, den finns till och med som bedövningssprej.
Everything, it seemed, came in Det verkade som om allt fanns i dozens of shapes and sizes. DF dussintals olika former och storlekar. Bankers come in all shapes. AS
Det finns alla slag av bankmän.
. Non-human subjects and the subject cline There are many semantically defined classes of subjects of komma which to various degrees deviate from our typical conceptualization of motion. One such class consists of natural forces such as rain, mist, wind and waves. These entities are relatively concrete but often lack a clear shape.
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.14 (88)
Åke Viberg
Då såg hon dimman komma över skogen. AL
Then she saw the mist coming over the forest.
The movement of air is often only indirectly experienced via the motion of other objects (leaves moved by the wind) or via a shift in temperature: Det kom in frisk luft. MR
Fresh air had come in.
Natural forces such as wind and rain can be conceptualized either as a kind of mass in motion or as a process, which is reflected in the alternation between expressions like the rain fell/came and it rained. This alternation can also be reflected in the translations (det blåste ‘it blew’): Det blåste genom hålet i köksväggen.
A breeze was coming through the hole in the kitchen wall.
In an example like she saw the mist coming, a human observer is explicitly mentioned and Origo can be identified as the place of the observer. In other cases, a potential human observer is usually presupposed and Origo is gradually extended towards something like ‘the realm of experience in general’. This notion of perceptual access is one aspect of what Lindner (1981) calls the Region of Interactive Focus (see below). One semantic class of subjects which is particularly interesting is represented by sense impressions such as light, sound, smell and heat/cold which are experienced as moving entities, even if such phenomena are much less tangible than three-dimensional animate beings and physical objects. The concept of motion also deviates from the prototype. Not all parts of the movement schema can be clearly identified. The source is often salient. We can often clearly identify the source of a light, a smell and a sound at a distance. We also experience that a sense impression somehow moves towards us and strikes one of our senses. This motivates the use of ‘come’, if Origo is identified as our realm of experience, what we can perceive. But the experience of motion is rather vague with sense expressions. Even if there is an expression indicating Path in the first two examples (‘through the window’, ‘through the fog’), it is difficult to form a clear picture of the complete path from Source to Goal. Furthermore, it is impossible for our sense organs to calculate the speed of such entities. Ljuset kommer genom fönsterrutan och ett mellanrum och sedan in i rummet. CE
The light comes through the window pane and a space in between and then into the room.
Genom dimman kom en lukt av dy Through the fog came a smell of och avlopp och gammal sur sälta earth and sewer and old sour salty pysande in mot mig. JMY steam.
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.15 (89)
The polysemy of the Swedish verb komma ‘come’
Dunsen kom väl från nån annan våning då. MG
The noise must have come from some other apartment.
Even if ‘come’ can be used in both Swedish and English, it is sometimes possible to use a peception verb as translation. In Swedish, it appears that höras ‘be heard, be audible’ is rather often used as a translation when sound is involved (The literal translation of the following example is: ‘A cheer could be heard from the living room’.): Ett hurrarop hördes från vardagsrummet.
A cheer came from the living room. ST
There are many complications which will not be gone into here. For example, light can be experienced both as a mass without a clear shape and as various types of countable entities with a clear shape such as a cone of light which can be experienced as moving along a path with an observable speed. The use of ‘come’ with a sense impression as subject represents a universal tendency. In some languages, the primary equivalents of perception verbs such as ‘taste’ and ‘smell’ are realized as ‘taste come’ or ‘smell come’. There are even languages such as Lezgian, where ‘hear’ is realized as ‘voice come’ (Viberg 1984, 2001). In an analysis of the Swedish verb slå ‘strike’, ‘hit’ (Viberg 1999), a scale called the subject cline was proposed showing the decreasing prototypicality of subjects of this verb. It is shown in slightly modified form in Table 5. It must be stressed that the scale in its present form is only meant to illustrate the basic idea. I hope to be able to work it out in greater detail in a more general publication on Swedish motion verbs. Table 5. The subject cline Humans ⇒ & animals
Vehicles ⇒
Natural ⇒ forces
Sense ⇒ impressions
Physical ⇒ events
Verbal ⇒ events
Mental events
Moving down the scale, both the concept of motion and the concept of Origo are gradually extended. Primarily, this process is a kind of continuous generalization rather than metaphorical transfer. Contrary to Mental events which will be described in a later section, Physical events can be observed directly with our senses but lack time-stability. When physical events are realized as nouns, ‘come’ marks the point in time when the event takes place and enters the realm of physical experience (Origo): Då kom explosionen. HL
Then came the eruption.
Swedish komma is sometimes translated with occur in examples of this type:
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.16 (90)
Åke Viberg
En stor omvälvning i människornas liv kom omkring år 500 före Kristus, när de lärde sig att utvinna järn ur den sjömalm som bildades på sjöarnas botten. AA
Another revolution in human life occurred about 500 BC, when iron was first extracted from the bog ore formed at the bottom of the lakes.
There are several related concepts with partly different characteristics. Periods of time in many respects resemble physical events. Concepts such as ‘spring’ and ‘summer’ are closely associated with typical changes in nature and certain cultural events: Våren kom och blev sommar. SCO Spring came and grew into summer. Sommaren kom och med den bondfamiljens årliga höjdpunkt, Kiviks marknad. RJ
Summer came and with it the highlight of the farming community – the annual market day at Kivik.
Certain processes such as cultural and technical innovations can move from various concrete places and are situated in time: För ungefär 5 000 år sedan kom jordbruket till vårt land. AA
About 5000 years ago agriculture was introduced into Sweden.
This example clearly refers to motion in physical space, even if the moving entity (agriculture) is a process of change and not a physical object. Various types of physiological changes in our body such as various types of pain can be localized to various parts of the body but are otherwise very similar to mental events, which will be treated in the next section: Men huvudvärken hade kunnat komma lika väl utan brännvin. LG
But the headache could just as easily have come without the alcohol.
. Mental expressions The following is a typical example of the last step of the subject cline: Nu kom de där tankarna. KE
Those thoughts were coming again.
In examples of this type, Origo has been extended further from the realm of physical experience into something like our inner experience or consciousness. According to Lindner (1981, 1982), this kind of cognitive access can be conflated with cases of perceptual access into a single region of interactive focus mentioned above, which is characterized as follows:
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.17 (91)
The polysemy of the Swedish verb komma ‘come’
a functional assembly which takes human interaction as essential and is organized around the way people canonically interact with things in the world around them – physically, socially, perceptually, cognitively, etc. Things (or people) located in this region can be in any of a cluster of states represented by this region: in use, prepared, active, mobile, agitated, cognitively or perceptually salient, existing, public, viable, known, and so on. (Lindner 1982: 317–18)
The concept of interactive focus also subsumes expressions related to the metaphor Existence Is Here discussed in 3.1. Di Meola (1994) shows that interactive focus explains a wide range of uses of German kommen, which tend to have rather close parallels in Swedish. The extensions of komma based on the interpretation of Origo are summed up in Table 6. A proposition syntactically realized as a that-clause is also a kind of mental subject. In the following two Swedish examples, the Experiencer appears in a prepositional complement marked with the preposition för (‘It came for her that-S’): Det kom för henne att hon skulle smyga tillbaka och se vad Birk hade för sej i hennes skog. AL
It had occurred to her that she could steal back and see what Birk was doing in her woods.
Men det var många timmar kvar till natten och Vargsången, därför kom det för henne att nu skulle hon göra det hon så länge hade tänkt. AL
But there were many hours before nightfall and the Wolf ’s Song, so it crossed her mind that now was the time to do what she had planned for so long.
Table 6. The extension of the deictic interpretation of Origo into the region of interactive focus
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.18 (92)
Åke Viberg
Two different expressions are used as translations in English. In the following two examples from English originals, Swedish uses expressions with slå ‘strike’ (’it struck her that-S) and falla ‘fall’: And this, it suddenly came to her, Och detta, slog det henne plötsligt, might well be the wages of sin. FW skulle mycket väl kunna vara syndastraffet. Except for the playwrights, none of Bortsett från dramatikerna var inga the great names in poetry that av de stora namn inom poesin som came quickest to mind were först föll i tankarna athenare. Athenian. JH In spite of the fact that there are close similarities across the languages in the sample at a general plane, there is an unusually wide range of variation in the translations of mental predications involving ‘come’. One characteristic of mental predicates is that they can either be Phenomenonbased, taking a mental entity as subject (as in the above examples), or Experiencerbased, taking the human Experiencer as subject. There is frequently an alternation between Phenomenon- and Experiencer-based constructions both within and between languages. In the following examples, Swedish has Experiencer-based expressions based on the verb få ‘get’ (‘I got the idea’), which correspond to Phenomenon-based expressions with ‘come’ in English and French: Jag fick tanken tidigt på morgonen. AP
The idea came to me in the morning.
Tout à coup, une idée lui vint, qui l’éblouit. CARR
Plötsligt fick han en idé som tycktes honom vara lysande.
Experiencer-based mental predicates with komma are shown in Table 7. The most frequent use of Swedish komma in a mental predicate is the set phrase komma ihåg ‘remember’, which consists of ‘come’ + the stressed particle ihåg composed of i ‘in’ and a more or less obsolete word for ‘mind’ håg. In present-day Swedish, the elements of the phrase do not form a coherent interpretation (‘I came into mind that-S’) but in Medieval Swedish there was a Phenomenon-based construction (‘Me came into mind that-S’) with the Phenomenon in nominative and the Experiencer in the dative (Söderwall: hugher 5). This use was based on the metaphors The Mind Is A Container For Objects and Ideas Are Moving Things. As a counterpart, there was also the now obsolete gå ur hågen (SAOB: håg 1b) ‘go out of mind, be forgotten’. In general it appears that modern Swedish favours Experiencer-based mental verbs. The dominant translation for komma ihåg is remember but even a number of synonyms are used such as recall.
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.19 (93)
The polysemy of the Swedish verb komma ‘come’
Sedan kom han ihåg kaffet och tömde det i ett enda drag. AP
Then he remembered his coffee and drank it down in one go.
Men hon kunde aldrig komma ihåg vad hon egentligen drömt. MG
But she could never recall exactly what she had dreamed.
As can be observed in Table 7, it is primarily verbs with a basic mental meaning that are used as translations of the Swedish expressions. The major exception is komma fram till (en slutsats) ‘arrive at a conclusion’ which is based on the metaphor Reasoning is Following a Path (in combination with the metaphor that equates Goal of motion and Result): Nobel ansåg sig redan i unga år ha kommit fram till en bestämd uppfattning om människonaturen. KF
Even as a young man, Nobel believed he had arrived at an immutable opinion regarding human nature.
But even in this case, there are alternative translations: Han analyserade läget och kom fram till att han var i behov av hjälp. KOB
He thought the matter over and concluded he needed help.
Table 7. Experiencer-based mental expressions with komma ‘come’ and their translations into English and French Swedish
komma ihåg
komma på’
komma överens
komma fram till
Various other expressions
English Total remember recall keep in mind Various think of find Various agree get along arrive at find think conclude
118 46 3 2 2 5 5 12 5 1 4 2 1 1 29
French Total
47
se souvenir
14
Various penser trouver Various s’entendre se mettre d’accord arriver à
6 3 2 12 1 1 1
7
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.20 (94)
Åke Viberg
Even though several of the expressions with komma forming mental predicates have lost much of their original motivation, the phrases are not completely lexicalized. In all the Experiencer-based expressions, the subject of komma lacks Agent-like properties. Komma ihåg can be compared to dra sig till minnes ‘pull (reflexive) to memory’, which describes an active effort to remember: Han försökte dra sig till minnes He tried to remember if he had seen om han hade sett de här skogarna i these woods in their younger days. deras yngre dagar. RR The expression komma överens, which can be translated ‘agree’, can be compared to a phrase with gå which also tends to be translated with ‘agree’: gå med på ‘get along with, agree’ (Viberg 2000). The phrase with gå always implies an active acceptance. Redan tidigare hade vi kommit överens om att texten skulle läras så fort som möjligt. IB
We had already agreed that the text should be learnt as soon as possible.
Sometimes the translation represents a reversal of the base-selection as in the following example where an Experiencer-based expression in the original is shifted to a Phenomenon-based construction in the translation: Ta som exempel den finaste tapeten damerna kan komma på, Kung Sveno, som hänger i Skattkammaren och är från 1560-talet. GAPG
Take, for example, the finest tapestry that comes to mind, King Sveno, which hangs in The Treasury and dates back to the 1560s.
. Verbal communication Verbal communication can be conceptualized either as a physical or as a mental event or as a combination of the two, which motivates placing it between these two types of events in the subject cline. In some cases, the sound impression is prominent. In the following English example, come is translated with låta ‘sound’, which, however, is partly extended to a mental verb (pat refers to the content rather than the sound): Wexford thought that came out rather too pat. RR
Wexford tyckte att det där lät alldeles för fixt och färdigt.
In the following example, the noun ‘lie’, which basically refers to verbal communication, primarily has a mental interpretation as a kind of impulse or idea: The lie came easily to her and was as easily accepted. PDJ
Lögnen kom av sig själv och de accepterade den utan vidare.
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.21 (95)
The polysemy of the Swedish verb komma ‘come’
The region of interactive focus is involved in the above examples. It is also the motivation for many expressions of other types such as the following ones which primarily refer to what becomes publicly accessible: Så kom nyheterna om den ryska invasionen, tanksen på Budapests gator.
Then came the news of the Russian invasion, the tanks in the streets of Budapest. BR
The expression komma till tals ‘come to speech, have the word’ has a similar motivation: I den tidens kommittéarbete kom redarna och bankcheferna direkt och personligen till tals. TR
In those days, shipowners and bankers addressed the committee directly and personally.
There are several different types of motivations for expressions related to Verbal communication. The Conversation-as-a-Journey metaphor (Lakoff & Johnson 1980) involves among other things the conceptualization of topics as Goals and the evolving discourse as motion. Lundagård in the following example is originally the name of a place, but komma in på ‘come in on’ in this case refers to a journal which is the topic of conversation: Det är omöjligt att skriva om Lundaanda utan att komma in på Lundagård, envist utkommande sedan 1920. LI
It is impossible to write about the Lund spirit without bringing up Lundagård which has persisted in coming out since 1920.
The Conversation-as-Journey-metaphor is also the motivation behind the following example (with komma tillbaka till ‘come back to’): Kungaparet kommer flera gånger tillbaka till vikten av att värna om traditioner och föra dem vidare. GAPG
Several times the Royal couple returns to the importance of standing by traditions and passing them on.
The same metaphor is also the motivation behind the English expression when it came to which is developing into a phrasal connective: Tydligen var Billy Polo väldigt ombytlig ifråga om lagbrott och hade aldrig brytt sig om att specialisera sig.
Apparently, Billy Polo was pretty shiftless when it came to breaking the law and had never even settled on an area of expertise. SG
När det gällde tekniska framsteg var William Ford ingen bakåtsträvare.
When it came to mechanical progress, William Ford was no Luddite. RL
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.22 (96)
Åke Viberg
. Grammatical meanings The verb komma has several grammatical meanings but only one is frequent, the use as a future marker, which will be treated in the next section. After that the other uses are briefly summarized in a separate section. The development of grammatical meanings of ‘come’ represents a universal tendency. According to Bybee et al. (1994), 10 of the 76 languages in their stratified sample had a future marker originating from a verb meaning ‘come’. A number of other grammatical meanings have also been documented for ‘come’ (see Lichtenberk 1991 for a detailed description of grammaticalization of ‘come’ and ‘go’ in Oceanic languages).
. The use as a future marker The major grammatical use of komma is as a future marker. The equivalent of ‘komma’ in English, German and French cannot be used with this meaning. German uses werden with the basic meaning ‘become’, whereas French uses an inflectional future marker, which alternates with the periphrastic future with aller ‘go’ as a translation in the corpus. Only the Finnish verb tulla ‘come’ can be used with a future sense as shown in the following example from a Finnish original: Vi kommer att använda pengarna för medicinska ändamål, Me tulemme käyttämään rahat lääketieteellisiin tarkoituksiin, (Salama) We will use the money for medical purposes. (My English translation) The Swedish future markers are described most thoroughly in Christensen (1997). See also Teleman, Hellberg and Andersson (1999, Ch. 31 §28–31). A brief description in English is given in Viberg et al. (1984, 9.2). The verb komma is primarily used as a future marker in its present tense form: Maria kommer att vinna (literally: Maria comes to win) ‘Maria will win’. The future with kommer att to a very high Table 8. English translations of Swedish kommer att as a future marker TEMPORAL:
Future
kommer + att VPInfinitive
Proportion of all uses of komma
will ‘ll would shall be going to be to Modals come to Present Various TOTAL 12,5%
142 20 5 7 13 2 11 2 5 14 221
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.23 (97)
The polysemy of the Swedish verb komma ‘come’
degree expresses prediction. Even if prediction is also the most frequent meaning of the British English future markers will, shall and be going to according to Coates (1983), Bybee (1988: 362), these markers can also introduce other shades of future meaning, whereas Swedish kommer att expresses the predictive meaning in a more or less pure form. The translations of kommer att as a future marker are shown in Table 8. The dominant translation is will (and its variant form ‘ll), which is used more than ten times as often as any other marker: Principen om enhällighet i fråga om nationellt viktiga utrikes- och säkerhetspolitiska beslut kommer med största sannolikhet att upprätthållas. IC
When it comes to foreign and security policy decisions of national importance, the principle of unanimity will in all probability be maintained.
Tio år är en lång tid, era barn kommer att vara vuxna, några kommer att ha tröttnat. GT
Ten years is a long time. Your children will be grown, some will have grown tired.
Hon kommer att stöna eller sucka eller vissla eller på något sätt demonstrera sitt motstånd. MR
She’ll moan and groan and whistle or in some way show her dislike.
When some other alternative than will is used, kommer att still basically expresses prediction even when some other semantic factor appears to motivate the English choice. According to Quirk et al. (1985, §4.43), the future marker be going to has two more specific meanings future fulfilment of present intention with personal, agentive subjects (When are you going to get married?) and future result of present cause with all types of subjects (It’s going to rain; She’s going to have a baby). These meanings motivate the choice of be going to in the following examples: “Det här kommer att reta livet ur Borka”, sa Mattis. AL
“This is going to plague the life out of Borka,” said Matt. (present cause)
“Nej,” sa vi, “vi kommer inte att nöja oss med det.” CE
“No,” we said, “we’re not going to make do with that.” (intention)
Actually, it is possible to find kommer att even in other examples where intention appears to be involved as in the following example: Som medlemmar kommer vi också And as members we will actively att aktivt verka för ett fortsatt work for closer relations between närmande mellan de baltiska the Baltic states and the EU. staterna och EU. MAU
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.24 (98)
Åke Viberg
This example is taken from an official speech by the then foreign minister of Sweden Margareta af Ugglas. Since the subject is in the first person plural, the example strictly speaking involves intention, but the use of the future marker based on tänka ‘think’ which is the clearest marker of an intentional future in Swedish would sound casual and insincere in this context. Even the use of the general future marker ska ‘shall’ which often introduces a modal component such as obligation or intention would sound less sincere. The use of kommer att makes the statement sound like a commitment due to the basic predictive meaning of this marker. Even if the present tense form kommer is primarily used when the verb serves as a future marker, Swedish komma can be used as a future marker in all its inflected forms. When the infinitive is used in combination with a modal verb such as kunna ‘can’, English appears primarily to use a single modal. (The phrase kan komma att is translated by may in 11 cases out of 16. This correspondence is not included in Table 7 which exclusively shows the translations of kommer att): Med ökade kunskaper kan kraven komma att formuleras på andra sätt. BJ
As our knowledge increases, these requirements may well be formulated in some other way.
. The ‘hindsight’ past future marker The verb komma can be used as a future marker even in its past form kom: Vi kan då konstatera att vid sekelskiftet hade den svenska rederinäringen en relativ position som den kom att behålla i 70 år. TR
We can observe that the Swedish shipping industry at the turn of the century occupied a position that it would retain for 70 years.
The retaining of the position is something which is situated in the future with respect to the turn of the century, which serves as a temporal reference point for the main narration. Simultaneously, the use of kom att as a past future marker presents this as a fact based on the knowledge that is available to the writer at encoding time. The use of komma as a past future marker has the same objective tone as the use of komma in ordinary future expressions. With hindsight, a future development in the past can be presented as a fact, whereas a future development with respect to ‘now’ can only be a prediction even if it is based on solid objective evidence. It appears that the other languages in the sample lack a clear exponent of this meaning. As shown in the following example, kom att in this use can be paraphrased or even left untranslated. In the following example, paraphrases are used which focus on the gradualness of change which happens to fit this particular example (grow in
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.25 (99)
The polysemy of the Swedish verb komma ‘come’
English and phrases meaning ‘with time’ and ‘little by little’ in German and French. The Finnish verb kiintyä ‘become attached to’ contains an inchoative suffix -y/ -u): Han kom att fästa sig vid fiskhandlarn. MF He grew to like the fish merchant. Mit der Zeit schloßer den Fischhändler ins Herz. Il s’attacha peu à peu au poissonnier. John kiintyi kalakauppiaaseen. There is, however, a rather close translation equivalent in German of the Swedish ‘hindsight’ past future marker, even if it does not appear in the still very restricted pilot corpus. The past form sollte of sollen ‘shall’ can be used as a past future marker to represent something which has turned out to be a fact (sometimes referred to as ‘the future of destiny’). The major translations in ESPC of kom att as a past future marker are shown in Table 9. The most frequent explicit equivalent is be to (in past form): Så utbildades ett världsligt frälse, en adel som kom att omfatta hela riket. AA
In this way a lay nobility was formed, an aristocracy which was to spread throughout the entire country.
However, be to only occurs as a translation in 7 out of totally 45 cases. The most frequent option is actually no translation (zero in the table). Stora områden som vi idag uppfattar som idylliska med låg gatubebyggelse kom på så sätt att räddas för eftervärlden. LI
Large districts with their small houses facing the street that we consider idyllic today thus were preserved for posterity.
Eftersom kyrka, stat och handel vid denna tid samarbetade kom Lund att bli en betydelsefull plats, som kunde få många människor att vilja flytta hit. LI
Since Church, State and trade at this time went hand in hand, Lund became an important place which caused many people to move here.
Table 9. Grammatical meanings of Swedish komma 2: Past future TEMPORAL: Past Future
kom att + VPInfinitive
be to come Zero Various TOTAL
7 5 23 10 45
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.26 (100)
Åke Viberg
The use of kom att as a factual past future marker does not appear to have a direct equvialent in English and since it represents a fine nuance, it is often left without a translation. The factual component is evident, if kom att is compared to other alternative past future markers in Swedish. The most general past future marker skulle ‘should’ usually has a modal element expressing the intention of the subject or some social obligation (e.g. a contract), as in the example: Peter skulle stanna i Moskva i två år. ‘Peter should stay in Moscow for two years.’ Another option in Swedish is to use the past form tänkte of tänka ‘think’ which expresses the intention of the subject Peter tänkte stanna i Moskva i två år. ‘Peter would/intended to stay in Moscow for two years.’ In both of these cases men det gjorde han inte ‘but he didn’t’ can be added without contradiction. This is not possible with the expression Peter kom att stanna i Moskva i två år, which describes what actually turned out to be the case, perhaps contrary to expectations, since a simple past tense also would report the situation as a fact: Peter stannade i Moskva i två år ‘Peter stayed in Moscow for two years.’ What is characteristic of the Swedish past future marker kom att is a double perspective. The situation is viewed as future from some observation point in the past. Simultaneously, it is presented as a fact by the author who has access to information about the actual outcome. In addition to kom att as a past future marker the combination skulle komma att is also found (8 occurrences in ESPC). Basically, this combination also reports a situation as a fact, which seems to be signalled by komma, but skulle appears to introduce an additional shade of meaning which remains to be characterized properly: Det var en sensation som skulle komma att nå ganska långt, förstod jag senare. POE It was a sensation which was going to be widely known, I realized later. L’événement allait faire sensation, et connaître un assez grand retentissement, je le compris plus tard. Se oli sensaatio jonka vaikutus ulottui aika pitkälle, sen ymmärsin myöhemmin. There are also a few examples in the corpus of the perfect of komma (har kommit ‘has come’) used as a future marker. Even in this case, the situation is regarded as a fact: Biståndet har dock volymmässigt kommit att domineras av ovan nämnda former. CO
However, in terms of volume, development assistance has been dominated by the above-mentioned forms.
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.27 (101)
The polysemy of the Swedish verb komma ‘come’
. The use as an inchoative and causative marker Sometimes when komma is combined with att + VPInfinitive , it has an inchoative rather than a future meaning. The two meanings can cooccur, but I am referring to the cases where a future, or future in the past, meaning is ruled out. In the ESPC, one phrase komma att tänka på (usually in the past tense) ‘come to think about’ accounts for more than half of the occurrences of this use. As can be observed in Table 10, the phrase is often translated by a simple cognitive verb (think, remember) without any direct marking of the inchoative meaning as in the following two examples: Siiri blev upprörd igen när hon kom att tänka på sitt misstag. AP
Siiri grew agitated again when she thought about her mistake.
Då kom han att tänka på Then he remembered the tow rope bogserlinan som de bundit honom they had tied him up with. med. KE However, the verb occur used in the following example has a dynamic component: Så kom han plötsligt att tänka på nånting och spände ögonen i Ronja. AL
Then something suddenly occurred to him, and he looked at Ronia, his eyes narrowed.
English come can also be used with a primarily inchoative meaning but there appear to be rather strong collocational restrictions on this use in both languages. English come and Swedish komma very seldom are used as translations when this function is involved: Gradually we came to know each other. BR
Så småningom lärde vi känna varandra. [‘learn (to) know’]
He’s coming to believe in human goodness in his old age. MA
Han börjar tro på människans godhet på gamla dagar. [‘begin (to) believe’]
Table 10. The use of komma with a primarily inchoative meaning INCHOATIVE: komma att tänka på komma + att VPInfinitive Other main verbs
think about remember Various come Zero Paraphrase Total
6 4 4 2 4 3 23
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.28 (102)
Åke Viberg
In Swedish, the use of komma as an inchoative marker is relatively restricted. The use of come + Adjective in expressions like come happy is not possible. Finnish has an even more extended use of tulla ‘come’ as an inchoative marker both with nouns and adjectives as complement. In the syntactic frame: NP + komma + NP + att + VPInfinitive , where att is the infinitive marker, the verb komma functions as a periphrastic causative marker. Lyktorna utanför kom snön att avteckna sig som guldkorn. GT
Outside, the lamps made the snow stand out like grains of gold.
Denna idé kom honom att bulta i sin kakelvägg med allt hårdare slag och med klappande hjärta. LG
This notion made him hammer at his tiled wall with even heavier blows and a pounding heart.
Komma in this use tends to be translated by the most general periphrastic causative verb in the other languages in the pilot corpus. English tends to use make, French faire ‘make’ and Finnish saada ‘get’. (For German, the material is too restricted to allow any general claim.) Solen kom från galet håll. Det ovana ljuset kom honom att känna sig utkastad och främmande. KE The sun was coming from the wrong direction, the unusual light making him feel strange and exposed. /—/ Das ungewohnte Licht bewirkte, daß er sich ausgestoßen und fremd fühlte. /—/ et cette lumière inhabituelle le fit se sentir exclu, étranger. /—/ Outo valo sai hänet tuntemaan olonsa hylätyksi ja vieraaksi. The use of komma as a periphrastic causative is a rather marked option. The verb få ‘get’ is the most frequent option in the syntactic frame VCausative + NP + att + VPInfinitive . In the ESPC corpus, there are 39 examples of this type, whereas there are only 7 examples with komma. There is also a clear functional contrast. The periphrastic causative få allows both intentional (i.e. agentive) and non-intentional subjects, whereas komma only allows a non-intentional subject. This is also reflected in the semantic class of the subject. With få as a periphrastic causative, there are approximately as many human subjects as there are inanimate subjects (primarily events), whereas all 7 subjects of komma are inanimate. With a human subject, both an intentional and a non-intentional reading are possible with få : Maria fick oss att skratta [lit. Maria got us to laugh] ‘Maria made us laugh’ (e.g. intentionally by telling a joke or unintentionally by saying something foolish). A human subject does not appear to be completely ungrammatical with komma, but to the extent that the following example is acceptable, it only allows a non-intentional reading
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.29 (103)
The polysemy of the Swedish verb komma ‘come’
Table 11. The use of komma as a periphrastic causative CAUSATIVE: komma + NP + att VPInfinitive
make Zero Paraphrase Total
5 1 1 7
?Maria kom oss att skratta. An example like Maria lyckades få oss att skratta ‘Maria managed to make us laugh’ is perfectly acceptable, whereas *Maria lyckades komma oss att skratta is clearly out. In the use as a periphrastic causative, komma has a nonintentional reading in accordance with the value of this semantic parameter in the basic use of the verb (see 2.1 above).
. Conclusion As stated in the introduction, the primary aim of this paper has been to identify the major meanings of Swedish komma from a crosslinguistic perspective showing what properties represent universal tendencies (to the extent these are known) and what properties appear to be more language-specific. The theoretical problem of representation concerning the relatedness of the various meanings and the problem of interpretation have been referred to in various places but will require a separate paper to be treated more extensively and systematically. In its basic use as a physical motion verb, Swedish komma is characterized by a wide extension of the deictic component which is often reduced to a general notion of goal-orientedness and end-point focus. This is correlated with the fact that other verbs of arrival are comparatively infrequent already in comparison to English but in particular with respect to French. There is also a wide range of abstract uses of komma based on the extension from a deictically defined reference point into the region of interactive focus in Lindner’s (1981, 1982) sense. With respect to abstract, often metaphorically based uses, there appears to be a rather high degree of conventionalization and idiomaticization. This necessitates very large corpora in order to study abstract uses systematically, since languages contrast with respect to the formal realization in spite of the fact that there are great similarities at the cognitive level containing the metaphors which motivate many phrasal combinations. Grammaticalization of ‘come’ has been attested in a wide range of languages, often with many parallelisms in semantic patterning. However, transparent grammatical uses of lexical verbs appear to develop comparatively rapidly on a historical time scale and can push cognates apart during the space of a few centuries as has happened with English come (or even German kommen) and Swedish komma. (A similar observation was made with respect to English go and Swedish gå in Viberg
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.30 (104)
Åke Viberg
2000). According to the Swedish historical dictionary (SAOB: komma), the use of komma as a future marker dates back to the 16th century. In certain respects, closely related languages such as Swedish and English may contrast noticeably with respect to the patterns of polysemy of cognate words even when the semantic extensions represent universal tendencies with parallels in several unrelated languages. Translation corpora are excellent sources of data for identifying major semantic contrasts as well as contrasts with respect to subtle shades of meaning.
References Altenberg, B., & Aijmer, K. (2000). The English-Swedish Parallel Corpus: A resource for contrastive research and translation studies. In C. Mair & M. Hundt (Eds.), Corpus linguistics and linguistic theory (pp. 15–33). Amsterdam and Atlanta: Rodopi. Burling, R. (1970). Man’s many voices. New York: Holt, Rinehart & Winston. Bybee, J. (1988). The diachronic dimension in explanation. In J. Hawkins (Ed.), Explaining language universals (pp. 350–378). Oxford & New York: Blackwell. Bybee, J., Perkins, R., & Pagliuca, W. (1994). The evolution of grammar. Tense, aspect, and modality in the languages of the world. Chicago: The University of Chicago Press. Christensen, L. (1997). Framtidsuttrycken i svenskans temporala system. (Expressions for the future in the Swedish tense system.) [Lundastudier i Nordisk språkvetenskap A52] Ph. D. thesis. Lund: Lund University Press. Coates, J. (1983). The semantics of modal auxiliaries. London: Croom Helm. Di Meola, C. (1994). Kommen und gehen. Eine kognitiv-linguistische Untersuchung der Polysemie deiktischer Bewegungsverben. Tübingen: Niemeyer. Fellbaum, C. (Ed.). (1998). WordNet. An electronic lexical database. Cambridge, MA: MIT Press. Fillmore, C. (1977). Santa Cruz lectures on deixis. Indiana University Linguistics Club. Ide, N., & Véronis, J. (1998). Introduction to the special issue on word sense disambiguation: The state of the art. Computational Linguistics, 24(1), 1–40. Kilgariff, A., & Palmer, M. (2000). Introduction to the special issue on SENSEVAL. Computers and the Humanities, 34(1/2), 1–13. Kuiper, A., & Merrifield, W. (1975). Diuxi Mixtec verbs of motion and arrival. International Journal of American Linguistics, 41(1), 32–45. Lakoff, G. (1987). Women, fire and dangerous things. Chicago: University of Chicago Press. Lakoff, G., & Johnson, M. (1980). Metaphors we live by. Chicago: University of Chicago Press. Lakoff, G., & Johnson, M. (1999). Philosophy in the flesh. The embodied mind and its challenge to western thought. New York: Basic Books. Langacker, R. (1988). A usage-based model. In B. Rudzka-Ostyn (Ed.), Topics in cognitive linguistics (pp. 127–161). Amsterdam: Benjamins. Lichtenberk, F. (1991). Semantic change and heterosemy in grammaticalization. Language, 67(3), 475–509.
TSL[v.20020404] Prn:4/11/2002; 11:47
F: PBNS23.tex / p.31 (105)
The polysemy of the Swedish verb komma ‘come’
Lindner, S. (1981). A lexico-semantic analysis of English verb particle constructions with out and up. University Microfilms International. Lindner, S. (1982). What goes up doesn’t necessarily come down: The ins and outs of opposites. Papers from the 18th Regional Meeting of Chicago Linguistic Society (pp. 305–323). University of Chicago. Newman, J. (1996). Give. A cognitive linguistic study. Berlin: Mouton de Gruyter. Pustejovsky, J. (1995). The generative lexicon. Cambridge, MA: MIT Press. Quirk, R., Greenbaum, S., Leech G., & Svartvik, J. (1985). A comprehensive grammar of the English language. London & New York: Longman. Ravin, Y., & Leacock, C. (Eds.). (2000). Polysemy. Theoretical and computational approaches. Oxford: Oxford University Press. SAOB. The Swedish Academy Dictionary. Available from http://spraakdata.gu.se/lb Söderwall. Dictionary of Swedish medieval language. Available from http://spraakdata. gu.se/lb Talmy, L. (1985). Lexicalization patterns: Semantic structure in lexical forms. In T. Shopen (Ed.), Language typology and syntactic description, Vol. 3 (pp. 57–149). Cambridge: Cambridge University Press. Teleman, U., Hellberg, S., & Andersson, E. (1999). Svenska akademiens grammatik. Uddevalla: Norstedts. Viberg, Å. (1984). The verbs of perception: A typological study. Linguistics, 21–1, 123–162. ——— (1993). Crosslinguistic perspectives on lexical organization and lexical progression. In K. Hyltenstam & Å. Viberg (Eds.), Progression and regression in language (pp. 340– 385). Cambridge: Cambridge University Press. ——— (1996). Crosslinguistic lexicology. The case of English go and Swedish gå. In K. Aijmer, B. Altenberg, & M. Johansson (Eds.), Languages in contrast. Papers from a Symposium on Text-based Cross-linguistic Studies. [Lund Studies in English 88.], (pp. 151–182). Lund: Lund University Press. ——— (1999). Polysemy and differentiation in the lexicon. Verbs of physical contact in Swedish. In J. Allwood & P. Gärdenfors (Eds.), Cognitive semantics. Meaning and cognition (pp. 87–129). Amsterdam: Benjamins. ——— (2000). The polysemous cognates Swedish gå and English go. Universal and language-specific characteristics. Languages in Contrast, 2(1), 89–115. ——— (2001). Verbs of perception. In M. Haspelmath, E. König, W. Oesterreicher, & W. Raible (Eds.), Language typology and language universals. Handbücher zur Sprachund Kommunikationswissenschaft 20.2 (pp. 1294–1309). Berlin & New York: Walter de Gruyter. ——— (2002). Polysemy and the cues for disambiguation across languages. The case of Swedish få and English get. In S. Granger & B. Altenberg (Eds.), Lexis in contrast (pp. 119–150). Amsterdam: Benjamins. Viberg, Å., Ballardini, K., & Stjärnlöf, S. (1984). A concise Swedish grammar. Stockholm: Natur & Kultur. Vossen, P. (Ed.). (1998). EuroWordNet: A multilingual database with lexical semantic networks. Special issue of Computers and the Humanities, Vol. 32, Nos. 2–3. Wilkins, D., & Hill, D. (1995). When “go” means “come”: Questioning the basicness of basic motion verbs. Cognitive Linguistics, 6(2/3), 209–259.
TSL[v.20020404] Prn:8/04/2002; 15:25
P II
Metaphor in contrast
F: PBNSP02.tex / p.1 (1)
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.1 (109)
Studying metaphors using a multilingual corpus Kay Wikberg Department of British and American Studies, University of Oslo
Introduction The study of metaphor combines very nicely semantics and pragmatics: semantics because metaphors convey non-literal meaning, a property they share with other figurative expressions and idioms, pragmatics because metaphors are instances of language use. If we exclude ‘dead’ metaphors like the brow of the hill and in her twilight days, metaphors are created at a given point in discourse. The focus in this paper will be on innovative metaphors; therefore lexicalized metaphors and idioms will be discussed only if they are used in the same context as another figurative expression or if they are made use of to create something new. I am going to say very little about quantities, more about methodology and the linguistic manifestations of metaphor. My aim is to find out what we can learn about metaphors, and their translations in a multilingual corpus by comparing the SL English versions with their renderings, particularly into Swedish and Finnish. I want to test two hypotheses. First of all, one might expect some variation in the extent to which metaphors are translated into the TL depending on a number of factors such as the complexity of the metaphor, distance between source language and target language, and the translator. The last factor has to be dropped in this context since there is no way of testing the effect of the translator. There is only one translator of each text and different translators for each text. As far as the complexity of a metaphor is concerned, its structure varies from a single word to a clause. As we shall soon see, it can even extend over a longer chunk of discourse. Another hypothesis that might be worth testing is whether the study of the metaphors in their co-text would make it possible to relate the use of specific metaphors to stylistic or rhetorical strategies. Too little is still known about why metaphors are used (however, cf. Low 1988; Cac-
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.2 (110)
Kay Wikberg
ciari 1998) or how they are to be described within text theory. Genre awareness is very important if we want to understand metaphor. Thus, literary metaphors are considered to have a predominantly expressive or evocative function (Steen 1994), which is supported by the data in my corpus. By comparison, metaphors in nonfiction tend to have an explanatory function. However, only a close linguistic analysis of metaphors in their wider context can help us to understand their various functional uses. The computer is not an obvious tool for metaphor research since it cannot generally identify metaphors in a text. It cannot detect semantically or pragmatically anomalous sentences, nor can it have any idea of the author’s metaphorical intention. However, provided we have access to interesting machine readable texts, the computer can be used to look for items that we know may be part of metaphorical expressions or expressions containing particular markers of figurative use, such as similes, as if (though)-clauses, and specific phrases. The computer can also provide concordance data to help us study the distribution of literal and non-literal meanings (cf. Deignan 1999) of polysemous words. Above all, it can provide data which otherwise would not have been available or which might have been timeconsuming to find. It can be used to obtain additional data from large corpora like the British National Corpus1 for English or Språkbanken at Gothenburgh2 for collocations and non-literal meanings. For Finnish I did not have access to such a resource.
Corpus The English-Norwegian Parallel Corpus (ENPC)3 consists of original texts in English and several other languages and the translated versions. The texts used for this study are fiction only. I ended up with an arbitrary selection of 11 extracts of 10,000–15,000 words from these books: Atwood, Margaret, Cat’s Eye Barnes, Julian, Talking It Over Brink, André, The Wall of the Plague Dahl, Roald, Matilda Drabble, Margaret, The Middle Ground Forsyth, Frederick, The Fourth Protocol Francis, Dick, Straight Grafton, Sue, “D” is for Deadbeat James, P. D., Devices and Desires King, Stephen, Cujo Townsend, Sue, The Queen and I
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.3 (111)
Studying metaphors using a multilingual corpus
These texts consist mostly of popular fiction written by bestselling authors, but there is great variation in the amount and types of metaphors in the various extracts. Some of the texts are particularly rich or interesting. Thus, one author who stands out as contributing image metaphors of great semantic insight is Margaret Atwood, although she does not use them very frequently in the studied extract. In contrast, the extract by Julian Barnes is rich in metaphors.
Some preliminary examples Let us begin by having a look at a metaphor (1) to see what the translations are like and how the metaphor is used in the English origin and in the target languages. The accent being referred to in this example from The Queen and I is that belonging to the Queen and what is being described is its effect (cf. bleed in the next sentence) on Jack Barker, the Prime Minister. (1) Her accent cut into Jack [Barker] like a crystal. (ST1.1.2.s13) [He half expected to bleed.] N Tonen hennes skar som diamant i ørene hans. S Hennes accent skar i Jack som en diamant i glas. F Hänen ääntämyksensä leikkasi Jackia kuin kristalli. G Ihr aristokratischer [sic] Akzent fuhr Jack durch Mark und Bein. P A maneira como pronunciava as palavras tinha em Barker um efeito semelhante ao de um gume afiado que lhe estivesse a cortar a carne. [N = Norwegian; S = Swedish; F = Finnish; G = German; P = Portuguese. The text in bold is the SL sentence. Square brackets are used to mark an adjacent sentence.] Comments. The optional like-construction adds a manner element for precision. The N/S/F translations preserve the metaphor although with slightly different means. German uses an idiom. The Norwegian version adds i ørene hans (“in his ears”), which is implicit in the original. The Swedish version supplies a possible missing element (i glas “in glass”), which turns the whole thing into an analogy. The Finnish version is nearest to the original. From a grammatical point of view, the interesting thing about (1) is that a material process verb cut into is used to convey intense perception (hearing). There is a similar example in: (2) Regret stabbed in again, a needle of grief. (DF1.2.s318) N Det meldte seg et nytt stikk av sorg. S Saknaden högg tag i mig igen och sorgen stack som nålar.4 F Kaipaus iski jälleen, murheen piikki.
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.4 (112)
Kay Wikberg
where the instrument needle is mentioned as well and a stronger but nearsynonymous word grief is used for regret. Although the Swedish version has two coordinated clauses, including a simile (som nålar = “like needles”), we can still say that the original metaphor is maintained. Like stabbed, the Finnish verb iski is intransitive, and the final noun phrase is an accurate translation. What (1) and (2) illustrate is the concretizing effect of metaphors: abstract and impalpable (2) domains can be expressed in terms of the concrete. Cut into and stab in their figurative senses are well established and found in dictionaries. On these grounds alone, then, (1) and (2) could be labelled ‘inactive’ metaphors (Goatly 1987), but the total impact of these metaphors in their co-texts is such that they deserve to be included as active metaphors.
Methods The study of metaphor in a contrastive perspective raises a number of theoretical and methodological problems. We shall begin with the role of the computer and then examine some theoretical issues. There are two possible approaches to the study of metaphor using the computer. I have already mentioned the first one, which presupposes that we already know what we are looking for, e.g. an item or construction which is a recognized figurative expression. Let us call this:
The item approach The item approach involves: – – –
searching for specific lexical items, e.g. words for body parts, animals, colours, etc in the SL (cf. Wilkinson 1993; Deignan 1995); examining their non-literal senses and the ways in which these have been used and translated into the TLs; searching for constructions signalled by special linguistic markers (like, as if (though), as it were, sort of,...) and their translations.
Since we know that many words with referents like those just mentioned are used extensively across languages, they are useful for contrastive purposes as well. Lakoff and Johnson’s (1980) conceptual approach is less suitable as a starting point since there may be hundreds of linguistic manifestations of each semantic domain. Kövecses (1991) points out quite correctly that one has to examine the conventionalized linguistic expressions that are related to a given semantic domain to find the metaphors. However, conceptual metaphors can often be referred to as a way of making sense of what otherwise may give the impression of being just disparate
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.5 (113)
Studying metaphors using a multilingual corpus
entities. For contrastive purposes they can also be used to illustrate that, even if the SL and the TL share the same conceptual metaphor, there may be differences in the way in which its elements are made use of in surface structure. The Lakoff and Johnson approach has actually been applied successfully to Finnish in Marjatta Lehtinen’s (1997) study of how English expressions of containment are translated into Finnish. Not surprisingly, it is the inessive case that predominates, which applies to temporal concepts as well. Such studies where it is possible to trace a relationship between a conceptual or cognitive category and structural properties which characterize a whole language are of course particularly interesting. The item approach, then, works nicely if we are looking for lexicalized metaphors and idioms. But we cannot possibly get at innovative metaphors in this way, as Deignan (1999) has pointed out. For this reason the approach adopted here was to start out from the actual texts, not from a list of ready-made items or images. Let us call this the textual approach.
The textual approach The tasks I set myself were: – – – – –
identify the metaphors by reading the texts and then find their translations using the alignment programme; transfer the examples to a database; compare the ST with the TL translations; analyse the linguistic structure of the metaphors; examine the metaphors in their contexts, e.g. find out if they are isolated phenomena, if there are other figurative devices in the immediate co-text which have a reinforcing effect, if the metaphors are repeated or extended.
Theoretical comments Anybody doing research on metaphor will have to make up his mind about what a metaphor is. Since ‘metaphoricity’ or ‘metaphorhood’ is a scalar concept, there will always be some difference of opinion and some subjectivity in deciding on where to place a given metaphor along such a scale. George Miller describes a metaphor as “a comparison statement with parts left out” (1993: 379). By that he also means that a metaphor can be seen as a reduced simile with like left out, a view which not everybody would accept. If we adopt Richards’s (1936) well known terms, Topic and Vehicle, the comparison can be expressed in the following way, using Miller’s formalization:
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.6 (114)
Kay Wikberg SIM [F (x), G (y)] Topic Vehicle
Figure 1. A similarity relation between Topic and Vehicle
sim stands for the relation of similarity between the two (sentential) concepts F(x) and G(y). The symbolisation implies that the concepts can be used to express relations between a predicate and an argument. In a simile, sim is typically expressed through a copula (is like). This is the simplest of Miller’s rules, but as it stands, it already captures some of the gist of metaphor and can be used in many cases as a peg to hang other things on. Steen (1999) uses it and elaborates on it in a paper where he puts forward a checklist for metaphor research. The missing element in Figure 1 is of course the Ground, which hinges on the matching of the relevant features shared by the two concepts within square brackets. Provided the metaphor is a straightforward one, the normal thing is for the Vehicle to give the new information by providing a bridge over Ground to the Topic. Psychological experiments show that people’s perceptions of the topics in metaphors change significantly more than their perceptions of the vehicles (Gibbs 1994: 239). This suggests that the amount of new information is higher in the Vehicle. If this is so, it must make a difference whether the Vehicle consists of just a noun phrase or, say, a verb + complementation. The question of information value is closely connected with degrees of metaphoricity. Whereas Black (1993) questions the feasibility of establishing reliable criteria for determining this notion, Cameron (1999) has recently tried to pin down such criteria. Goatly (1997) also presents ‘clines of metaphoricity’ where ‘approximate versus distant similarity’ seems to be a basic one. According to Cameron, for a metaphor to be considered ‘active’ or ‘original’, it would have to express high degrees of (here slightly modified): – – –
–
incongruity between Topic and Vehicle; vitality, i.e. form a novel link between Topic and Vehicle; cognitive demand: a more complex dimension having to do with both Topic and Vehicle, their abstraction, generality and complexity; connotative power in the Vehicle.
These are by no means clear-cut concepts but are welcome as a list to refer to in further research. Incongruity is a sort of lack of fit or semantic mismatch. Richards talked about ‘tension’, which could be the consequence of such incongruity. The second notion, ‘vitality’, is related to frequency, the fact that the Topic-Vehicle link-
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.7 (115)
Studying metaphors using a multilingual corpus
age is an unusual constellation. Cognitive demand can be traced to Relevance Theory (Sperber & Wilson 1986), according to which complex and abstract metaphors may require additional cognitive effort to be understood. This might be true of poems by Milton and Shelley which have a very high density of metaphors, but as far as spoken discourse is concerned, according to Gibbs (1994: 232), “the psychological research . . . clearly shows that listeners do not ordinarily devote extra processing resources to understanding metaphors compared with more literal utterances.” Finally, connotative power is connected with “culturally shared associations” elicited by an item or an image. Thus in this sentence: (3) It was like shaking hands with a Playtex rubber glove.
(SG1.3.s181)
it is not surprising that the tradename Playtex is omitted in translation since it lacks the connotative power it may have in an Anglo-American context. Another thing is that if somebody’s hands feel like rubber gloves, it does not matter much where the gloves come from anyway. Vehicles vary a great deal in terms of incongruity in relation to the Topic or in relation to the wider context. A polysemous word used figuratively (June was a whirlwind) would have a low degree. An example of a high degree of incongruity can be found in a passage where a character’s “theory of life” is introduced: (4) Life is like invading Russia (JB1.1.s336). A blitz start, massed shakos, plumes dancing like a flustered henhouse; a period of svelte progress recorded in ebullient despatches as the enemy falls back; then the beginning of a long, morale-sapping trudge with rations getting shorter and the first snowflakes upon your face. The enemy burns Moscow and you yield to General January, whose fingernails are very icicles. Bitter retreat. Harrying Cossacks. Eventually you fall beneath a boy-gunner’s grapeshot while crossing some Polish river not even marked on your general’s map. Comments. Unlike (3), the opening sentence in (4) is a generic sentence which requires further development. As it stands, Life is like invading Russia is a simple comparison. It is just that the complement contains a proposition and that the image continues over a whole paragraph – with embedded similes and metaphors in turn. What we are dealing with here is a complete scenario taken from a description of Napoleon’s campaign against Russia. This is then an example of an extended metaphor. It illustrates the fact that copula constructions can be much more complex and cognitively richer than those normally discussed in the literature, i.e. A is B (Man is a wolf ), a formula which may have been reinforced by the way in which metaphorical concepts are represented in the Lakoff and Johnson tradition (cf. love is war).
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.8 (116)
Kay Wikberg
Approaches to metaphor in translation We shall now switch to the translation aspect of metaphor. This is best done by analysing further examples and discussing some relevant theoretical problems in the process. It is hardly surprising that, when Cameron’s categories apply to the full, i.e. when the Vehicle term refers to a very unusual concept, representing a high degree of incongruity and vitality and with considerable connotative power, this may cause difficulty to the translator. Surprisingly, there are very few clear-cut instances in the present data of such a metaphor. (5) is a case where the Swedish translator gave up for no apparent reason and where Finnish and German have paraphrases. The passage is a description of a ghost or a similar creature in a cupboard: (5) his breath a thin winter-whistle in his throat N Pusten lød som en hes fløyte i halsen hans S untranslated F hengitys takellellen G er wagte kaum zu atmen
(SK1.1.s26)
The impact of this nonce metaphor can be contrasted with what the German translation means literally: “He hardly dared to breathe.” There is a wide gulf between such a paraphrase and the original. By contrast, the Norwegian paraphrase renders the original excellently (“The (his) breath sounded like a hoarse flute in his throat”). The Finnish takellellen sounds almost as if the ghost were stammering. The example supports the view that metaphors in fiction are used for emphasis, to express evaluation, and to evoke images which stir the reader’s imagination. The examples given so far have already indicated some issues that crop up in translation. Originally I intended to organize my database according to a suggestion by Dobrzynska (1995): – – – –
M→M M1 → M2 M → Paraphrase M→Ø
identical images in SL and TL different metaphors SL metaphor not translated
However, it turned out that there were rather few translations that did not belong to the first category, and therefore this classification was not particularly useful for my purposes. Another thing was that even if a SL metaphor was rendered as an isomorphic metaphor in the TL, there could still be semantic differences or differences in the realization of the conceptual metaphor (see example (10) below), which meant that the translation was not completely equivalent to the English original. The use of a paraphrase for a metaphor is not an obvious choice either since paraphrasability may not even be possible.
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.9 (117)
Studying metaphors using a multilingual corpus
Dagut criticizes those who argue that “metaphor is no problem for translation” and stresses the fact that translatability depends on: (1) the particular cultural experiences and semantic associations exploited by [any given SL metaphor], and (2) the extent to which these can, or cannot, be reproduced non-anonomously in TL, depending on the degree of ‘overlap’ in each particular case. (1976: 32)
Dagut here points at a major problem, of which there was a lot of evidence in my data. He also discusses different categories of metaphor, focusing on metaphors which gradually enter the lexicon and contribute to semantic expansion. A current example from my data would be the metonym topless holidays, which has clearly extended its original meaning as described even in the most up-to-date dictionaries. Surprisingly, neither the Swedish nor the Finnish translator coped with this example satisfactorily. The Swedish translator dropped it altogether, whereas the Finnish version used the word nakuloma (lit. “naked holiday”), which makes one think of ‘nudism’. German, finally, had Oben-ohne-Ferien.
Analysis The most interesting findings have to do with the structure of metaphors and their textual use. In these fiction texts 25% of the metaphors were expressed through similes (like- or as if/though-constructions), or metaphor in combination with similes. (6) What we share, Jon and I, may be a lot like a traffic accident, but we do share it. (MA1.2.1s116) We are survivors, of each other. We have been shark to one another, but also lifeboat. N Vi er overlevende etter hverandre. Vi har vært hai for hverandre, men også livbåt. S Vi är överlevande, från varandra. Vi har varit haj för varandra, men också livbåt. F Me olemme eloonjääneitä, toistemme haikaloja, mutta myös pelastusveneitä. Comments. This is an extended image made up of a combination of metaphor and similes. First of all, there is the traffic accident [dangerous event] which ‘Jon and I’ have both survived, but they have also survived each other in spite of having behaved like sharks to each other. Shark stands for [aggressive] and [dangerous], but an important point about the interpretation of shark here is that although they had been shark to one another, they could have been so to different degrees: one could have been more shark than the other. From a linguistic point of view shark and lifeboat are interesting singular uses of the nouns, used in Swedish and Norwegian
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.10 (118)
Kay Wikberg
as well, whereas the Finnish translation is more logical in having the plural. A special feature that applies to idioms in Finnish is that nouns are often in the plural form (ottaa loparit plur < F lopputili “final payment”, antaa potkut plur “gives by the sack”) compared with English (Niemi 1998). (7) shows that metaphor (humanizing), idiom (E bear a grudge – S hysa agg till, F kantaa kaunaa) and simile (E as though – S som om – F kuin) can profitably be combined and should therefore – from a textual point of view – be considered together: (7) [It was dusk when the furniture van drew up outside Number Nine Hellebore Close. The Queen looked stonily at her new home.] The house looked grimly back through the gloom, as though it bore a grudge. (ST1.1.3.s3) N Huset stirret bryskt tilbake gjennom tussmørket, som om det nærte uvilje mot henne. S Huset stirrade bistert tillbaka genom dunklet, som om det hyste agg till henne. F Talo mulkoili takaisin tuimana ja synkkänä coordinated adj phrase, essive kuin olisi kantanut kaunaa. Comments. The need for a textual perspective is also evident if we look at the immediately preceding co-text. Thus gloom is a stronger version of dusk but could also allude to the Queen’s mood. The as though-clause with its idiom reinforces the impact of the main clause at the same time as it extends the humanizing metaphor. Grudge is a noun that only seems to survive in this idiom. The Finnish conditional kuin olisi kantanut renders the hypothetical meaning of the subclause. Whereas the Swedish version is a literal rendering, Finnish uses an adjective (synkkänä) for through the gloom and thus assigns the quality of ‘gloom’ to the house. (8) is another example which has to be seen in its contextual framework: (8) [A representative of the London office of the Bank of Tokyo said yesterday:] “The pound is a goldfish swimming in a tank of piranhas.” (ST1.1.9.s179) N “pundet svømmer som en gullfisk i et basseng med pirajaer.” S “Pundet är en guldfisk som simmar i en tank med pirayor.” F “Punta on kultakala, joka ui piraija-akvaariossa.” Comments. The quotation is part of a longer passage about the state of the pound containing several metaphors such as POUND SAVAGED (“criticized brutally”) (pound) suffering a brutal attack a savage beating
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.11 (119)
Studying metaphors using a multilingual corpus
Both translations preserve the metaphor although with slightly different linguistic means: the postmodifying ing-clause is translated into a relative clause, which is a regular change when translating from English into both Swedish and Norwegian. In Finnish a premodifier is possible but would be stylistically awkward (piraijaakvaariossa uiva kultakala “in a piranha acquarium swimming goldfish”) and a relative clause is therefore used here as well. Another fact about Finnish is the compound piraija-akvaario; owing to the lack of prepositions, prepositional phrases have to be rendered by other means, as in this instance from Atwood’s Cat’s Eye: (9) [we are a far cry from picket fences and white curtains,] here in our lagoon of postwar mud. (MA1.2.4.s71) N her i vår lagune av etterkrigssøle. S här i vår lagun av efterkrigslera. F täällä sodanjälkeisessä kurapoukamassamme. Comments. The girl whose voice we are hearing is describing her new home in Toronto. The house, still in an awful state of disrepair, stands in a muddy part of the city (mud has just been mentioned a number of times), and the period is immediately after WWII. Lagoon, the Vehicle, is the incongruous element. Since we tend to associate a lagoon with positive features like [peaceful], [calm], [blue], and [romantic] there seems to be an ironic element here; it sounds as if the house had got stuck in the mud. Apart from the native word poukama, Finnish has the loanword laguuni, but the compound kuralaguuni (lit. “dirt lagoon”) could hardly be used seriously. There are many examples of metaphors like this forming networks of related images which stretch over paragraphs or long chunks of text. For translation research this obviously raises the question of maintaining coherence but also the problem of dealing with lexical items, idioms and metaphors which represent in part different associative networks in the SL and the TL. There are examples where the elements of a conceptual metaphor are realized by slightly different means in the TL. An example of such a metaphor is words = liquid, as in (10) his talk washes over her, a broad and placid stream with occasional eddies of humour or gentle cynicism. (ABR1.1.1.s498) N Lekende lett skyller praten hans inn over henne, en bred og rolig strøm der det av og til er hvirvler av humor eller mild kynisme. S Lätt och ledigt sköljer hans ord över henne i en bred och maklig flod med virvlar av humor eller mild cynism här och där. F Paulin puhe huuhtoutuu hänen ylitseen vaivattomasti, leveänä ja levollisena virtana, jonka pinnan obj (“whose surface”) jokin leikkisä tai lempeän kyyninen adj p pyörre silloin tällöin rikkoo (“breaks”). G Mühelos geht sein Erzählfluß über sie dahin, ein breiter, friedlicher
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.12 (120)
Kay Wikberg
Strom, nur gelegentlich aufgelockert durch Strudel von Humor und sanftem Zynismus. Finnish is unique in spelling out the surface element of the metaphor, which is a way of coping with the long prepositional phrase: washes stream eddies [surface (of stream)] huuhtoutuu virta pyörre sing pinnan accusative At the lexical level there are changes of word class as well (humourN → leikkisäAdj “playful, humorous”; cynicismN → kyyninenAdj “cynical”). Finally, if we leave innovative metaphors aside for a moment, Newmark (1985: 305) mentions the difficulty of reproducing “one-word metaphors where the sense is an event or quality rather than an entity.” His example is elbow one’s way, which, according to him, cannot be reproduced “into any foreign language, unless the metaphor is literalized.” Although this particular item can be rendered by equivalent means in Swedish (armbåga sig fram), denominal, lexicalized verbs of this kind were found in the corpus and certainly provide support for Newmark’s observation: (11) and, breasting a small ridge, he saw. . . (PDJ3.1.4.s10) N Da han kom over en liten bakkekam S och när han körde över ett lågt backkrön (“when he was driving over”) F Noustuaan pienelle harjanteelle (“when he had reached”) G Als er über eine kleine Anhöhe fuhr (12) As the Jaguar crested the headland (PDJ3.1.6.s1) N Da Jaguaren kom kjørende over toppen av neset (“came running over”) S Samtidigt som Jaguaren nådde (“reached”) uddens högsta punkt F Samalla kun Jaguar nousi (“reached”) niemen laelle allative G Als der Jaguar die Hügelkuppe überquerte (13) So I Nureyeved the front steps and flowed through the door (JB1.3.s288) N Så nureyevet jeg opp yttertrappen og fløy inn S Varför jag tog yttertrappan i ett Nurejevskt språng prep phrase F Minä tanssahtelin ulkoportaat ylös (“danced lightly”) G Daher nurejewte ich über die Vortreppe und glitt Comments. According to The New Oxford Dictionary of English, breast V and crest V are synonymous here (“to reach the top of ”) but since breast expresses [instrument] and crest [top of], the two verbs are not always completely equivalent (cf. OED: crest = 3. To reach the crest or summit of (a hill, rising ground, wave, etc.); breast = 1. trans. To apply or oppose the breast to (waves, wind, a steep ascent); to stem, face, meet in full opposition. to breast a fence, horse, etc.:
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.13 (121)
Studying metaphors using a multilingual corpus
to mount by springing so as to bring the breast over.) Thus, crested can hardly be used in I breasted the dark like a swimmer (bnc-fu5 1331) where breast is a verb of movement. With its abstract object breast is more metaphorical in this use. The most metaphorical of these denominal verbs is undoubtedly Nureyeved (“danced/moved like Nureyev up”), not found in the OED, but rendered in both Norwegian and German by equivalent verbs. I certainly find the Norwegian innovation more marked then the original. Thanks to its rich morphology Finnish has a way of getting around the problem of conveying the cultural allusion, instead stressing the repeated light steps of the dancing movement, which the other languages cannot convey by lexical means. What these examples indicate, though, is that semantically rich denominal verbs as in (11) and (12) are difficult to render into other languages. However, this does not exclude compensation using other linguistic means in the TL, as in (13).
Conclusion The data in this investigation show that innovative metaphors caused relatively little trouble to the translators. Such metaphors are expressed using established words but combining them in novel ways, or as Miller (1993: 397) puts it: In novel metaphors, it is not the meanings of the words that change, but rather our beliefs and feelings about the things that the words refer to.
Thus, it is not very surprising that the metaphors in this corpus are generally rendered isomorphically. From a contrastive point of view the differences that were found were mostly due to what one might expect from the structural contrasts between the three languages. Owing to the distance between English and Finnish, Finnish as a TL displays more differences than Swedish in the ways it realizes conceptual metaphors. The contrasts that could be found between the SL and the TLs and that were not due to the structural constraints of the TLs mostly had to do with the lexical level and were similar to other instances of information loss or explicitation found in translations in general. The contrasts could also have to do with variations in metaphorical structure which are possible without generating an entirely new metaphor in the TL. As we all know, novelists deal with themes of general human interest. This means that they make use of a vast range of different metaphors to achieve poetic effects, to enliven the text, and to describe elements of their textual worlds which would not have been possible with straightforward literal language (‘the inexpressibility hypothesis’ (Paivio & Walsh 1993: 309)). It follows that the vast majority of the innovative metaphors in these fiction texts are not used for plot-advancing pur-
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.14 (122)
Kay Wikberg
poses. Instead they have a number of other functions, many of which are attitudinal and interpersonal in character.
Notes . I am grateful to the Department of English at Zurich University for allowing me to use their BNC web query system. . Website http://spraakbanken.gu.se/lb/konk . A description of the corpus is found at http://www.hf.uio.no/iba/prosjekt . A more correct Swedish translation would have been ‘högg till i mig’.
References Black, M. (1993). More about metaphor. In Ortony (Ed.), (pp. 19–41). Cacciari, C. (1998). Why do we speak metaphorically? Reflections on the functions of metaphor in discourse and reasoning. In A. N. Katz, C. Cacciari, R. W. Gibbs, & M. Turner, Figurative Language and Thought (pp. 119–157). New York & Oxford: Oxford University Press. Cameron, L., & Low, G. (Eds.). (1999). Researching and Applying Metaphor. Cambridge: Cambridge University Press. Cameron, L. (1999). Identifying and describing metaphor in spoken discourse data. In Cameron & Low (Eds.), (pp. 105–132). Dagut, M. B. (1976). Can ‘metaphor’ be translated? Babel, 22(1), 21–33. Deignan, A. (1995). Collins Cobuild English Guides 7: Metaphor. London: HarperCollins. Deignan, A. (1999). Corpus-based research into metaphor. In Cameron & Low (Eds.), (pp. 177–199). Dobrzynska, T. (1995). Translating metaphor: Problems of meaning. JoP, 24, 595–604. Gibbs, R. (1994). The Poetics of Mind: Figurative Thought, Language and Understanding. Cambridge: Cambridge University Press. Goatly, A. (1997). The Language of Metaphors. London and New York: Routledge. Kövecses, Z. (1991). Happiness, A definitional effort. Metaphor and Symbolic Activity, 6(1), 29–46. Lakoff, G., & Johnson, M. (1980). Metaphors We Live By. Chicago & London: The University of Chicago Press. Lehtinen, M. (1997). Differences in imagery: translating English expressions of containment into Finnish. In Kinga Klaudy & János Kohn (Eds.), Transferre Necesse Est (pp. 257–262). Budapest: Scholastica. Low, G. D. (1988). On teaching metaphor. Applied Linguistics, 9(2), 125–147. Miller, G. (1993). Images and models, similes and metaphors. In Ortony (Ed.), (pp. 357– 400). The New Oxford Dictionary of English. (1998). Oxford: Clarendon Press.
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS24.tex / p.15 (123)
Studying metaphors using a multilingual corpus
Newmark, P. (1985). The translation of metaphor. In W. Paprotté & René Dirven (Eds.), The Ubiquity of Metaphor. Metaphor in Language and Thought (pp. 295–326). Amsterdam/Philadelphia: John Benjamins. Niemi, J., Nenonen, M., & Penttilä, E. (1998). Number as a marker of idiomaticity. In T. Haukioja (Ed.), Papers from the 16th Scandinavian Conference of Linguistics. Turun yliopiston suomalaisen ja yleisen kielitieteen julkaisuja 60 (pp. 293–304). Turku: Turun yliopisto. OED. (1994). The Oxford English Dictionary on CD ROM. 2nd ed. Ortony, A. (Ed.). (1993). Metaphor and Thought (2nd ed.). Cambridge: Cambridge University Press. Paivio, A., & Walsh, M. (1993). Psychological processes in metaphor comprehension. In Ortony (Ed.), (pp. 307–328). Richards, I. A. (1936). The Philosophy of Rhetoric. London: Oxford University Press. Sperber, D., & Wilson, D. (1986). Relevance: Communication and Cognition. Oxford: Blackwell. Steen, G. (1994). Understanding Metaphor in Literature. London & New York: Longman. Steen, G. (1999). Metaphor and discourse: Towards a linguistic checklist for metaphor analysis. In Cameron & Low (Eds.), (pp. 81–104). Wilkinson, P. R. (1993). Thesaurus of Traditional English Metaphors. London & New York: Routledge.
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.1 (125)
Cross-language metaphors Conceptual or pragmatic variation? Andreas Musolff Department of German, University of Durham
Metaphors play a central role in public discourse, concretising complex political processes in the form of stereotypical schemas that shape the media agenda. Cognitive theory, as developed by George Lakoff and Mark Johnson, provides a framework to analyse the conceptual aspect of imagery by analysing metaphor as a mapping from a conceptual “source domain” onto a “target domain”. However, this model does not account for the argumentative function of metaphors, which is of crucial importance as regards their use in public discourse. This paper analyses a sample of media texts, in which family imagery is employed to argue EU policy issues. It concludes that the cognitive approach needs to be complemented by a pragmatics-oriented perspective, which views the “source domain” as a set of (context-specific) presuppositions rather than logically implied “entailments” (Lakoff).
.
Public discourse metaphors and cognitivist theory
The use of metaphors in public discourse has troubled philosophers of language and semanticists for some time; one of the most famous – and most criticised – comments, for instance, is John Locke’s condemnation of non-literal language use in his Essay Concerning Human Understanding: (1) (. . . ) if we would speak of Things as they are, we must allow, that all the Art of Rhetorick (. . . ), all the artificial and figurative application of Words Eloquence hath invented, are for nothing else but to insinuate wrong Ideas, move the Passions, and thereby mislead the Judgement; and so indeed are a perfect cheat: (. . . ) they are certainly, in all Discourses that pretend to inform or instruct, wholly to be avoided. (Locke 1979: 508)1 More recently, George Lakoff (1996) has claimed that the ‘prototypical’ imageschema of the family governs the conceptualisation of social and moral problems
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.2 (126)
Andreas Musolff
in the United States of America. According to Lakoff, the nation-as-family schema portrays the government as a parent and the citizens as children and thus “allows us to reason about the nation on the basis of what we know about a family” (Lakoff 1996: 155). There are two competing basic versions of the central metaphor: the Strict Father model and the Nurturant Parent model, which “induce” two corresponding “unconscious” patterns of moral belief systems, which, in turn, yield conservative and liberal world views (ibid.: 37, 155). This interpretation of metaphor as a basis for political worldviews is based on one of the most fundamental tenets of cognitivist metaphor theory, namely that metaphorical language is only “a reflection of metaphorical thought” (Lakoff & Johnson 1999: 123), or, in the words of an earlier publication by Lakoff, “the locus of metaphor is not in language at all, but in the way we conceptualize one mental domain in terms of another” (Lakoff 1993: 203). In the same publication from 1993, Lakoff explains that whilst linguistic metaphor use may vary in accordance with context conditions, the underlying conceptual mappings are governed by an “Invariance Principle” which stipulates that “mappings preserve the cognitive topology (that is, the image-schema structure) of the source domain, in a way consistent with the inherent structure of the target domain” (ibid.: 215). Furthermore, the use of metaphors not only reflects certain ‘underlying’ concepts that govern the language users’ understanding of the world, but it also “inherits” a host of further, hidden schemas and metaphorical mappings implicit in the source-target domain mapping (ibid.: 222–225). The metaphorical mapping is thus thought to select correspondences that are already inherent in the (pre-existing) source and target domains; the construction and usage of metaphors is therefore restricted to choosing from a set of pre-established correspondences (ibid.: 210). At first sight, this strong correspondence thesis seems to be weakened in Lakoff and Johnson (1999: 71), when the authors acknowledge that “rich” domains of experience require “more than one metaphorical mapping”. Thus, one might think, the deterministic “preservation” of cognitive “topologies” can be broken up: one target domain can have various source domains and, accordingly, various correspondences that allow different mappings. However, unless one wants to allow for contradictory target domain concepts, the source domains must still be sufficiently similar to provide correspondences for one and the same target domain, which leaves only little room for variation. On account of the strong link that cognitivism assumes between metaphor and thought, it seems to fit the task of analysing metaphors in public discourse particularly well. Furthermore, Lakoff ’s and Johnson’s strong hypotheses concerning the structure of the conceptual “mappings” that are at the core of every metaphor use, should allow testing against empirical data. In his 1996 book mentioned above, Lakoff indeed tries to demonstrate this link by indicating the ideological implications of the family metaphor; however, the evidence from discourse data that he
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.3 (127)
Cross-language metaphors
provides is limited. It consists mainly of a short list of idiomatic phrases, such as: founding fathers, father of his country, Uncle Sam, Big Brother, fatherland, its sons going to war (Lakoff 1996: 153–156); these are supplemented by quotations from popular writings on political and educational issues as well as research literature on socialisation research, political theory and public administration. These treatises bear out his statements regarding the ideological divide over moral issues, which can be linked to the Strict Father model and the Nurturant Parent attitudes in education; but these sources hardly count as evidence of an unconscious conceptual framework based on metaphors. Far from being unconscious or hidden, they are fully explicit and elaborate and do not contain any conspicuous imagery. In order to prove or demonstrate that the family metaphor underlies public perception and understanding of political issues, Lakoff ’s argument would need to be supported by empirical data of language use which show that the posited metaphorical schemas are indeed representative for the public debate. It is here that corpus-based studies would seem to be able to provide an empirical confirmation or test-case for the hypotheses of cognitive analysis. The following comparison of family imagery in British and German public discourse aims at contributing to the exploration of the feasibility of such a corpus-based approach in metaphor analysis.
. EU family imagery in British and German media Family metaphors are by no means a special feature of US political discourse but, as the etymologies of terms such as patriotism or fatherland show, they belong to a common stock of political metaphors, which have been used in Western culture since antiquity. It is thus not surprising to find them in public debates about European politics, for which a corpus has been compiled as part of a collaborative research project on linguistic manifestations of attitudes towards Europe in Britain and Germany.2 The metaphor corpus contains 2132 entries of passages (amounting to 500,000 words), drawn from 28 British and German broad sheet newspapers and magazines published during the 1990s.3 Within the corpus, seven thematic domains of metaphorical meaning have been identified on the basis of recurrence, frequency and explicit comment: 1. general transport; 2. specific modes of travel; 3. geometric and architectural structures; 4. social groupings (which contain family and love metaphors as sub-domains); 5. life, birth and health, strength and size; 6. competition, sports and war; 7. show and theatre. Some of these metaphor themes of Euro-discourse have been the object of previous case-studies;4 however, most of these analyses focused on individual prominent metaphors and their international echo (e.g. Gorbachev’s slogan of the Common European House) rather than on dis-
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.4 (128)
Andreas Musolff
tribution patterns in the use of metaphors belonging to specific thematic domains across different languages. The corpus contains 88 metaphors whose “source domain” includes notions of family, marriage and love, and whose ‘target domain’ concerns political issues of the European Union. Only a few of these texts can be related to the issues of family morality or family ethos (which Lakoff concentrates on in his 1996 book), namely passages depicting individual nations as ‘problem’ or ‘lost’ children that might or should return into the fold of the EU family: (2) Die Slowakei bleibt das Sorgenkind der europäischen Familie. (Die Welt, 13 March 1998) [Slovakia remains the problem child of the European family.] (3) Britain’s European Union partners yesterday feted the new government’s return to Brussels with a warmth which would not have disgraced the biblical welcome accorded the prodigal son. (The Guardian, 6 May 1997) (4) We must offer all Balkan nations, including Serbia, a place in the family of Europe. (The Independent, 21 May 1999) [author: the then leader of the British Liberal Democrats, Paddy Ashdown]. In these text passages, norms of solidarity and mutual obligations within the European family are invoked to argue about the treatment of specific countries that are seen as individual children. There are seven such examples altogether in the corpus. By contrast, there are 22 child-parent metaphors that refer to the new common currency, the euro, as the baby of EU-politics in the 1990s, as in the following examples: (5) In the long gestation of Europe’s Economic and Monetary Union – conceived in Maastricht 1991, to be delivered in Frankfurt 1999 – it suddenly seems likely this week that the anxious parents, Germany and France, are expecting a soft baby euro. The pangs of pregnancy have never coincided so painfully in both countries. (The Guardian, 30 May 1997) (6) Der Euro wird zunehmend zum ungeliebten Kind. Keiner will sich mehr mit ihm sehen lassen. (Die Welt am Sonntag, 30 April 2000) [The euro is fast becoming an unwanted child. Nobody wants to be seen with it.] The euro-child is mostly depicted as being in grave danger. In Germany, the debate focused on a specific variant of the metaphor, i.e. that of a premature, and thus dangerous birth of the euro-child. The most prominent and successful user of the premature birth argument was the opposition’s chief contender for the gen-
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.5 (129)
Cross-language metaphors
eral elections of autumn 1998, the Social Democrat Minister President of Lower Saxony, Gerhard Schröder. Sensing an increase in popular scepticism towards the new currency, Schröder launched his election campaign in March 1998 with an emphatic warning that a hastened introduction of the euro would deliver a sickly, premature baby: (7) Einen Tag, nachdem die Europäische Kommission in Brüssel den Startschuß für den Euro (. . . ) gegeben hat, meldet sich der Kanzlerkandidat Gerhard Schröder: Die Währungsunion komme überhastet und führe zu einer kränkelnden Frühgeburt, moniert er. (Die Welt, 27 March 1998) [One day after the European Commission has given the green light for the euro (. . . ), the candidate for the Chancellorship, Gerhard Schröder decries EMU as coming too soon and producing a sick, prematurely born child.] (8) (. . . ) in Leipzig wiederholte Schröder seine Diagnose einer Frühgeburt, die gepflegt werden müsse. “Das können wir am besten”, sagte er (. . . ). (Süddeutsche Zeitung, 3 April 1998) [ . . . in Leipzig, Schröder reiterated his diagnosis of a premature child that needed extra care. ‘It’s us (= SPD) who are best equipped to provide that care’, he said.] The incumbent Chancellor Kohl noticed the appeal of his opponent’s imagery and, in the context of a debate in parliament, attempted to counter-attack the premature birth-statement by denouncing it as isolating Germany among the euro-fathers of the other (predominantly Social Democrat) EU governments: (9) Im Interview läßt er [= Schröder] wissen, das Einheitsgeld sei “eine kränkelnde Frühgeburt”. Genüßlich verliest Kohl die lange Liste europäischer Sozialdemokraten (. . . ) – alle “Väter dieser angeblichen kränkelnden Frühgeburt”. (Die Zeit, 29 April 1998) [In an interview Schröder announced that the single money was a ‘sickly, premature child’. Now, Kohl delights in reading out a long list of European Social Democrats (. . . ) – all of whom are among the ‘fathers of this supposedly sick child’.] Judging by the election result, however, Kohl’s defence did not convince the voters – after all, Schröder had been careful not to criticise the euro birth in principle but mainly its timing. This left him a chance to turn his sceptic-sounding warning of a premature child into a commitment to provide special care for the baby once he was in charge. When confronted with his premature baby statement in an interview after his election victory, Schröder assumed the role of the responsible father:
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.6 (130)
Andreas Musolff
(10) Spiegel: Sie übernehmen den Vorsitz im Rat der EU in einem Augenblick, in dem das historisch einzigartige Experiment des Euro anläuft – eine “Frühgeburt”, wie Sie meinten. Immer noch skeptisch? Schröder: Wir müssen den Euro zu einem Erfolg machen. (Der Spiegel, 1/1999) [Spiegel: You are taking over the EU presidency just at the start of the historically unique experiment of the euro – a ‘premature child’, as you called it. Are you still sceptical? Schröder: We must make the euro a success.] One striking feature of the euro-family is that (apart from a few sex-neutral references to parents or godparents of the euro)5 it seems to have only fathers as parents. In addition to the notion of multiple fathers (as in example (9) above), individual politicians are praised for their euro-fatherhood, and in the last two years of his Chancellorship, Helmut Kohl, in particular, was promoted to the status of ‘grand old’ patriarch of the EU, as in examples (11) and (12): (11) The father of the euro [= Kohl] is an historic figure (. . . ). (The Guardian, 19 February 1998) (12) Maastricht wird (. . . ) “unumkehrbar”, wie Helmut Kohl so gerne und oft sagt. Von allen Gründervätern, die im Dezember 1991 die Währungsunion aushandelten, hat er als einziger politisch überlebt. Im Kreise der EU-Mächtigen gilt der Deutsche längst als Patriarch, als Pate, als Paterfamilias. (Die Zeit, 29 April 1998) [Maastricht becomes (. . . ) ‘irreversible’, as Helmut Kohl loves to remind us time and again. Of all the founding fathers that negotiated EMU in December 1991, he is the only one who has survived politically. Among EU-leaders, the German chancellor is acknowledged as the patriarch, the godfather, and the head of the family.] In contrast to the surfeit of fathers, there are no references to Euro-motherhood in the corpus – but for one oddly far-fetched example of a reader’s letter to Die Welt, attacking the common currency for depriving people of their mother currency; this is seen as equivalent to stealing their mother tongue: the latter ‘theft’ is assumed to ‘leave them speechless’, the former ‘robs them of a part of their sovereignty’.6 Thus, with the exception of example (5), which has France and Germany as euro-parents (without identifying who is mother or father, though), the ‘typical’ metaphorical European family consists of eleven or fifteen fathers (depending on whether Euroland or the full EU constitutes the family), who have one child, e.g. the euro currency or the EU itself (or its political project of “ever closer integration”). It seems counter-intuitive to assume, following the cognitivist doctrine, that this family sce-
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.7 (131)
Cross-language metaphors
nario is based on some “prototypical” source domain concept that “allows us to reason” about international relationships in Europe “on the basis of what we know about a family” (Lakoff 1996: 155). Rather, the source domain of a family is adapted to fit the target domain concept, in our case a community of eleven or fifteen states. The fact that, according to the formulations in the corpus, these parents are all male does not present in any way a problem for understanding the family metaphor – the journalists who use this image obviously rely on the readers’ acquaintance with (if not, perhaps, acceptance of) its underlying sexist political stereotype.
. Couples in EU debates The multiple euro-parentage is not the only strange feature of the EU family. Another complicating factor is the existence of at least one privileged couple within the family, i.e. France and Germany, as in example (5). Of the altogether 55 metaphorical references to couples/marriage/wedding in the corpus, 25 have the bilateral relations between France and Germany as their target domain. The British press focuses on the marriage troubles of the Franco-German couple, particularly with a view to the possibility of Great Britain replacing one partner or bringing about a ménage à trois. Thus, when Jacques Chirac was elected French president, The Guardian asked rhetorically: (13) The tricolour flies over a new tenant in France’s Elysée Palace today, but what of the eternal triangle at Europe’s core? Will Jacques Chirac be forced by the responsibility of power to go beyond the vagueness of his election statements and declare his true loyalty? Is the “Franco-German couple” really the “heart of the European Union”, as he said during the campaign, or are Britain’s Conservatives right to imagine they have found a fellowsceptic to flirt with? (The Guardian, 18 May 1995) The next occasion for a reconfiguration of national partnerships seemed to present itself when, in June 1997, Gerhard Schröder, in an early attempt to question Helmut Kohl’s Euro-policies, stated that he was looking forward to a future triangle of social democratic governments in Britain, France and Germany, which would make job creation a new priority in the EU. The original quotation, based on the juxtaposition of the geometrical/technological concepts of triangle and axis,7 did not include any marriage imagery; however, this did not prevent the British press from taking the opportunity to imagine a love triangle emerging from the ruins of the broken Franco-German marriage – only to complain later (after Schröder had been elected) that he had ‘reneged on his promise’:
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.8 (132)
Andreas Musolff
(14) Germany’s chancellor-in-waiting seemed to herald a dramatic shift in alliances, arguing that the Franco-German marriage was now over. Gerhard Schröder, leader of the opposition Social Democrats, argued in favour of a menagé [sic] à trois, involving London, Paris and Bonn. (The Independent, 17 June 1997) (15) The Franco-German relationship is no longer what it was. (. . . ) During his election campaign, the French were upset by Mr Schröder’s suggestion that their hitherto exclusive partnership should be opened up to form a ménage à trois with the British. To France’s relief, there has been no further mention of this distinctly touchy issue. Was a political Viagra pill taken this week in Potsdam to give the ageing relationship a new fizz? Maybe, maybe not. Talk of a possible new Paris-Bonn-London triangle in Europe is nothing new: every time, over the past decade, that a new president or prime minister has taken over in France, he briefly – and in the end unsatisfactorily – flirts with the Euro-sceptical British, only to fall back in relief on the old liaison with Germany. (The Economist, 5 December 1998) The last quote in particular demonstrates the argumentative potential of the nations-as-lovers metaphor in Euro-debates. The Franco-German relationship is presented as that of a long-standing couple who have grown out of their romantic phase but on the other hand are still able to put some new fizz into their marriage and have so far overcome the dangers of extramarital flirtations. The Economist presents the newest developments in EU politics as the latest instalment in a longrunning soap opera, with the regular couple, France and Germany, as the audience’s old favourites. This scenario is also present in German marriage metaphors, i.e., despite repeated warnings of its imminent break-up, the Franco-German couple seems to survive its partnership crises. What distinguishes German from British uses – apart from a complete lack of interest in a ménage à trois involving Britain on the part of the German media – are a tendency to personalise the couple by metonymically identifying it with the pairs of post-war national leaders, and a rather blunt explicitness about the marriage partners’ respective roles: (16) Die jüngere deutsch-französische Geschichte ist eine Geschichte von Paarbeziehungen: Charles de Gaulle und Konrad Adenauer, Georges Pompidou und Willy Brandt, Valéry Giscard d’Estaing und Helmut Schmidt, François Mitterrand und Helmut Kohl. (. . . ) Auch die Europäische Union ist ein Erfolg solcher Zweisamkeiten. Doch heute fehlt das Paar an der Spitze (. . . ). (Die Zeit, 14 November 1997) [The recent history of Franco-German relations has been a story of pair-
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.9 (133)
Cross-language metaphors
ings: de Gaulle and Adenauer, Pompidou and Brandt, Giscard d’Estaing and Schmidt, Mitterrand and Kohl. (. . . ) The European Union is also based on the success of these pairs. But these days, there is no such couple of leaders . . . ] (17) Nach dem Prestige-Krieg um die Euro-Bank wird sich Frankreich besinnen müssen. (. . . ) die Bonner müssen den französischen Freunden einschärfen, daß eine Ehe nicht bestehen kann, in der der Schwächere das Regiment zu führen trachtet. (Süddeutsche Zeitung, 8 May 1998) [After the prestige-war over the ECB, France will have to reconsider its position. (. . . ) Bonn must remind the French friends of the fact that a marriage cannot survive in which the weaker partner tries to dominate.] In contrast to the frequent appearances of the Franco-German couple in the corpus, there is just one instance of a British newspaper waxing lyrical about an “AngloGerman love-in” during the first year after the change of government in Britain (The Guardian, 21 March 1998). Apart from this singular reference, no other bilateral love-relationships between member states in the EU family are mentioned. However, there are numerous examples, both in the German and British parts of the corpus, of a love or marriage relationship between one country and the EU as a whole, or between a national currency and the common currency. Significantly, negative scenarios of the marriage being in danger or even being dissolved, or of an end to the honeymoon have a particularly high frequency in the British sample: (18) The pound’s shotgun separation from the exchange rate mechanism is proving painful for both Britain and the rest of Europe. The two-year marriage itself was unhappy (. . . ). As in most marriage breakdowns, there have been faults on both sides. Sterling and the German mark (. . . ) were always going to be uneasy bedfellows in a system that relied for its stability on a single anchor. (The Guardian, 2 March 1993) (19) The reality behind the will-they-won’t-they, pre-nuptial dances among aspirant members of Europe’s monetary union club is that as long as the present economic slowdown doesn’t turn into a full-blown recession the project will probably go ahead. (The Guardian, 27 January 1996) (20) Some sceptics believe, or hope, that the strains imposed by a single currency would be too great, and that those in the first wave, having acted in haste, would be left to regret at leisure. (The Independent, 21 October 1996) (21) The Government’s hopes of a prolonged honeymoon in Europe were dashed last night as the Chancellor, Gordon Brown, fought a losing battle against exclusion from the new single currency policy-setting group. “The
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.10 (134)
Andreas Musolff
euro is a monetary marriage, and in a marriage you do not allow others into the bedroom,” said the French finance minister, Dominique StraussKahn, after a twelve-hour meeting of European finance ministers broke up last night. “Those who share the same money will have more intimate relations than the others.” (The Guardian, 2 December 1997) Here again, ‘ontological correspondences’ between source and target domains must not be interpreted too rigidly. Indeed, the exact nature of the relationships in the love/marriage scenario seems to be shifting: whereas in examples (18) and in the first sentence of (21), the marriage-at-risk scenario concerns the bilateral relationship Britain/Pound Sterling-EU/EMU, the imagery of examples (20), (21) and the French minister’s quotation in (22) refers to a multiple marriage among all euro partners, leaving it open whether they are all getting married to each other or each of them to the euro. Any ‘prototypical’ notions of marriage or divorce in Western culture (which, according to a strong cognitivist theory, would be entailed by the source domain) are flouted – however, the metaphor is by no means difficult to comprehend. On the German side, the issue of a possible postponement of the euro introduction was discussed in terms of a debate on whether the euro-countries needed a period of engagement before exchanging marriage vows. In 1997, the Minister President of Saxony, Kurt Biedenkopf (CDU), stirred a political row by suggesting a five-year period engagement for all euro-candidates so as to avoid marital escapades and the danger of an acrimonious divorce: (22) Biedenkopf (. . . ) plädiert für (. . . ) eine “Verlobung” von fünf Jahren, damit die Staaten beweisen, daß sie es ernst meinen mit der wirtschaftspolitischen Tugend (. . . ) Eine Verlobung, in der man ohne Seitensprünge die Treue beweist, ist besser als eine Heirat unter Kalenderdruck. Denn eine Scheidung kann mörderisch sein, das Klima zwischen den ExPartnern bösartig vergiften. (Süddeutsche Zeitung, 20 September 1997) [Biedenkopf (. . . ) argues in favour of (. . . ) a five year ‘engagement’ period so that member states can prove their economic virtue (. . . ) Such an engagement, in which marital fidelity can be proved by abstention from affairs, is better than a hastily started marriage. For a divorce can be disastrous, poisoning the relationship between the ex-lovers.] Biedenkopf ’s argument rests on the assumption that all euro candidate nations should participate in the engagement. The resulting bizarre scenario of a simultaneous engagement over five years among EU currencies/economies lacks any grounding in prototypical experience or knowledge – just as the image of an elevenfold marriage – but was also easily understood (leading to a series of further engagement-metaphorical arguments in the German press). Again, the source domain is
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.11 (135)
Cross-language metaphors
adapted freely to the intended target meaning: the notion of an engagement seems to lend itself, so to speak, to describing an extension of a decision period, despite the ‘literal’ improbability of the details of such a scenario in the case of EMU. What we find in the corpus data, are nation-specific argumentation trends rather than any “conceptual” or “cognitive” differences. Whilst the British media comment almost triumphantly on apparent marriage problems of the FrancoGerman couple that might lead to a break-up or a gradual cooling down of the partnership, giving Britain the chance to flirt or establish a ménage à trois, the German press sees such developments as a worrying threat to the Union, which has to be combated. As regards the relationship of one’s ‘own’ nation to the EU and the euro, German sources only once mention the possibility of a divorce (i.e. as a horrorscenario, which was to be prevented by way of a long engagement and thorough examination of all euro marriage candidates), whereas metaphors of Britain’s divorce and separation from ERM and the EU, or at least, an end of its honeymoon with Europe appear frequently in the British sample and are complemented by images of the rejected or disappointed lover (i.e. Britain as the flirting partner spurned by France). It is not the source domain of family and love imagery that is different in the two national debates but rather the use that is made of particular source domain aspects and their argumentative exploitation to argue for or against specific target concepts.
. Metaphor and public discourse When looking at a corpus of metaphors rather than a few ‘chosen’ instances that illustrate a preconceived theoretical position, we can see that cognitivist metaphor theory, and in particular Lakoff ’s “Invariance Principle” is faced with the problem of improbable or contradictory source domain constructions. Whereas the metaphors of political thinking in the US (as Lakoff presents them) or the everyday “metaphors we live by” (Lakoff & Johnson 1980) fall into neat, systematic patterns and display an impressive array of “entailments”, the (still relatively small) sample of love/marriage metaphors drawn from media texts dealing with EU politics has shown that the use of its source domain features in public discourse is heavily ‘slanted’ and structured to fit specific communicative (and political) interests in particular discourse communities – so much so that almost any distortion or eccentric configuration seems possible. Similar results can be obtained for the other main metaphor fields of Euro-discourse (cf. Musolff 2000). Although these findings do not invalidate the fundamental cognitivist insight into the conceptual function of metaphor, they put into question the assumption of a strict congruence between the ‘cognitive topologies’ of source and target domains in every metaphor-
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.12 (136)
Andreas Musolff
ical mapping. The appeal of metaphors depicting the euro as a love child of eleven fathers or the Franco-German couple as turning into a ménage à trois with Britain does not lie in “ontological correspondences” of a prototypical marriage or family and EU politics but is related to the pragmatic purposes of the specific texts in which they are used, i.e. in most of our cases, arguments about the introduction of the euro and about the relationships among EU member states. As long as these argumentative purposes are met by a specific metaphorical formulation, the improbability or counter-intuitive status of its source domain schema does not seem to matter much – indeed, the greater the distortion is, the more colourful and persuasive the metaphor seems to become. In order to deal with this empirical aspect of metaphor use, I propose to apply Toulmin’s (1958) model of practical (as opposed to formal) argumentation to the concept of source and target domains. According to Toulmin (1958: 100– 107), arguments derive their power of providing (seemingly) convincing evidence from demonstrating that the correct pathway (metaphorically speaking) has been followed to arrive at a specific conclusion. The crucial step in this argumentative progress is the link between the “data” and the “conclusion”, which is established by an argumentative “warrant” that is backed by further data and arguments, which are specific to a particular domain of discourse and reasoning, i.e. in Toulmin’s terminology, they are “field-dependent” (ibid.: 104).8 Warrants have the task “to register explicitly the legitimacy of the [argumentative] step involved and to refer it back to the larger class of steps whose legitimacy is being presupposed” (ibid.: 100); due to this ‘grounding’ in presupposed steps of argumentation and further backing data, “statements of warrants are (. . . ) hypothetical, bridge-like statements” (ibid.: 105). If we conceive of metaphorical formulations as arguments in Toulmin’s sense, the source domain aspects would have to be thought of as some (special) kind of data that are linked – via the metaphorical mapping – to a conclusion concerning the target domain. Such a ‘metaphorical warrant’ only works on the basis of the “field”-specific assumption that the language-users find it at all useful or plausible to take the relevant source domains, e.g. everyday concepts of family or marriage, as starting points for developing models of international relationships. The source domain concept is then not to be understood as a fixed, finite concept with a well-defined number of entailments but as a rather vague presupposition that can be adapted and moulded to fit the conclusion in any way necessary. Such semantic adaptability and flexibility of course rules it out as a logically fully ‘valid’ type of data or information – but who, after all, would (mis-)take talk of a ‘currencymarriage’ to be literally true? The flexibility of a metaphorical warrant is nothing irregular, but is one manifestation (among many) of the hypothetical and fielddependent nature of warrants in practical arguments, which rules out formal, logical ‘entailments’ in Lakoff ’s sense. This concept of a presupposition-based map-
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.13 (137)
Cross-language metaphors
ping process solves the difficulties of cognitivist metaphor theory: the “ontological correspondences” between source and target domains are assumed or suggested correspondences; they are not ‘automatically derived’ from any premises with the certainty of an algorithm. What matters is not whether they are logically correct but whether they can be made use of to ‘back up’ the metaphorical warrant leading to a contextually relevant conclusion that could not be reached without this particular type of source domain “data”. What then, is the “field”-specific argumentative warrant that family/love metaphors provide? In my view, this field of imagery is characterised by an implicit appeal to notions of solidarity, mutual responsibility and emotional closeness which language users tend to associate with (non-metaphorical) family and love relationships; these include the notions about parent-child relationships that Lakoff focuses on in his 1996 book. Whenever politicians and journalists speak about the common currency as a child that has a difficult birth and needs special care, or whenever the marital fidelity between France and Germany or among the EU/Euroland states is disputed, the idea of such a special mutual bond and responsibility is evoked – however, without any logical necessity that would make it an ‘entailment’. Instead, these notions of ‘solidarity’ that are presupposed in metaphors of the EU-as-a-marriage/family/love relationship function as hypothetical suggestions that can be endorsed or rejected, depending on argumentative purposes that the speaker wants to achieve within a particular political context. This shift from a view of metaphor as containing logical “entailments” to a pragmatic emphasis on its function as a presupposition that informs specific argumentative conclusions opens the way to a more plausible interpretation of public discourse metaphors than the sole emphasis on the conceptual role of metaphors. From a cognitivist point of view, the only conclusion to be drawn regarding the use of love, marriage and family metaphors in British and German Euro-debates is that their “source domain” is more or less the same, e.g. the troubled birth of the euro, the troubled marriage between France and Germany, or the engagement, honeymoon and divorce scenarios concerning the euro partners. In an argumentation-oriented analysis, however, typical trends in the two national debates can be identified in the corpus – in our case: a dominance of marriage-/family-sceptical statements on the British side vs. more ‘faithful’ invocations of family bliss on the German side – which can be related to specific political agendas and media interests in the two countries, e.g. the relative weight of eurosceptic and pro-integrationist attitudes towards the EU.
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.14 (138)
Andreas Musolff
Notes . For recent interpretations of Locke’s apparent antipathy to metaphor cf. Bertau (1996: 85– 87); Goatly (1997: 1). . The project has been conducted jointly by the German Department at the University of Durham and the Institut für Deutsche Sprache in Mannheim, Germany, and has been funded by the British Council and the German Academic Exchange Service (DAAD) under the Anglo-German Research Collaboration programme. For research reports cf. Musolff, Good, Wittlinger and Points (2001) and the project web-site “www.dur.ac.uk/SMEL/depts/german/euro-arc/htm”. . The corpus is accessible at the internet web-site “www.dur.ac.uk/SMEL/depts/german/ Arcindex.htm”; for an overview cf. Musolff (2000). . Cf. e.g. Chilton and Ilyin 1993; Chilton and Lakoff 1995; Schäffner 1993, 1996; Musolff 1997; Mautner 1997. . Cf. e.g. quotation (5) above, identifying Germany and France as the euro-parents, and The Guardian, 9 January 1991: “When the ‘hard ecu’, Britain’s alternative to the Delors plan for European Monetary Union, was first presented to its prospective godparents last June it was an unappealingly premature baby. All 11 of them treated the infant with the mixture of embarrassment and derision accorded to nature’s regrettable errors.” . Cf. Die Welt, 19 August 1997: “Wie der Mensch sich als Staatsbürger mit seiner Muttersprache identifiziert, so identifiziert er sich auch mit seiner Mutterwährung. Beraubt man ihn der ersteren, macht man ihn sprachlos, beraubt man ihn der letzteren, so nimmt man ihm einen Teil seiner staatlichen Souveränität”. . Cf. The Independent, 17 June 1997: “Gerhard Schröder, the opposition politician expected to challenge Helmut Kohl in next year’s elections, said: ‘The days when the French President and the German Chancellor could decide everything that went on are over. The FrancoGerman axis must be transformed into a triangle involving London.”’ . As a further complication, “qualifying” specifications and conditions for “rebuttal” are included to determine the precise epistemic value of a conclusion, but these need not concern us here.
References Bertau, M.-C. (1996). Sprachspiel Metapher. Denkweisen und kommunikative Funktion einer rhetorischen Figur. Opladen: Westdeutscher Verlag. Chilton, P., & Ilyin, M. (1993). Metaphor in Political Discourse: The Case of the ‘Common European House’. Discourse and Society, 4 (1), 7–31. Chilton, P., & Lakoff, G. (1995). Foreign Policy by Metaphor. In Ch. Schäffner & A. Wenden (Eds.), Language and Peace (pp. 37–59). Aldershot: Dartmouth. Goatly, A. (1997). The Language of Metaphors. London/New York: Routledge.
TSL[v.20020404] Prn:3/09/2002; 11:47
F: PBNS25.tex / p.15 (139)
Cross-language metaphors
Lakoff, G. (1993). The contemporary theory of metaphor. In A. Ortony (Ed.), Metaphor and Thought (2nd revised edition) (pp. 202–251). Cambridge: Cambridge University Press. Lakoff, G. (1996). Moral Politics: What Conservatives Know That Liberals Don’t. Chicago/London: University of Chicago Press. Lakoff, G., & Johnson, M. (1980). Metaphors we live by. Chicago: University of Chicago Press. Lakoff, G., & Johnson, M. (1999). Philosophy in the Flesh. The embodied Mind and its Challenge to Western Thought. New York: Basic Books. Locke, J. (1979 [1689]). An Essay Concerning Human Understanding. Oxford: Oxford University Press. Mautner, G. (1997). Der britische Europa-Diskurs: Reflexion und Gestaltung in der Tagespresse. Unpublished Habilitationsschrift. Vienna: Wirtschaftsuniversität. Musolff, A. (1997). International metaphors: bridges or walls in international communication? In B. Debatin, T. R. Jackson, & D. Steuer (Eds.), Metaphor and Rational Discourse (pp. 229–237). Tübingen: Niemeyer. Musolff, A. (2000). Mirror Images of Europe. Metaphors in the public debate about Europe in Britain and Germany. Munich: Iudicium. Musolff, A., Good, C., Wittlinger, R., & Points, P. (Eds.). (2001). Attitudes towards Europe – Language in the unification process. Aldershot: Ashgate. Schäffner, Ch. (1993). Die europäische Architektur – Metaphern der Einigung Europas in der deutschen, britischen und amerikanischen Presse. In A. Grewenig (Ed.), Inszenierte Kommunikation (pp. 13–30). Opladen: Westdeutscher Verlag. Schäffner, Ch. (1996). Building a European House? Or at Two Speeds into a Dead End? Metaphors in the Debate on the United Europe. In Musolff, Schäffner, & Townson (Eds.), Conceiving of Europe – Unity in Diversity (pp. 31–59). Aldershot: Dartmouth Publishers. Toulmin, S. (1958). The Uses of Argument. Cambridge: Cambridge University Press.
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.1 (141)
A contrastive cognitive perspective on Malay and English figurative language Jonathan Charteris-Black University of Surrey
The aim of this paper will be to illustrate how cognitive semantics may be used to compare the relationship between thought, symbol and referent in two rather unrelated languages: English and Malay. A comparison of the thoughts underlying figurative phrases will be illustrated with reference to the source domain of a human body part: the foot. I will attempt to compare the evidence found in figurative phrases for the typical thoughts associated with this body part. Thoughts will be described in terms of ‘conceptual keys’ and also in terms of the connotations that are typically evoked by figurative phrases. For this purpose two large corpora are employed. Conceptual keys capture both metaphoric and metonymic conceptualisations and explain the relatedness of various linguistic metaphors and linguistic metonyms. The most effective way of testing the validity of a particular conceptual key is the extent to which it is able to relate a number of figurative phrases to a common idea. To the extent that conceptual keys can do this they are valuable notions for describing, classifying and explaining clusters of figurative language. Evidence for connotations will be provided by examining typical verbal contexts of figurative phrases for the English data and by consulting native Malay speakers for the Malay data.
.
Introduction: Comparative investigation of literal and figurative language
There are two ways of using words and phrases: in a literal and congruent way or in a figurative and incongruent way. Their literal senses are those that normally belong to them; this is usually established with reference to their etymology, frequency or their default sense: i.e. that which they have when they occur out of any context (cf. Cruse 2000: 199–200). However, the creative language user may remove words from their original contexts of use and place them in phrases that are applied to
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.2 (142)
Jonathan Charteris-Black
novel contexts, thereby giving them fresh senses. This is usually what is meant by figurative language. The most important types of figurative language are metaphor (cf. Black 1962) and metonymy (cf. Panther & Radden 1999). We know when comparing languages that there are evidently differences in their grammatical, phonological and lexical patterns; however, an issue that remains unclear is the extent to which there is variation in the underlying concepts that are expressed. Views of language that argue in favour of its role as an encoder of culture emphasise the uniqueness of concepts to a language whereas those that view language as encoding common human experience argue in favour of the universality of the underlying concepts that are expressed. A better understanding of figurative language as can be gained by comparing the figurative expressions of two languages provides an important source of insight into these issues. Figurative phrases in particular provide linguistic evidence of conceptualisation. If figurative language creates a source of tension between the literal and the novel contexts of words, then the extent to which resolution of this tension requires knowledge of the cultural system in addition to knowledge of the linguistic system provides evidence that concepts are dependent on culture as well as on language. I will use the term cultural resonance to refer to the set of beliefs, knowledge and values without which figurative language cannot be fully understood. In this paper I hope to provide some insight into the cultural resonance of some Malay and English figurative phrases in which the human body part ‘foot’ occurs. In other papers I have examined the cultural resonance of other parts of the human body in either Malay (cf. Charteris-Black 2000, on mata ‘eye’), or English (cf. CharterisBlack 2001b, on ‘blood’) or, as here, in both languages (cf. Charteris-Black 2001a, on tangan ‘hand/arm’ and ‘hand’).
. Theoretical context: Cognitive semantics A central tenet of cognitive semantics is that figurative language is motivated in so far as there are relationships that exist between the meanings of language units that could not otherwise be accounted for with reference to the syntax, grammar and lexicon alone. As Croft (1993: 336) notes: ‘One of the central tenets of cognitive semantics is that the meaning of words is encyclopaedic: everything you know about the concept is part of its meaning’. This view has emerged in reaction to what we might call the dictionary view: the view that language exists independently of the contexts in which it is used. The notion of conceptual metaphor has emerged as a theoretical construct that can relate a range of surface forms of metaphor to a common underlying idea or deep level based on experience – thereby showing them to be motivated
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.3 (143)
A cognitive perspective on figurative language
rather than arbitrary (Lakoff & Johnson 1980). A conceptual metaphor is a protometaphor (a term preferred by Goatly 1997) that explains the relatedness of linguistic metaphors – i.e. that they exist as part of a system rather than independently of each other. A conceptual metaphor is a formal statement of the idea that is hidden (though implied) in the linguistic metaphor and can be inferred from a number of linguistic occurrences. It is, then, an explanatory construct as it represents relationships of figurative meaning that would not otherwise be represented. It does this by making a formal statement of the underlying grounds of the metaphor: it therefore articulates the general knowledge structures that underlie linguistic metaphor. The most effective way of testing the validity of a particular conceptual metaphor is the effectiveness with which it is able to relate a number of conventional linguistic metaphors to a common idea. To the extent that conceptual metaphors can do this they are valuable notions for describing, classifying and explaining clusters of figurative language. However, the term ‘conceptual metaphor’ overlooks the fact that figurative language is frequently metonymic as well as metaphoric. The need for a more general notion to describe the conceptual level when comparing figurative expressions in different languages can be illustrated by considering the expressions in example (1): (1) The pot calling the kettle black. The prawn does not know of the hump on its back. The camel cannot see his own hump.
(English) (Malay) (Moroccan Arabic)
In each case these expressions refer to a similar situation: that of when a person criticises others for the same faults that he or she possesses. In the case of the first expression, this is done with the use of an anthropomorphic figure that involves a metonym kitchen artefact for person; in the other two the metonym is animal for person. In all three cases evaluation is based on figures that refer to culturally familiar phenomena. We can summarise the conceptual basis as follows: 1. METONYM: INANIMATE AGENT OR NON-HUMAN FOR HUMAN AGENT 2. METAPHOR: ATTRIBUTES OF AGENT ARE BAD (= NEGATIVE) 3. OUR KNOWLEDGE OF THE INANIMATE OR NON-HUMAN AGENT’S LEVEL OF CONSCIOUSNESS IMPLIES THAT THE HUMAN AGENT IS NOT AWARE THAT HE POSSESSES THESE BAD ATTRIBUTES. Apparently equivalent figurative expressions such as these demonstrate how linguistic encodings reflect cultural knowledge. These expressions are resonant within a culture because they encode particular evaluations that are determined by the value placed on a particular attribute or entity. Understanding the evaluation relies on general cultural knowledge frameworks. For example, interpreting the evaluation in the figurative phrase the pot calling the kettle black presumes a metaphor
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.4 (144)
Jonathan Charteris-Black
BLACK IS BAD; in some cultures black may not be negatively evaluated. Similarly, the other two figurative phrases presume a metaphor: PHYSICAL IRREGULARITY IS BAD – in other cultures physical irregularities (e.g. long necks, or short feet) may be taken as signs of beauty. Interpretation of the above expressions relies on prototype associations; for example, the negative connotation of the Moroccan Arabic expression depends on encyclopaedic knowledge of how camels behave in relation to humans. In the case of the example from English, prior to electric kettles, kettles were similar in appearance and function to pots; for example they shared the same colour because they were heated by fire. However, we can see that to capture the cultural knowledge of these figurative expressions from a comparative point of view we need a notion that incorporates metaphor with metonymy and encyclopaedic knowledge to provide a representation of the conceptual level. For this purpose I propose the use of the term ‘conceptual key’. A conceptual key explains which of the ideas associated with the source are to be transferred to the metaphorical target irrespective of whether these are metaphoric, metonymic or based on encyclopaedic knowledge. In a sense a conceptual key is the reverse of linguistic metaphor or metonym; for whereas figurative language creates an incongruity between the original and novel contexts, between the vehicle and the topic, a conceptual key resolves this incongruity irrespective of the figurative process involved. It does this by making a formal statement of the underlying grounds of the figure: it therefore articulates what in normal language interpretation relies on inferencing by making explicit the general knowledge structures that are implicit in linguistic metaphor and metonymy. Conceptual keys therefore make explicit a systematic set of semantic relationships that occur in figurative language. Clearly, to do this we need representative samples of figurative language and as Ungerer and Schmid (1996: 119) suggest ‘The conclusion from a cognitive perspective is that the metaphors that have been unconsciously built into the language conventions are the most important ones’. A major reason for using large corpora of language to investigate figurative language is, therefore, to establish which are the most important metaphors, metonyms etc. with reference to some criteria other than intuition (Biber et al. 1998; Stubbs 1996). In this study I will use two corpora: the Bank of English and a Malay corpus owned by the Dewan Bahasa dan Pustaka (or Malay language planning agency).
. Methodology This study was based on the analysis of figurative expressions that include the body part term foot in English and kaki ‘foot/leg’ in Malay. All English expressions containing foot were selected from the Cobuild Dictionary of Idioms and all Malay
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.5 (145)
A cognitive perspective on figurative language
expressions containing kaki were selected from the Malay from Hasan Muhamed Ali (1996) and Masri (1997). Occurrences of each expression were then identified in two corpora: the Bank of English held by Collins at the University of Birmingham and a Malay corpus owned by the Malay Language Agency (Dewan Bahasa) in Kuala Lumpur. The English corpus contains 12 separate sub-corpora and 323 million words and the Malay corpus is comprised of 25 million words taken from 3 sub-corpora: books, newspapers and magazines. Conceptual keys were determined by introspection and were based on the criteria of resolving the incongruity of original and novel contexts of use. Connotation was established by comparison of verbal contexts in both the English and the Malay corpus and, in addition, for the Malay phrases, by a questionnaire given to 58 native speakers of Malay. They were asked to classify 25 figurative phrases presented in short contexts as either negative, positive or neutral in their expressive meaning. Resonance was calculated by multiplying the number of types for each conceptual key for which evidence was found in the corpus by the sum of the number of tokens for these types. Each of the figurative phrases that occurred in the corpus and were classified as sharing a conceptual key contributes towards the quantitative measure of conceptual resonance of that key. I propose that this use of a corpus enable us to combine data on both lexical productivity and frequency to produce a measure of resonance that can be used to compare figurative phrases in different languages. Though it should be remembered that variations in the number and frequency of figurative phrases might be part of the overall differences in the lexicons of the languages.
. Findings Table 1 summarises the evidence for the conceptual keys and resonance of figurative expressions in Malay and English for the ‘foot’. In Malay a single conceptual key accounts for over 90% of the resonance of this body part, whereas in English there are three conceptual keys that may be considered important. There is little overlap in the conceptual senses of these figurative expressions in the two languages (much less than is reported for ‘hand’ in Charteris-Black 2001a). There is also little overlap in connotation with the most resonant conceptual key in English having a positive connotation while the most resonant conceptual key in Malay has a negative connotation.
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.6 (146)
Jonathan Charteris-Black
Table 1. Summary of conceptual keys and connotations: Kaki foot Conceptual key
Types
TO BE, TO DO, OR TO BE ADDICTED TO DOING, SOMETHING NOT GOOD BODY PART STANDS FOR ITS FUNCTION METONYM: SIZE STANDS FOR THE OBJECT APPEARANCE BASED SIMILE TOTAL
Sum of tokens
Resonance (Tokens X types)
%
Connotation
5,454
90.4
NEGATIVE
18
MALAY 303
3
97
291
4.8
1
271
271
4.5
NEUTRAL/ NEGATIVE NEUTRAL
1
15
15
0.3
NEUTRAL
22
688
6,031
100
ENGLISH THE FOOT FOR CONTROL LIFE IS A JOURNEY THE FOOT IS VULNERABLE TOTAL
15
1,342
20,130
47
POSITIVE
10 10
1,474 833
14,740 8,330
34 19
NEUTRAL NEGATIVE
35
3,649
43,200
100
. Analysis of Malay data Because of the importance of a single conceptual key in Malay I will only analyse one conceptual basis. The most common figurative sense of kaki is a way of liking something that is so intense that it overcomes the will of the individual to resist it. It can therefore be translated into English as addiction or obsession; I use these terms because they convey an important negative connotation that I will argue is central to the phraseological meaning of kaki. The full findings for this conceptual key are found in Table 2. In the last column of Table 2 the findings for the native informant data on expressive meaning are shown. The phraseological sense of kaki refers to social behaviour that is negatively evaluated, this can be represented by the conceptual key: to be, to do, or to be addicted to doing, something not good. This is an interesting semantic divergence from compound word senses, none of which contain any negative evaluation; compound nouns primarily exploit the spatial or functional senses of kaki; for example kaki gunung refers to the foot of a mountain. The disparity be-
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.7 (147)
A cognitive perspective on figurative language
tween phraseological and lexical meaning helps account for the discourse function of figurative language which Cameron and Low (1999: 86) describe as follows: Not only does metaphor shield a proposition from direct discourse, as nothing literal has been said, but it has the inestimable advantage of combining the fact that the speaker cannot be held responsible for the message with the flagging of the fact that there is a message being conveyed which cannot be discussed openly. Table 2. Malay: Kaki = to be, to do, or to be addicted to doing, something not good (n = 18 types) Figurative phrase
Translation
Kaki lawan Foot fight Goyang kaki To shake foot/leg Kaki judi Foot/leg gamble Kaki perempuan Foot/leg woman Patah kaki Broken foot/leg
Kaki botol Foot/leg of the bottle Kaki pukul Foot strike Kaki minum Foot/leg drink Kaki gaduh Foot/leg quarrel Kaki kuda Foot/leg horse Kaki bola Foot/leg ball
Book
News
Mag
Tot.
Freq.
To enjoy fighting
122
2
2
126
5.04
NEGATIVE
Someone who lives comfortably but is apathetic Someone who has a gambling habit Someone who is crazy about women Someone who no longer has the means to do something Someone who drinks heavily
31
8
6
45
1.8
NEGATIVE
9
24
1
34
0.96
NEGATIVE
2
8
4
14
0.56
NEGATIVE
5
5
3
13
0.52
NEGATIVE
8
2
3
13
0.52
NEGATIVE
Bouncer
3
7
0
10
0.40
NEGATIVE
Someone who is crazy about drinking A hooligan
6
2
1
9
0.36
NEGATIVE
7
1
0
8
0.32
NEGATIVE
8
0
0
8
0.32
NEGATIVE
0
6
1
7
0.28
NEGATIVE
Someone who loves betting on horses To be crazy about football
Connotation
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.8 (148)
Jonathan Charteris-Black
Table 2. (continued) Figurative phrase
Translation
Book
News
Mag
Tot.
Freq.
Kaki wayang Foot/leg screen Kaki joli Foot/leg an easy liver Kaki temberang Foot nonsense Kaki bangkang Foot oppose Kaki betina Foot female Kaki bodek Foot/leg support Kaki seribu Foot/leg one thousand TOTAL
Connotation
Someone who adores films A spendthrift
2
2
1
5
0.25
NEGATIVE
4
0
0
4
0.16
NEGATIVE
To love idly chatting To like protesting
2
1
0
3
0.12
NEGATIVE
1
0
0
1
0.04
NEGATIVE
womaniser
1
0
0
1
0.04
NEGATIVE
Flatter someone
0
1
0
1
0.04
NEGATIVE
Run away because one is afraid
1
0
0
1
0.04
NEGATIVE
212
69
22
303
I suggest that the highly productive metaphorical meaning of kaki in Malay phraseology is a culturally resonant way of talking about and evaluating behaviour that is not approved of in the prevailing belief system. Moreover, the selection of metaphor shields the speaker from responsibility for this evaluation. Further evidence of this motivation can be found in the use of kaki in what have been referred to as ‘hybrid creations’; these are defined as follows: The hybrid creation is a lexical form in which are combined both native and imported elements. Hybrid creations involve a kind of “reverse substitution”, in which loan morphemes are filled into native models. Heah Lee Hsia (1989: 202)
In the newspaper section of the corpus, we find that a number of these hybrid creations are used to fill semantic gaps to describe patterns of contemporary social behaviour that may be negatively evaluated according to the dominant belief system: (2) Kita harap ia memberi pengajaran kepada “kaki Internet” supaya tidak menyalahgunakan kemudahan berkenaan. We hope he give instruction on “foot Internet” in order to not abuse new appeal. ‘We hope he gives instruction on “Internet addiction” so as not to misuse something novel and appealing.’
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.9 (149)
A cognitive perspective on figurative language
(3) Jangan sekali-kali menjadi “kaki gosip” Ini mudah. Never excessively become “foot gossip” this easy. ‘We advise against addiction to gossip – this is all too easy to do.’ (4) Kalaupun ada mereka yang disebut oleh senator tadi, mungkin sebilangan kecil bapa-bapa yang kaki judi, kaki ‘enjoy’ atau kaki perempuan. Although there are they who be mention by senator just now, perhaps every small father-father who foot gamble, foot enjoy or foot woman. Although there are those who have just been referred to by the Senator, perhaps he was referring to every gambler, hedonist and playboy. Clearly, the use of figurative expressions with a modifying verb or noun has a strong pragmatic force in communicating the writer’s negative evaluation of certain types of social behavior. These include adultery, gambling and other behaviour that is forbidden by Islam and therefore is socially disapproved. It is common in Muslim societies for religious values to take on a strong socio-cultural role and this is manifested in Malay figurative language. The productivity of this type of evaluation is evidence of its cultural resonance. For example, we have an extension of the meaning of ‘addiction’ to behavior that is not prohibited or negatively evaluated; however, people’s attitude towards it is such that it recalls the behavior of those who are addicted to forbidden types of behavior. A good example of this is kaki bola – foot ball – ‘to be crazy about football’; while enjoyment of football is not sinful in itself, over-indulgence can have a negative connotation. We may also recall the use of fan in English; this term originated in fanatic but now the metaphorical meaning is less active. I would suggest, nevertheless, that it is dormant and can readily be reactivated when in collocation with words such as hooligan. We have further evidence of the productivity of pragmatically motivated cultural resonance when the conceptual key TO BE ADDICTED TO DOING SOMETHING NOT GOOD is extended to TO BE SOMETHING NOT GOOD, as in kaki gaduh – foot quarrel – ‘a hooligan’ and kaki pukul – foot strike – ‘a bouncer’. I suggest that cultural resonance of this body part can be traced to a sociocultural meaning related to orientation in Malay culture. In traditional views of the social order that which is atas (above or on top) signifies high status – by analogy with the raja or king – and that which is bawah (below) – as with the common people – signifies low status. While spatial metaphor does not in itself signify a negative evaluation in Malay it can do so when it is associated with the lower part of the human body. This is evidence of a culture specific adaptation of the crosslinguistic image schema UP IS GOOD, DOWN IS BAD. When this is applied to the domain of the human body those parts that are at the top when we are standing up are positively evaluated while those parts which are lower are negatively evaluated. We can find evidence for this cultural model in the characteristic behaviour
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.10 (150)
Jonathan Charteris-Black
in Malay society: the shoes are obligatorily removed from the feet on entering a house and it is considered impolite to show the soles of your feet; for example, by placing them on a stool. Address titles for sultans include reference to the speaker not being fit to kiss the dust in front of the sultans’ feet. While the social practices may be associated with traditional society, I believe that the phraseological productivity of kaki – as evident from its use in contemporary hybrid creations – provides linguistic evidence for the resonance of this cultural model. In terms of expressive meaning we will find that English figurative expressions in which foot occurs can have either a positive, a negative or a neutral expressive meaning; this contrasts with kaki which – as we have seen from Tables 1 and 2 above – usually has a negative connotation. In English the only phrase that seems to have a negative evaluation based on the body part alone is the expression my foot. What is of particular interest is that the evaluation for the same body part can be primarily positive or negative according to the conceptual key. This suggests that, in English, evaluation originates in the part of the figurative expression that modifies the body part; this contrasts with Malay where evaluation was found to reside in the body part term itself and therefore this can be described as the figurative head. This variation in location of the figurative head is potentially important from a contrastive perspective as it assists in interpreting the expressive meaning of figurative expressions.
. Analysis of English data As indicated above, there are three important conceptual keys for the English data; I will analyse each of these in their order of importance as measured by their resonance.
. The foot for control The full findings for this conceptual key in English are shown in Table 3. The conceptual key FOOT FOR CONTROL has an exclusively positive connotation; this can be seen both from the lexical context of the phraseological units for example in verbs such as find, land. Typical high frequency collocations imply slow progress towards a satisfactory goal within a sports context as in example (5): (5) of his team, initially unsure but gradually finding its feet. His goal using 54 different players England gradually found their feet to help erase unselfishly next to him and Butt gradually finding his feet, United managed (Bank of English)
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.11 (151)
A cognitive perspective on figurative language
Table 3. English: Foot for control (n = 15) Figurative phrase
Conceptual key
To find one’s feet Put your feet up Put your foot down Stand on your own (two) feet Vote with one’s feet To fall/land on one’s feet Think on one’s feet A foot in the door The boot/shoe is one the other foot Have one’s feet on the FOOT FOR CONTROL ground Get one’s feet under the table To cut the ground/pull the rug from under someone’s feet A foot in both camps/ A foot in each camp Get one’s feet on the ground The ball is at someone’s feet TOTAL
Total Frequency Connooccurrence per million tation words 232 232 215 167
0.70 0.70 0.65 0.51
POSITIVE POSITIVE POSITIVE POSITIVE
165 81
0.50 POSITIVE 0.25 POSITIVE
59 55 31
0.18 POSITIVE 0.17 POSITIVE 0.10 *
28
0.08 POSITIVE
23
0.07 POSITIVE
21
0.06 *
20
0.06 *
12
0.04 POSITIVE
1
0.003 POSITIVE
1,342
*Depends on context
There is also a positive connotation of words associated with foot by a relation of contiguity such as ground (to get in on the ground floor, to hold one’s ground, groundswell, to make ground and to break new ground) and ball as in: (6) The ball is at his feet. He has been successful in his
(Bank of English)
Connotation is also positive in start/keep the ball rolling, to play ball, on the ball, the ball is in your court. There is perhaps a common conceptual basis in the notion of preserving or winning territory in which ball, ground and foot are all conceptually associated. This can be distantly related to the positive associations of communal
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.12 (152)
Jonathan Charteris-Black
tribal hunting ritual that probably accounts for the almost universal popularity of competitive sports in which territorial control is vital (such as football and rugby). Many English figurative phrases containing ‘foot’ refer to physical gestures. Indeed non-verbal communication is a very good example of the way that our encyclopaedic or schematic knowledge of the world associates particular types of action or behaviour with the mental state which is antecedent to and motivates that form of behaviour. By a process of metonymic extension the material behaviour can be taken to signify the mental state that preceded it. For example, we know that if we drag our feet we are reluctant to walk and if we put our feet up it is because we are feeling relaxed. In each of these the physical action stands for an underlying affective cause or state of mind. In fact, we have mental scripts for a number of things people do when they are relaxed. However, we select a particular salient part of this image in order to represent the whole. A culturally salient example is selected and given a generic application. What is interesting is that verbal figurative units may have a strong expressive meaning because of this schematic knowledge.
. The foot is vulnerable The full findings for this conceptual key in English are shown in Table 4. English figurative phrases that generally convey a negative evaluation originate in a conceptual association between the foot and its metonymic extensions (walkTable 4. English: The foot is vulnerable (n = 10) Figurative phrase To foot the bill Shoot yourself in the foot To get cold feet Put your foot in it To have feet of clay Not put a foot wrong Dead on one’s feet Put your foot in your mouth Bound hand and foot Caught on the wrong foot TOTAL *Depends on context
Conceptual key
THE FOOT VULNERABLE
Total occurrence
Frequency per million words
Connotation
288 244 116 98 32 20 19 7
0.87 0.74 0.35 0.30 0.10 0.06 0.02 0.02
NEGATIVE NEGATIVE * NEGATIVE NEGATIVE POSITIVE NEGATIVE NEGATIVE
7 2 833
0.02 0.006
* NEGATIVE
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.13 (153)
A cognitive perspective on figurative language
ing, standing) with vulnerability. In these cases it is clear that the foot is associated with types of actions that are misjudged as in put your foot in your mouth, put your foot in it, and shoot yourself in the foot or with failure to act as in to have feet of clay. Here there is basic bodily knowledge that if there is a miscalculation involving the foot, one is likely to fall down and possibly inflict injury on oneself. This contrasts with the positive associations of moving confidently on the feet: this is the bodily experience that underlies the positive evaluation implied by the conceptual key FOOT FOR CONTROL. In effect we know that our vertical positions can be an advantage as long as they are sustained but can be a source of danger if there is any impediment to our balance. In figurative terms the foot can function as a metonym for controlled movement or for uncontrolled movement, movement is then extended metaphorically to domains other than the physical such as general abstract statements for mental control or vulnerability. It is the extension of the notion of loss of control leading to vulnerability that typically shows in the collocations shown in example (7): (7) said the Government would not allow taxpayers to foot the bill for the legal economical for the community the taxpayers who foot the bill? If so much palaces belong to the Queen, but taxpayers and tourists foot the bill for of security for the M3 extension, taxpayers are likely to foot the bill. The might get somewhere. After all, the taxpayers foot the bill for most academic (Bank of English) In the Bank of English corpus there were 27 lines in which taxpayers was found in one of the three slots prior to the phrase foot the bill and it seems that the rhetorical intention is to represent the taxpayer as vulnerable to the whim of government. In the case of bound hand and foot evaluation depends clearly on the context. While we might expect it to be negative – on the basis of its literal meaning, implying restriction on freedom – as in example (8): (8) an obscure junta of Labour agitators, bound hand and foot to the will their that he had an idea. In a land bound hand and foot by petty regulations (Bank of English) it is more commonly positive implying that restriction can be constructive as in example (9): (9) perspicacious old masterpiece. Bound hand and foot by my curiosity it was wise to hand the Party over bound hand and foot to the goodwill and laws, she wrote, the aboriginal is bound hand and foot by tradition; beyond descended from Marcus himself. I’m bound hand and foot to this place – its (Bank of English)
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.14 (154)
Jonathan Charteris-Black
This pattern of a positively evaluated subject of the figurative unit implies that it is beneficial to the patient to be constrained. We should also note that a number of figurative units in this group typically have abstract subjects as in example (10): (10) The government got cold feet, but Wilson persuaded the Scottish the government also got cold feet about Faradays. But CEST persevered the government may be getting cold feet. Also, President Bush isn’t budging did not heed. If government got cold feet over making him a London incumbent, the Government has begun to get cold feet. The new regulations have been (Bank of English) This is a case where we need to distinguish between the text producer’s evaluation and that of the subject; for example, in these cases the writer may be approving of the government’s negative evaluation of a situation.
. Life is a journey The full findings for this conceptual key in English are shown in Table 5. Finally, I will consider the use of another conceptual key LIFE IS A JOURNEY to imply general evaluative statements as regards human emotional relationships. ‘Foot’ in these phrases is a metonym for ‘travel’ and we can see that the phrases typically contain a verb of motion (e.g. start, drag, follow, sweep); we may infer that motion is conceptualised as a source of experience. The association between the physical act of travelling and an accompanying mental state is conveyed in expressions such as footloose and fancy free implying a carefree mental state and to get itchy feet implying a desire to travel based on a state of frustration or boredom with one’s current situation. Since in life – as in journeys – we can have good or bad experiences, figurative expressions in this group can communicate either a positive or a negative evaluation. There is an extension from knowledge of bodily experience to general human relationships. Experiences can be either positive as when we put our best foot forward or negative as when we start off on the wrong foot. This is indicated by lexical content such as right/wrong, since it is not the body part term that indicates the nature of the evaluation we can say that the accompanying lexis comprises the figurative heads of these figurative expressions. However, it seems that the journey metaphor is typically a positive one: we can see this in the most frequent figurative expressions to follow in someone’s footsteps; the most common adjectives in the slot preceding footsteps in the Bank of English were: famous, distinguished, hallowed,
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.15 (155)
A cognitive perspective on figurative language
Table 5. English: Life is a journey (n = 10) Figurative phrase Follow in someone’s . . . footsteps Drag one’s feet One foot in the grave Sweep someone off their feet Get one’s feet wet Put your best foot forward Footloose and fancy free Get itchy feet Start off/get off on the right foot Start/get off on the wrong foot TOTAL
Conceptual key
LIFE IS A JOURNEY
Total Frequency Connooccurrence per million tation words 728
2.20 POSITIVE
385 170 58
1.17 NEGATIVE 0.51 * 0.18 POSITIVE
39 33
0.12 POSITIVE 0.10 POSITIVE
22
0.07 POSITIVE
21 10
0.06 * 0.03 POSITIVE
8
0.02 NEGATIVE
1,474
*Depends on context
successful, musical and illustrious. On this basis we can say that the text producer offers a positive evaluation of the life achievements of an individual and implies that these relate to a hereditary influence (the most common noun collocates were parents and fathers). The iconic image for LIFE IS A JOURNEY is therefore one in which a person is seen as following the footsteps left by a parent (typically a father).
. Conclusion In this paper I have provided a number of illustrations of how meanings are conveyed in figurative expressions in Malay and English. The aim here has been to show how figurative expressions can be demonstrated to be systematic – irrespective of language – and how the figurative meaning can be represented by identifying underlying metaphors and metonyms that I have described as conceptual keys. We have also seen the need to examine the lexical content of a figurative expression as expressive meaning can be conveyed through the body part term or through the semantic field of the modifying verb, adjective or noun. Anatomical knowledge of the functionality of parts of the human body may reflect universal tendencies in human conceptualisation in cases where the body part constitutes the figurative
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.16 (156)
Jonathan Charteris-Black
head. In cases where the figurative head is in the modifier we have found that the evaluation tends to be culture specific. We can see this by comparing the Malay goyang kaki – swing leg – ‘freebooter’ which has a negative connotation implying apathy or laziness with the English to put your feet up which has a positive connotation implying intentional relaxation. The figurative head in both cases is in the modifiers (goyang ‘swing’ and ‘put up’) and these have different connotations in the two languages. There are two components to encyclopaedic knowledge: first universal knowledge of the biological functions of the human body and secondly knowledge of how people behave in particular socio-cultural contexts: are social practices, gestures, body parts etc. approved or otherwise? We have seen this with the exclusively negative connotation of ‘foot’ in Malay. This has been accounted for with reference to specific anthropological practices (cultural or social). Identifying the figurative head is therefore crucial to interpreting the meaning of figurative units. The cultural resonance of figurative expressions should be interpreted as the outcome of both their conceptual and their expressive content. In this respect there is support for the view that conceptualisation is to some extent experientially grounded but also has culture specific dimensions. However, the connotations of cultural meaning only become evident through examination of both verbal context and native informant interpretation.
Acknowledgement The author would like to express his thanks to the Dewan Bahasa dan Pustaka for access to their corpus, to Cobuild for access to the Bank of English, and to the British Academy for a research grant that permitted the author to collect data in Malaysia.
References Biber, D., Conrad, S., & Reppen, R. (1998). Corpus Linguistics. Cambridge: CUP. Black, M. (1962). Models & Metaphors. Ithaca, New York: Cornell University Press. Cameron, L., & Low, G. (1999). Metaphor. Language Teaching, 32, 77–96. Charteris-Black, J. (2000). Figuration, lexis and cultural resonance: A corpus based study of Malay. Pragmatics, 10 (3), 281–300. Charteris-Black, J. (2001a). Cultural resonance in English and Malay figurative phrases: The case of ‘hand’. In J. Cotterill & A. Ife (Eds.), Language Across Boundaries. Proceedings of the 33rd BAAL annual meeting (pp. 151–170).
TSL[v.20020404] Prn:4/11/2002; 9:34
F: PBNS26.tex / p.17 (157)
A cognitive perspective on figurative language
Charteris-Black, J. (2001b). Blood sweat and tears: A corpus-based cognitive analysis of ‘blood’ in English Phraseology. Studi Italani di Linguistica Teorica e Applicata: Italian Studies on Theoretical and Applied Linguistics (pp. 273–287). Croft, W. (1993). The role of domains in the interpretation of metaphors and metonymies. Cognitive Linguistics, 4, 335–370. Cruse, D. A. (2000). Meaning in Language: An Introduction to Semantics and Pragmatics. Oxford: OUP. Moon, R. et al. (Eds.). (1995). The Cobuild Dictionary of Idioms. London: Harper Collins. Goatly, A. (1997). The Language of Metaphor. London: Routledge. Hasan Muhamed Ali (1996). Malay Idioms. Singapore & Kuala Lumpur: Times Books International. Heah Lee Hsia, C. (1989). The Influence of English on the Lexical Expansion of Bahasa Malaysia. Kuala Lumpur: Dewan Bahasa dan Pustaka. Lakoff, G., & Johnson, M. (1980). Metaphors We Live By. Chicago: University of Chicago Press. Masri, S. (1997). Simpulan Bahasa Melayu. Shah Alam: Perbit Faja Bakti Sdn. Bhd. Panther, K., & Radden, G. (Eds.). (1999). Metonymy in Language and Thought. Amsterdam/Philadelphia: Benjamins. Stubbs, M. (1996). Text and Corpus Analysis. Oxford: Basil Blackwell. Ungerer, F., & Schmid, H. J. (1996). An Introduction to Cognitive Linguistics. London & New York: Longman.
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.1 (159)
Metaphorical expressions in English and Spanish stock market journalistic texts Anna Espunya and Patrick Zabalbeascoa Universitat Pompeu Fabra, Barcelona, Spain
.
Introduction1
This paper reports the results of a study of metaphorical expressions used in nonspecialised journalistic texts that deal with the topics of investment and the stock exchange markets in two languages, British English and Castilian Spanish. The study is the initial phase of a larger project that aims to describe translational strategies (and their patterns) for metaphorical expressions. It is a well accepted methodological premise of translation studies that any description of translated texts should be made in relation to the writing conventions of the target language. A metaphorical expression is a linguistic expression that contains a metaphor. We define metaphor as a tripartite relation between two separate entities and a property – or set of properties – that they share, according to the intention and/or interpretation of the metaphor. Traditional terms for each of the two entities (Richards 1965; Leech 1969) are object – the entity referred to in the text, i.e. the actual unconventional referent of the metaphorical expression – and vehicle – the entity in terms of which the object is referred to in the text, i.e. the conventional referent of the metaphorical expression. The sense (or ground) is the property or set of properties of the vehicle attributed to the object by an interpretive operation of analogy, similarity, etc. Newmark’s (1981) approach to the translation of metaphorical expressions maintains that the type(s) of solutions to be considered by the translator will depend on their degree of conventionality, which is measured by a typology that ranges from dead metaphorical expressions (highest conventionality, attested in the dictionary as literal, primary senses) to novel ones (lowest conventionality) with several middle types such as stock metaphors (expressions that have by convention a metaphorical meaning) and clichés/idioms (expressions whose
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.2 (160)
Anna Espunya and Patrick Zabalbeascoa
metaphorical meaning is their only possible meaning and thus their literal one). Newmark’s approach rests on the assumption that metaphorical conventionality is a static feature of every word or expression, not of each instance in a text of a given word or expression. The view that metaphorical expressions are lexical items which can be classified, in isolation, according to their vitality runs counter to the observation that even highly conventionalised expressions and dead metaphors may receive a metaphorical interpretation in a given text. During the language acquisition process, be it mother tongue or foreign language, expressions that fully competent speakers no longer treat as metaphorical are revitalised (Lakoff & Johnson 1980; Kronfeld 1981). Moreover, resuscitation of dead metaphors is found in literary texts, especially poetry (Goatly 1997) but also in journalism and advertising, just to mention the most obvious instances. The process by which a highly conventionalised expression is interpreted as having a conventional meaning is called passive construal, whereas the process by which the reader participates actively in its interpretation is called active construal (Kronfeld). If we accept that the life of a metaphor is a dynamic feature, then we must accept that metaphors are given (or denied) life in texts (see also Snell-Hornby 1995), or to be even more precise, in the interpretation of the texts, and sometimes only in the author’s intent. This has dramatic consequences for any attempt at detecting and studying metaphorical expressions in automated corpora. Further support can be found for our argument if we look at metaphors through the parameter of function. A textual approach indeed provides an empirical methodology for a comparative study of the kind we have set out to do. As shown by research on the popularisation of science, metaphors go beyond the aesthetic and the emotive functions. To understand them and to understand how function might determine the way a metaphorical expression is translated we need to analyse their textual environments. Thus, we fully agree with SnellHornby (1995: 58) since “abstract rules” seems to imply “rules that are not sensitive to context”: Whether a metaphor is ‘translatable’, how difficult it is to translate, how it can be translated and whether it should be translated at all cannot be decided by a set of abstract rules, but must depend on the structure and function of the particular metaphor within the text concerned.
. The corpus We chose a corpus of popularisation texts because we wanted to see metaphorical expressions with as many functions as possible. The topic of stock exchange
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.3 (161)
Metaphorical expressions in stock market journalistic texts
analysis is predictable in its contents in the sense that the situations described are very limited in range: the value of stocks may grow or diminish, and this can happen suddenly or more or less gradually. Investors usually take a limited number of different attitudes toward risk (going from cautious to reckless) and their possible range of reactions is largely just as limited when facing the changes in the stock values (from satisfaction to disappointment). It was our assumption that limited content-type would promote a larger number of writing strategies, including the use of metaphorical expressions. The paper presents data extracted from a pilot corpus of 18 texts, 9 English texts from the weekly magazine The Economist and 9 Spanish texts extracted from two Spanish publications, the weekly Actualidad Económica and the daily Expansión. Although the publications are not exactly parallel, the articles have been chosen with that criterion in mind. The Economist covers general information for an international readership but has specialised sections on business, economy and finances. Actualidad Económica focuses on economy and finance plus investment advice, and has a more limited target readership. However, the fact that it publishes translated articles from The Economist on a regular basis gave it parallel status. To complement the corpus with texts on international topics, we selected features from Expansión, which has a broader coverage of them.
. The functions of metaphorical expressions . A typology of functions Metaphor as a semantic relation that involves two entities synthetically combines a considerable amount of meaning in a single expression. No wonder then that metaphorical expressions fulfil many textual functions simultaneously at different levels. Borrowing the terminology from Halliday and Hasan (1976), metaphorical expressions may contribute at the same time to all of the metafunctions of texts: ideational, interpersonal and textual. Goatly (1997: 149) attempts to organise them according to his analysis of different genres. To these, Goatly adds the combination of ideational plus interpersonal functions that results in the metaphorical calls to action or problem-solving. By this he recognises the difficulty in assigning clear-cut borders between levels. Our analysis suggested other combinations between ideational and textual, such as the effects of filling lexical gaps (ideational) on text cohesion (textual). One of the findings of our analysis is that the label “textual structuring” must be further specified.
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.4 (162)
Anna Espunya and Patrick Zabalbeascoa
Table 1 Ideational (understanding the environment): a. Shaping ideology b. Lexical gap-filling
c. Explaining and modelling d. Reconceptualising e. Arguing by analogy and /or false (?) reasoning
Interpersonal (“acting on Textual (“providing resources others in the environment”) to ensure that what is said relates to the co-text/context”) f. Expressing emotional attitude g. Decorating, disguising (euphemism)
k.
Text-structuring
l.
Enhancing memorability, foregrounding and informativeness h. Exaggerating, hyperbole m. Fiction
i. Cultivating intimacy j. Playing games, humour
. Functions of metaphors in high-level popularisation of the stock exchange The texts in our corpus share some of the characteristics of the popularisation of science such as the importance of the ideational function (introducing and explaining concepts, comparing and contrasting situations, etc.) combined with a hierarchical relation between an expert writer and a lay reader. But since they are published in the weekly press, they have similar constraints to those of newspaper reports, i.e. the need to attract and maintain their readership, by entertaining it while keeping it informed. Metaphor is one of the mechanisms deployed to fulfil these requirements. In the texts that we have analysed, most metaphorical expressions perform several functions at the same time, or rather, they combine functions of each metalevel.
.. Ideology Our generalisations regarding the ideological function of metaphor will be formulated in the framework provided by the experientialist view of the human conceptual system proposed, among others, by Lakoff and Johnson (1980). In this cognitivist view, metaphor is mediated by the physical restrictions of human experience, which in turn is influenced by culture. The assumption is that different cultures may “live by” different metaphors. Not surprisingly, we found that the discourse about the stock market in both cultures yields the metaphors “more is up” and “good is up”. The terminology of the field clearly expresses value increase as upward movement and value loss as downward movement, both in English (rise, peak, fall, slump, collapse) and in Spanish (subida ‘rise’, caída ‘fall’, tocar techo ‘reach the ceiling’). So do non-terms
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.5 (163)
Metaphorical expressions in stock market journalistic texts
(balloon, high-flying, stratospheric levels, escalada ‘escalation’, valoraciones de vértigo ‘vertigo-giving valuations’).2 We have also observed several differences related to other metaphors. The most significant is that Spanish texts have many more expressions referring to the history of stocks, markets, currencies, etc. The texts very often look at the evolution of those elements over time and Spanish has expressions to refer to this as if the entities were moving along a path and covered distances: trayectoria ‘trajectory’, recorrer ‘to cover a distance’, marcha ‘march’, evolución ‘evolution’, perder terreno ‘to lose ground’. In Lakovian terms, the metaphors underlying the Spanish expressions are “stocks (currencies, markets) are objects in space”, “stocks are moving objects”, “time is stationary and stocks move through it”, “value is stationary and stocks move through it”. A few examples are provided in (1). (1) a.
En su corta trayectoria bursátil, Prisa ha acumulado una ganancia del 18,22%. ‘In its short stock market trajectory, Prisa has accumulated an 18.22% gain.’ b. Tecnocom fue capaz de multiplicar por diez su precio en tan sólo diez sesiones, en la marcha bursátil más increíble que se recuerda en nuestro mercado. ‘Tecnocom was able to multiply its price by ten in just ten sessions, in the most incredible stock exchange march anyone can remember in our market.’ c. Independent News, que tiene un recorrido del 142 por ciento hasta su precio objetivo. ‘Independent News, which has a 142 percent [distance] to cover to get to its target price.’ d. La falta de entusiasmo es consecuencia del nuevo rumbo del Nasdaq. ‘The lack of enthusiasm is a consequence of the Nasdaq’s new course.’
Seeing the stock market as an object in motion allows for the occurrence of metaphors highlighting other aspects of the motion or the object (halting the motion, engines, impulse, etc.; see the extended metaphor in example (11)). Another difference is that whereas in English the markets tend to be referred to metonymically as places, in Spanish they are more frequently personified, or at least seen as agents or experiencers (see (2) vs. (3)). This does not exclude the converse (personification of the markets in English and metonymical reference in Spanish) as the fragments in (4) and (5) illustrate. (2) a.
The Nasdaq: from the condemned home of sleazy practices to the shining temple of [. . . ]
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.6 (164)
Anna Espunya and Patrick Zabalbeascoa
b. Nasdaq, home to all that is best, brightest and outrageously overvalued. c. If there is life yet in the floor [. . . ] (3) a.
Las subidas de los tipos de interés no sientan bien a los mercados de renta variable. ‘The rises in interest rates do not go down well with the stock markets.’ b. Desde la crisis de abril, la bolsa española ha ido más de la mano de Wall Street que de sus colegas europeas. ‘Since the April crisis, the Spanish stock market has gone hand in hand with Wall Street rather than with its European colleagues.’ c. El dinero ha decidido ver y esperar a que se aclare un poco la situación. ‘The money has decided to see and wait (wait and see) for the situation to clear up a bit.’
(4) a.
Nasdaq is rapidly taking its technology gospel to the corners of the Earth b. one exchange proposing to marry another c. the floor may yet be allowed to die.
(5) El pesimismo se ha instalado en los mercados de renta variable. ‘Pessimism has settled in the stock markets.’
.. Filling lexical gaps The terminology of the field is made up of many dead metaphors. The senses of the words in their contexts are considered literal, they are unambiguous and belong to a set of items that are shared by a professional community. The following list, by no means a comprehensive one, provides parallel instances of English and Spanish terms.3 (6) floor/parqué bearish/defensivo rise/subida crash/desplome flotation/salida a bolsa flow (of money)/entrada
bull market/línea alcista fall/caída volatility/volatilidad behaviour (of shares)/comportamiento exposure/tomar posiciones or posicionarse portfolio/cartera
Other English terms with metaphorical origin are ‘deep’, ‘liquid’ (denoting two distinct features of a market), ‘emerging’ [markets], ‘data-mining’. Other Spanish terms born out of metaphor are castigo ‘punishment’, corrección ‘amendment’ (both mean a reduction in a stock’s value following a steady rise), trayectoria ‘course’ or ‘trajectory’.
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.7 (165)
Metaphorical expressions in stock market journalistic texts
Although terms are considered to be inactive metaphorical expressions, the most recent ones are still perceived as metaphorically active and as such may appear enclosed in quotation marks. This is the case of ‘emerging markets’ and ‘the new economy’. The introduction of existing terminology and concepts to the reader is one of the roles of popularising genres and is usually signalled by metalinguistic expressions (known as, lo que se conoce como). Two examples from our corpus are provided in (7). (7) a.
Value investors [those who want to buy cheap shares] have lost ground to a school of investing based on [. . . ] and to another fashionable group of large investors known as momentum investors,4 whose bets are made solely on the direction in which a market is heading. b. Lo normal es que los inversores institucionales no intervengan en la gestión de las compañías en las que participan. Más bien optan por lo que se conoce como “votar con los pies”, es decir, venden sus acciones si no están de acuerdo con el equipo directivo. ‘The normal thing is for institutional investors not to intervene in the management of those companies where they hold stocks. Rather they choose what is known as “voting with their feet”, that is they sell their stocks if they disagree with the management team.’
Together with specialised terms we sometimes found contextual synonyms or quasi-synonyms of a metaphorical nature, instances of what Goatly calls “diversification of metaphors”, where the same object is expressed through different though related vehicles. In some cases the occurrence of synonyms and quasi-synonyms seems to respond to the same gap-filling function for established terms. Support for this hypothesis stems from this observation that the same items occur in several texts. Quasi-synonyms differ in some semantic component which presumably needs to be presented as new information. It is precisely this semantic component that constitutes the ground of the metaphor, i.e. the sense in which the word must be understood as applying to the concept. This is so even if the expressions are inactive metaphorical expressions. Table 2 provides instances of terms, their nearsynonymous variants and the semantic component made prominent, following the definition of the words in the Oxford Concise Dictionary. Synonyms and quasi-synonyms may satisfy another need, lexical cohesion without lexical repetition. The texts under study are in danger of multiple repetition of certain lexical items, probably because of the limited variety in the type of events described. Different languages and different writing and journalistic traditions may differ in the degree to which repetition is tolerated. Spanish is said to
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.8 (166)
Anna Espunya and Patrick Zabalbeascoa
Table 2 Term
Variant
Semantic component in focus (sense or ground of the metaphor)
Go up/rise
Boom (a booming stock-market)
Go up/rise
Surge (exchanges, uncontrollable)
Go up/rise Go up/rise Go down/fall Go down/fall
Balloon (the market value) Peak (the Nasdaq) Plunge (prices) Slump (shares, slow)
Go down/fall Go down/fall Go down/fall Go down/fall
Dip “buying the dips” Sink Collapse Land (in the bottom)
Sudden activity or development Move suddenly, powerfully, lack of control “High in the air”, swell Reach highest value, quality Violently or suddenly Great and prolonged fall in price Below a surface or level Slowly Sudden, “giving in” Reach the lowest point
be much less tolerant than English, which means that we should expect a higher presence of avoidance strategies than in English. A measurement of repetition was out of the scope of this paper, but it is our intuition that repetition is avoided even in the English texts, maybe because the likelihood of ambiguity is lessened by the abovementioned small number of possible events or scenarios. Therefore, synonyms and quasi-synonyms appear in segments where an old discourse referent is recovered for updating, e.g. through modification or contrast. We consider this a text-structuring function and we discuss it in more detail in Subsection 3.2.4. English examples are provided in (8): (8) a.
The shares of one of America’s biggest firms [. . . ] dropped sharply [. . . ] But Nasdaq contains many big technology firms [. . . ]. Shares in most of these firms slumped too. b. Nasdaq has fallen sharply [l.5]. Softbank and Hikari Tsushin have hit downward trading limits. c. Wit capital has fallen [. . . ] DLJ Direct has sunk from [. . . ]
A group of expressions that stands apart is that of “popular terms”, with informal variants of institutional names as the most common instance. The findings – listed in (9) – were 7 for English and 1 for Spanish, which suggests that English has indeed a higher tolerance for the occurrence of items of mixed registers. (9) a.
antitrust bloodhounds, i.e. members of the antitrust commission in the US b. ‘the supervisory watchdog’
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.9 (167)
Metaphorical expressions in stock market journalistic texts
c. d. e. f.
the Securities and Exchange Commission, the main market watchdog moneymen, i.e. investors the greenback, i.e. the dollar bill el billete verde, i.e. the dollar bill (original italics, marking non-literal use, i.e. the thousand-peseta bill) g. The pensions system cannot seriously be overhauled without mature and efficient capital markets. This is the main reason to think that the roller-bladers will get their way over the central planners. [China stockmarkets & reform proposals: ‘the roller-bladers’ refers to the reformist economy planner who goes roller-skating in his leisure time] These items are of various origins, some are based on analogy (9 a, b, c), others on metonymy (9 d, e, f, g). What they have in common is that they refer to imaginable things and therefore “have a particular vitality” (Goatly 1997: 164). They have conventional senses (attested in dictionaries) and are inactive metaphorical expressions, with the exception of (9g), which illustrates that the mechanism for metaphor creation (here through derivational morphology) is productive. These “popular terms” are very efficient shorthands because they fulfil several functions at the same time: they enhance memorability, while adding humour and therefore developing intimacy with the reader, and ultimately they convey ideology (in the roller-skaters example, enhanced locomotion as a symbol of modernity and reform, in contrast with lack of movement by traditional central planners).
.. Explanation and modelling Explaining concepts is one of the goals of the popularisation texts in our corpus. One way to achieve this is by drawing analogies between the concepts that the author wishes to explain and concepts or situations that are known to the reader. An example is provided in (10). (10) La operativa de entrar y salir dentro de una jornada, o en plazos inferiores a tres días, es un deporte intelectual que requiere entrenamiento. Hay que vencer al mercado en las condiciones más desfavorables: un entorno cambiante en que los protagonistas de un día son los bajistas del siguiente. Y hay que hacerlo con poco dinero, ya que son las instituciones las únicas con poder para aupar o hacer caer a un valor. Agilidad y una disciplina espartana, para evitar tener pautas de conducta distintas cuando se pierde que cuando se gana, es fundamental [. . . ] ‘The operating policy consisting in buying and selling within a day, or in periods smaller than three days is an intellectual sport that requires training. One must beat the market in the most unfavourable conditions: a changing
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.10 (168)
Anna Espunya and Patrick Zabalbeascoa
environment in which one day’s protagonists are the next day’s downgoers. One must do it with little money, since the institutions are the only ones with the power to raise or sink a stock. Agility and Spartan discipline, to avoid having different behaviour patterns when losing than when winning, are fundamental.’ Analogy involves equivalence in the relationships between the parts of each concept. In (10) same-day buying and selling operations are compared to sport. The analogy is sustained by several metaphorical expressions that cover several aspects of sport within the same paragraph: it requires training, its goal is to beat an opponent (the market), one can win or lose, it requires agility and Spartan discipline from the player. The excerpt also illustrates the interaction between the ideational level and the text structuring level. The development of an explanation by analogy goes hand in hand with the establishment of the appropriate lexical network or chain (in the previous example, marked by our bold). Lexical networks are the manifestation of lexical cohesion and help the reader interpret words and expressions by reference to other words and expressions in the surrounding sentences (Halliday & Hasan 1976). Lexical networks can be established between expressions with literal readings and also between metaphorical expressions. (10) illustrates a subcase of the latter, called metaphor extension, where metaphorical expressions that make up the network develop several aspects of the same vehicle. Metaphor extension tends to reactivate the life of metaphors. The use of extended metaphor for the purposes of explaining through analogy occurs only once in the English corpus (it involves 4 metaphorical expressions out of 137) and twice in the Spanish texts (involving 6 metaphorical expressions in 78). The notion ‘extended metaphor’ presupposes the author’s intent, which is easier to identify in paragraphs that have specific functions such as explanation of a concept. In those cases, extension takes place locally within a text fragment. But it may be the case that the chain of metaphorical expressions does not play a clearly identifiable ideational role, or that the metaphor extension takes place globally – in the whole text – though not locally – without concentrating on any specific text fragment. These are cases where other metaphorical expressions may intervene, giving rise to metaphor mixture. Those mixtures may be successful if each new metaphor reinforces the ones already introduced, or they may fail and be judged as sloppy style, if the new metaphor is inconsistent in some way with the previous ones. Consider the following example, a set of excerpts from a rather long text on the future of the Spanish stock market, presented here in table format (the column on the left shows the line numbers, the excerpts are given in the right column). The line numbers are a good indication of the distance separating the various related expressions. Its back-translation is provided at the end of the paper in an appendix.
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.11 (169)
Metaphorical expressions in stock market journalistic texts
(11) l. 20–23
Este paquete de medidas, unido a la polémica en torno al presidente de Telefónica [han puesto el motor de la bolsa en punto muerto]1. l. 40 ¿Por qué este [cruce de caminos]1? l. 71–73 Las dos filiales de la operadora [. . . ] también han [perdido terreno]1. l. 186–188 Lo cierto es que en España el mercado [se ha paralizado]1. [. . . ] En esta situación ¿qué actitud están tomando los inversores extranjeros? l. 193–199 Hasta marzo han sido compradores [. . . ]. Pero en las últimas semanas han decidido [parar las máquinas]1. Aunque la sensación general es que no se ha producido [una desbandada]2, algunos de estos fondos sí que han reducido su posición en la bolsa española y la han sustituido por otros mercados europeos. l. 211–219 Si algo caracteriza la forma de actuar de los fondos de inversión y de pensiones internacionales es que, aunque no son especulativos, sí son volubles: las oportunidades de inversión no suelen durar mucho y, una vez alcanzados sus objetivos de revalorización, [echan a volar en busca de valores frescos]2. l. 234 and 250–254 No hay ningún inversor igual a otro. [. . . ] Pero, sea como sea, se han convertido en [los motores que mueven los mercados]1 al alza o la baja. The main feature of the text is the coexistence of two metaphors that have foreign investors in the Spanish stock market as their object. Foreign investors are at the same time the engines that move the Spanish stock markets (enclosed in square brackets followed by #1) and a flock of birds that are ready to fly away once they cease to find the market attractive (enclosed in square brackets and followed by #2). On a first reading, we may conjecture that the author intended to create two extended metaphors, one of them (#1) mainly for text structuring purposes, the other (#2) for value judgement. Whether their coexistence draws a consistent picture of foreign investors may depend on the reader’s capacity or willingness to find coherence in the message. But if we look closely at the expressions marked #1, we find slight variations in the image of the metaphor: poner el motor de la bolsa en punto muerto ‘set the stock exchange markets in neutral’ is an automobile metaphor (in Spanish, car engines are referred to as motor, in singular), whereas in parar las máquinas the image is one of a ship’s engines (in Spanish a ship engine is máquinas, in plural). To a certain extent, the author was not relying on his
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.12 (170)
Anna Espunya and Patrick Zabalbeascoa
choice of metaphorical expressions to draw a precise analogy to explain stock market processes. The authors probably only meant the lexical metaphorical network as a cohesion mechanism, or maybe only as a decorative element. Mixed metaphors can occur in what appears to be the same lexical network. The coherence of a text is then in jeopardy, unless, of course, the writer does not intend the reader to process the expressions as metaphorical. As Goatly observes, “serious newspapers, with their informative rather than entertainment priorities, generally discourage metaphorical processing” (p. 302). (12) provides good illustration. (12) Title: “Los portales chinos navegan contracorriente en el Nasdaq” [. . . ] La falta de entusiasmo por la oferta de Sohu es consecuencia del nuevo rumbo del Nasdaq [l. 68]. ‘The Chinese portals sail against the current (i.e. upstream) in the Nasdaq [. . . ] The lack of enthusiasm toward the Sohu public offer is a consequence of Nasdaq’s new course.’ In the title, Nasdaq is rendered as a river where companies (boats?) navigate and where a specific Chinese company is moving against the current. In the last paragraph of the article, however, Nasdaq is depicted itself as a boat that has taken a new course. To add to the confusion the Chinese company develops an Internet portal (a metaphorical term denoting a static object, hard to imagine as navigating along a river). The choice seems to have fallen on semantically related items even if they were used in textually inconsistent ways.
.. Text structuring Goatly restricts his definition of textual structuring to the parameter of lexical cohesion. Metaphor extension is one of its manifestations and we saw how it can serve ideational purposes like drawing analogies to explain concepts. In this section we focus on the specific types of relations between metaphorical expressions that not only contribute to lexical cohesion but are also interwoven with the organization of information in the text. The first type of relation is one of contrast or comparison of referents with respect to some semantic component. In the English corpus it accounts for 18% of the occurrences of metaphorical expressions. Two instances (13 and 14) from one text provide good examples of what we mean by contrast or comparison. (13) The flood of money into equities has slowed to a trickle. Compared with March, new money invested in equity mutual funds fell by half last month [. . . ]
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.13 (171)
Metaphorical expressions in stock market journalistic texts
(14) Last year, the market capitalisation of Charles Schwab surpassed that of Merrill Lynch, an event that was interpreted as a sign that a new order had risen on Wall Street. In March, just before Nasdaq began to crumble, Merrill’s marked capitalisation peaked at $44 billion and Charles Schwab’s at $57 billion. The recent tumult has singed Merrill (now valued at $38 billion) but scorched Schwab (value: $35 billion). In both instances the metaphorical expressions have the same object (or kind of object) and the vehicles are lexical items that denote different intensities of the same process. While they share semantic features – [liquid, volume, moving, speed], in (13) and [affected by burning], in (14) – at least one of those features is gradable, it is precisely the degree which is in focus. In (13) note that ‘the flood of money’ is a highly conventional – and thus, in principle, inactive – metaphorical expression that recovers information introduced in the previous discourse. The resultative combination V+Prep ‘slow to NP’ expresses part of the ground of the ‘flood’ metaphor in the money domain, and the term of comparison with ‘trickle’. The latter expresses the same concept as ‘flood’ but with a lower degree of intensity, which is the new information to be provided. The ground of the ‘trickle’ metaphor is given in the next sentence. Information progresses from old and highly conventional to new and less conventional; the metaphor is preserved and at the same time the first member of the pair regains its old metaphorical life. In the case of (14) the focus is on the higher intensity of the damage suffered by one of two companies by certain events in the stock market. From the discourse point of view, it is interesting to note that the information provided early in the paragraph led the reader to the opposite expectations. The physical vehicle of the pair singe-scorch provides a climax which is necessary to focus on the extent to which those expectations were unjustified. In this case the ground of the metaphorical expressions is not made explicit, although the interpretation of ‘singe’ is conventionally available (the idiom ‘to singe one’s wings’ is attested in the Oxford Concise Dictionary with the meaning ‘take harm in a venturesome attempt’). The syntactic structure where the two expressions are embedded – the VPs are parallel constructions linked by the contrastive ‘but’ – helps to establish the relationship between the conventional ‘singe’ and the less conventional ‘scorch’. It could be objected that these examples are from a text with a special sensitivity to the metaphorical potential latent in conventional expressions. However, the trend for metaphorical expressions to appear in pairs in contrastive environments extends to all of the texts in the English corpus and with different rhetorical aims, such as exposition of opposing concepts, illustrated in (15).
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.14 (172)
Anna Espunya and Patrick Zabalbeascoa
(15) Strictly speaking, the equity premium is a forward-looking concept: the extra returns, on average, that investors in shares can expect to earn. But many studies of the equity premium have estimated expected returns using backward-looking performance data: the extra returns that investors actually earned. As far as Spanish is concerned, there are differences in two respects. First of all, we did not find pairs where both members were metaphorical in nature; rather, we found pairs consisting of one item with a metaphorical reading and another that would normally have a literal reading. We found no instance of a discourse fragment where the discourse focus was on the contrast between the grounds of metaphorical expressions. Here lies an important difference between the two corpora: the Spanish writers do not ask their readers to unravel the metaphorical relationship for rhetorical purposes, just as they do not employ puns or word-play, as will be discussed later. The second difference was quantitative. The pairs with contrastive relationships accounted for 11.5% of the occurrences of metaphorical expressions. Other uses linked to cohesion outside contrastive contexts are stylistic synonymy (retaking old themes to add new information about them), as the pair downturn-landing in the excerpt (16) below, and rephrasing previous information for the purpose of summarising. (16) Besides stockmarket gyrations, the prolonged caution about equities seems to reflect worries about the expected downturn of the American economy. However “hard” the landing turns out to be, it will have an effect on corporate profits. Finally, we also found instances (17) where highly conventionalised metaphorical expressions recovered their metaphorical vigour via intra-sentential cohesion relations such as selectional restrictions between predicate and argument. (17) a. The [Chinese] market has recently been Asia’s only raging bull. b. Su debut en bolsa llegó teñido por un negro período de suscripción [. . . ] ‘Its debut in the market came tinged by a black (i.e. rough) subscription period’ In (17a), ‘bull’ is the term for markets with rises in stock prices. The adjective ‘raging’, which collocates with the literal reading, prompts a second interpretation. Similarly, the adjective negro means ‘gloomy, rough’. The participle collocates with the literal meaning ‘black’. There are slight differences between English and Spanish. In English occurrences of metaphorical expressions related to cohesion in non-contrastive environ-
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.15 (173)
Metaphorical expressions in stock market journalistic texts
ments account for 20% of the total; in Spanish, for 38%. Metaphorical expressions in the Spanish texts appear mostly in the same contexts as the English ones, forming lexical chains with other lexical items; but whereas in English metaphorical expressions tend to appear after the lexical item with the literal reading, in Spanish we found several cases where they occur in the reverse order, “cataphorically”. Up to this point we have discussed the ability of metaphorical expressions to enter lexical cohesion relations. But according to our observations, metaphorical expressions participate in text structuring functions of other kinds. If we extend text structuring to include text composition, we can account for the number of metaphorical expressions that appear in opening and closing paragraphs of texts, and in opening and closing sentences of certain paragraphs (26% of all occurrences in the English corpus, 19% in the Spanish one). In these instances the textual structuring level interacts quite closely with the interpersonal level. Most metaphorical relations have abstract objects and physical vehicles, i.e. an abstract entity is talked about in physical terms. On the assumption that physical concepts are easier to grasp than abstract ones, it is to be expected that writers will use metaphorical expressions as interpersonal devices to attract and maintain the reader’s attention. Certain parts of texts seem to require such resources to a higher degree: titles, subtitles, opening paragraphs, the first sentence of any paragraph, closing paragraphs and their closing sentences, and a few others. The expected function of the title and the first paragraph is to introduce the subject, to direct the reader’s attention to the text, and to set its tone. The introduction of the subject matter requires the activation of a given area of the reader’s background knowledge before the topic can be set. This is done by making reference to known information (linguistically marked with definite NPs) by means of words and expressions used conventionally. The kinds of metaphorical expressions to be expected in the first paragraph of a text are conventional ones and clichés. However, the text structuring function must be combined with the interpersonal, and the demands may be quite different. If the reader has to be actively involved, one type of involvement may be derived from a context where conventional metaphorical expressions and clichés are returned to life. This can be achieved by various means, among them the recurrence of the same metaphor through different metaphorical expressions, and word-play on metaphorical expressions. Consider the passages in (18) and (19). (18) Title: “Online broking: Fizz, pop” Disaster, for an Internet broker, is not when the market goes down; it is when the wires stop pulsing. In the current slow slump in share prices, that is just what has happened. As if a switch had been thrown, the buzz has gone from online trading. [new paragraph] The result is that the tech bubble is deflating not with a bang
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.16 (174)
Anna Espunya and Patrick Zabalbeascoa
and a panic sell-off, as has been feared, but with a whimper. [there have been no mass redemptions nor outright dumping of portfolios] (19) From the condemned home of sleazy stockbroking practices to the shining temple of cutting-edge technology-company shares in only five years, Nasdaq’s image has undergone quite a transformation. In (18) the metaphorical potential of the word ‘bubble’ (whose occurrence in half of the texts would suggest its status as a term or at least as a cliché) is activated in the title by the words ‘fizz’ and ‘pop’ (noises that involve the outflow of air). In the first paragraph it appears as the subject of the predicate ‘deflate’ accompanied by the manner adjuncts containing ‘bang’ and ‘whimper’, the noises that are expected when air goes out of an object. The word ‘bubble’ occurs in other texts in combination with the alliterative partner ‘burst’. Activation of inactive metaphors grabs the reader’s attention and directs it to some textual purpose. We thus disagree with Goatly, who considers grabbing the reader’s attention a “decoration” function. The metaphorical expressions in example (19), which is the opening paragraph of a text, realise a combination of functions: both ‘sleazy’ and ‘shining’ are hyperboles; the former is a euphemism (as it avoids making reference to the specific accusations), and between the two there is a relation of word play (alliteration). Both hyperbole and alliteration serve the attention-grabbing function expected of opening paragraphs in the journalistic genre to which the text belongs. Other examples of reactivated clichés and dead metaphorical expressions that occur in opening paragraphs are (20) and (21). (20) The wild ride in the world’s stockmarkets grows ever wilder. Time and again, plunging prices have caused investors to ask if this is the crash, only for the market to turn around. (21) La bolsa vive un momento difícil. De las 35 empresas del Ibex, 15 están en pérdidas y el revés de los valores tecnológicos ha dejado huérfano al mercado. ‘The stock market is going through a difficult period. Out of 35 Ibex companies, 15 are sustaining losses, and the setback of the technological values has orphaned the market.’ Note in (20) how the interaction of the expressions ‘wild ride’, ‘ever wilder’, ‘plunging prices’ (alliterated) and ‘crash’, all of them with vehicles that denote sudden and uncontrollable movement, results in the reactivation of each individual expression, which in isolation would be considered as highly conventionalised clichés or even terms. Similarly in (21), several expressions with vehicles denoting loss (pérdidas ‘losses’, huérfano ‘orphan’) or negative situation (vivir un momento difícil ‘go
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.17 (175)
Metaphorical expressions in stock market journalistic texts
through a difficult period’, el revés ‘the setback’), get their metaphorical potentials reactivated. The difference in frequency of metaphorical expressions in openings and closures may be attributed to different writing conventions in the two cultures. In Spanish, loading an introductory paragraph with metaphors is a much less frequent resource for attracting the reader than in English. In general, the Spanish journalistic prose prefers background-giving introductions. The link between writer and reader is formed by establishing a common ground: the writer refreshes known information for the readers, and uses current clichés familiar to them through other media products. Only after the reader has been updated is the topic of the text introduced. Metaphorical expressions are used as formulae for text-structuring purposes at other points in a text. Among them is introducing a list of new discourse referents. (22) Although Microsoft may be more truculent than most, it is not unique in having characteristics that could attract an antitrust investigation: [. . . ]. Look across the technology landscape at leaders such as Intel (chips), Oracle (database management), Cisco (routers), and Sun Microsystems (servers). All these [. . . ] Note also how the imperative draws the reader in the text, certainly part of the interpersonal function. Metaphor reactivation is not a necessary condition, at least for certain text structuring functions. The Spanish texts analysed do not abound in metaphorical expressions and even fewer of them get back their metaphorical life. However, the Spanish texts are full of clichés and idioms performing the same functions. They can be considered formulae for signalling certain invariable textual functions, for example presenting an entity that stands out, e.g. the company that pays the highest dividends to its investors, as in (23), and for stating the existence of a choice of alternatives, as in (24), parallel to (22) above: (23) a.
Aumar es la verdadera estrella. ‘Aumar is the true star (i.e. tops the list).’ b. La deuda que presenta mejor cara es la norteamericana. ‘The bonds with the best face (i.e. the most attractive)’ c. Cinco valores clave que suponen más del 40% del Ibex tienen la palabra. ‘Five key companies which amount to more than 40% of the Ibex have the word (i.e. have the floor).’
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.18 (176)
Anna Espunya and Patrick Zabalbeascoa
(24) Las previsiones [. . . ] y la ganancia adicional [. . . ] amplían el abanico de posibilidades. ‘The forecasts [. . . ] and additional gains [. . . ] widen the fan of possibilities (i.e. widen the range of possibilities).’
.. Conveying attitudes Metaphor, or rather the use of metaphor instead of a literal formulation may signal the author’s attitude toward the situation being described, as illustrated in (25). (25) Si algo caracteriza la forma de actuar de los fondos de inversión y de pensiones internacionales es que, aunque no son especulativos, sí son volubles: las oportunidades de inversión no suelen durar mucho y, una vez alcanzados sus objetivos de revalorización, echan a volar en busca de valores frescos. ‘If anything characterizes the way international investment and pension funds act, it is the fact that, even though they are not speculative, they are indeed voluble: investment opportunities do not last long, and once their goals regarding value increase have been attained, they fly away in search of fresh stocks.’ Foreign investors are rendered as birds in the predicate though not explicitly in the subject. Rather than directly asserting their lack of commitment to the companies and markets from which they draw profits, the author proceeds metaphorically. Even if it is done through humour, the metaphor conveys a value judgment. Clichés and idioms may be equally effective. In fact, they help to achieve familiarity or intimacy with the reader, especially because of their frequency and colloquial nature. (26) Undertakings China made to get into the World Trade Organization will allow foreigners to underwrite Chinese securities. But regulators acknowledge that this is a fat lot of good if they cannot distribute them too. [new paragraph] Yet all this has not, so far, had foreign houses leaping up and down for joy.
.. Humour and games Undoubtedly, humour has an interpersonal function: a sense of community with the writer develops when the reader accepts the challenge to decode the word-play on a metaphor. The journalistic genre may resort to this in English, but not in Spanish. Humour is an essential feature in many of The Economist articles and is provided mainly through the interplay between the main text and the visual elements – pictures and drawings – that accompany it. Captions are usually exploited
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.19 (177)
Metaphorical expressions in stock market journalistic texts
as bridges between the two. Given a key word in the main text, the picture may represent a different sense of that polysemous word, often a literal sense. The picture may be the instrument for reactivation of a dead metaphorical expression by realising the vehicle of the expression visually. Consider the examples in (27). (27) a.
Article on Yale University’s Endowment Fund manager: picture of the manager with the caption well-endowed. b. Article on the effects of antitrust measures, entitled Monopoly money: drawing of the “Monopoly” boardgame. c. Sentence in an article on the NYSE: “It may not only be its dogged adherence to a trading floor or its system of specialists that knocks the NYSE off its perch as the world’s leading exchange.” Picture: of the NYSE president holding a dog. Caption: dog days for Grasso.
In (27a), from a text on David Swensen, Yale University’s Endowment Fund manager, The Economist exploits the polysemy of the word ‘endowment’ (money or property; talent) in its morphological variant ‘endowed’ (passive), to comment on his performance. In (27b) the text dealing with monopolistic practices is illustrated with a drawing depicting the game Monopoly with its board, tokens and players, the latter depicted as the prototypical tycoons. In (27c) each instance of the form ‘dog’ takes on a different sense. Even without a visual counterpart, puns contribute to the reactivation of dead metaphors and establish a closer tenor with the reader, as in (28) on the demise of online brokerage. (28) As if a switch had been thrown, the buzz has gone from online trading [l. 4]. One thing that traditional stockbrokers such as Merrill have always been good at is holding a client’s hand during tough times. Try doing that with a mouse [l. 56–57].
. Summary and conclusions English and Spanish have been compared with respect to the way metaphorical expressions are put to use at the three metalevels of discourse: ideational, interpersonal and textual. Both languages have terminology of metaphorical origin. Both languages use quasi-synonyms of terms of metaphorical origin either to highlight a certain component of the term, or in order to achieve lexical cohesion without resorting to repetition. In both languages, metaphors based on analogy are used to introduce or explain specialised concepts, and to express irony or value judgement.
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.20 (178)
Anna Espunya and Patrick Zabalbeascoa
English has many metaphorical relations based on abstract concepts that can be referred to by words that refer to physical experiences. This is exploited in popularising genres such as those represented in our corpus in two ways: firstly, metaphorical relations between an abstract object and a concrete vehicle lower the formality of the language, an almost inevitable feature of the register forced by the subject matter. Secondly, because of their complex nature, metaphorical expressions tend to contain a large amount of information in a small amount of linguistic material and thus enhance conciseness. Probably the clearest cases are “popular terms”, colloquial shorthands for institutional titles, which can even be created ad hoc to refer to entities introduced in the discourse which are new to the reader. English exploits metaphorical relations for rhetorical purposes, i.e. the identification of a metaphorical relation is used to focus on particular semantic components, to establish contrasts, for word games and other means of achieving familiarity with the reader. In Spanish, a less formal register is achieved not only by metaphorical expressions, but also by other mechanisms, which tend to be lexical (highly conventionalised idioms), or syntactic (special word order, reflecting prosodic patterns of focus and other features of speech, as opposed to writing; coordination rather than subordination, etc.). Metaphorical expressions are not used for rhetorical purposes, and except for instances of metaphor extension where the analogical basis of the metaphor serves a clear ideational, explanatory purpose, they are not expected to be interpreted as metaphors. Consequently, English and Spanish differ in the range of reactivation mechanisms available for inactive metaphorical expressions. Both languages share metaphor extension (with or without rhetorical purpose) as the main factor. Undoubtedly a cluster of metaphorical expressions, no matter how conventionalised they may be in isolation, achieves a certain lexical prominence in the text. Even the members of small groups like pairs reinforce each other’s reactivation capacity. Linguistic devices that direct the reader’s attention to the form rather than the content – wordplay such as alliteration and puns (visual and textual); manipulation of idioms or popular sayings – can help bring metaphorical life back to highly conventionalised expressions. The difference between English and Spanish lies in the writing conventions for journalistic genres and how they strike the balance between information and entertainment. Focus on form, except for the carefully signalled introduction of neologisms, is not a device commonly allowed in many genres and publications in Spain. The counterpart to metaphor activation is metaphor deactivation. Clusters of metaphors can have a “priming” effect on certain lexical components shared by all of the individual items. The lack of common elements leads to metaphor mixture. Lexical networks which exclude all readings but the conventionalised one are, naturally, adverse environments for reactivation of metaphorical life.
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.21 (179)
Metaphorical expressions in stock market journalistic texts
One of the conclusions to be drawn from this paper concerns the usefulness of contrastive text studies for the field of translation studies. The idea that there are differences in how metaphorical expressions are used in text structuring, and that those differences affect registers and rhetorical functions implies that the translation of metaphor cannot be approached exclusively as a lexical phenomenon or even a literary or cultural one. Translators will have to consider metaphor as yet another linguistic mechanism and know how it is used in their source and target languages. The notion that the life of metaphors is a dynamic feature, measurable only in each context, dependent on the function of each metaphorical expression, has consequences for approaches that subordinate translation solutions to a static degree of conventionality.
Appendix l. 20–23
This bundle of measures, together with the controversy surrounding the President of Telefónica [have set the engine of the stock exchange in neutral]1. l. 40 Why this crossroads1? l. 71–73 The two subsidiaries of the company [. . . ] also have lost ground1. l. 186–188 Certainly in Spain the market [has become paralysed]1. [. . . ] In this situation, what attitude are the foreign investors taking? l. 193–199 Until May they were buyers [. . . ]. But in the last weeks they have decided to [stop the engines]1. Although the general feeling is that there has not been [a mass exodus]2, some of those funds have indeed reduced their positions in the Spanish market and have replaced them with other European markets. l. 211–219 If anything characterises the way international investment and pension funds act, it is the fact that, even though they are not speculative, they are indeed voluble: investment opportunities do not last long, and once their goals regarding value increase have been attained, they [fly away in search of fresh stocks]2. l. 234 and 250–254 No two investors are alike. [. . . ] However, they have become [the engines that move the markets upwards or downwards]1.
TSL[v.20020404] Prn:5/09/2002; 9:40
F: PBNS27.tex / p.22 (180)
Anna Espunya and Patrick Zabalbeascoa
Notes . The research reported in this paper is part of a project financed by the Spanish Government, code PB98-1062-C04-01. The paper was presented at the Second International Conference on Contrastive Semantics and Pragmatics, held in Cambridge in September 2000. . The discourse field of the stock exchange and economics in general would allow for an excellent opportunity to test the experientialist hypothesis, given that other Spanish-speaking countries have different economic systems. A member of the audience at SIC-CSP2000 suggested that texts written in other countries might not realise the “good is up” and “more is up” metaphors. We would rather hypothesise that they might apply the same metaphors to other entities and actions. . Our description would not be invalidated by the discovery that some of the Spanish terms are loanwords. We are concerned with how metaphors are used and activated in texts. It is at that semantic and textual level that we would be concerned with calques. . Our bold. From this point onwards, boldface signalling the metaphorical items under discussion is ours.
References Allen, R. E. (1991). The Concise Oxford Dictionary of Current English (8th ed.). Oxford: Clarendon. Goatly, A. (1997). The Language of Metaphors. London: Routledge. Halliday, M. A. K., & Hasan, R. (1976). Cohesion in English. London and New York: Longman. Kronfeld, Ch. (1981). Novel and Conventional Metaphors. A Matter of Methodology. Poetics Today, 2(1), 13–24. Lakoff, G., & Johnson, M. (1980). Metaphors We Live By. Chicago: The University of Chicago Press. Leech, G. (1969). A Linguistic Guide to English Poetry. Harlow: Longman. Newmark, P. (1981). Approaches to Translation. Oxford: Pergamon. Richards, I. A. (1965). The Philosophy of Rhetoric. Oxford: Oxford University Press. Snell-Hornby, M. (1995). Translation Studies: An Integrated Approach. Amsterdam: John Benjamins.
TSL[v.20020404] Prn:8/04/2002; 15:26
F: PBNSP03.tex / p.1 (1)
P III
Cross-cultural pragmatics and speech acts
TSL[v.20020404] Prn:4/11/2002; 9:37
F: PBNS28.tex / p.1 (183)
Directions of regulation in speech act theory Susumu Kubo Matsuyama University, Japan
Directions of fit In Searle and Vanderveken’s (1985) speech act theory,1 the correspondence relation between words and world (or objects) is properly explained in terms of the theory of directions of fit that consists of four and only four types of directions of fit, namely words-to-world, world-to-words, double, and empty directions of fit.2 For instance, assertives such as stating or claiming have words-to-world direction of fit and represent how things are. Commissives such as promising or swearing and directives such as ordering or requesting have world-to-words direction of fit and get the world to be transformed by the future course of action either by the speaker (commissives) or by the hearer (directives). Declaratives such as appointing and naming have double direction of fit and get the world to match the propositional content by saying that the propositional content matches the world. And expressives such as thanking and welcoming have empty direction of fit and express mental states of the speaker about the state of affairs represented by the propositional content. In contrast, the relations between the participants in a dialogue have not been studied properly even if their importance has been emphasized.3 They are treated as if they were the features of either preparatory conditions or modes of achievement of single illocutionary acts.4 For instance, in Vanderveken (1990), those features of illocutionary acts are described in the description of the following illocutionary force naming verbs: (1) a.
Reject/Refusal: Like acceptances, rejections and refusals have the additional preparatory conditions that one has the option of accepting or rejecting/refusing. [p. 185]
TSL[v.20020404] Prn:4/11/2002; 9:37
F: PBNS28.tex / p.2 (184)
Susumu Kubo
b. Direct: Most actual directive forces have a special mode of achievement of their illocutionary point in that generally it is clear that the hearer either has or has not the option of refusal. [p.189] c. Request: A request is a directive illocutionary act that allows the option of refusal. [p.189] d. Invite: There is an option of refusal in this mode of achievement. [p.191] e. Order: To give an order is to demand of the hearer that he do something while invoking a position of authority over him (special mode of achievement),5 while to issue a command is just to give an order from a position of authority. [p.194] Each condition in (1), however, specifies the relation between participants in a dialogue that reflects the success and the satisfaction of the dialogue, because both one’s option of acceptance/refusal and one’s invocation of a position of authority are guaranteed if and only if he is somehow in a superior position to the other speaker. In other words, these kinds of conditions have their significance only if they are studied as the features of dialogue but not those of a single utterance or illocutionary act. In the next section, I would like to demonstrate a complementary theory to the theory of directions of fit, namely directions of regulation that will explain the relation between the participants in a dialogue. However, the introduction of new theory does not mean the unnecessary complication or multiplication of speech act theory. In the third and the fourth section, I will show how this theory fits in the traditional framework of speech act theory and how it helps us to formulate the conditions of success and satisfaction of a dialogue and extend speech act theory to the theory of dialogue. The last section is for summation and a few remarks on relevant linguistic phenomena.
Directions of regulation What is regulation? In ordinary dialogues,6 we usually find types of regulatory actions performed by the participants. Let us observe several dialogues such as a pleading-dialogue, a threatening-dialogue, a mutual-retort-dialogue and a suggestion-acceptance dialogue, in turn and see how we regulate ourselves to other speakers in dialogues. The pleading-dialogue in (2) shows how the pleader regulates herself to the pleaded. (2) “Take them back,” he commanded, his voice dripping with ice. “Now.” “Oh, sir,” pleaded Grace through the lump rising in her throat.
TSL[v.20020404] Prn:4/11/2002; 9:37
F: PBNS28.tex / p.3 (185)
Directions of regulation in speech act theory
“She only got here.” She felt far worse for Annie than she did for herself, although it was her own neck and her own job that were on the line here. [Annie, 52] In this context, Grace, the pleader, pleads with her master, the pleaded, not to make her take Annie back home without regard to her own risk, in least expectation of her master’s generous response because she believes that her master is a tycoon and does not listen to other’s opinion. As a reply, Grace has to face her master’s silence that signifies his arrogant ignorance.7 In other words, it is only the pleader that regulates herself to another speaker, but there is no regulatory action from the pleaded. In contrast, in a threatening-dialogue in (3), in reply to the threat, Pepper’s silence signifies the obedience to the threatener hiding in the laundry cart. It is common knowledge that silent obedience is a kind of regulation from the inferior to the superior.8 Whenever a speaker threatens another speaker to do something, it is not the threatener but the threatenee that has to regulate himself to the other person in a dialogue. (3) Pepper thrust her lower lip out. “I’m gonna tell.” “And I’ll rearrange your teeth,” threatened the laundry cart.
[Annie, 14]
Then, in a mutual-retort-dialogue such as (4), either of the participants does not listen to his/her opponent’s pretenses, but contends that only his claim is correct and tries to refute the other. Therefore, neither of them allows his opponent’s option of refusal and no regulatory action is taken between them. (4) “You always think you’re right!” Ryuji had said in the end. Sachiko had retorted, “You’re the one who’s always right!”
[Knock, 59]
Lastly, examine a suggestion-acceptance dialogue in (5). (5) “Let’s go for a walk,” he suggested. What he needed was a breath of air. “Okay,” said Annie, standing up immediately. [Annie, 94] In cases like this, whenever a dialogue is negotiative, both parties of the dialogue intend to regulate themselves to reach their common goal.9 As stated above, regulation is directional. From a logical point of view, it is the relation between two different entities, namely the speaker and the hearer, so the number of combination of the types is the square of 2. That is to say, as in the case of directions of fit, there are four and only four types of directions of regulation. They are the unidirectional regulation from the 1st speaker to the 2nd speaker, the unidirectional regulation from the 2nd speaker to the 1st speaker, the
TSL[v.20020404] Prn:4/11/2002; 9:37
F: PBNS28.tex / p.4 (186)
Susumu Kubo
null-directional regulation between the 1st speaker and the 2nd speaker, and the bi-directional regulation between the 1st speaker and the 2nd speaker.10 In this way, the theory of directions of regulation contributes to the typology of dialogues or minimal unit of discourses.11 What we need to do is generalize the condition under which the directions of regulation obtain.
Definitions of directionalities of dialogues12 In conversation, participants develop their talk by selecting their optimal action to take at the very moment of the interaction.13 The choice of the direction of regulation is also a type of the optimal action. Thus, in the theory of dialogue, it would be correct to assume that every competent speaker knows types of directions of regulation, presupposes one of them whenever s/he participates in a dialogue, and performs the regulation act according to the selection of the type. Based upon these assumptions, the directional properties of dialogues are defined consecutively in (6) and (7). (6) a.
A dialogue is unidirectional from the 1st speaker to the 2nd speaker if and only if it is an open dialogue and only the 1st speaker regulates himself to the other speaker in the dialogue. b. A dialogue is unidirectional from the 2nd speaker to the 1st speaker if and only if it is an open dialogue and only the 2nd speaker regulates himself to the other speaker in the dialogue. c. A dialogue is null-directional if and only if it is a closed dialogue and neither the 1st speaker nor the 2nd speaker regulates himself to the other speaker in the dialogue. d. A dialogue is bi-directional if and only if it is a closed dialogue and both the 1st speaker and the 2nd speaker regulate themselves to each other in the dialogue.
(7) a.
The 1st speaker regulates himself to the other speaker if and only if the 2nd speaker has the option of refusal to the master illocutionary act by the 1st speaker in the dialogue.14,15 b. The 2nd speaker regulates himself to the other speaker if and only if the 2nd speaker does not have the option of refusal to the master illocutionary act by the 1st speaker in the dialogue. c. Neither the 1st speaker nor the 2nd speaker regulates himself to the other speaker in the dialogue if and only if neither of them has the option of refusal to the master illocutionary act by the other speaker.
TSL[v.20020404] Prn:4/11/2002; 9:37
F: PBNS28.tex / p.5 (187)
Directions of regulation in speech act theory
d. Both the 1st and the 2nd speaker regulate themselves to each other if and only if both of them have the option of refusal to the master illocutionary act by the other speaker. The definitions in (7) clearly show how directions of regulation and features allocated to the preparatory conditions in traditional speech act theory are co-related. Then, the definitions of a closed and an open dialogue are given in (8). (8)16 a.
A dialogue is closed if and only if it requires a particular illocutionary act with which its master illocutionary act is always in an adjacency pair relation. b. Otherwise, a dialogue is open.
Directions of regulations and types of dialogues In this section, for explication, English dialogues are studied according to the types of directions of regulations.
1st-to-2nd unidirectional type In English dialogues, members of this type are very few. They are pleading, reguesting, imploring, entreating, appealing and supplicating dialogues.17 Pleadingdialogues in (9) show how the pleader regulates herself to the pleaded. (9(=2))“Oh, sir,” pleaded Grace through the lump rising in her throat. “She only got here.” She felt far worse for Annie than she did for herself, although it was her own neck and her own job that were on the line here. [Annie, 52] In the context (9), this dialogue is 1st-to-2nd unidirectional because it is a closed dialogue consisting only of the 1st speaker’s utterance without any direct verbal reply from the 2nd speaker, and the 2nd speaker has the option of refusal to the master illocutionary act by the 1st speaker in the dialogue.
2nd-to-1st unidirectional type Warning-dialogue in (10) is an example of a dialogue that belongs to this type since it satisfies the conditions to be of this type. (10) “You’d better not sound an alarm or anything. That would only hasten your end,” the woman warned him, her pistol at the ready.18 Komazawa switched off the stereo. [Knock, 90]
TSL[v.20020404] Prn:4/11/2002; 9:37
F: PBNS28.tex / p.6 (188)
Susumu Kubo
This dialogue is 2nd-to-1st unidirectional because it is an open dialogue consisting only of the 1st speaker’s utterance without any direct verbal reply from the 2nd speaker, and the 2nd speaker does not have the option of refusal to the master illocutionary act by the 1st speaker in the dialogue.
Null-directional type Members of the dialogue of this type are limited to a group of dialogues that share the property of mutual-denial or rejection such as a mutual-retort dialogue in (11). (11(=4))“You always think you’re right!” Ryuji had said in the end. Sachiko had retorted, “You’re the one who’s always right!”
[Knock, 59]
This dialogue is null-directional because it is a closed dialogue and neither of the participants has the option of refusal.
Bi-directional type Other dialogues belong to this type. Let us observe a suggestion-acceptance dialogue in (12) and a begging-acceptance dialogue in (13). (12) “Look, won’t you have some coffee?” he suggested. “It probably won’t be as good as La Mer’s but. . . ” “Thanks, I will.” Akiko nodded and sat down in a nearby chair. [Knock, 81] (13) “Sing something, Annie,” she begged. “Please? Sing ‘Minnie the Moocher.’ Please, Annie. It will help me sleep.” Annie laughed and nodded.19 [Annie, 10] These dialogues are bi-directional because they are closed dialogues and both the 1st speaker and the 2nd speaker regulate themselves to each other in the dialogue. In the latter case, the 2nd speaker responds non-verbally with laughing and nodding, which function as regulation.
Conditions of success of a dialogue As Vanderveken (1990: 26) defines them, “the conditions of success of an illocutionary act are the conditions that must obtain in a possible context of utterance in order that the speaker succeeds in performing that act in that context”. Then, in the same manner, the conditions of success of a dialogue are defined as the conditions that must obtain in a possible context of dialogue in order that the speaker succeeds in performing the master illocutionary act in that dialogue. For example, pleading-dialogue is successful if and only if the participants of the dialogue
TSL[v.20020404] Prn:4/11/2002; 9:37
F: PBNS28.tex / p.7 (189)
Directions of regulation in speech act theory
presuppose the addressee’s superior position over the addresser’s and presuppose the latter regulates himself to fulfill the dialogical goal and its master illocutionary act, namely, pleading, is successful. Warning-dialogue is successful if and only if the participants of the dialogue presuppose the addresser’s superior position over the addressee’s and presuppose the latter regulates himself in order to fulfill the dialogical goal and its master illocutionary act, warning, is successful. Mutual-retortdialogue is successful if and only if the participants of the dialogue presuppose that the regulation is not performed by themselves and its master illocutionary act, retort, is successful. Suggestion-acceptance dialogue is successful if and only if the participants of the dialogue presuppose that they regulate themselves in order to carry out the discursive goal of the dialogue and one of its master illocutionary acts, suggestion, is successful.20 Therefore in general, (14) a.
A dialogue is successful if and only if its direction of regulation and its master illocutionary act are both successful. b. A direction of regulation in a dialogue is successful if and only if it is presupposed by the participants of the dialogue.
For instance, (15) Pleading-dialogue is successful if and only if the participants of the dialogue presuppose its 1st-to-2nd unidirectional regulation and its master illocutionary act, pleading, is successful. And its master illocutionary act, pleading, is successful if and only if: (16) a.
The 1st speaker pleads the acceptance of the propositional content of the act by the 2nd speaker (illocutionary point), b. The 1st speaker speaks humbly and enthusiastically, paying his respect to the 2nd speaker (mode of achievement), c. The 1st speaker presupposes that (a) his proposal is not agreeable/happy for the 2nd speaker, (b) the 2nd speaker is able to accept his proposal, but it is quite unlikely that he will do it (preparatory conditions), d. The 1st speaker earnestly wishes the 2nd speaker to accept his plea (sincerity condition).
Likewise, (17) Warning-dialogue is successful if and only if the participants of the dialogue presuppose its 2nd-to-1st unidirectional regulation and its master illocutionary act, warning, is successful. And, warning is successful if and only if:
TSL[v.20020404] Prn:4/11/2002; 9:37
F: PBNS28.tex / p.8 (190)
Susumu Kubo
(18) a.
The 1st speaker makes the second speaker commit to his future course of action (illocutionary point), b. Propositional content is something realized by the 2nd speaker in his future course of action (propositional content condition), c. The 1st speaker presupposes that the 2nd speaker is capable of obeying the warning (preparatory condition), d. The 1st speaker intends to make the 2nd speaker obey his warning (sincerity condition).
Then, (19) Mutual-retort-dialogue is successful if and only if the participants of the dialogue presuppose its null-directional regulation and its master illocutionary act, retort, is successful. And, retort is successful if and only if: (20) a.
The speaker intends to represent how things are in the world (illocutionary point), b. The speaker has the perlocutionary intention of rebutting what his opponent has said (mode of achievement), c. The speaker presupposes what his opponent has said is wrong (preparatory condition), d. The speaker firmly wants his claim to be accepted by his opponent (sincerity condition).
Lastly, (21) Suggestion-acceptance dialogue is successful if and only if the participants of the dialogue presuppose its bi-directional regulation and its master illocutionary act, suggestion, is successful. And, suggestion is successful if and only if: (22) a.
The 1st speaker lets the 2nd speaker accept the propositional content of the act (illocutionary point), b. Propositional content is what will happen in the future course of action (propositional content condition), c. The acceptance of his suggestion is good for the 2nd speaker (preparatory condition), d. The 1st speaker hopes that the 2nd speaker accepts his proposal (sincerity condition).
TSL[v.20020404] Prn:4/11/2002; 9:37
F: PBNS28.tex / p.9 (191)
Directions of regulation in speech act theory
Conditions of satisfaction of a dialogue The conditions of satisfaction of an illocutionary act are conditions that must obtain in the world of a context of utterance in order that that act be satisfied in that context (Vanderveken 1990: 27). For instance, an illocutionary act such as promising is satisfied if and only if the person who promised keeps it. An illocutionary act such as ordering is satisfied if and only if the person who is ordered obeys it. Similarly, the conditions of satisfaction of a dialogue are conditions that must obtain in the world of a context of dialogue in order that that dialogue be satisfied in that context. For instance, pleading-dialogue is satisfied if and only if the act of regulation from the person who pleads to the person who is pleaded is performed and the latter accepts the plea from the former. Warning-dialogue is satisfied if and only if the act of regulation from the person who is warned to the person who warns is performed and the former takes warning from the latter. Mutual-retort-dialogue is satisfied if and only if no act of regulation is performed between the participants in the dialogue and the retorted acknowledges the retorter’s statement as true. Then, suggestion-acceptance dialogue is satisfied if and only if both participants of the dialogue perform acts of regulation one after another, and the person who is suggested accepts the suggestion offered by the person who suggests. Consequently, the condition of satisfaction of a dialogue is defined in general as shown in (23). (23) a.
A dialogue is satisfied if and only if the act of regulation of the dialogue and its master illocutionary act are both satisfied. b. An act of regulation of the dialogue is satisfied if and only if the participants of the dialogue perform it.
Thus, the conditions of satisfaction of pleading-dialogue, warning-dialogue, mutualretort-dialogue and suggestion-acceptance dialogue are given from (24) to (27). (24) Pleading-dialogue is satisfied if and only if the 1st speaker performs its 1st-to-2nd unidirectional regulation act and its master illocutionary act, pleading, is satisfied. (25) Warning-dialogue is satisfied if and only if the 2nd speaker performs its 2nd-to-1st unidirectional regulation act and its master illocutionary act, warning, is satisfied. (26) Mutual-retort-dialogue is satisfied if and only if the participants of the dialogue perform its null-directional regulation act and its master illocutionary act, retort, is satisfied. (27) Suggestion-acceptance-dialogue is satisfied if and only if the participants of the dialogue perform its bi-directional regulation acts and its master illocutionary act, acceptance, is satisfied.
TSL[v.20020404] Prn:4/11/2002; 9:37
F: PBNS28.tex / p.10 (192)
Susumu Kubo
Here, the satisfaction of the illocutionary act of suggestion need not be included in the condition of satisfaction of suggestion-acceptance-dialogue, since it is presupposed in the satisfaction of the illocutionary act of acceptance.21 Also, the conditions of satisfaction of their master illocutionary acts are given below. (28) Pleading as an illocutionary act is satisfied if and only if the 2nd speaker accepts it. (29) Warning as an illocutionary act is satisfied if and only if the 2nd speaker obeys it. (30) Retort as an illocutionary act is satisfied if and only if it is acknowledged as true. (31) Acceptance as an illocutionary act is satisfied if and only if the 2nd speaker carries it out.
Conclusion In this paper, in the conviction that human actions are under regulation among them, major tools of speech act theory are reconsidered in the hope that the theory can be developed into the theory of dialogue. In the first section, it was claimed that features inherent to dialogue and those inherent to single illocutionary acts must be distinguished among the components of conditions of success. Modes of achievements and preparatory conditions were under scrutiny. It was argued that modes of achievements and preparatory conditions representing the relation between the participants in the dialogue must be treated as the features inherent to dialogues but not to single illocutionary acts. Then, in the following section, a new theory of directions of regulation that is a complementary theory to the theory of directions of fit was proposed to describe properly the relations between participants in dialogues. In the third and fourth sections, the conditions of success and satisfaction of dialogues were respectively formulated. On the one hand, the participants of dialogues are presumed to presuppose one of the directions of regulation in the conditions of success of a dialogue. On the other hand, they are presumed to perform one of the acts of regulation in the condition of satisfaction. This paper has so far been confined to the construction of a general and philosophical theory of dialogue without direct commitment to linguistics. Thus, it appears that the new theory of directions of regulation has nothing to do with linguistics. However, the opposite is true. If one analyzes expressions such as ‘if you like’, ‘if you please’ and ‘if you don’t mind’, one will soon notice that they are regulation-
TSL[v.20020404] Prn:4/11/2002; 9:37
F: PBNS28.tex / p.11 (193)
Directions of regulation in speech act theory
based expressions. They do not only represent the regulation from the 1st speaker to the 2nd speaker, but also invite the regulation from the 2nd speaker. Similar expressions are all around us.22 Moreover, these expressions are related to speakers’ mental and social attitudes, thus are far from irrelevant to the theory of politeness. Therefore, the new theory will help us to introduce a new perspective to the analysis of attitudes as well as the analysis of politeness. It will also help us to construct the reasoned dictionary of regulation-based expressions that would be the revision of the reasoned dictionary of illocutionary force naming verbs. The more we cultivate, the more we will gain.
Sources for examples [Annie] Fleischer, L. (1982). Annie. New York: Ballantine Books. [KNOCK] Hoshi, S. (1984). There was a knock. Tokyo: Kodansha International
Notes . They call it ‘the general semantics of success and satisfaction’. . ‘Illocutionary point, direction of fit, and sincerity condition seem to me the most important’ (Searle 1979: 5). For further information, see Vanderveken (1990). . ‘Like Francis Jacques I believe that the relationship of interlocution between speakers and hearers is as important as the traditional relation of correspondence between words and things in the philosophy of language in general and in discourse theory in particular’ (Vanderveken 1999: 10). . See Searle (1979: 5–6). Searle considers differences in the status of participants in a dialogue as the feature that ‘corresponds to one of the preparatory conditions’. . If a speaker X invokes a position of authority over another speaker Y, only if X is in a position of authority over Y. . A dialogue is a minimal unit of discourse consisting of consecutive verbal or non-verbal utterances performed by two different speakers. Please see Footnote 11. . We have to also note that it is in dialogue that a non-verbal indirect speech act has its significance in communication. . Silent obedience is an indirect non-verbal speech act, as well. . This is what Vanderveken (1994) calls discursive goal in his theory of discourse. . When a bi-directional regulation is successful, we call it co-regulated. . According to the general theory of discourse cultivated in Vanderveken (1999), ‘a discourse is not a sequence of individual illocutionary acts, but is a sequence of dialogues that is in turn a sequence of illocutionary acts. A dialogue is a minimal unit of a discourse and an
TSL[v.20020404] Prn:4/11/2002; 9:37
F: PBNS28.tex / p.12 (194)
Susumu Kubo
illocutionary act is a minimal unit of a dialogue. Thus from logical point of view, a minimal discourse consists of a single illocutionary act.’ . Directions of regulation are irreducible to directions of fit since there is no one-toone correspondence between them. In consequence, directions of regulation have to be independently studied. Compare them in Table 1. . The speaker develops his talk in the recognition of the affordance that is the selection of the optimal action to take at the very moment of the interaction. We conceive that interlocutors themselves function as their environment from which they are affected. As for the original idea of affordance, see Gibson (1979). . The notion of a master illocutionary act is the same as Vanderveken’s master speech act. For details, see Vanderveken (1994, 1999). . A speaker X has the option of refusal to the master illocutionary act performed by another speaker Y if and only if X is in the superior position to Y in the dialogue. . The name of an open dialogue is given according to the name of its master illocutionary act selected by the major speaker in the interaction. In contrast, the name of a closed dialogue is given according to the name of pair illocutionary acts in the interaction such as question-and-answer dialogue and request-and acceptance/rejection dialogue. . Sometimes, apology is classified in this class: the 1st speaker apologizes to the 2nd speaker for having made a fatal mistake, expecting in vain a word of pardon from the 2nd speaker who badly suffered from the 1st speaker’s mistake. . Not all quoting verbs specify the master illocutionary act of the quoted utterances in the quotation. For instance, ‘many speech act verbs such as announce, interject, and shout do not name an illocutionary force because they do not carry any restriction as to the illocutionary point or refer only to the feature of the utterance act’ (Vanderveken 1990: 167). . Here, we have to note that the 2nd speaker accepted the 1st speaker’s begging nonverbally. . The condition of success of acceptance needs not be included in the condition of success in a dialogue, since it is presupposed in part by the speaker who performs the illocutionary act of suggestion. In general, in a dialogue consisting of two consecutive illocutionary acts, the preceding act must be successful in order that the dialogue be successful.
Table 1. Directions of regulation, directions of fit and their master illocutionary acts Directions of regulation
Directions of fit
Types of master illocutionary acts
[1st→2nd] [2nd→1st]
[World→Words] [World→Words] [Double] [Words→World] [Empty] [World→Words] [Empty]
Commisives/Directives Commissives/Directives Declaratives Assertives Expressives Commissives/Directives Expressives
[Empty] [Double]
TSL[v.20020404] Prn:4/11/2002; 9:37
F: PBNS28.tex / p.13 (195)
Directions of regulation in speech act theory . In general, in a dialogue consisting of two consecutive illocutionary acts, the following act must be satisfied in order that the dialogue be satisfied. . I name these expressions ‘regulation force markers’ after ‘illocutionary force markers’, the typology of which will be constructed naturally in due course.
References Gibson, J. J. (1979). The Ecological Approach to Visual Perception. Boston: Houghton Mifflin. Searle, J. (1979). Expression and Meaning. Cambridge: Cambridge University Press. Searle, J., & Vanderveken, D. (1985). Foundations of Illocutionary Logic. Cambridge: Cambridge Univ. Press. Vanderveken, D. (1990). Meaning and Speech Acts. Cambridge: Cambridge Univ. Press. Vanderveken, D. (1994). Principles of Speech Act Theory [Cahiers d’Épistémologie, 9402]. Montréal: Université du Québec à Montréal. Vanderveken, D. (1999). Illocutionary Logic and Discourse Typology [Cahiers d’Épistémologie, 9912]. Montréal: Université du Québec à Montréal.
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.1 (197)
On Japanese ne and Chinese ba Mutsuko Endo Hudson and Wen-ying Lu Michigan State University
.
Introduction
The sentence-final particles ne1 in Japanese and ba in Chinese share many functional similarities, but exhibit some interesting differences. The objective of the present study is to demonstrate that a multi-dimensional discourse model such as Schiffrin’s (1987) is useful to locate the sources of the similarities and differences between them. The organization of the article is as follows. In Section 2, we explain the functions of ne and ba separately. In Section 3, we discuss the similarities and differences between ne and ba, and then analyze them in terms of information ownership. In Section 4, we present our analysis by applying Schiffrin’s (1987) model to ne and ba, and in Section 5, a brief summary and future directions.
. Functions of Japanese ne and Chinese ba . Functions of Japanese ne The Japanese particle ne is generally regarded as serving two main functions, namely, to solicit agreement or confirmation when appended to a statement, and to soften the tone of a sentence when appended to request, cohortative, intention and so forth, as shown in examples (1a–c).2 (1) a. [soliciting agreement] Kore oishii ne. this tasty-is b. [soliciting confirmation] A: Satoo-san desu ne? Ms. Sato cop
‘This tastes good, doesn’t it.’
‘You are Ms. Sato, right?’
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.2 (198)
Mutsuko Endo Hudson and Wen-ying Lu
B: Ie, watashi wa Katoo desu. no I top Kato cop
‘No, I’m Kato.’
c. [softening] Nihongo hanasoo ne? Japanese speak-coh
‘Let’s speak Japanese, OK?’
Ne is one of the most frequently used sentence-final particles in Japanese conversation (Maynard 1989), and, arguably, one of the most analyzed from various perspectives in modern Japanese linguistics. For example, from the performative perspective, ne functions to provide ‘rapport’ (Uyeno 1971). From the conversation management perspective, its use helps ‘intensify the level of involvement between conversation participants’ (Maynard 1989), it indicates ‘affective common ground’ (Cook 1992), and also acts as a turn-management device (Tanaka 2000). From an informational perspective, it expresses information shared by the speaker and the hearer (Kamio 1990, 1994); and from a cognitive perspective, it serves modalitylike functions of ‘estimate [yosoku] and assumption [sootei]’ (Takubo & Kinsui 1996). Nonetheless, the analysis of ne provided by Martin (1975: 914–917) from a descriptive perspective still holds true and serves as the basis for many of the analyses today. Kindaichi (1957: 170) observed that ‘[t]he speakers of Japanese hate to let a sentence end on a note of finality’, and this, according to Martin, ‘accounts for the tendency to attach a FINAL PARTICLE to impart some additional hint of the speaker’s attitude toward what he is saying...’ In the case of ne and its variants, he explains, they ‘soften a statement and invite confirmation on the part of the hearer’ (‘don’t you think?’, ‘n’est-ce pas?’), and are used to ‘involve both speakers and hearers in what is being said, as English speakers often do with you know, you see... etc.’ and by inserting ‘a vocative (Now, Tom, ...) or some generalized substitute for a vocative such as my dear and my friend.’ He also notes that a sentence ending in ne is often replied to by another with ne. Examples (2a, b) are from Martin (1975: 917).3 (2) a. Sayonara nee. Goodbye b. Samui ne. Soo ne. cold-is so
‘Goodbye now.’ ‘It’s cold, isn’t it.’ ‘Yes, it is, isn’t it.’
The functions of all three instances of ne in examples (2a, b) seem to be those of providing rapport, intensifying involvement between conversation participants who are on common ground affectively, and signalling a possible turn change, rather than expressing shared information or estimate/assumption.
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.3 (199)
On Japanese ne and Chinese ba
. Functions of Chinese ba Let us now turn to the functions of the Chinese particle ba. Like ne, it has long been claimed in the literature (e.g. Chao 1968) that its main functions are to solicit agreement/confirmation and to soften the tone of a sentence. The latter occurs often at the end of ‘request, advice, command, consultation, consent or agreement’ (Beijing Language Institute 1990). See examples (3a–c). (3) a. [soliciting agreement] Zhege hao chi ba. this good eat
‘This tastes good, doesn’t it.’
b. [soliciting confirmation] A: Ni shi Wang Taitai ba? ‘You’re Mrs. Wang, right?’ you cop Wang Mrs. B: Bu shi. Wo shi Gui Taitai. ‘No, I’m Mrs. Gui.’ no I cop Gui Mrs. c. [softening] Women shuo Zhongwen ba. we speak Chinese
‘Let’s speak Chinese, OK?’
Ba also adds a note of ‘supposition’ (uncertainty) and ‘compliance’ (DeFrancis 1963), as shown in examples (4a, b). (4) a. [supposition/uncertainty] Ta shi Meiguoren ba. he cop American b. [compliance] Hao ba. fine
‘I suppose he’s an American.’
‘It’s all right (with me).’ ‘Fine!’
Li and Thompson (1981: 307), likening the semantic function of ba to that of the Don’t you think so? or Wouldn’t you agree? type of question in English, state: ‘... ba has the effect of soliciting the approval or agreement of the hearer with respect to the statement to which ba is attached...’. This is why ba often occurs ‘as the marker for first person plural commands.’4 According to them (p. 308), when the subject of a command is the inclusive we, the sentence expresses the meaning ‘let’s...’, and when the subject is the second person, the sentence expresses advice (cf. ‘advisative particle’ in Chao 1968). Ba is also used in a sentence with the first person subject when the speaker has some reason to ‘request the hearer to agree’. (5) a. Women zou ba. we go
‘Let’s go!’
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.4 (200)
Mutsuko Endo Hudson and Wen-ying Lu
b. Ni he shui ba. you drink water
‘Why don’t you drink some water?’
c. [after being repeatedly toasted at a banquet] Wo he ban bei ba. ‘I’ll drink half a glass, OK?’ I drink half glass
. Comparison of ne and ba . Similarities Let us now compare ne and ba. There are at least three similarities. First, their main functions are the same, namely, to solicit agreement/confirmation when attached to a statement, and to soften the tone of other types of sentences such as request/invitation and intention. Examples (6c) were given earlier, but are repeated below for convenience. (Hereafter, ‘J’ marks Japanese sentences, and ‘C’ Chinese.) (6) a. [soliciting agreement] J: Kore oishii ne. this tasty-is C: Zhege hao chi ba. this good eat b. [soliciting confirmation] J: Satoo-san desu ne? Sato-Ms. cop C: Ni shi Wang Taitai ba? you cop Wang Mrs.
‘This tastes good, doesn’t it.’ ‘This tastes good, doesn’t it.’
‘You’re Ms. Sato, right?’ ‘You’re Mrs. Wang, right?’
c. [softening] J: Nihongo hanasoo ne? ‘Let’s speak Japanese, OK?’ Japanese speak-coh C: Women shuo Zhongwen ba. ‘Let’s speak Chinese, OK?’ we speak Chinese Second, both ne and ba are used in a close relationship, and their use adds a tone of familiarity and friendliness. They are not common when there is social and/or psychological distance between the speaker and the hearer or when one is to sound scientific or objective. Third, neither ne nor ba is felicitous in sentences like (7J, C), as the FIRST urging by the speaker to wake up someone asleep or unconscious. In both Japanese and Chinese, the sentences would be felicitous only if the addressee were lying awake
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.5 (201)
On Japanese ne and Chinese ba
or half-awake in bed, and not completely asleep or unconscious. In the latter case, another sentence-final particle yo would be used in Japanese. (7) [as the FIRST urging to someone asleep or unconscious] J: Oki-nasai { #ne/yo/ø}. ‘Get up { #OK?/OK?/ø}’ get up-com C: Qichuang { #ba/ø}. ‘Get up { #OK?/ø}’ get up Strictly speaking there is no reason for the function of softening not to work with the addressee who is asleep or unconscious. This, therefore, suggests that both ne and ba always ask for and expect some kind of reaction from the addressee. Such examples also suggest that, in exploring the use of sentence-final particles in either language, it is not enough just to look at the pragmatic domains such as softening, but that we must look into cognitive domains as well.5
. Differences We now turn to differences between ne and ba, of which there are at least six. First, in soliciting agreement, sentences with ba, like (8C), have much broader usage than those with ne, like (8J). For example, (8C) can be used in all of the situations (8a– d), while (8J) is possible only in situations (8a) and (8b). (8) J:
Kono uchi ii desu ne. this house good-is pol C: Zhe fangzi bucuo ba.
a. b. c. d.
‘This house is nice, isn’t it?’ ‘This house is nice, isn’t it?’
this house not-bad by a visitor to the house owner by a prospective buyer to a co-buyer by a realtor to a prospective buyer by a house owner to a visitor
Sentence (8C) would likely be uttered in situation (8a) when it was the visitor who had recommended the house to the owner, and in situation (8d) it may sound as though the speaker is showing off. The Japanese counterpart with ne, or (8J), is infelicitous in situations (8c) and (8d). In both situations, the speaker must say (9J) with deshoo (‘probably’; tentative form of the copula desu)6 because s/he is ‘guessing’ about the hearer’s assessment (Akatsuka 1990), and is ‘imposing his/her information on the hearer’ (Kimura & Moriyama 1992). Sentence (9J) used in situation (8d) would sound as if the speaker (‘house owner’) is bragging, just as the Chinese counterpart (8C) would.
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.6 (202)
Mutsuko Endo Hudson and Wen-ying Lu
(9) [in situations (8c) and (8d)] J: Kono uchi ii deshoo? this house good-is
‘This house is nice, isn’t it?’
(10C) is a similar example taken from Li and Thompson (1981: 309–310), in which the ba-version ‘attempts to solicit agreement from the hearer’ and sounds ‘accommodating and conciliatory’, and the ø-version is a statement in which the speaker shows his/her ‘anger’ while defending the referent.7 (10) C: Ta bu hui zuo zhe-yang-de shi { ø/ba }. s/he not know do such things ‘S/he wouldn’t do such things { ø!/don’t you agree? }’ J: Ano hito ga sonna koto suru wake nai { #ne?/deshoo?/yo }. that person nom such things do not ‘S/he wouldn’t do such things { #would s/he?/don’t you agree?/ I tell you! }’ In Japanese, a meaning similar to the ba-version in (10C) would have to be expressed with deshoo (‘probably’), and not with ne, as shown in (10J). This is because, like in (9J), the speaker is imposing his/her opinion (information) on the hearer. The ø-version implying the speaker’s anger would be expressed with yo. Examples like (9J, C) and (10J, C) suggest, again, the importance of considering cognitive domains and information status in analyzing sentence particles. Second, when soliciting confirmation, too, ne and ba are not completely alike. For example, (11J) with ne is felicitous, but not (11C) with ba, though it would be appropriate in other situations when the speaker genuinely seeks confirmation. Notice that in the Japanese version (11J) the speaker does not expect to be contradicted; ne is attached because the information belongs to the hearer’s ‘territory’, as will be discussed in Section 3.3. (11) J:
A: Katoo desu. Kato cop B: Katoo-san desu ne? Kato-Ms. cop C: A: Wo shi Gui Taitai. I am Gui Mrs. B: Gui Taitai { ø/#ba }? Gui Mrs.
‘I’m Ms. Kato.’ ‘Ms. Kato?’ ‘I’m Mrs. Gui.’ ‘Mrs. Gui?’
Ba, but not ne, can be used in a situation like (12) below, where the speaker checks the hearer’s understanding as s/he develops his/her narrative. In Japanese, deshoo (‘probably’) must be used, as shown in (12J).
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.7 (203)
On Japanese ne and Chinese ba
(12) J:
Nanno kanno yuttemo, atashi Nihonjin { #desu ne?/deshoo? } Dakara ... whatever you may say I Japanese cop so ‘Whatever you may say, I’m Japanese, you see? So, ...’ C: Shuo lai shuo qu, wo shi Riben-ren ba, ... whatever you may say ‘Whatever you may say, I’m Japanese, you see? ... ’
The third difference concerns the function of softening the tone of a command. While ba can occur freely, ne can only be used in a ‘gentle’ command (...nasai), not in a ‘rough’ command. See examples (13J, C) and (14Ja, C). (13) [request/invitation] J: Kite ne? come-req C: Qing lai ba. please come (14) [command] Ja: Kinasai ne? come-com Jb: Koi {#ne/yo }? come-com C: Lai ba. come
‘Please come, OK?’ ‘Please come, OK?’
‘Come, OK?’ [gentle command] ‘Come! OK?’[rough command] ‘Come! OK?’
In Japanese, yo must be appended, in place of ne, to soften the tone of a rough command, as shown in (14Jb). Yo, unlike ne, never functions to seek agreement or confirmation, but only to assert the speaker’s information in a personal way, when appended to a statement. In Chinese, not only are there NOT the different levels of command that Japanese has, (14C) can actually be interpreted as an order or request, although the force of the command is not as strong as simple Lai! (‘Come!’). The fourth difference is that, as indicated in (15C, J), ba can and often does express speaker’s reluctant compliance, but ne does not.8 In Japanese, again, yo would express such an attitude. Notice that the two Japanese sentences in (15J) would be well-formed in other situations, as shown in (16Ja, b). (15) [expressing speaker’s reluctance] J: Ii {#ne/yo }. Ikinasai { #ne/yo }. ‘Okay. (If you have to,) Go!’ good go-com C: Hao ba. Qu ba. ‘Okay. (If you have to,) Go.’ good go (16) Ja: Ii ne? ‘OK?; You follow me?; Understand?’ good-is
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.8 (204)
Mutsuko Endo Hudson and Wen-ying Lu
Jb: Ikinasai ne? go-com
‘Go, OK?’
The fifth difference is that, as seen before, ne is possible with certain set phrases, but not ba. See examples (17a, b). (17) a. J:
Arigatoo ne. thank you C: Xiexie { ø/#ba }. thank you b. J: Sumimasen ne. / Gomen ne. sorry (for) sorry (inf) C: Duibuqi { ø/#ba }. sorry
‘Thank you, dear.’ ‘Thank you.’ ‘Sorry, dear.’ ‘Sorry.’
The Chinese counterpart of Martin’s (1975) example (2), repeated as (17cJ), would be possible, but would indicate the speaker’s reluctance. (17) c. J:
Sayonara nee. goodbye C: Zai qian ba. goodbye
‘Good-bye now.’ ‘OK, (if you have to go) good-bye.’
The sixth and the final difference is that ne is ubiquitous in Japanese conversation, and can be used even when responding to sentences already appended with ne, as pointed out in Martin (1975: 917). Chinese ba is not used when responding to another ba sentence. (18) J:
A: Oishii ne. tasty-is B: Un, oishii { ø/ne }. yes tasty-is C: A: Zhege hao chi ba. this good eat B: Hao chi { ø/#ba }. good eat
‘It’s tasty, isn’t it.’ ‘Yes, it is { ø/isn’t it }.’ ‘It’s tasty, isn’t it.’ ‘Yes, it is { ø/#isn’t it }.’
We contend that the contrast exhibited in (17) and (18) is due to the fact that, unlike ne in Japanese, Chinese ba does not express empathy, rapport, or involvement. This is consistent with the fact that ba is not usually used in academic lectures, as it would signal that the teacher is uncertain about the material s/he is presenting. In contrast, ne is used commonly by teachers in the classroom (e.g. Ohta 1994). Recall that ba can express supposition, but ne does not; the meaning ‘probably’ is expressed by deshoo in Japanese.
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.9 (205)
On Japanese ne and Chinese ba
We have seen above that ba in Chinese often behaves like ne, but sometimes like yo, and at other times like deshoo in Japanese. While the main functions of both ne and ba are to solicit agreement/confirmation and to soften the tone, their specific behavior differs in complex ways. In the next section, we discuss the effects of information ownership, in an attempt to identify the possible causes of the behavior of ne and ba.
. Comparison from the perspective of information ownership Kimura and Moriyama (1992)9 compare sentence-final markers in Japanese and Chinese from the point of view of information ownership, more specifically, whether or not the speaker relies on (or considers) the hearer’s information. Table 1 summarizes their analysis of ne, ba, and daroo (plain form of deshoo). Table 1. Speaker’s reliance on hearer’s information Speaker relies on hearer’s information
Speaker has firm information (= non-question)
Speaker does not have firm information (= question)
Yes
Marked statement (e.g. ...ne, daroo ‘probably’, ba) Unmarked statement
Unmarked question (= typical question) Marked question (e.g. ... daroo ka, kashira)
No
In their view, the basic meaning of ba is ‘suspension of a true-or-false judgment or avoidance of conclusion’, and when the speaker does not rely on the hearer’s information, ba expresses ‘supposition’ [suiryoo], whereas when s/he does, it expresses ‘seeking of agreement or confirmation’ [dooi, kakunin no yookyuu]. They provide (19C) and (20C) as respective examples. (19) C: A: Mingtian tianqi zenme yang? tomorrow weather how B: Mingtian xia yu ba. tomorrow fall rain (20) C: ... Hen liang ba. ... very shiny
‘What’s tomorrow’s weather going to be?’ ‘It’ll probably rain tomorrow.’ ‘... It’s shiny, you see? ...’
According to them, such ba sentences as (20C) are uttered with the speaker assuming that the hearer possesses necessary knowledge or information with which to respond, and in that sense, ba explicitly marks speaker’s reliance on hearer’s information. The Japanese counterparts of (19C) and (20C) are (19J) and (20J), respectively, expressed not with ne, but with deshoo (‘probably’), which is a marker that assumes the same type of information ownership as does ba.
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.10 (206)
Mutsuko Endo Hudson and Wen-ying Lu
(19) J: A: Ashita no tenki wa doo desu ka? ‘What’s tomorrow’s weather tomorrow’s weather top how cop q going to be?’ B: Ashita wa ame deshoo. ‘It’ll probably rain tomorrow.’ tomorrow top rain (20) J: ... Hikatteru deshoo? ... ‘... It’s shiny, you see? ...’ shiny-is Kimura and Moriyama (1992) posit two types of deshoo; one used to express supposition and the other used to request confirmation, the latter of which is further subcategorized into two types, one that imposes the speaker’s information and the other that asks the hearer for information. In terms of the pronunciation, deshoo expressing supposition has falling intonation, and the one requesting confirmation has rising intonation. In sentences like (9J), (8C) and (10J, C), repeated below as (21J, C) and (22J, C), respectively, both deshoo and ba function to impose the speaker’s judgment on the hearer and to asks for a reaction at the same time. In all of (21J, C)–(22J, C), the speaker may be understood to be demanding that the hearer have identical information to that of his/her own. (21) J:
Kono uchi ii deshoo? this house nice C: Zhe fangzi bucuo ba. this house nice
‘This house is nice, isn’t it?’ ‘This house is nice, isn’t it?’
(22) J:
Ano hito ga sonna koto suru wake nai deshoo? that person nom such things do ext not ‘S/he wouldn’t do such things, don’t you agree?’ C: Ta bu hui zuo zhe-yang-de shi ba. s/he not know do such things ‘S/he wouldn’t do such things, don’t you agree?’
Likewise, in sentences (12J, C), repeated below as (23J, C), the speaker imposes his/her own information (rather than judgment) on the hearer and asks for a reaction, which ‘you see?’ does in English. (23) J:
Nanno kanno yuttemo atashi Nihonjin deshoo? ... whatever you may say I Japanese ‘Whatever you may say, I’m Japanese, you see? ...’ C: Shuo lai shuo qu, wo shi Riben-ren ba, ... whatever you may say I cop Japanese ‘Whatever you may say, I’m Japanese, you see? ... ’
According to Kimura and Moriyama (1992), ne, on the other hand, is a marker which shows that the speaker relies (or pretends to rely) on the hearer’s informa-
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.11 (207)
On Japanese ne and Chinese ba
tion, albeit temporarily, and demands that the hearer’s information be revealed in the discourse. In other words, the use of ne requests a reaction from the addressee regarding the status of the information in question. Note that, in Japanese, the speaker cannot appear to be ignoring the hearer’s knowledge/information. For example, to someone who the speaker suspects knows the time, one cannot say (24Ja), but must say (24Jb), with ne. This is because, as Kamio (1990) observes, a marker requesting agreement, or ne, is obligatory when the speaker anticipates that the hearer agrees. Ba is optional in such a case, as shown in (24Ca, b). (24) J:
Moo 5-ji desu { a. # ø/ b. ne }. ‘It’s already 5 o’clock { #ø/isn’t it }.’ already 5-o’clock cop C: Yiding 5-dian zhong le { a. ø/ b. ba }.‘It’s already 5 o’clock { ø/isn’t it }.’ already 5-o’clock
Ne is also obligatory when the information in the sentence largely belongs to the hearer, though there is no such restriction for ba. As a result, (25Ca), devoid of ba, is well-formed, but not (25J), missing ne. Note that, if ba is added, as in (25Cb), it would mean ‘I suppose you like Japanese tea’. (25) J:
(Anata wa) Nihoncha ga suki desu { a. # ø/ b. ne }. you top Japanese tea nom fond cop ‘You like Japanese tea { a. #ø/ b. don’t you }.’ C: Ni xihuan Ribencha { a. ø/ b. ba }. you like Japanese tea ‘You like Japanese tea { a. ø/ b. I suppose }.’
Some uses of ne assume no knowledge of the hearer, as in (26J), what Kamio (1990) calls ‘optional ne’. Interestingly, ba can be used similarly. (26) J:
A: Ikura desu ka. how much cop q B: 200-en desu ne. 200-yen cop C: A: Duoshao qian? how much money B: Liang kuai qian ba. 2-dollars money
‘How much is it?’ ‘It’s 200-yen, I’d say.’ ‘How much is it?’ ‘It’s 2 dollars, I’d say.’
Kamio (1998) observes that optional ne like the one in (26J) is felicitous when the information is NOT readily accessible to the speaker. If it is, ne would sound odd, as shown in (27J). Ba would be odd in the same situation, as shown in (27C). (27) J:
A: (Anata no) Onamae wa? you gen name top
‘What’s your name?’
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.12 (208)
Mutsuko Endo Hudson and Wen-ying Lu
B: Yooko desu { ø/# ne }. Yoko cop C: A: Nin qui xing? you pol name B: Wo xing Gao { ø/# ba }. I name Gao
‘It’s Yoko, { ø/# I’d say }.’ ‘What’s your name?’ ‘My name is Gao, { ø/#I’d say }.’
We contend that the respondents to questions in sentences like (26J, C) with ne and ba, respectively, are engaged in soliloquy, asking ‘confirmation’ from themselves, so to speak, rather than from another person. In Kimura and Moriyama’s (1992) framework, the difference between ne and ba is as follows. As there appears to be no recognition in Chinese of the hearer’s ‘information territory’, the speaker’s knowledge/information serves as the standard in making remarks. In Japanese, on the other hand, the speaker must always pay attention to the hearer’s knowledge/information. And the difference between ne and deshoo is that, with ne the speaker assumes no informational gap between his/her own and that of the hearer, simply anticipating agreement from the hearer, while with deshoo s/he does. Kimura and Moriyama’s (1992) account based on information ownership is important as it clarifies some of the ambiguous notions in past analyses. For example, their analysis supports and gives reasons for Uyeno’s (1971) claim that ne and its variants ‘imply that the option of judgment on the given information is left to the addressee’ and ‘...the appropriate use of these particles [like ne] reflects the speaker’s consideration of the addressee...’, that is, ‘rapport’. It also provides new insight into the similarities and differences between ne and deshoo in Japanese.
. Ne and ba analyzed in Schiffrin’s (1987) discourse model We now analyze ne and ba in Schiffrin’s (1987) discourse framework. The model consists of five planes: Exchange Structure (turns), Action Structure (acts such as request and suggestion), Ideational Structure (propositions/meanings), Participation Framework (speaker-hearer and speaker-utterance relations), and Information State (speaker/hearer knowledge and meta-knowledge). Local coherence in discourse is achieved by joint efforts of integrating ‘knowing, meaning, saying and doing’. In Schiffrin’s model, ne and ba are similar in Ideational Structure, Exchange Structure, and Action Structure, but different in others. They are similar in Ideational Structure as neither carries propositional meaning, and in Exchange Structure they both signal a possible turn change. In Action Structure the core action ne and ba express is the same; that is, soliciting a reaction from the hearer.
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.13 (209)
On Japanese ne and Chinese ba
It is reasonable to assume that such consideration or respect for the hearer’s knowledge/information on the part of the speaker results in a softening effect in requests, invitations and commands, and seeking agreement and confirmation in others. In Information State ne and ba behave similarly in some ways, but differently in others. They are similar in that they both indicate that the speaker believes that the hearer agrees with the content of the sentence, which explains the infelicity of the sentences uttered to an addressee who is asleep or unconscious, as in (7J, C), who by definition is unable to do so. Their behavior differs, however, depending on whether or not the speaker imposes his/her judgment/information on the hearer; it is possible with ba, but not with ne. In Participation Framework, too, there are some differences and similarities. Regarding the speaker-hearer relation, ne and ba are similar in that they can both be used regardless of the relative status of the speaker and the hearer (which is not the case with ci, the comparable element in Korean). Japanese ne, however, can only be used in a friendly relationship where there is “affective common ground” (Cook 1992) shared by speaker and hearer. In contrast, the usage of ba is broader. This explains why ne is possible in requests, invitations, and gentle commands, etc., but not in rough commands. In terms of the speaker-utterance relations, one difference is that ba can and often does express reluctant compliance, while ne does not. The foregoing discussion is summarized in Schiffrin’s five-plane model presented in Table 2. Table 2. Ne and ba in Schiffrin’s framework (Sp: speaker; Hr: hearer; S: sentence) Ideational Action Exchange Information State Structure Structure Structure J No ne propositional meaning C SAME ba
Solicits reaction from Hr SAME
Signals possible turn change SAME
Participation Framework
Sp believes Hr agrees; Friendly relation; Does not impose Sp’s Sp ≥ or < Hr; Does not judgment/information express reluctant compliance Sp believes Hr agrees; Any close relation; May impose Sp’s Sp ≥ or < Hr; May express judgment/information Sp’s reluctant compliance
. Summary and future directions Above we have presented similarities and differences between Japanese ne and Chinese ba, and analyzed the particles from a multi-dimensional perspective using Schiffrin’s (1987) framework. Admittedly, many areas were only briefly treated in the present study, and many more related topics remain to be investigated. There
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.14 (210)
Mutsuko Endo Hudson and Wen-ying Lu
are two approaches that may prove fruitful in the future. One is an analysis in a similar vein that includes other sentence-final particles, such as yo and yone in Japanese, and ma and ne in Chinese (on ne, see Lee-Wong 2001, for example). Another is a comparison of similar sentence-final elements in Chinese, Korean and Japanese. Korean ci seems to serve the same main functions as those of ne and ba, and share many features with ba, but it exhibits several unique features as well (Hudson, et al., 2001). We hope to have shown that ne and ba, which are similar and yet subtly different, cannot be explained satisfactorily unless we examine them in various domains, such as the functional, the interactional and the informational, as is the case with many other linguistic phenomena; for example, Japanese conjunctions (Hudson 1998) and adverbs (Hudson 2001).
Notes . Ne used to solicit backchannels after phrases is not treated in this study. Also, nee is assumed to be an elongated variant of ne. . The following abbreviations are used in the examples: COH: cohortative; COM: command; COP: copula; EXT: extended predicate; FOR: formal; GEN: genitive; INF: informal; NOM: nominative; POL: polite; Q: question; REQ: request; TOP: topic. Tones are not marked in the Chinese examples. . Martin (1975: 917) provides the following translations for (2a, b), respectively: ‘G’bye now.’; ‘It’s cold, isn’t it. Yes, it is.’ . In fact, A New Practical Chinese English Dictionary (1972) gives the definition of ba simply as ‘a particle used after an imperative sentence’. . This is in line with the position taken by Takubo and Kinsui (1996) regarding Japanese ne, though their proposal does not specifically refer to the addressee’s state of consciousness. . For convenience, I treat the variants deshoo, desho, daroo and daro together as deshoo in the discussion to follow. . Contrary to their claim, however, W.-Y. Lu judges the version without ba to be possible as a neutral statement. . Interestingly, the sentence-final element ci in Korean, many of whose functions are similar to those of ne and ba, can also express the speaker’s reluctance. . This work is written in Japanese. The English translation is ours.
References Akatsuka, N. (1990). On the meaning of daroo. In O. Kamada & W. Jacobsen (Eds.), On Japanese and How to Teach It (pp. 67–75). Tokyo: Japan Times.
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.15 (211)
On Japanese ne and Chinese ba
Beijing Language Institute. (1990). Practical Chinese Reader, Elementary Course: Book 1. Boston: Cheng & Tsui. Chao, Y. R. (1968). A Grammar of Spoken Chinese. Berkeley: University of California Press. Cook, H. M. (1992). The Sentence-final particle ne as a tool for cooperation in Japanese conversation. In H. Hoji (Ed.), Japanese/Korean Linguistics, Vol. 1 (pp. 29–44). DeFrancis, J. (1963). Beginning Chinese. New Haven: Yale University Press. Hudson, M. E. (1998). So? (On Japanese connectives sorede, dakara, and ja). In N. Akatsuka et al. (Eds.), Japanese/Korean Linguistics, Vol. 7 (pp. 59–75). Stanford: Center for the Study of Language and Information. Hudson, M. E. (2001). Disukoosu no wakugumi ni yoru hukushi no bunseki [Analysis of adverbs in a discourse framework]. In M. Minami (Ed.), Linguistics and Japanese Language Education [Gengogaku to Nihongo Kyooiku], Vol. 2. Tokyo: Kuroshio. Hudson, M. E., Lu, W.-Y., Ma, C.-H., & Park, O.-S. (2001). On Japanese ne, Korean ci, and Chinese ba. A paper presented at Georgetown University Round Table on Languages and Linguistics, Georgetown University. Kamio, A. (1990). Joohoo no nawabari riron [The Theory of Territory of Information]. Tokyo: Taishukan. Kamio, A. (1994). The theory of territory of information: The case of Japanese. J. of Pragmatics, 21, 67–100. Kamio, A. (1998). An analysis of Japanese ne in terms of the theory of territory of information. In N. Akatsuka, et al. (Eds.), Japanese/Korean Linguistics, Vol. 7 (pp. 231– 242). Kimura, H., & Moriyama, T. (1992). Kikite joohoo hairyo to bunmatsu keishiki [Consideration of hearer’s information and sentence-final forms]. In Y. Ookoochi (Ed.), Nihongo to Chuugokugo no taishoo kenkyuu ronbunshuu [Articles on contrastive studies of Japanese and Chinese], Vol. 2 (pp. 3–43). Tokyo: Kuroshio. Kindaichi, H. (1957). Nihongo [The Japanese Language]. Tokyo: Iwanami. Lee-Wong, S. M. (2001) Coherence, Focus and Structure: The Role of Discourse Particle Ne. Pragmatics, 11, 139–153. Li, C. N., & S. Thompson, A. (1981). Mandarin Chinese: A Functional Reference Grammar. Berkeley, University of California Press. Liang, S. C. (Ed.). (1972). A New Practical Chinese English Dictionary. Taipei: Far East Book. Lu, W.-Y. (1999). A Pragmatic study of the sentence-final particle ba in Mandarin Chinese. A paper presented at the Linguistics Colloquium, Michigan State University. Martin, S. (1975). A Reference Grammar of Japanese. New Haven: Yale University Press. Maynard, S. K. (1989). Japanese Conversation: Self-contextualization through Structure and Interactional Management. Ablex, Norwood. Ohta, A. S. (1994). Socializing the expression of affect: An overview of affective particle use in the Japanese as a foreign language classroom. Issues in Applied Linguistics, 5 (2), 303–325. Schiffrin, D. (1987). Discourse Markers. New York: Cambridge University Press. Takubo, Y., & Kinsui, S. (1996). Fukusuu no shinteki ryooiki niyoru danwa kanri [Discourse management in terms of multiple mental domains]. Ninchi Kagaku [Cognitive Science], 3 (3), 59–74.
TSL[v.20020404] Prn:3/09/2002; 12:10
F: PBNS29.tex / p.16 (212)
Mutsuko Endo Hudson and Wen-ying Lu
Tanaka, H. (2000). The particle ne as a turn-management device in Japanese conversation. J. of Pragmatics, 32 (8), 1135–1176. Uyeno, T. (1971). A Study of Japanese Modality: A Perfomative Analysis of Sentence Particles. Ph.D. dissertation, University of Michigan.
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.1 (213)
‘I am asking for a pen’* Framing of requests in black South African English Luanga A. Kasanga University of the North, South Africa
Introduction Requests are a ubiquitous speech act in everyday interaction. The understanding and production of requests are governed by social rules of context and situation which take into account social variables (social distance, status or power relationships) and the degree of imposition. The system of requests can be quite elaborate, as shown, for example, in Ervin-Tripp’s (1976) outline of a syntactic framework of American English directives. As Wierzbicka (1985: 150–154) observes, the preferred use of interrogative forms and the heavy restriction on the use of imperative forms constitute “striking linguistic reflexes” of the “socio-cultural attitude” of native speakers of English. However, usage across native dialects, let alone new varieties, of English may vary considerably. Hodge (1990), for example, found variation in requesting behaviour between two samples of native speakers of English (Australian and South African English respectively) living in Tasmania, Australia. He attributed this variation to the differences in the socio-political and cultural backgrounds of the two sets of native speakers. An even more marked difference occurs between request realizations in native and nonnative forms of English. Although a full discussion of the implications for inter-dialect communication of too much variation between varieties of English is beyond the scope of this article, it is worth mentioning its negative implications: variation may, sadly, lead to pragmatic failure which can obfuscate the pragmatic meaning of utterances and can subsequently lead to misunderstanding or a breakdown of communication. The latter may, in turn, result in “negative face”. Resentment, ethnic (crossgroup) stereotyping, and negative labelling (de Kadt 1998) have been cited as some
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.2 (214)
Luanga A. Kasanga
of the unfortunate consequences, ultimately, of such marked variation, especially in multi-ethnic or multi-racial different-language same-culture interaction. The perception of “pragmatic failure” may result from either mis-communication or mis-understanding, either having the potential of causing embarrassment, or worse still, disharmony. Chick (1986) discovered that mis-communication constituted a source of friction in the workplace and in educational settings (see also Gumperz, Jupp, & Roberts 1979) and he (Chick 1985) believed that negative labelling contributes to the perpetuation of discrimination in South Africa. This article, using the contrastive approach to intercultural communication (see Clyne 1994: 3), is an analysis of requests in Sepedi, an African language predominantly spoken in the Northern Province of South Africa, and black South African English (hereafter BSAfE), the variety of English typically spoken by black South Africans, sometimes referred to in the literature (e.g. Ntlhakana 2000) as “People’s English”. This variety displays features ubiquitously unique to speakers of African languages. It is hypothesized that the differences between preferred requesting behaviours of speakers of native forms of English (e.g. South African English) and speakers of BSAfE should be accounted for in terms of choices rather than “incompleteness” of learning as some (e.g. de Klerk 1996) have suggested on the basis of perceived differences across generations which may rather be more idiosyncratic or stylistic than dialectal. Data from several studies (Kasanga 1998; Makalela 1999; Mothoa 2001) point to much greater similarities than differences across the spectrum of sub-varieties of what is assumed to be a common, unique variety, that is BSAfE. The “acquisitional” stance seems to be discounted by the finding that speakers of BSAfE can display much greater awareness than is generally thought of requesting strategies frequently found in other varieties of English but which they seldom used (see Kasanga 1998).
Background and theoretical framework English is spoken natively only by a small number of South Africans, while it is spoken as a second language (L2) by the majority of the population whose first language (L1) is any of the African languages or Afrikaans. The post-apartheid language policy of “open multilingualism” which considers all eleven official languages (i.e. Afrikaans, English, Ndebele, Sepedi or Northern Sotho, Sesotho or Southern Sotho, Siswati, Xitsonga, Setswana, Tshivenda, Xhosa, & Zulu) as equal has not, however, encroached on the “high language” status of English in its diglossic relationship with the other ten official languages. English is openly considered by all as a “code-in-between”– a concept proposed by Kamwangamalu (1998). English has inevitably been “loaded” over time by features from other languages and has,
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.3 (215)
Framing of requests in black South African English
thus, given rise to distinct varieties, such as Afrikaans English (e.g. Chick 1991), South African Indian English (see Mesthrie 1992), and black South African English – sometimes referred to as “South African black English”. In the World Englishes paradigm, credited to Kachru (1992, 1997) and developed by others (Platt et al. 1984; Smith 1992), these varieties are regarded as independent varieties with established norms rather than deviations from native-speaker norms. BSAfE is the main focus of this article. As was mentioned earlier, this new variety of English displays features ubiquitously unique to speakers of African languages. BSAfE in this paper subsumes all sub-varieties of BSAfE, documented – for example Zulu English (e.g. de Kadt 1992a) – or assumed. Indeed, it has become increasingly evident that African languages (see de Kadt 1992b; Gough 1995) share broad socio-pragmatic and pragmalinguistic features. Studies of BSAfE have focused mainly on syntactic or grammatical, lexicosemantic, and, to a lesser extent, stylistic and phonological aspects of the variety (e.g. Buthelezi 1995; Gough 1996a, b; Lanham 1995; Makalela 1999; Wissing 1987). Earlier studies (e.g. Finn 1986; Mawasha 1984) almost exclusively considered BSAfE from a “deficit theory” perspective because its unique features were taken as not more than “deviations” from the native form norms. These studies belong broadly to interlanguage pragmatics – defined by Kasper (1995: 1) as “an intersection of pragmatics and the study of second language acquisition”. Studies in this direction generally take an acquisitional stance by looking at data as representative of the learners’ developing (i.e. unstable, transient) pragmatic competence in English, i.e. their developing abilities to comprehend, produce, and acquire a pragmatically “appropriate” target language. Recent empirical research has, however, shown that what might be construed as transitional knowledge is, instead, L2 users’ deliberate choices and expression of preferences presumably to suit their own variety of English. This finding, thus, discounts “deficit theory” assumptions. For example, a study of requests in English by L2 speakers (Kasanga 1998) found evidence of greater awareness by L2 users of native speaker pragmatic forms. The analysis of the requests revealed the following: in their everyday, natural, face-to-face interaction students frequently used direct request formulae, as attested in observational data, i.e. data collected in naturallyoccurring, unelicited exchanges, whereas in a less natural context, as found in data elicited by means of a discourse completion task (DCT), they tended to produce comparatively more indirect request formulae. This mismatch is graphically shown in the illustration below (Figure 1)1 and in Appendix A. In a preliminary study (Kasanga 1997), which consisted of observational data only, it had been found that L2 users were making use of only a small range of request strategies. A conclusion, with hindsight, unwarranted, was drawn about the limited knowledge by the L2 users of the full gamut of request formulae available in English. However, the ex-
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.4 (216)
Luanga A. Kasanga 60
Percentage
50 40 30 20 10 0 MD
EP
HP
LD
WS
QP
SH
MH
Direct vs. indirect requests Observational requests
DCT-elicited requests
Figure 1. Observational vs. DCT data (CCSARP framework)
amination of further data analysis found a wider spread of uses across categories of requesting formulae. These somewhat perplexing results warranted the reformulation of the initial hypothesis: the unexpected discrepancy in the two sets of data was assumed to be an indication of the users’ heavy reliance on a transfer of strategies from their first language (L1) in the production and use of speech acts in BSAfE. One of these strategies was the use of translation equivalents from their L1s. It was, therefore, necessary to undertake a contrastive study of requests in BSAfE and one or more than one of the African languages presumed to be the source of transfer of strategies. This need was the spur to the study reported in this article.
Methodology The study follows on empirical research globally within a cultural difference model. This type of research consists of studying contrastively speech acts in two sets of (usually distant) languages. In South Africa, research within this paradigm has focussed on the contrastive analysis of speech acts in English and African languages – with Zulu and Xhosa having the lion’s share (see, for example: Chick 1996; de Kadt 1992a, 1995; Gough 1996a, b; Wood 1992). From a number of studies involving English and other languages (e.g. Blum-Kulka 1987; Lwanga-Lumu 1999), the notion of “universality” in contemporary theories of politeness, such that posited by Brown and Levinson (1987) of the link between indirectness and
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.5 (217)
Framing of requests in black South African English
politeness has now been contested. Studying contrastively pragmatic features of two sets of languages is theoretically justified, following the empirical evidence of form-function correlation as the potential locus for the occurrence of phenomena such as borrowing, transfer, and interference (see Prince 1992). It was, thus, hypothesized that, for example, the pervasive occurrence of performatives of the type I am asking for a pen in BSAfE was evidence of transfer of linguistic structures from African languages. As Gough (1995) and LwangaLumu (1999) found in their empirical studies, performatives and other “direct requests” are commonly used in African languages to express “positive politeness” or deference. For example in Sepedi, requests are more often than not “explicit performatives” such as example (1) below: (1) Ke kgopel-a ranta (Sepedi) I ask-tense marker one rand (I am asking for one rand) To empirically test the assumption about transfer of strategies from African languages into English as an explanation of the nativization of the latter into BSAfE (see also de Kadt 1992a, for a similar assumption between Zulu and Zulu-English), a contrastive study of requests in BSAfE and Sepedi, an African language widely spoken in the Northern Province of South Africa, was undertaken at the University of the North. The choice of an academic setting as the main research site is justified both by the multilingual and multiracial composition of the student and staff population and the observed cases of pragmatic failure by many students, misunderstanding or mis-communication resulting from cross-cultural differences, or the use of different varieties of English. It, thus, offered an ideal setting for the present study. Besides, BSAfE is mainly an “educated variety” of English, which is best researched in an academic setting. A multi-method strategy was adopted to suit the aims and particular needs of the study: DCT, a native speaker judgment politeness scale, and a limited number of face-to-face interviews were used. The collection of the Sepedi and BSAfE requests was done by means of DCTs (see Appendices B & C) completed by 89 final-year students of English. Sepedi was selected as the only African language for comparison with English, given its wide use. Besides, very few of the students involved in the study chose to complete the DCT in Setswana also made available. An interval of seven weeks was allowed between the administering of the English and Sepedi DCTs to discourage rehearsal and possible translation from one language to another. Roughly the same number of students completed both DCTs during lecture time. The completion was timed to ensure that all the tasks were completed and to elicit only those requests which could come naturally in the flow of thought.
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.6 (218)
Luanga A. Kasanga
The main assumption was that the propensity to use explicit performatives in Sepedi as the most favoured request formulae would be mirrored in the data in BSAfE by the pervasive use of explicit performatives and, to a lesser extent, hedged performatives. On the other hand, requests further down on the scale of directness which, by virtue of their structure, are reputed to be “indirect”, would be seldom used in BSAfE. Native speaker judgment scales were used to measure the politeness level of the Sepedi requests. Interviews served to supplement and clarify in case of discrepancy the native speaker judgments of politeness.
Results and discussion Sepedi requests Because the BSAfE results were the same as those in a previous study (see Kasanga 1998), it was decided to focus on the Sepedi data whose results are shown in Table 1. Table 1 shows aggregate scores and percentages across the four tasks in the DCT. It must, however, be pointed out that there were differential patterns across tasks in the occurrence of different types of requests, the discussion of which is beyond the scope of this article. Results below were compared to the observational data (because they are more a reflection of the BSAfE speakers’ choice and preferences). While in the Sepedi data there was much greater spread across the scale than in the BSAfE data, there is one significant matching feature in both cases: explicit performatives in both the Sepedi and BSAfE data outnumber all other types of requests. A partial conclusion was, therefore, drawn: the preferred use of explicit performatives in BSAfE is the result of transfer from Sepedi. Because the suggestion of the preference of explicit performatives over other forms of requests comes from evidence from Xhosa (Gough 1995), a representative of the Nguni language group, and from Sepedi, which belongs to the Sotho language group, it may tenTable 1. DCT Sepedi requests (CCSARP framework) Request type Mood derivables Explicit performatives Hedged performatives Locution derivables (obligation statements) Want statements Suggestory formulae Query preparatories Strong hints Mild hints
Number
%
42 164 4 7 54
12.7 49.4 1.2 2.1 16.3
26 35
7.8 10.5
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.7 (219)
Framing of requests in black South African English
tatively be concluded that the preferred use of explicit performatives is shared by African languages spoken in South Africa in general. This finding seems to concur with recent studies such as those by Makalela (1999) and Mothoa (2001).
Politeness coding of Sepedi requests: Implications for BSAfE In attempting to identify similarities and dissimilarities between Sepedi and BSAfE, the concern was to determine the level of politeness-deference represented in each type of requests. It was not possible to have politeness coded for BSAfE requests. Given the predominance of requests which, in the Cross-Cultural Speech Act Realization Patterns Project (CCSARP) framework (see Blum-Kulka, House, & Kasper 1989), would be considered as direct requests (i.e. explicit performatives), it was surmised that in Sepedi (and in BSAfE indeed), they have a high deferencepoliteness level. Therefore, a test was necessary. In the present study, 28 fairly common requests from the Sepedi data were randomly selected for closer examination and analysis. A 5-point scale of politeness was suggested and the list of selected statements from three different DCT episodes was submitted for judgment to five native speakers of Sepedi, three of whom well versed in African linguistics. On the scale, 1 represents the most polite request, 5 the least polite; 3 is, therefore, considered as “neutral”, as it were, because it is a mid-point from both ends of the scale. The purpose was to identify how politeness/deference is expressed in Sepedi and how these politeness strategies may have shaped pragmatic usage in BSAfE. The expression of politeness or deference hinges on social factors already referred to as power or status, distance, and the weight of imposition. Three possibilities have been postulated by Scollon and Scollon (1995: 46) on the basis of: (i) the existence or not of power or status differential between the two interactants; (ii) the level of familiarity, i.e. how well interactants know each other; and (iii) the level of the perceived imposition of the request. These possibilities can be diagrammatically presented as in Figure 2. As shown in the matrix, there is no power (P) or status differential in both deference-politeness (C) and solidarity-politeness (D) as both types of politeness are determined respectively by the presence and absence of familiarity (F), whereas hierarchical politeness (A and B) is determined by power or status, regardless of the level of familiarity. The episodes differed on the basis of these variables (see Appendix B). In deciding on the placement of each request on the politeness scale, the level of agreement or disagreement was taken into account. There was a large measure of agreement among the native-speaker judges. The few cases on which they differed markedly were discarded from the sample. A clear majority (minus any “neutral”) tipped towards either side of the politeness scale. In a case of a split judgment, i.e. a
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.8 (220)
Luanga A. Kasanga + +
–
P
A Hierarchical politeness
C Solidarity-politeness
B Hierarchical politeness
D Deference-politeness
F –
Figure 2. Matrix representation of different politeness systems (Scollon & Scollon 1995)
case in which there was an equal number of choices on each side of the scale, the request was discarded. To delineate borderline cases, a sixth native speaker was called upon to arbitrate. The final tally, done after the borderline cases were submitted to the sixth judge, yielded 23 clear-cut requests from the original 28 (see Appendix C). A summary of the results is graphically shown in Table 2. Mood derivables and strong hints, exemplified respectively in (2) and (3), and rated 4.2 and 4, can be considered as the least felicitous, taking into account the power or status and familiarity. (2) Stapolara hle! Stapler, please! (3) Ga le na stapler naa? [SH] neg-you have stapler by the way? Do you, by any chance, have a stapler? As expected, mood derivables and strong hints represent a relatively small proportion (respectively 12.7% and 10.5%) of all the requests elicited in the Sepedi DCT study (see Table 1 above). It is assumed that in natural everyday interaction, both types of requests will be dispreferred, especially where deference-politeness is required. The assumption seems to hold true for strong hints only in the BSAfE study in which they constituted less than a percent (0.4%) of all the requests colTable 2. Rating (average) of selected Sepedi requests Request
average rating
Mood derivable Strong hint Query preparatory Explicit performative
4.2 4 1.5 1.3
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.9 (221)
Framing of requests in black South African English
lected in naturally-occurring interaction. As to mood derivables, it is not clear why they were the second most preferred type of requests in both the observational BSAfE study (42.7%) and the DCT data (16.7%). Indeed, this preference does not tally with the corresponding result in the Sepedi data. It can, however, be surmised that the mismatch may be accounted for by the discrepancy in the source of requests: the high proportion of mood derivables in the BSAfE study was found in the observational, i.e. naturally-occurring, data, whereas the result in the Sepedi was elicited. A more conclusive finding could only be confirmed if both data came from naturally-occurring interaction. Query preparatories in Sepedi were rated high on average on the politeness scale by almost all the native-speaker judges. According to one of the native speaker judges, “verbosity” has a mitigating effect on the weight of imposition that a request might carry. Hence, the use of query preparatories in contexts such as the student-lecturer episode in this study, may have been preferred. Surprisingly, in the BSAfE naturally-occurring data, they occurred minimally, at less than 1% (0.1) of all requests recorded. In contrast, in the DCT BSAfE data, they ranked highest (52.4). It may be speculated here that the low number of query preparatories may be due to limitations in performance: the possibility of performance anxiety or fright has been suggested as a possible explanation (M M Mohlake 1999: personal communication), given the differential of status between the speaker (students) and hearer (lecturers in all the situations in the study). Hence, a discrepancy was found between the actual requests (in the observational study) and those produced in the simulated dialogues. In contrast, explicit performatives stand out as the type of request which mirrors the phenomenon of transfer. Both in BSAfE and in Sepedi (see respectively Appendix A and Table 1) they outnumber all other requests. Besides, in the Sepedi data, they were rated highest (average = 1.3) on the politeness scale by all nativespeaker judges. Even in the BSAfE observational data, they constitute the highest number of requests recorded. On the face of the fact that the most frequently used English request formulae in the BSAfE study was of the the type I am asking for a pen, it can be concluded that translation strategies used in the L2, English, have shaped the requesting behaviour displayed in BSAfE. Closer examination of some of the explicit performatives, however, warrants a few remarks. Firstly, the following explicit performative was rated 5, i.e. among the least polite requests: (4) Stapler seo, ke a se kgopela. Stapler that, I it ask-pres. I am asking for (want to use) that stapler.
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.10 (222)
Luanga A. Kasanga
At first sight, it is not significantly different from the explicit performative in example (1), which is generally the most preferred request. Only after an observation by one of the native-speaker judges on another request (see example 5 below) did it become clear that there was an element in the proposition acting as an “aggravator”. (5) Ke kgopela gore le nthuše ka seteipolara seo sa lena ke swariše matlakala. I ask-pres that you me help-PRES with stapler that of yours I hold-pres together paper. I am asking if you can help me with (lend me) that stapler of yours so that I can staple these/the sheets of paper. According to the informant, the demonstrative seo (that), depending on the tone of delivery, may determine the level of politeness of the request. It is not surprising that the politeness ratings of the above request varied widely (average rating = 3.2): intonation could not be inferred from the written data submitted to the informants. Likewise, the request in (4) above was uncharacteristically rated very low on the politeness scale (average rating = 4.2). A second important remark is on the use of honorifics, such as mohlomphegi (sir, my lord) which is a term of address indicating a high level of respect or deference, used as an alerter in one of the requests as shown in (7) below: (6) Ke kgopela stapler. I ask-pres for stapler. I am asking for the/a stapler. (7) Mohlomphegi, ke kgopela stapler. Sir, I ask-pres for stapler. Sir, I am asking for the/a stapler. As was pointed out by one native-speaker judge, the use of this honorific may have a strong mitigating effect on most types of requests, including those that would otherwise have been rated as least polite. This mitigating effect of the honorific was confirmed by another native speaker. It must be noted, however, that the rating for both examples (6) and (7) was close. A logical conclusion can be drawn from the findings on seo and mohlomphegi, namely that the coding of politeness in Sepedi, and perhaps in most African languages, is subtly achieved by devices other than indirecteness. Native speakers of English generally prefer indirect requests to direct ones in the form of interrogatives (see Wierzbicka 1985) in most non face-threatening situations. In contrast, speakers of Sepedi would prefer the use of a more direct form of request – for example, explicit performatives. One implication of this difference is that the variety of English shared or preferred by Sepedi speakers has developed a request system in which direct requests of the explicit performative type predominate.
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.11 (223)
Framing of requests in black South African English
The finding on the special devices used in Sepedi and the propensity of speakers of BSAfE to use what speakers of native forms of English would find too direct formulae has both theoretical and methodological implications. Theoretically, there are limitations to indirectness as an important criterion in request realization both in BSAfE and in Sepedi – or African languages in general (see Gough 1995; Lwanga-Lumu 1999). In terms of methodology for the study of requests – and other speech acts – it is reasonable to ask whether the classification of requests ought to be based on their syntactic (sentential) structure or the level of politeness (which excludes indirectness). This finding points to the limitations of the CCSARP, developed for the study of European languages (mainly English, French, German), a point already raised in the literature (see, for example, de Kadt 1995).
Conclusion The study reported in this article aimed to account for some perplexing findings from a previous study (Kasanga 1997). Observational data suggested limited knowledge by the L2 users of the full gamut of request formulae available in English. Further examination of elicited data, however, found a wider spread of reported awareness across categories of requesting formulae. The preliminary conclusion drawn from an acquisitional perspective was revised accordingly. A hypothesis was therefore formulated in favour of dialectal choice, in the direction of a nativized variety of English, most likely shaped by the socio-pragmatic rules of the host culture and the pragmalinguistic features of the L1s (i.e. African languages). A contrastive study was, therefore, subsequently undertaken. The main finding of the study is the significant differences in request realization both sociopragmatically and pragmalinguistically between, in the one hand, Sepedi and BSAfE, and on the other hand, standard (native speaker English). In standard native forms of English (especially British English), indirectness is paramount in achieving politeness (see Wierzbicka 1985), whereas in Sepedi (and, by the same token, in BSAfE), indirectness is not a determining factor. The expression of deference-politeness depends rather on subtle mitigating devices, such as the use of honorifics. Likewise, devices such as the demonstrative seo (that) may have an aggravating effect on request realization. This finding calls for further study to propose a classification framework more accommodating than the CCSARP. In closing, it is perhaps more realistic to assume the occurrence of explicit performatives in BSAfE to the detriment of more indirect requests which predominate in standard English usage as an indication that BSAfE is an institutionalized variety of English, rather than a transitional one. The choice of a World Englishes
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.12 (224)
Luanga A. Kasanga
perspective seems more appropriate in understanding speech act realization than an interlanguage one.
Notes * This is a revised version of a paper read at SIC-CSP 2000, Newnham College, Cambridge, 11–13 September 2000, under a slightly different title, “ ‘I am asking for a pen’: Framing of requests in English by Sepedi speakers”. . MD = mood derivable; EP = explicit performative; HP = hedged performatives; LD = locution derivable; WS = want statement; QP = query preparatory; SH = strong hint; MH = mild hint (see Appendix A for definitions and examples of request types).
References Blum-Kulka, S. (1987). Indirectness and politeness in requests: Same or different? Journal of Pragmatics, 11, 131–114. Blum-Kulka, S., House, J., & Kasper G. (Eds.). (1989). Cross-Cultural pragmatics: Requests and apologies. Norwood: Ablex. Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in language usage. Cambridge: Cambridge University Press. Buthelezi, G. (1995). South African black English: Lexical and syntactic characteristics. In R. Mesthrie (Ed.), Language and Social History: Studies in South African Sociolinguistics (pp. 242–250). Cape Town: David Philip Publishers. Chick, J. K. (1985). The interactional accomplishment of discrimination in South Africa. Language in Society, 14(3), 299–326. Chick, J. K. (1986). Intercultural miscommunication as a source of friction in the workplace and in educational settings. In G. Kendall (Ed.), Education and the Diversity of Cultures (pp. 25–36). Pietermaritzburg: University of Natal Press. Chick, J. K. (1991). Sources and consequences of miscommunication in Afrikaans EnglishSA English encounters. In J. Cheshire (Ed.), English Around the World (pp. 446–461). Cambridge: Cambridge University Press. Chick, J. K. (1996). English in interpersonal interaction in South Africa. In V. de Klerk (Ed.), Varieties of English Around the World: Focus on South Africa (pp. 269–284). Amsterdam & Philadelphia: John Benjamins. Clyne, M. G. (1994). Inter-cultural communication at work. Cultural values in discourse. Cambridge: Cambridge University Press. de Kadt, E. (1992a). Politeness phenomena in South African black English. In L. F. Bouton & Y. Kachru (Eds.), Pragmatics and Language Learning (Vol. 3, pp. 103–116). UrbanaChampaign, IL: University of Illinois at Urbana-Champaign. de Kadt, E. (1992b). Requests as speech acts in Zulu. South African Journal of African Languages, 12(3), 101–106.
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.13 (225)
Framing of requests in black South African English
de Kadt, E. (1995). The cross-cultural study of directives: Zulu as a non-typical language. South African Journal of Linguistics, Supplement 27, 45–72. de Kadt, E. (1998). Keeping the kitchen clean: Towards an analysis of English-medium interactions between black people and white people in post-apartheid South Africa. Multilingua, 17(2/3), 249–276. de Klerk, V. (1996). Encounters with English over three generations in a Xhosafamily: For better or for worse? In E. W. Schneider (Ed.), English Around the World (pp. 97–118). Amsterdam: John Benjamins. Ervin-Tripp, S. (1976). Is Sybil there? The structure of some American English directives. Language in Society, 5, 25–66. Finn, S. M. (1986). The abuse of English by Vista students. English Usage in Southern Africa, 17(1), 1–8. Gough, D. (1995). Some problems for politeness theory: Deference and directness in Xhosa performative requests. South African Journal of African Languages, 15(3), 123–125. Gough, D. (1996a). Black English in South Africa. In V. de Klerk (Ed.), Varieties of English Around the World: Focus on South Africa (pp. 53–78). Amsterdam and Philadelphia: John Benjamins. Gough, D. (1996b). Thinking in Xhosa and speaking in English: The theory and practice of contrastive analysis. South African Journal of Applied Language Studies, 4, 2–19. Gumperz, J. J., Jupp, T., & Roberts, C. (1979). Crosstalk. London: National Centre for Industrial Language Training. Hodge, F. K. (1990). Cross-cultural study of request realization patterns across two groups of native speakers of English. South African Journal of Linguistics, 8(3), 121–128. Kachru, B. B. (1992). Models for non-native Englishes. In B. B. Kachru (Ed.), The Other Tongue: English Across Cultures (2nd edition) (pp. 48–74). Urbana: University of Illinois Press. Kachru, B. B. (1997). World Englishes 2000: Resources for research and teaching. In L. E. Smith & M. L. Forman (Eds.), Literary Studies East and West. World Englishes 2000 (pp. 209–251). Honolulu, Hawai’i: University of Hawai’i Press. Kamwangamalu, N. M. (1998). ‘We-code’, ‘they-codes’, and ‘codes-in-between’: Identities of English and codeswitching in post-apartheid South Africa. In N. M. Kamwangamalu (Ed.), Aspects of Multiligualism in Post-apartheid South Africa. Special issue of Multilingua, 17(2/3) (pp. 277–298). Berlin & New York: Mouton de Gruyter. Kasanga, L. A. (1997, July). Requests in English. How second-language users cope. Paper read at the 25th annual conference of the South African Association for Language Teaching (SAALT), University of Natal, Pietermaritzburg. Kasanga, L. A. (1998). Requests in English by second-language users. ITL Review of Applied Linguistics, 119/120, 123–153. Kasper, G. (1995). Interlanguage pragmatics. In J. Verschueren, J.-O. Ôstman, & J. Blommaert (Eds.), Handbook of Pragmatics (pp. 1–17). Amsterdam/Philadelphia: John Benjamins. Lanham, L. W. (1995). The communicative effectiveness of stress and intonation in BSAE. In L. W. Lanham, D. Langham, A. Blacquiere, & L. Wright (Eds.), Getting the Message in South Africa. Intelligibility, Readability, Comprehensibility (pp. 42–58). Howick: Brevitas.
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.14 (226)
Luanga A. Kasanga
Lwanga-Lumu, J. C. (1999). Cross-cultural contrastive analysis of request directness levels. Southern African Journal of Applied Language Studies, 7(1), 88–106. Makalela, J. L. (1999). Institutionalized black South African English. The NAETE Journal, 13, 58–71. Mawasha, A. L. (1984). Your English, our English: English in Black South Africa. English Usage in Southern Africa, 15(2), 12–18. Mesthrie, R. (1992). English in Language Shift: the History, Structure and Sociolinguistics of South African Indian English. Cambridge: Cambridge University Press. Mesthrie, R. (1996). Putting some linguistics into applied linguistics: A sociolinguistic study of left dislocation in South African Black English. SPIL PLUS [Linguistics for the Language Professions 3], 29, 260–283. Mothoa, S. (2001). Grammatical features of black South African English; As exemplified in written samples of Ndebele and Northern Sotho first-language speakers. Unpublished Master’s dissertation, University of the North. Ntlhakana, P. (2000). People’s English. English Today, 16(2), 11–17. Platt, J., Webber, H., & Ho, M. L. (1984). The New Englishes. London: Routledge & Kegan Paul. Prince, E. F. (1992). On syntax in discourse in language contact situations. In C. Kramsch & S. McConnell-Ginet (Eds.), Text and Context. Cross-disciplinary Perspectives on Language Study (pp. 98–112). MA and Toronto: D C Heath and Company, Lexington. Scollon, R., & Scollon, S. W. (1995). Intercultural Communication. Oxford: Blackwell. Smith, L. E. (1992). Spread of English and issues of intelligibility. In B. B. Kachru (Ed.), The Other Tongue: English Across Cultures (2nd edition) (pp. 75–90). Urbana: University of Illinois Press. Wierzbicka, A. (1985). Different cultures, different languages, different speech acts. Journal of Pragmatics, 9, 145–179. Wissing, R. J. (1987). Language contact and interference in the acquisition of English proficiency by Bantu-speaking students. Unpublished Master’s dissertation, University of South Africa. Wood, M. (1992). Expressing gratitude in Zulu: A speech act study emphasizing communicative competence. In R. K. Herbert (Ed.), Language and Society in Africa (pp. 265–276). Johannesburg: University of the Witwatersrand Press.
Definition/realisation method Requestive force derives from grammatical mood of verb, or sentence type, of the locution; often imperatives; also infinitive forms and elliptical sentences. Use of performative verb (ask, request) (indicative) to explicitly name the illocutionary force Performative verb (ask, request) (naming of illocutionary force) is modified by hedging expressions (modal verb: would like to/must/have to; or verb of intention want to) States the obligation of the hearer to carry out the act; requestive force indicated by use of obligation modals, (have to, must, should, ought to)
Obs. 296
Could you display your ID document? Your ID document is not on the desk. We were asked to check candidates’ identity.
2
3
1
I want/wish/would like you to – display your ID document. How about displaying your – ID document?
You’ll have to display your ID document. 8
I’m asking you to display your 379 ID document. I would like to ask you to display your ID document 5
Example(s) Display your ID document.
3
16
88
–
21
–
12
–
DCT 28
*Ervin-Tripp’s (1976: 29ff.) “need statements” (used interchangeably with “want statement”) would include more than one category in the CCSARP classification, such as: “want statements” ( I want you to check the requirement for stairs), “mood derivables” (I’ll have a beer) , and suggestories (Why don’t you go ahead now to dinner. . . ). Her category “embedded imperatives” includes the CCSARP’s “suggestories” (Why don’t you open the window?), “hedged performatives” (Could I trouble you to open the window?), and “preparatories” (The line is busy; would you care to wait?).
Expression of speaker’s wish/desire for the hearer to carry out the act; use of “want” or “desire”verbs (want/wish) or modals (would like) Suggestory formula Illocutionary intent is phrased as a suggestion by means of a framing formula; contains a suggestion to do something (How about/Why. . . not) Preparatory (Query Reference to a preparatory condition for feasibility of the requested preparatory) action (e.g. ability, possibility, willingness), very often in question form. Strong hint Illocutionary intent not immediately derivable from the locution, but the locution refers to elements of intended illocution or proposition. Mild hint No elements of immediate relevance to the intended illocution are present in the locution; interpretable by context.
Locution derivable (Obligation statement) Want statement*
Explicit performative Hedged performative
Request type Mood derivable
Appendix A: Observational vs. DCT-elicited data requests (CCSARP framework) TSL[v.20020404] Prn:4/11/2002; 9:40 F: PBNS30.tex / p.15 (227)
Framing of requests in black South African English
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.16 (228)
Luanga A. Kasanga
Appendix B: Requesting/asking in Sepedi THIS IS NOT A TEST! THEREFORE, THERE ARE NO CORRECT OR INCORRECT ANSWERS. IT IS A BRIEF SURVEY OF YOUR USE OF/PREFERENCES FOR FORMULAE USED IN DAILY REQUEST ROUTINES IN SEPEDI. PLEASE WRITE DOWN WHAT YOU THINK YOU WOULD NATURALLY SAY: DO NOT MAKE UP FORMULAS WHICH YOU DO NOT USE IN REAL SITUATIONS. 1.
Imagine you have loose sheets of an assignment which you must submit shortly. You want to ask for a stapler from a lecturer you do not know well. This is what you say in Sepedi: 1.1 You: Thobela [Mohlomphegi] Lecturer: Thobela. You: (asking for the stapler in Sepedi)
What you wrote above might perhaps be only one (the most common) way of asking for something (for example, a stapler) from an unfamiliar lecturer. Can you list below (from 1.2 to 1.4) one or many more formulas in Sepedi in the same situation? IF YOU CANNOT THINK OF FIVE OTHER DIFFERENT WAYS, PLEASE STOP WHERE YOU CAN. 1.2 You: Lecturer: You:
Thobela [Mohlomphegi] Thobela. (asking for the stapler in Sepedi)
1.3 You: Lecturer: You:
Thobela [Mohlomphegi] Thobela. (asking for the stapler in Sepedi)
1.4 You: Lecturer: You:
Thobela [Mohlomphegi] Thobela. (asking for the stapler in Sepedi)
2.
You go to the police station to collect an important document which you had left hours before to be certified. What would you say in Sepedi in the following dialogue to request the certified document? 2.1 (You greet the duty officer, then you request your document by saying to him/her in Sepedi)
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.17 (229)
Framing of requests in black South African English
You: How else would you request your document? Please fill in dialogues 2.2 to 2.4 if possible. IF YOU CANNOT THINK OF FIVE OTHER DIFFERENT WAYS, PLEASE STOP WHERE YOU CAN. 2.2 (You greet the duty officer, then you request your document by saying to him/her in Sepedi) You: 2.3 (You greet the duty officer, then you request your document by saying to him/her in Sepedi) You: 2.4 (You greet the duty officer, then you request your document by saying to him/her in Sepedi) You: 3.
You are at the Department of Home Affairs to see a relative, Jane Khuto, who works there. Because it is your first visit, you do not know how to get to her office. You see a middle-aged Sepedi-speaking clerk at a reception desk. What would you say to her in Sepedi to ask for directions to your relative’s office? 3.1 You ask: Clerk: You:
Ke lebati la mafelelo go la nngele gona mo pathising ye.
Other formulas in Sepedi [please fill in dialogues 3.2 to 3.4] you might consider using to ask the same lady for directions. IF YOU CANNOT THINK OF FIVE OTHER DIFFERENT WAYS, PLEASE STOP WHERE YOU CAN. 3.2 You ask: Clerk: You:
Ke lebati la mapelelo go la nngele gona mo pathising ye.
3.3 You ask: Clerk: You:
Ke lebati la mapelelo go la nngele gona mo pathising ye.
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.18 (230)
Luanga A. Kasanga
3.4 You ask: Clerk: You:
Ke lebati la mapelelo go la nngele gona mo pathising ye.
4.
You drive out of a private car park and show your ticket to the park attendant. S/he asks for R1 as your parking fee. Imagine what she would say to you in Sepedi to tell you to pay R1: 4.1 Park attendant: Other formulae in Sepedi you can imagine that s/he would use to ask for the R1 charge. [please fill in dialogues 4.2 to 4.4]. IF YOU CANNOT THINK OF FIVE OTHER DIFFERENT WAYS, PLEASE STOP WHERE YOU CAN. 4.2 Park attendant: 4.3 Park attendant: 4.4 Park attendant: THANK YOU FOR YOUR CO-OPERATION.
Appendix C: Requesting/asking in English THIS IS NOT A TEST FOR WHICH YOU CAN EARN MARKS! THEREFORE, THERE ARE NO CORRECT OR INCORRECT ANSWERS. IT IS A BRIEF SURVEY OF YOUR PREFERENCES FOR FORMULAS USED IN DAILY REQUEST ROUTINES. PLEASE WRITE DOWN WHAT YOU THINK YOU WOULD NATURALLY SAY: DO NOT MAKE UP FORMULAS WHICH YOU DO NOT USE IN REAL SITUATIONS. PLEASE WRITE LEGIBLY 1.
Imagine you are running a small project and you want to ask lecturers you don’t know well to complete your questionnaire. In the dialogue you are going to read, write down what you would say to the lecturers you meet in the corridor, after the greetings already provided. Fill in the missing part of your next turn in the dialogue, i.e. what you would naturally say: 1.1 You: Good morning [sir/madam] Lecturer: Good morning. You:
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.19 (231)
Framing of requests in black South African English
What you wrote above is perhaps only one (the most common) way in which you might ask an unfamiliar lecturer to complete the questionnaire. If you have one or many more formulas in the same situation, please list them below. 1.2 You: Lecturer: You:
Good morning [sir/madam] Good morning.
1.3 You: Lecturer: You:
Good morning [sir/madam] Good morning.
1.4 You: Lecturer: You:
Good morning [sir/madam] Good morning.
2.
You go to see one of your lecturers in his/her office to collect your marked assignment. What would you say in the following dialogue to request your assignment? 2.1 (You enter; after greeting the lecturer, you request your assignment by saying) You:
How else would you request your assignment? 2.2 (You enter; after greeting the lecturer, you request your assignment by saying) You:
2.3 (You enter; after greeting the lecturer, you say to him/her) You:
2.4 (You enter; after greeting the lecturer, you say to him/her) You:
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.20 (232)
Luanga A. Kasanga
3.
You went to Home Affairs to see a relative who works there. You do not know how to get to his/her office. You see a middle-aged lady who apparently works there and you ask her for directions to your relative’s office. What do you say to her? 3.1 You: Lady: You:
It’s the last door on your left down the corridor.
Other formulas you might consider to use in asking the same lady for directions? 3.2 You: Lady: You:
It’s the last door on your left down the corridor.
3.3 You: Lady: You:
It’s the last door on your left down the corridor.
3.4 You: Lady: You:
It’s the last door on your left down the corridor.
THANK YOU FOR YOUR CO-OPERATION
Appendix D: Politeness rating of selected Sepedi requests [1 = most polite; 5 = least polite] (1) Nthuše ka stapler. [MD][4] Me you help-pres with stapler. Help me with (lend me) the/a stapler. (2) Stapolara hle! [MD] [5] Stapler, please!
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.21 (233)
Framing of requests in black South African English
(3) Nkadime seteipolara. [MD] [?] Me lend-imper stapler. Lend me the/a stapler. (4) Le ka se nthuše ka stapler? [QP] [?] You can me help with stapler? Can you help me with (lend me) the/a stapler? (5) Ga le na stapler naa? [SH] [4] neg-you have stapler by the way? Do you, by any chance, have a stapler? (6) Ke kgopela stapler. [EP] [2] I ask-pres for stapler. I am asking for the/a stapler. (7) Mohlomphegi, ke kgopela stapler. [EP] [1] Sir, I ask-pres for stapler. Sir, I am asking for the/a stapler. (8) Ke adima stapler. [EP] [2] I borrow-pres stapler. I am borrowing the/a stapler. (9) Re kgopela stapler ge ekaba le na le sona. [EP] [1] We ask-pres stapler if you you be-pres with it. I’d like to use your stapler, if you have got one. (10) Re le kgopela sekgomantšhi sa maphephe. [EP] [1] We you ask-pres the thing hold together for paper. I would like to use the/a stapler (thing which holds together paper). (11) Stapler seo, ke a se kgopela. [EP] [5] Stapler that, I it ask-pres. I want to use that stapler. (12) Ke kgopela le nkadime stapler. [EP] [1] I ask-pres you me lend stapler. Can I ask if you can lend me the/a stapler? (13) Ke kgopela gore le nthuše ka seteipolara. [EP] [1] I ask-pres that you me help-pres with stapler. I am asking if you can help me with (lend me) the/a stapler. (14) Ke kgopela gore le nthuše ka seteipolara seo sa lena ke swariše matlakala. [EP] [?] I ask-pres that you me help-pres with stapler that of yours I hold-pres together paper.
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.22 (234)
Luanga A. Kasanga
I am asking if you can help me with (lend me) that stapler of yours so that I can staple these/the sheets of paper. (15) Ke kgopela gore le nkadime stapler. [EP] [1] I ask-pres that you me lend stapler. I am asking if you can lend me the/a stapler. (16) Ke kgopela gore le nkadime stapler ke swariše matlakala a. [EP] [2] I ask-pres that you me lend stapler I hold-pres together paper. I am asking if you can lend me the/a stapler. (17) Re kgopela le re adime stapler ge le na le sona. [QP] [1] We ask-pres you we lend stapler I hold together paper this. We (I) are (am) asking if you can lend us (me) the/a stapler so that I can staple these sheets of paper. (18) Ke be ke re le ka se kgone go nkadima stapler naa? [QP] [1] I be-past I say-pres you cannot lend stapler by the way? I was wondering whether you could not lend me the/a stapler. [I was wondering: could you not lend me the/a stapler?] (19) Le ka nthuša ka stapler? Le na le stapler naa? [QP] [?] You can me help with stapler? You be-pres with stapler by the way? Can you help me with (lend me) a stapler? By the way, have you got one? (20) Ke kgopela thušo; ke nyaka go kgomagantšha mapheephe. [QP] [2] I ask-pres help; I want that to hold together paper. Can you help me? I need a stapler. (21) Jane Khuto o šomela kae? [MD] [4] Jane Khuto, she works where? Where does Jane Khuto work? (22) Le ka ba le tseba mošomi yoo ba mmitšago Jane Khuto ke kgopela nomoro ya phaphoše. [QP] [2] You if you know-pres worker who they call Jane Khuto I ask-pres number of office. Would you know where the office of (the worker called) Jane Khuto is? (23) Ke kgopela thušo, le ka mpotša moo Jane Khuto a šomago gona? [EP] [1] I ask-pres for help, you can me tell where Jane Khuto she work-pres where? I am asking for help, can you tell me where Jane Khuto works? I am asking for help, can you show me Jane Khuto’s office?
TSL[v.20020404] Prn:4/11/2002; 9:40
F: PBNS30.tex / p.23 (235)
Framing of requests in black South African English
(24) Nka hwetša kae motho yo ba rego ke Jane Khuto? [MD] [?] I can find where person that they call her Jane Khuto? Where can I find (the worker called) Jane Khuto? (25) Ke kgopela thušo. Ke nyaka Jane Khuto o šoma gona mo, e ka ba ofisi ya gage e mokae? [QP] [1] I ask-pres help. I want Jane Khuto that she work-pres here office of hers is where? Can I ask you to do me a favour? Where is Jane Khuto’s office? (26) Ga se la bona dimpampiri tšeo di setifailwego? [MD] [4] You not you see-past paper that they certity-past? Didn’t you see the documents that were certified? (27) Nka hwetša document yela? [MD] [4] I can find document that? Can I have/get that document? (28) 28) Ke be ke kgopela gore le ka se nthuše ka di certified document tšela tša ka naa? [QP] [1] I be-past I ask-pres that you can not me help with the certified document finish-past you by the way? I was wondering if you could not let me have the certified document, if it’s done.
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.1 (237)
Cultural scripts for French and Romanian thanking behaviour Tine Van Hecke University of Antwerp, Belgium
. Introduction Wierzbicka (1991, 1992, 1996) proposes the cultural script model in order to explore, from a maximally neutral and culture-independent perspective, both differences in ways of communicating and the underlying differences in attitudes and values. The term cultural script stands for the outlining of the tacitly accepted cultural rules which determine and influence people’s behaviour in a given society. The cultural scripts are formulated in what Wierzbicka calls a natural semantic metalanguage (NSM), a simple and straightforward language which contains only semantic primes representing universal human concepts. The use of the NSM for the cultural scripts is motivated by Wierzbicka’s criticism of three characteristics of traditional cross-cultural pragmatic investigation: 1. The use of general notions such as “directness”, “indirectness”, “solidarity”, “intimacy”, “spontaneity”, etc. without explaining what they exactly mean; 2. Ethnocentric perspectives, as Wierzbicka reproaches for instance politeness theory of Brown and Levinson; 3. Universalist presumptions, as Wierzbicka reproaches for instance Searle’s speech act theory. It should be stressed that the cultural scripts which will be proposed for French and Romanian thanking behaviour do not describe how all French and Romanian people thank or think about thanks, but rather that they represent norms that French and Romanian people are familiar with. In her semantic dictionary of English speech act verbs, Wierzbicka (1987: 214– 215) proposes definition (1) for the verb to thank:
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.2 (238)
Tine Van Hecke
(1) Thank a. I know that you have done something that is good for me b. I say: I feel something good towards you because of that c. I say this because I want you to know what I feel towards you d. I assume that you would want to hear me say this to you This all-round definition applies as well to the French and Romanian speech act verbs remercier and a mul¸tumi. However, in order to account for some differences between French and Romanian thanking behaviour, I will have to reduce it in some cases, and to further develop it in others.1
.
The perfunctory use of merci
The act of thanking is typically based on two mental states – gratitude and indebtedness – for which I will adopt the general definitions (2) and (3) inspired by Wierzbicka’s definition (1): (2) Indebtedness: I know that you have done something good to me (3) Gratitude: I feel something good towards you because you have done something good to me Of course, thanks are not necessarily defective if S doesn’t feel grateful or indebted. Bach and Harnish (1979: 52–54) distinguish between genuine and perfunctory thanks. By the former, S expresses gratitude and the intention that H believes that S is grateful to him, whereas by the latter, S expresses the intention that his utterance satisfies the social expectation to express gratitude at being benefited (without actually expressing gratitude). Perfunctory thanks are issued routinely or as a formality. They are generally regarded as an act of courtesy. Eisenstein and Bodman (1993) note that Americans characterize a sequence like (4) (4) Thanks. – Thank you. Have a nice day. as virtually automatic, as a social amenity, rather than an expression of gratitude. Rubin identifies it as a bald thank-you. Hymes (1971: 69) considers British thank you “on its way to marking formally the segments of certain interactions, with only the residual attachment to ‘thank-
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.3 (239)
Cultural scripts for thanking behaviour
ing’ in some cases”. Coulmas (1981: 91) notes a similar use of Australian English thank you, namely as a kind of verbalized punctuation mark of interaction. To sum up, we can sketch a hierarchy of the uses of thanking formulas in function of the effective presence of feelings of gratitude and indebtedness: (5) a. b. c. d.
genuine thanking perfunctory thanking mere (bald) thanking punctuation-thanking
It is said that French people often and easily say merci. As witness, an authentic declaration of a Spanish student sejourning in France, cited in Kerbrat-Orecchioni (1996: 89): Quand les Français te passent l’eau, le sel, il faut que tu dises à toute heure “merci”; si tu ne le dis pas, ils te remarquent. . . Je sais pas moi, si je le fais chez moi, ils rient, “tu pourrais me passer l’eau, s’il te plaît – merci”, chez moi ils se moquent et nous maintenons le même respect, non?
This observation could as well have been uttered by a Romanian. In the interaction between Romanian family members and close friends, thanks don’t occur as in French society. Moreover, inadvertent use of thanks in family circles is likely to suggest a greater distance and a lack of warmth and intimacy (cf. Apte (1974: 81–82) refering to Marathi and Hindi society). Obviously, the less frequent use of thanks among family members and close friends doesn’t mean that they feel less grateful or indebted. It is just that in French society, the (5b)-, (5c)- and (5d)-use of merci is more frequent. Merci can be used as a mere thanking, and even as a punctuationthanking. Appearing frequently in closing sequences, it is often postponed to the departure greeting, apparently without a well-determined propositional content: (6) Jonas ajouta un merci à son salut.
(Frantext.merci.2: 770)
(7) Au revoir et merci.
(Frantext.merci.2: 369)
(8) A bientôt, et merci.
(Frantext.merci.2: 488)
But particularly frequent is the perfunctory use of merci, as it seems alluded to in expressions such as s’en aller sans même dire merci, i.e. “it doesn’t mean a big thing to say merci, but it’s the least you should say”: (9) Il vient manger, et s’en retourne sans même dire merci. (Frantext.merci.2: 560) (10) Sans rien dire, toujours sympa, même pas merci, rien, il va s’enfermer dans la chambre de Jacqueline. (Frantext.merci.2: 371)
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.4 (240)
Tine Van Hecke
(11) [. . . ] mon père s’est éloigné, sans dire au revoir, sans un merci. (Frantext.merci.2: 688) The perfunctory use of merci can be translated in the NSM by omitting in Wierzbicka’s definition (1) the (c)-component which presupposes that S feels grateful and/or indebted: (1 ) a. I know that you have done something that is good for me b. I say: I feel something good towards you because of that d. I assume that you would want to hear me say this to you
. Indebtedness Although indebtedness is in the first place an assertive mental state, i.e. being aware of the fact is that one owes something to somebody, it can hardly be dissociated from a commissive mental state to discharge oneself of a duty, which can be split up into (12) and (13): (12) I must acknowledge my debt (13) I must compensate for my debt Romanians particuliarly feel the need to discharge themselves of their debt when accepting a benefit. Moreover, an offer or an invitation may be refused, because of the fear that one would be unable to compensate it, i.e. to give or to do something in return. So I tentatively propose script (14) for the Romanian concept of indebtedness:2 (14) a. I know: you have done something good to me b. I want you to know that I know this (= I want to acknowledge my debt) c. I want to do something good to you because of this (= I want to compensate for my debt) d. I feel bad if I cannot do this In comparison with the Romanian concept of indebtedness, in the French one, the components (c) and (d) are more likely to be absorbed by the component (b). So, for the French concept of indebtedness I would rather propose script (15): (15) a. I know: you have done something good to me b. I want you to know that I know this (= I can compensate my debt acknowledging it)
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.5 (241)
Cultural scripts for thanking behaviour
What could be the explanation of the difference between the French and the Romanian concept of indebtedness? The fact is that the way of evaluating the debt of S is essentially different in French and Romanian society. In both societies S evaluates his debt with the parameters such as utility of the benefit for S, cost for H, social rank of S and H, habituality of the benefit, etc. (cf. Held 1996). But regardless of these parameters, the evaluation of the debt is strongly influenced by a general cultural creed about benefits, which I will state and illustrate later. For the moment, I would like to resume in advance by saying that French people have a rather optimistic view of the debt contracted by the acceptance of a benefit, whereas Romanian people are especially preoccupied by the giving of a counter-gift to the benefactor: (16) French society: a rather optimistic view of debt (17) Romanian society: importance of a counter-service Assuming that the statements (14)–(17) are not completely way out, I will show that they have repercussions on the forms and the use of thanking formulas in both languages. Before listing these repercussions, I would like to stress two points: a.
I don’t claim that the features of thanking formulas I will mention confirm the statements (14)–(17); only that there is a correspondence between the two. b. It is a well known fact that frequency of occurrence of politeness formulas and their literal-meaningfulness are often inversely related: the more a certain formula is used, the less it is meant literally, and the more the pragmatical meaning – i.e. the meaning of the speech act it serves to perform – prevails upon the literal one. However, at least with regard to thanking formulas, the literal meaning is never completely lost: it remains a clue of the attitude the user of the formula thinks he should adopt towards the fact that he received a benefit.
. Repercussion of French and Romanian attitude to indebtedness on thanking formulas and responders . Standard responders Except the formula je vous/je t’en prie, French responders typically deny, or play down, the existence of the object of gratitude, and consequently the debt of the thanker:
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.6 (242)
Tine Van Hecke
(18) de rien / il n’y a pas de quoi / ce n’est rien / c’est vraiment peu de chose / c’est tout naturel / pas de merci Slama-Cazacu (1983) quoted pentru pu¸tin as the most frequent responder in Romanian. Nowadays, however, the most frequent responder is – maybe along with n-ave¸ti / n-ai pentru ce – cu (mult˘a) pl˘acere (“pleasure”), i.e. instead of turning down the object of gratitude, S recognizes its existence.3
. “(Ever)lasting” gratitude and indebtedness (19) Je vous suis reconnaissant (+det.). (20) V˘a/θti sunt recunosc˘ator (+det.). (21) Je vous suis redevable (+det.). (22) Je vous suis obligé (+det.). (23) V˘a sunt îndatorat (+det.). The assertion of S’s gratitude (19)–(20) or indebtedness (21)–(23) doesn’t occur frequently in French nor in Romanian. In our corpora it occurs however more frequently in Romanian than in French. Moreover, our Romanian corpus contains several occurrences in which S assures his benefactor of an eternal, permanent, everlasting gratitude or indebtedness. An example: (24) Permanent îmi amintesc cu nepieritoare recuno¸stin¸ta˘ de sfaturile pe care Excelen¸ta Voastr˘a a avut bun˘avoin¸ta s˘a mi le împ˘art˘as¸ easc˘a în memorabilele zile dinaintea plec˘arii mele din Berlin. (w.rom.392) Permanently I remember with imperishable gratitude the advice Your Highness had the good will to give me during the unforgettable days before my departure from Berlin. Also peculiar for the Romanian occurrences in which the thanker expresses his gratitude or indebtedness is the use of the verb a r˘amâne (“to remain”), instead of a fi (“to be”). In the example (25), the expression a r˘amâne îndatorat supplements the proper thanks. This makes it sound as a commissive speech act, rather as just an expressive speech act of manifesting S’s indebtedness: S engages himself verbally to a virtual act of compensating H: (25) – Atunci, spuse medicul, totul e-n regul˘a. Nu mai ave¸ti nevoie de mine. – Nu, s¸ i v˘a mul¸tumesc foarte mult. Ce v˘a datorez? – Nimic. – Cum nimic? – L˘asa¸ti, cu alt˘a ocazie. Oricând ave¸ti nevoie, veni¸ti pe la mine.
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.7 (243)
Cultural scripts for thanking behaviour
– V˘a mul¸tumesc foarte mult s¸ i v˘a r˘amân îndatorat. – La revedere s¸ i s˘an˘atate. (B˘aie¸su I., Chi¸timia, p. 92)4 – So, said the doctor, everything is alright. You don’t need me anymore. – No, and I thank you very much. How much do I owe you? – Nothing. – What do you mean, nothing? – It’s alright, leave it for next time (lit.: for another occasion). Whenever you need me, come and see me. – I thank you very much, and I remain indebted to you. – Good bye, and I wish you health. The line of thought underlying the use of this expression could be described as in (26): (26) a. b. c. d. f.
I think: you’ve done something good to me I think: if I could, I would do something good to you because of this I cannot do this now I think: I will do it when I can I want you to know that I think this
It may be that S uses the expression v˘a r˘amân îndatorat/recunosc˘ator without thinking (d), and even without thinking (b). Nonetheless, the fact that he uses this expression shows that he realizes that he is expected not to forget his indebtedness. The evocation of the intention to reward the benefactor, occurs also frequently in indirect requests: (27) Am 28 de ani, sunt din Bucure¸sti s¸ i caut pe cineva care m˘a poate ajuta s˘a pot pleca în SUA sau Canada, eu neavând rude acolo sau bani pentru a putea pleca pe cont propriu. Eventualului sponsor doresc s˘a-i mul¸tumesc de pe acum s¸ i s˘a-l asigur c˘a nu va regreta nici o clip˘a ajutorul pe care mi-l va acorda. (w.rom.217) I am 28 years old, I am from Bucharest and I am looking for somebody who could help me to get to the USA or Canada, since I don’t have family over there, nor money to get there. I wish to thank in advance the eventual sponsor and to assure him that he won’t regret a moment the help he will give me. (28) V-am r˘amâne îndatora¸ti dac˘a a¸ti putea ad˘auga în paginile dumneavoastr˘a un link c˘atre site-ul firmei noastre. (w.rom.528) I would remain indebted if you could add in your pages a link to the site of our firm.
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.8 (244)
Tine Van Hecke
. Verbal counter-gifts If one is unable to reward the benefactor by a physical action, one can give a spiritual counter-gift. In our Romanian corpus thanks are often followed by a wish for the benefactor: the thanker wishes him success, health, that he or his children may stay alive, that God may bless him, or reward him for the benefit: (29) Activitatea desf˘as¸ urat˘a de Dv. consider c˘a poate fi de mare ajutor lumii bursei. V˘a mul¸tumesc pentru datele trimise s¸ i v˘a urez succes în viitor! (w.rom.497) I consider that your activity is very helpful for the market world. I thank you for the data you have sent me and wish you success in the future! (30) Margareta, te invit s˘a vii s˘a cân¸ti s¸ i în Elve¸tia, pentru Românii de aici s¸ i pentru Elve¸tieni. Mul¸tumesc cu s˘an˘atate s¸ i î¸ti doresc din tot sufletul la mul¸ti ani! (w.rom.183) Margareta, I invite you to come and sing also in Switzerland, for the Romanians here and the Swiss. I thank you and wish you with the whole of my heart health and many more years! (31) V˘a mul¸tumim c˘a ne citi¸ti, v˘a mul¸tumim pentru vorbele bune, dar mai ales pentru dorin¸ta dvs. de a participa, acum, în pragul Cr˘aciunului, la sprijinirea atâtor neferici¸ti. Dumnezeu s˘a v˘a r˘aspl˘ateasc˘a înmiit s¸ i s˘a v˘a g˘asi¸ti acolo unde a¸ti plecat, patria pe care o merita¸ti! (w.rom.364) We thank you for reading us, we thank you for the good words, but most of all for your wish to participate, now, in the eve of Christmas, to the supporting of so many unlucky people. May God reward you a thousand times and may you find there where you went, the fatherland you deserve! Slama-Cazacu’s (1985: 44) observation about the typical Romanian obligatory strings of politeness formulas in dialogue sequences, is still up to date: “It is still the use (and it was a general obligation) to answer by a PF [politeness formula] to the answer of a partner whom you ask about his/her age, children, professional status, a new acquisition (new house or clothes, a big animal – cow, horse, pig – recently bought etc.); after that, another F follows, of thanks.” This thanking is systematically dubbed by the return of the wish. (32) is an example of the corpus of Slama-Cazacu (1985: 44): (32) – Ai copii? – Da. – Câ¸ti? – Trei. - S˘a-¸ti tr˘aiasc˘a. – Mul¸tumesc, s¸ i-ai dvs. – Câ¸ti ani au? – Cel mare, zece ani. – Mul¸ti-nainte. – Mul¸tumesc, s¸ i dvs. la fel. – Have you (any) children? – Yes. – How many? – Three. – I wish you joy of them (ad.lit. May they live (well) [to you]). – Thank you, the same to yours.
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.9 (245)
Cultural scripts for thanking behaviour
– How old are they? – The oldest is ten. – May he reach old age ((ad. lit. (May he live still) many more (years)). – Thank you, (I wish) the same to you.
. A fi recunosc˘ator/a mul¸tumi meaning to “compensate” It is not unusual that a mul¸tumi, recuno¸stin¸ta˘ and a fi recunosc˘ator are used metonymically to mean the physical act of compensating the benefactor – instead of the speech act of thanking, respectively the feeling of gratitude and the assertion of being grateful: (33) [A doctor coming home catches in his apartment the woman from next door, with his dirty linen in her arms:] – De ce trebuie s˘a-mi speli dumneata rufele mele? url˘a el. Te-am rugat eu? De ce? Ea deschise u¸sa s¸ i încerc˘a s˘a fug˘a, dar el o opri. – Nu pleci pân˘a nu-mi r˘aspunzi! – Vream s˘a v˘a fac s¸ i eu un bine, ca vecin˘a, m˘arturisi ea speriat˘a. – Da’ de ce s˘a-mi faci un bine? – B˘arbatu-miu a zis c˘a fiindc˘a i-a¸ti scos piatra de la rinichi s¸ i fiindc˘a n-a¸ti vrut s˘a lua¸ti nimic, zicea s˘a v˘a sp˘al rufele. . . C˘a eu tot sp˘al. . . Noi nu v-am fost recunosc˘atori pân˘a acum cu nimic. . . (B˘aie¸su I., Treizeci ¸si opt cu doi, p. 32) – Why do you have to wash my clothes? he cried out. Did I ask you to do so? Why? She opened the door and tried to slip away, but he held her. – You won’t leave until you answer me! – I wanted to do something good for you, as your neighbour, she confessed anxiously. – But why should you do something good for me? – My husband said that since you took out his kidney stone and you didn’t want to accept anything, he told me to do your laundry. . . I do our wash anyway. . . Until now we haven’t been grateful to you with anything. . . (34) Cei care au primit un sprijin moral sau material de la guvernul Antall, acum se simt datori s˘a fie recunosc˘atori fa¸ta˘ de cei care i-au ajutat. (w.rom.742) Those who got a moral or material support from the government Antall, now feel obliged to be grateful to those who helped them. (35) Face valv˘a in prezent, arestarea chirurgului din Tecuci. Motivul? Cel bine cunoscut de acum, luarea unor bani (200.000 lei) de la un pacient. [. . . ]
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.10 (246)
Tine Van Hecke
am fost s¸ i am r˘amas un adversar declarat al tuturor celor ce accept˘a [. . . ] pentru mit˘a s˘a-¸si fac˘a datoria. [. . . ] gestul unui medic [. . . ] care cade întrun astfel de p˘acat, mi se pare odios. Si, ¸ mi se pare [. . . ] în plus odios, atunci când a¸sa-zis˘a recuno¸stin¸ta˘ a pacientului este provocat˘a, condi¸tionat˘a de c˘atre medic [. . . ]. (w.rom.743) A presently vexing question is the arrest of the surgeon of Tecuci. Why? The well-known reason of accepting money (200.000 lei) of a patient. [. . . ] I have always been and I still am a declared adversary of all those who accept to do their duty for a bribe. The gesture of a doctor [. . . ] who commits such a sin, seems awful to me. And, I regard it as more awful still, if the so-called gratefulness of the patient has been elicited, conditioned by the doctor [. . . ]. (36) [From a letter of a Romanian to a Dutchman:] V˘a mul¸tumim foarte mult pentru lucrurile primite, care mi se potrivesc foarte bine. Nu s¸ tiu în ce împrejurare v-a¸s putea mul¸tumi. Dar poate, cândva, o s˘a pute¸ti veni în România s¸ i poate atunci voi avea posibilitatea s˘a v˘a mul¸tumesc pentru ceea ce a¸ti f˘acut pentru mine. I thank you very much for the things you have sent me, and which fit very well. I don’t know in what circumstances how I could ever thank you. But maybe, sometime, you will be able to come to Romania, and maybe then I will have the possibility to thank you for what you have done for me.
. Historical explanation? Since, in identical situations, people behave in identical ways, the differences in the French and Romanian conceptualization of indebtedness should be explained historically, rather than by saying that we are dealing with inherent features of French and Romanian people. As one possible explanation I propose the following: When institutions such as administration of justice, health services, education, and other public services function normally, that is, when their behaviour corresponds to what people consider to be judicious and reliable, people feel respected and supported by them. This allows a heart patient to think: “It’s only natural that the health service takes care of me”; or a traveller who has just been informed that his place on the plane has been double-booked: “It’s only natural that the air company solves my problem and offers me another ticket”; or a student who is given by a professor a detailed bibliography for his Master’s thesis: “It’s only natural that the professor helps me”; and in addition these three persons may think: “I don’t have to feel
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.11 (247)
Cultural scripts for thanking behaviour
indebted to the surgeon, the employee of the air company, or the professor, since this benefit is my right and their duty”. Consequently, it is enough to say something like “merci”. If, on the contrary, institutions do not function as people expect them to do, the norms concerning institutional rights and duties get blurred too. This is, it seems, what has been going on to a certain extent in Romanian society in the last decades of the 20th century: it has not always been considered as obvious that an institution – or concretely, the personnel working in an institution – provide you with something that, as a matter of fact, it is your right to get. Even if the personnel are wanting to help you, they may be unable to do so (because of the disfunctioning of the institution). As a result, if someone obtains what he has asked for, he will be likely to feel indebted (not to the institution, but to the personnel who made an extra effort). This is why the Romanian cardiac patient, not taking it for granted that he has been operated on in optimal conditions, is willing to reward the surgeon who took care of him; the Romanian traveller who obtains a place on the overbooked plane, will be ready to compensate the employee of the air company; the Romanian student, knowing about the unfair loan of his professor, will feel the need to insist on his gratitude for the fact that the professor sacrificed some of his time for him. Those examples reflect French and Romanian people’s attitude to institutional benefits, but it seems that, in the course of time, this attitude has grown out to a general cultural creed, which I would venture to articulate for French society as in (37), and for Romanian society as in (38): (37) People can and want to help me. (38) People help me, if they want to help me, and if they can help me.
. Final questions 1. The aim of the NSM and the cultural scripts is to avoid as much as possible an ethnocentristic approach in the comparison of cultural norms and values. Therefore, if we want to formulate cultural scripts for other cultures, we must articulate them – as Wierzbicka (1991: 10) puts it – in terms that are both “theirs” and “ours”. The NSM contains only one pair of evaluators: good/bad. But isn’t it dangerous to translate cultural norms and values in terms of good and bad? Surely, the sense of the term norm comprises good and bad; and the negation in the NSM allows us to nuance the evaluation along the scale (39):
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.12 (248)
Tine Van Hecke
(39) good – not bad – not good – bad The problem that remains is that, by the lack of other evaluators, good and bad must do duty for considerably different evaluators, such as usual/unusual, necessary/optional, face flattering/face attacking, etc. No doubt, these more specific evaluators are in the mind of the researcher trying to formulate cultural scripts. Consequently, the risk is that, even if the terms good and bad could possibly be considered as both “theirs” and “of the researcher”, the choice of the use of one of them remains for the researcher to make, who, as Wierzbicka (1991: 9) admits, cannot place himself outside all cultures. 2. Although the cultural script approach is in the first place a descriptive one, it would considerably gain in utility if it could predict the use of certain formulas. It is obvious that the NSM, because of its very restricted lexicon, cannot do this: it would be indispensable to introduce terms such as woman, man, child, small/great benefit, . . . The reason why those words are not included in the NSM-lexicon is that they are not universal, or involve too many connotations when they are translated into other languages. Wierzbicka claims that the primes are translatable into all languages and are at least partial equivalents. Rather than hanging on to the chimera of equivalents, wouldn’t it be more fruitful to push the compromise a bit further and enrich the NSM lexicon with items allowing us to predict the use of formulas? As an example, I propose (40), a use-script for the Romanian s˘aru’mâna (lit.: “I kiss your hand”): (40) Use-script for s˘aru’mâna Typical objective: thanking/greeting + showing respect for H Typical identity of S and H: a man/a child to a woman a child to an adult (but not, for instance, a woman to a man) Typical relation between S and H: (family), (closed) Typical occurrence: after a meal: “— pentru mas˘a” Specifically: greeting/thanking to a priest/a monk Such a guideline could be useful not only for foreigners, but even for some Romanians, as is shown in a piece of evidence (41), the answer of a Romanian lady in one of my questionnaires on Romanian politeness formulas:
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.13 (249)
Cultural scripts for thanking behaviour
(41) În ce fel de circumstan¸te folosi¸ti formula s˘aru’ mâna? – Copil fiind, îi salutam astfel pe cei mari. Când am crescut, au existat unele situa¸tii genante, deoarece eu continuam s˘a le spun b˘arba¸tilor “s˘aru’ mâna”, iar ei, considerându-m˘a deja domni¸soar˘a, îmi r˘aspundeau la fel. In what circumstances do you use the formula s˘aru’mâna? – As a child, I used it as a greeting to adults. When I grew up, there have been some embarrassing moments, because I continued to say “s˘aru’mâna” to men; and regarding me already as a lady, they answered the same to me.
Notes . This study is based upon a corpus collected during about six years (1994–2000), including data from the French databanks Frantext (Institut National de la Langue Française) (encoded Frantext.merci.2), Europresse (CEDROM-SNi), occurrences of thanks on the World Wide Web (encoded w.rom), dialogue fragments of French and Romanian literature, field notes of naturally occurring discourse, and written responses from native speakers to questionnaires. . Cf. the cultural script Wierzbicka (1991) proposes for the Japanese kansha suru. . Slama-Cazacu notified cu pl˘acere as being a rather new formula. Other responders, beside pentru pu¸tin and cu pl˘acere, are: pentru nimic (notified by Slama-Cazacu as being less frequent); s˘a v˘a/s˘a-¸ti fie de bine (after a meal) (“may it be good for you); s˘a tr˘ai¸ti/s˘a tr˘aie¸sti (especially at the receipt of an alm); s˘a fii s˘an˘atos (regional (Slama-Cazacu 1983: 301)). . Notice that the doctor in his reply L˘asa¸ti, cu alt˘a ocazie (“don’t mention it, (you will compensate me) on another occasion”) – which is actually not an unusual answer to a compensating-offer – first refuses the offer, but then immediately recalls his refusal.
References Apte, M. L. (1974). “Thank you” and South Asian Languages: A Comparative Sociolinguistic Study. International Journal of the Sociology of Language, 3, 67–89. Bach K., & Harnish, R. M. (1979). Linguistic communication and speech acts. Cambridge: MIT Press. Coulmas, F. (1981). Poison to your soul: Thanks and apologies contrastively viewed. In F. Coulmas (Ed.), Conversational Routine (pp. 69–91). The Hague: Mouton. Eisenstein, M., & Bodman, J. (1993). Expressing Gratitude in American English. In G. Kasper & Sh. Blum-Kulka (Eds.), Interlanguage Pragmatics (pp. 64–81). Oxford: Oxford University Press.
TSL[v.20020404] Prn:4/11/2002; 9:42
F: PBNS31.tex / p.14 (250)
Tine Van Hecke
Held, G. (1996). Two polite speech acts in contrastive view: Aspects of the realization of requesting and thanking in French and Italian. In M. Hellinger & U. Ammon (Eds.), Contrastive Linguistics (pp. 363–384). Berlin: Mouton de Gruyter. Hymes, D. (1971). Sociolinguistics and the Ethnography of Speaking. In E. Ardener (Ed.), Social Anthropology and Language. London: Tavistock Publications. Kerbrat-Orecchioni, C. (1996). La conversation. Paris: Editions du Seuil. Slama-Cazacu, T. (1983). Formule de polite¸te în limba român˘a: structur˘a s¸ i func¸tii. II. Studii ¸si Cercet˘ari Lingvistice, 14 (3), 229–236. Slama-Cazacu, T. (1985). The Concept of Politeness and its Formulas in the Romanian Language. In J. Fishman (Ed.), Festschrift for Ch. Ferguson. Berlin: Mouton de Gruyter. Wierzbicka, A. (1987). English Speech Act Verbs. A Semantic Dictionary. Sydney: Academic Press. Wierzbicka, A. (1991). Cross-Cultural Pragmatics. The Semantics of Human Interaction. Berlin: Mouton de Gruyter. Wierzbicka, A. (1992). Semantics, Culture, and Cognition: Universal Human Concepts in Culture-Specific Configurations. New York: Oxford University Press. Wierzbicka, A. (1996). Contrastive sociolinguistics and the theory of “cultural scripts”: Chinese vs. English. In M. Hellinger & U. Ammon (Eds.), Contrastive Sociolinguistics (pp.313–344). Berlin: Mouton de Gruyter.
TSL[v.20020404] Prn:3/09/2002; 12:20
F: PBNS32.tex / p.1 (251)
Sociocultural variation in native and interlanguage complaints Ronald Geluykens and Bettina Kraft Universität Münster, Germany
Introduction Complaints as Face-Threatening Acts (FTAs) This paper reports on a fieldwork investigation into the linguistic realization of complaints in English and, to a lesser extent, German. The basic assumption is that this linguistic realization depends on the degree to which complaints threaten ‘face’, the term ‘face’ being defined as “the positive social value a person effectively claims for himself by the line others assume he has taken during a particular contact. Face is an image of self delineated in terms of approved social attributes (. . . )” (Goffman 1967: 5). This notion of face was further developed by Brown and Levinson (1987), who make a distinction between positive and negative face. The former is defined as “the positive consistent self-image or ‘personality’ claimed by the interactants” (Brown & Levinson 1987: 66); the latter as “the basic claim to territories, personal preserves, rights to non-distraction” (ibid.). From this perspective, complaints can be classified as acts which intrinsically threaten face, or face-threatening acts (FTAs). Given their high degree of intrinsic face-threat (cf. infra), complaints certainly deserve more attention (but see Boxer 1993; Frescura 1995; Trosborg 1995). On the whole, complaints have received less attention in the interlanguage pragmatics literature than some other FTAs such as, say, requests or apologies (e.g. Blum-Kulka et al. 1989; Trosborg 1987). Complaints can be characterized as follows: “In the speech act of complaining, the speaker (S) expresses displeasure or annoyance – censure – as a reaction to a past or an ongoing action, the consequences of which are perceived by S as affecting her unfavorably. This complaint is usually addressed to the hearer (H) whom the S holds, at least partially, responsible for the offensive action” (Olhstain & Weinbach 1993: 108). In terms of the main face type threatened
TSL[v.20020404] Prn:3/09/2002; 12:20
F: PBNS32.tex / p.2 (252)
Ronald Geluykens and Bettina Kraft
by complaints, it would probably be fair to say that the main face-threat is to the hearer rather than to the speaker (in that the speaker holds the hearer responsible for the censured action), and that it is mainly the hearer’s positive face which is being threatened (in that the hearer runs the risk of having his/her positive self-image negatively affected by the complaint). This view is also held by Brown and Levinson, who claim that complaints “show that S[peaker] has a negative evaluation of some aspect of H[earer]’s positive face” (Brown & Levinson 1987: 66). Crucially, for our purposes, a correlation is claimed to exist between the degree of face-threat of a particular FTA, and the degree of indirectness with which it will be realized (see also Geluykens 1999). Roughly speaking, the more risk of face loss an FTA potentially involves, the less direct a speaker can afford to be, since being too direct may result in coming across as impolite. The least direct way of realizing an FTA is ‘off-record’, with the writer not going on record as having performed the FTA, but only hinting at it. Conversely, FTAs with a relatively low risk of face loss will be realized in a fairly direct (or ‘on-record’) manner, since employing too much indirectness will make the FTA look more face-threatening than it actually is. At its most direct, the FTA will be expressed ‘bald-on-record’, i.e. without any politeness strategies whatsoever (e.g. formulating a request as an imperative). In this empirical investigation, we will investigate complaint strategies in native English, native German, and German-English Interlanguage (i.e. English produced by native speakers of German). By collecting controlled elicitation (DCTs; cf. below) data from these three different languages, we will be in a position to carry out a contrastive pragmatic analysis of complaint strategies. It is important to stress from the outset that we do not equate complaints with ‘speech acts’, in the sense that one complaint equals one utterance. Rather, a complaint strategy ‘unit’ may (but need not) consist of more than one utterance (cf. (1) and (2) below for examples); in such cases, we can make a distinction between the complaint proper, or main complaint strategy, and one or more supportive moves. The latter form an integral part of the complaint FTA, and will be examined separately below.
Research questions Complaints can be argued to be an inherently weighty type of FTA. As Trosborg points out, “[t]hese acts are by their nature designed to cause offence and they are therefore highly threatening to the social relationship between speaker and hearer” (Trosborg 1995: 311). As a result, one would expect them to be realized with a fairly high degree of indirectness. Our first assumption, then, is that a high proportion of complaints will be performed off-record, or at the very least that they will often be accompanied by downgrading material. Secondly, we will implicitly be examining intercultural variation, in the sense that we are dealing with native German data as
TSL[v.20020404] Prn:3/09/2002; 12:20
F: PBNS32.tex / p.3 (253)
Sociocultural variation in native and interlanguage complaints
well as native (British) English and IL English data. It remains to be seen to what extent complaining behaviour is different in Germany and Great Britain. Thirdly, and more importantly, we will investigate pragmatic aspects of Interlanguage English. By carrying out a contrastive analysis between these IL data and the native language as well as the target language, we will be able to examine systematically any differences in complaint realization, and evaluate to what extent such differences might be attributable to pragmatic transfer from the interlanguage speakers’ L1. The final dimension, and perhaps the most central one to our investigation, is that of socio-cultural variation. Several social variables may potentially have an impact on the realization of complaint FTAs. The variables which especially interest us here are, firstly, gender (we will concentrate exclusively on the gender of the speaker, although the gender of the addressee is also potentially relevant), and, secondly, social distance (i.e. the degree of familiarity between speaker and addressee). With regard to gender, despite the fact that there is quite some evidence of genderspecific verbal behaviour (cf. Coates 1986, 1998; Holmes 1995; Tannen 1994 for recent overviews), there has been little work on gender-specific realizations of FTAs, especially in the context of interlanguage pragmatics (but see Blum-Kulka et al. 1989, among others, for some discussion). As far as social distance is concerned, Brown and Levinson (1987) explicitly claim that it is one of the three factors influencing the weightiness of an FTA (the two others being, respectively, the power balance between speaker and speaker, and the absolute ranking of impositions in the particular culture).
Data-collecting method The data-gathering method employed here is that of Discourse Completion Tasks (cf. Blum-Kulka & Olshtain 1984; Blum-Kulka et al. 1989). While this method certainly has certain drawbacks – the non-naturalness of the task for respondents not being the least of them – these are outweighed by the advantages, the major one being that DCTs allow us to gather a body of comparable data for the three ‘languages’ (i.e. including IL English), collected under controlled circumstances, which can be used as input for our contrastive analysis. As pointed out before, two sociocultural variables were deemed relevant, and are therefore built into the DCT: sex of the speaker and social distance. The former is an obviously binary variable, while for social distance we have created a threefold distinction into intimates (i.e. people you know as friends), acquaintances (i.e. people you hardly know), and strangers (i.e. people you do not know at all). Combining these variables gives six possible scenario types, illustrated by the six tokens in (1) below (cf. Kraft 1999 for more details):
TSL[v.20020404] Prn:3/09/2002; 12:20
F: PBNS32.tex / p.4 (254)
Ronald Geluykens and Bettina Kraft
(1) 1. Your best friend Jane has borrowed your favourite (and very expensive) leather jacket. When she returns it, you discover a big hole in the back. You say to her: [Female addressee/Intimate] 2. The night before your oral exams the neighbour next-door is throwing a noisy party. You go over and say to him: [Male addressee/Acquaintance] 3. You are meeting with your friend George in a pub. When he arrives twenty minutes after the appointed time you say: [Male addressee/Intimate] 4. You are at an expensive restaurant. After you have waited three quarters of an hour for your meal, it turns out to be cold and much too salty. You stop the waiter and say to him: [Male addressee/Stranger] 5. While you are seeing a film in a cinema, the woman next to you continues to make loud noises with a plastic bag from which she is eating s.th. You lean over and say: [Female addressee/Stranger] 6. There is a new student in your class. In order to help her, you have lent her some books but told her you needed them back soon. Today you meet her again and find that she has forgotten to bring the books with her. You say: [Female addressee/Acquaintance] German L1 respondents were, of course, given a German version. Scenarios were randomly ordered, and differently ordered versions were distributed. A grand total of 247 DCTs was analyzed; of these, 84 were native German (GL1), 79 native English (EL1), and 84 German-English Interlanguage (EIL). Care was taken to maintain a balance between male and female respondents (125 vs. 122, respectively). All respondents were university students, in order to keep the groups as homogeneous and comparable as possible.
Empirical analysis Length of complaint The first variable investigated is the length of the complaint (i.e. the total length of the complaint strategy, including additional supportive moves). Whereas both the English and German complaints have a mean length of 8 words, the Interlanguage complaints’ mean length is 14 words. This difference in length is due to two factors. First of all, the number of main complaint strategies employed is higher: 312 for the IL data, versus 231 and 235, respectively, for the L1 German and L1 English data. Secondly, there is a higher number of supportive moves in the IL data, viz. 617,
TSL[v.20020404] Prn:3/09/2002; 12:20
F: PBNS32.tex / p.5 (255)
Sociocultural variation in native and interlanguage complaints
versus 441 and 398, respectively, for the native German and English complaints. Two examples will illustrate this: (2) Hi would you mind keeping it down a small bit I have exams tomorrow. (EL1) (3) Excuse me, but I’ve got my oral exams tomorrow and I would appreciate it if you turn down the music a bit. And, please, be more quiet. I’ll invite you for a beer tomorrow night. (EIL) Whereas the native English example contains only one complaint strategy plus one supportive move, the interlanguage example contains several of both, resulting in a much longer complaint response. This confirms the “waffle” phenomenon reported in other studies.
Type of complaint strategy The second linguistic variable under analysis is the main complaint type employed by respondents. In order to keep things simple, we have classified complaints into two major categories. The first category is labelled ‘on-record’, and contains all instances in which the speaker explicitly signals that he/she is performing a complaint. Types of on-record strategies include: expression of disappointment, toneddown accusation, and explicit criticism, among others (after Trosborg 1995). Examples from the data are: (4) I’m really disappointed. I have been waiting here for over 20 minutes. (EIL) (5) Hey, you really get me into trouble [sic].
(EIL)
The second cetagory includes all complaints which are not on-record. In most of these cases, the speaker performs the complaint ‘off-record’, i.e. without going on record as having formulated a complaint, as in (6): (6) Ahm, Jane, what happened to my jacket?
(EIL)
Off-record complaints include hints, expressions of worry, use of irony, and questions about the (cause of the) situation. In example (6), the speaker merely hints at there being something wrong, by virtue of asking a question about the state of the jacket; no overt complaint is used. Given that complaints have a fairly high degree of face-threat, one would expect the majority of them to be performed offrecord. However, the speaker may also opt out of the complaint althogether, as in (7), where the following meta-comment was provided by a respondent: (7) I wouldn’t go over but use my ‘ear plugs’.
TSL[v.20020404] Prn:3/09/2002; 12:20
F: PBNS32.tex / p.6 (256)
Ronald Geluykens and Bettina Kraft
Table 1. Type of primary complaint strategy (in %)
Off-record On-record Total
German L1
English L1
Interlanguage G-E
59.4% 40.6% 100% (231)
65.1% 34.9% 100% (235)
34.1% 65.9% 100% (312)
Such opting out was not expressly disallowed in our DCTs, and, though it rarely occurs, should not be ignored. Since this opting-out strategy is about the most offrecord a speaker can get, we have grouped these instances together with off-record strategies in Table 1. Although the above table does not incorporate gender-related variation, we should point out here that our results, such as they are, do not support the stereotypical claim that women are less ‘direct’ than men. On the contrary, in each of the three languages, there is a slightly higher proportion of on-record complaints used by women than by men. Figures are, respectively, 46.9% versus 34.3% for native German, 36.9% versus 32.8% for native English, and 72.1% versus 59.6% for IL English. Crucially, Table 1 confirms our hypothesis that complaints are often performed off-record, at least for the native language data. Both in L1 English and L1 German, there is a majority of off-record complaints. Furthermore, there is no significant difference between EL1 and GL1 in this regard. The same does not hold for the IL English data, however: about two thirds (65.9%) of the IL complaints are onrecord formulations. Clearly, non-native speakers tend to be more direct in their complaining behaviour, which may lead to face-saving problems.
Types of supportive moves In order to facilitate contrastive analysis, we have subdivided supportive moves into three major categories, according to the type of impact they have on the perceived face-threat of the FTA: Solidarity-enhancing, Neutral, and Confrontational. Solidarity-enhancing moves encompass apologies, signs of acquiescence, and showing understanding. Neutral supportive moves are, for instance, those suggesting an alternative, pointing to one’s own wants, or requests for conciliation. The category of confrontational moves, finally, contains orders, warnings, insults/swearing, and threats to abandon the situation. Examples of each category can be found in (8), (9), and (10), respectively: (8) I’m usually the one who’s late. You’re lucky. But the first drink is on you. (EIL)
TSL[v.20020404] Prn:3/09/2002; 12:20
F: PBNS32.tex / p.7 (257)
Sociocultural variation in native and interlanguage complaints
(9) You don’t seriously expect me to eat this. I didn’t wait 45 minutes for something as cold as this!
(EIL)
(10) If you’re still hungry you’d maybe better go outside and finish your noisy meal. (EIL) These examples contain, respectively, an explicit show of understanding, a pointing towards the speaker’s own wants, and an explicit instruction to the hearer. In terms of face-saving work, solidarity-enhancing supportive moves are obviously more useful than neutral moves, while one could argue that confrontational moves have the opposite effect of increasing the degree of directness of the FTA. Table 2 below contains an overview of supportive moves in the three languages; results are given (in percentages) for women (F) and men (M) combined (cf. the middle column in each language group) as well as separately (cf. other columns). Several conclusions can be drawn from these figures. First of all, as was already pointed out, the IL data contain a much higher proportion of supportive moves on the whole (N = 617; cf. bottom row in Table 2). Furthermore, when one looks at the difference between F and M speakers in absolute numbers (cf. second row from bottom in Table 2), it emerges that there is a higher number for F speakers in each of the three language groups. However, this difference only reaches a significant level in the IL English data (370 vs. 247). Thirdly, neutral moves are by far the most frequent type of supportive move encountered in the data; this holds across the board, for all subcategories. Finally, at least in two out of three of the language groups, there is a higher frequency of confrontational moves for female speakers than there is for male ones; this holds for both GL1 and EIL complaints. The same does not hold for EL1 complaints, however, where male speakers use more solidarity-enhancing supportive moves. These gender-related results deserve more attention than we can give them here, but, for the time being, we can conclude that the relationship between gender and complaining behaviour is indeed a complex one. Table 2. Types of supportive moves accompanying complaint strategies (in %)
Solidarity
German L1 F
M
13.7
12.1
English L1 F 9.0
12.9 Neutral
80.6
74.0
5.7
80.1
13.9
[226]
15.3
24.3
28.7 26.5
74.0
73.0
10.9
10.7
[221]
2.7
13.4 8.0
[177] [398]
57.9 65.5
10.8 [215]
[441]
M
77.1
9.8 Total
English IL F
12.1
77.3 Confronting
M
[370]
[247] [617]
TSL[v.20020404] Prn:3/09/2002; 12:20
F: PBNS32.tex / p.8 (258)
Ronald Geluykens and Bettina Kraft
Use of downgraders and social distance In order to investigate the effect of social distance, we have concentrated on the occurrence of downgraders or hedges. Firstly, we checked the frequency of occurrence of ‘please’ and its German equivalent ‘bitte’ (see also House 1989 for a detailed study on requests); results can be found in the upper half of Table 3 below. Secondly, we did the same for the apologizing expressions ‘Excuse me’/ ‘(I am) sorry’ and their German equivalents ‘Entschuldigung’ / ‘Es tut mir leid’ (cf. lower half of Table 3). Since both types of downgrading are relatively frequent in our data, they readily lend themselves to quantitative evaluation. Naturally, other downgraders occur in the data. In fact, although a full discussion and overview of all possible downgraders is outside the scope of this paper, a few examples will make clear that downgraders can be combined within one complaint: (11) Please be a little quieter thanks.
(EL1)
(12) Entschuldigen Sie bitte. Das Essen ist kalt und wohl etwas versalzen. (GL1) In (11), apart from ‘please’, the speaker also uses ‘a little’ to quantify the degree of quiet he would like. Likewise, in (12), apart from apologizing and using ‘bitte’, the speaker employs the downtoning adverbs ‘wohl etwas’. Such downgraders, labelled ‘hedges’ by Brown and Levinson, “may be used to soften FTAs of suggesting or criticizing or complaining, by blurring the speaker’s intent” (Brown & Levinson 1987: 117). Turning now to Table 3, there is a clear correlation between degree of intimacy on the one hand, and frequency of occurrence of certain downgraders on the other. To be more precise: the higher the social distance between speaker and addressee, the higher the number of downgraders. With regard to the use of apologies, there is a clear decrease in all three languages, apologies in all three cases being most frequently used with strangers, less so with acquaintances, and least of all with Table 3. Correlation between social distance, gender, and use of downgraders
‘Please’/‘bitte’ GL1 EL1 EIL Apologies GL1 EL1 EIL
Strangers
Acquaint.
Intimates
TOTAL
F
M
33 45 63
46 48 56
2 4 4
81 97 123
50 44 71
31 53 52
14 19 77
5 7 30
– 1 6
19 27 115
12 16 75
7 11 40
TSL[v.20020404] Prn:3/09/2002; 12:20
F: PBNS32.tex / p.9 (259)
Sociocultural variation in native and interlanguage complaints
intimates. Note also that apologies are much more frequent in the IL data (N = 115) than they are in the native language data. When one looks at results for ‘please’/‘bitte’, results are less clear-cut: there is a marked difference between intimates on the one hand, and strangers and acquaintances on the other hand, but no significant differences are found between the categories of strangers and acquaintances (in fact, in the German native data, ‘bitte’ occurs slightly more often in interactions with acquaintances than with strangers). Once again, this downgrader is more frequent in the IL data than in the native data, though not very significantly so in this case. We also had a look at the influence gender (once again, of the speaker) has on the use of downgraders. As can be observed in the final two columns in Table 3, there is a tendency towards increased downgrading with female speakers. However, this does not apply across the board, in that ‘please’ occurs less frequently with male speakers than with female ones in the native English data. This is, to some extent, in line with what was found in the previous section, where there was a higher degree of correspondence between IL English and L1 German than between L1 English and the other two languages. One might wonder here, then, whether there are systematic cross-linguistic differences between English and German, and whether the findings for IL English are due to the fact that the IL speakers transfer the strategies they employ in their German L1. One final word needs to be said about the overall frequency of downgraders (i.e. excepting the ones already mentioned). Apart from the fact that other downgraders occur far less often, it should be pointed out here that the interlanguage data differ from the native data, not so much with regard to the relative frequencies (there being no significant difference between the three languages), but with regard to the number of different types of downgrader employed. Whereas, in both the English and German L1 data, speakers use 8 different types of downgraders, the interlanguage speakers only use 4 types. In other words, the range of downgraders is much narrower in IL English; the only ones occurring are ‘a little bit’, ‘pretty’, ‘rather’, and ‘quite’. The same holds, incidentally, for upgrading expressions such as ‘awfully’ and ‘really’ (or, in German, ‘echt’ and ‘unbedingt’).
Conclusion Several conclusions can be drawn from this contrastive investigation. First of all, a marked difference was found between native and interlanguage complaints, in several respects. IL complaints are, on the one hand, more verbose than native ones, in that they contain more main strategies as well as more supportive moves; moreover, IL speakers use a higher proportion of apology-like downgraders. On
TSL[v.20020404] Prn:3/09/2002; 12:20
F: PBNS32.tex / p.10 (260)
Ronald Geluykens and Bettina Kraft
the other hand, however, IL complaint strategies tend to be more direct (on-record) than native ones, resulting in a higher level of potential face-threat. We should also point out here that, at least with regard to the linguistic variables investigated here, little or no difference was found between native English and native German. Secondly, with regard to gender, this study has only considered the impact the gender of the speaker has on the choice of complaint strategies. We can only draw some very tentative conclusions. With respect to the type of complaint strategy employed, there appears to be a slight tendency towards the increased use of onrecord strategies with female speakers; this is true for the interlanguage complaints as well as for the two types of native ones. This runs perhaps slightly counter to expectations. On the other hand, however, fewer confrontational supportive moves are used by female speakers, at least in the German L1 and English IL data. On the whole, female speakers also use slightly more downgrading expressions, although results here are far from clear-cut and do not hold across the board. Thirdly, and finally, the factor social distance clearly has a significant impact on apology realization, at least with regard to the use of certain mitigating strategies. In particular, the use of ‘please’/‘bitte’ and the use of apologizing expressions such as ‘excuse me’/‘Entschuldigung’ appear to be influenced by the level of familiarity between speaker and hearer. The correlation here is a fairly simple one: the higher the social distance between the interactants, the more frequently these downgraders are used. In terms of possible ways to extend the current research in a meaningful way, two avenues immediately spring to mind. Primo, we are convinced that the addition of a corpus of spontaneous, uncontrolled data would be useful. Not only would this ensure that complaints can be analyzed in a more natural context, it would also allow us to gain more insight into the sequential organization of complaints and their relationship to the turn-taking system (as demonstrated by Schegloff 1987). It would also give us the opportunity to investigate prosodic characteristics of complaints. Secundo, we are planning to expand this type of contrastive analysis to other types of L1 and interlanguage. The most obvious extension that springs to mind here is the use of English-German IL. In Kraft (1999), the first step towards such an broader perspective was already undertaken. Whatever the case may be, we hope to have shown here, despite this study’s limitations, that systematic contrastive analysis can yield new insights into pragmatic aspects of interlanguage.
References Blum-Kulka, S., House, J., & Kasper, G. (1989). Cross-Cultural Pragmatics: Requests and Apologies. Norwood, NJ: Ablex.
TSL[v.20020404] Prn:3/09/2002; 12:20
F: PBNS32.tex / p.11 (261)
Sociocultural variation in native and interlanguage complaints
Blum-Kulka, S., & Olshtain, E. (1984). Requests and apologies: A cross-cultural study of speech act realization patterns (CCSARP). Applied Linguistics, 5(3), 196–213. Boxer, D. (1993). Social distance and speech behavior: The case of indirect complaints. Journal of Pragmatics, 19, 103–125. Brown, P., & Levinson, S. C. (1987). Politeness. Some Universals in Language Usage. Cambridge: Cambridge University Press. Coates, J. (1986). Women, Men and Language. London: Longman. Coates, J. (1998). Language and Gender. Oxford: Blackwell. Frescura, M. (1995). Face orientations in reacting to accusatory complaints: Italian L1, English L1, and Italian as a Community Language. Pragmatics and Language Learning, 6, 79–104. Geluykens, R. (1999). Face-threatening acts in business English: Directness vs. indirectness. In J. Van der Auwera et al. (Eds.), English as a Human Language (pp. 195–200). München: Lincom Europa. Goffman, E. (1967). Interaction Ritual: Essays On Face-to-Face Behavior. New York: Anchor Books. Holmes, J. (1995). Women, Men and Politeness. London/New York: Longman. House, J. (1989). Politeness in English and German: The functions of please and bitte. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-Cultural Pragmatics: Requests and Apologies (pp. 96–119). Ablex, Norwood, New Jersey. Kraft, B. (1999). Pragmatische Aspekte Interkultureller Kommunikation. University of Münster, Unpublished M.A. Dissertation. Olshtain, E., & Weinbach, L. (1993). Interlanguage features of the speech act of complaining. In S. Blum-Kulka & G. Kasper (Eds.), Interlanguage Pragmatics (pp. 108–123). Oxford: Oxford University Press. Schegloff, E. A. (1987). Some sources of misunderstanding in talk-in-interaction. Linguistics, 25, 201–218. Tannen, D. (1994). Gender and Discourse. Oxford: Oxford University Press. Trosborg, A. (1987). Apology strategies in natives/non-natives. Journal of Pragmatics, 11, 147–167. Trosborg, A. (1995). Interlanguage Pragmatics. Requests, Complaints and Apologies. New York: Mouton de Gruyter.
TSL[v.20020404] Prn:3/09/2002; 12:22
F: PBNS33.tex / p.1 (263)
A cross-cultural study of requests* The case of British and Japanese undergraduates Saeko Fukushima Tsuru University, Yamanashi, Japan
Introduction In the literature, it has been said that an implicit way of communication is valued in collectivist cultures and an explicit way of communication is valued in individualist cultures. The Japanese, who are considered to be collectivists, have been said to be “indirect” (e.g., Clancy 1986; Lebra 1976; Nakane 1970; Okabe 1983; Yamada 1994). However, some recent studies suggest that such stereotypical notions do not always hold true. For example, Miller (1994), comparing Japanese and American English, claims that it is too simplistic to simply claim that the Japanese are indirect and Americans are direct and that both direct and indirect communication styles are appropriate behaviour for both Japanese and Americans. She (Ibid.: 53) admits that there is a difference between the Japanese and Americans in what they value, the former prioritising social harmony, the latter individuality, and these are reflected in communication. But both cultures also value “saving face.” Rose (1996), reviewing previous studies on communication styles, especially in Japanese and American English, offered some counter-examples to the prevailing stereotypes. Similarly, the results obtained by Rinnert and Kobayashi (1999), who collected naturally occurring requestive hints in Japanese and English, showed that Japanese speakers used direct requests more frequently than conventionally indirect requests, in contrast to English speakers, who showed the opposite tendency. Keeping the above reviews in mind, in this paper, I attempt to make a crosscultural comparison between British English, a representative of communication in an individualist culture, and Japanese, a representative of communication in a collectivist culture. This study focuses on requests, investigating whether there are any differences between British English and Japanese in their choice of requesting strategies and whether their choice of requesting strategies is influenced by the
TSL[v.20020404] Prn:3/09/2002; 12:22
F: PBNS33.tex / p.2 (264)
Saeko Fukushima
variables proposed by Brown and Levinson (1987: 76). According to Brown and Levinson (Ibid.), the assessment of the variables, i.e., the social distance of S and H, the relative power of S and H, and the absolute ranking of imposition, contribute to the seriousness of an FTA, and thus to a determination of the level of politeness with which, other things being equal, an FTA will be communicated. Before going into the details of this study, I will briefly discuss requests and point out some salient features of Japanese and British cultures in the next two sections.
Requests Definition of requests A request has been defined by several researchers. It is the speech act which makes somebody (the hearer) do something (see Searle 1969: 66; Tracy et al. 1984: 514; Trosborg 1995: 189) for the benefit of the speaker (see Trosborg 1995: 187; Rinnert & Kobayashi 1999: 1185). A request is an act which S believes that H is able to do; and which it is not obvious that H will do in the normal course of events or of H’s own accord (Searle 1969: 66). Finally, a request is different from a command in the sense that it recognises the hearer’s right not to comply (see Tracy et al. 1984: 514).
Request strategies The classification and terminology applied to request strategies differ from researcher to researcher. For example, Rinnert and Kobayashi (1999) use “Direct, Conventionally indirect and Non-conventionally indirect” and Trosborg (1995: 192–209) uses four categories, i.e., indirect requests, conventionally indirect requests (hearer-oriented conditions), conventionally indirect requests (speakeroriented conditions) and direct requests, classifying each category into strategies. Blum-Kulka and House (1989: 123–124), using the CCSARP coding scheme (Blum-Kulka et al. 1989: 278–281), classify five strategies as impositives, two strategies as conventionally indirect requests and two as hints or nonconventionally indirect requests. Since most of the studies are variations of Brown and Levinson (1987), I will use the following, based on Brown and Levinson’s (1987: 68–70) strategies for doing FTAs. 1. Direct requests (On record without redress) e.g. Give me an aspirin, please. 2. Conventionally indirect requests (On record with redress) e.g. Can you give me an aspirin?
TSL[v.20020404] Prn:3/09/2002; 12:22
F: PBNS33.tex / p.3 (265)
A cross-cultural study of requests
3. Off-record requests e.g. I’ve got a splitting headache. (The examples are taken from Sifianou 1995: 244).
Japanese culture and British culture Several researchers have tried to explain collectivism and individualism (see Hofstede 1991: 51; Kim et al. 1994: 2, for example). It is difficult to define them briefly, but the salient features of collectivist and individualist cultures can be summarised as follows (see Triandis 1994: 184–186). In collectivist cultures, people care for the group and at the same time they are cared for by the fellow group members. Thus, good human relationships are important, and people try to maintain good interpersonal relationships, by paying attention to such contextual factors as, status, age, etc., and try to act accordingly. As members of a collectivist culture have more shared knowledge than those in an individualist culture, meaning can be communicated implicitly. Subtlety is valued and much meaning is conveyed by inference. By contrast, in individualist cultures, the interests of individuals are important. People are expected to take care of themselves, and independence and privacy are valued. People do not have as much shared knowledge as those in collectivist cultures do, and as a result, explicit communication is necessary. Group identity is also differently perceived in individualist and collectivist cultures. Ting-Toomey (1989: 353) refers to Triandis et al. (1986) who note that while the boundary conditions between ingroups and outgroups are fairly diffused and loosely structured in individualist cultures, the boundary conditions between ingroups and outgroups, and also between memberships in various ingroups, are more sharply defined and tightly structured in collectivist cultures. These features are reflected in communication style. Scollon and Scollon (1995: 134) note that: In an individualistic society, groups do not form with the same degree of permanence as they do in a collectivist society. As a result, the ways of speaking to others are much more similar from situation to situation, since in each case the relationships are being negotiated and developed right within the situation of the discourse. On the other hand, in a collectivist society, many relationships are established from one’s birth into a particular family in a particular segment of society in a particular place. These memberships in particular groups tend to take on a permanent, ingroup character along with special forms of discourse which carefully preserve the boundaries between those who are inside members of the group and all others who are not members of the group.
TSL[v.20020404] Prn:3/09/2002; 12:22
F: PBNS33.tex / p.4 (266)
Saeko Fukushima
In short, forms of discourse used to those who are inside the group and forms of discourse used to those who are not the members of the group differ in a collectivist culture, whereas in an individualist culture, forms of discourse do not differ greatly from situation to situation. Based on Hofstede’s (1991: 53) list of individualism index values, we may be able to characterise British culture as an individualist culture and Japanese culture as a collectivist culture, although Japanese culture is not an extreme case of the latter, and it has shifted towards greater individualism than before under the influence of economic growth and change.
Data collection Subjects One hundred and twenty-one British and one hundred and thirty-three Japanese undergraduates served as the subjects for this study. As it was intended to have comparability between the subjects in British and Japanese cultures as much as possible, the subjects were confined to undergraduates.
Instruments Eight request situations, which were identified as comparable between British and Japanese cultures (see Fukushima 2000: 146–149) were used (see Appendix). Some attention was given to identifying comparable situations, because I believe that it is important to ensure such comparability of the situations used in any crosscultural research when different cultures are involved. A written questionnaire with requesting choices in the mother tongue of the subjects was used.
Procedures for situational assessment The subjects were asked to read the written situational descriptions and to rate the following on the five-point scale: – – –
power difference between S and H; social distance between S and H; and the degree of imposition of the requested act.
TSL[v.20020404] Prn:3/09/2002; 12:22
F: PBNS33.tex / p.5 (267)
A cross-cultural study of requests
Procedures for requesting strategies The subjects were presented with a multiple-choice questionnaire with requesting strategies and asked to choose one requesting strategy out of the following three: – – –
direct requests; conventionally indirect requests; and off-record requests.
In this study, the subjects were asked to choose a request strategy, not its linguistic realisation, because it is difficult to compare linguistic realisations across languages, especially when comparing English and Japanese, whose linguistic systems are totally different. Given such linguistic differences, it cannot be assumed that similar linguistic realisations in different languages and cultures have the same pragmatic or social meanings. According to Marquez-Reiter (2000), for example, conventional indirectness in English and Spanish does not mean the same thing with regard to compliance with requests. There are also other problems with obtaining elicited data because, as BargielaChiappini and Harris (1996: 636) point out, the greatest limitation of previous studies is the difficulty of obtaining adequate data. For instance, the use of written questionnaires to elicit spoken data has a limitation. By focusing on the strategy choice rather than linguistic realisation in this study, it is hoped that such limitations can be reduced.
Data analyses and results Situational assessment The assessments by British subjects and Japanese subjects on power differences between S and H, social distance between S and H, and the degree of imposition of the requested act were analysed by conducting t-tests. In the assessment of power difference between British and Japanese subjects, there were significant differences in Situations 1, 3, 6 and 7.1 In those situations, power differences were assessed as bigger by Japanese subjects than by British subjects. In the assessment of social distance between British and Japanese subjects, there were significant differences in Situations 1, 3, 5, and 7.2 Japanese subjects assessed social distance as bigger than British subjects in those situations. In the assessment of the degree of imposition of the requested act between British and Japanese subjects, there were significant differences in Situations 2, 3, 4 and 6.3 Japanese subjects assessed the degree of imposition of the requested act as higher than did British subjects in Situations 2 and
TSL[v.20020404] Prn:3/09/2002; 12:22
F: PBNS33.tex / p.6 (268)
Saeko Fukushima
3, while in Situations 4 and 6, British subjects assessed it as higher than Japanese subjects.
Requesting strategies In the analyses of requesting strategies, Mann-Whitney U tests, which are nonparametric tests, being suitable for ordinal scale data and comparing two groups, were used. There were significant differences in the choices of requesting strategies between British and Japanese subjects in Situations 4, 5, 6 and 8.4 In all these situations, British subjects selected higher-numbered choices. Since the choices of requesting strategies were arranged from (1) direct requests; (2) conventionally indirect requests and to (3) off-record requests, the above results indicate that British subjects chose more indirect requests than Japanese subjects in these situations.
Correlation between situational assessment and requesting strategies Spearman’s Rank-order was used in order to investigate whether there were correlations between the results of situational assessment of the three variables and the choices of requesting strategies. The analysis showed that between the assessment of power difference and the choice of requesting strategies, there were positive correlations in Situations 6 and 7.5 The choices of requesting strategies were arranged from (1) direct requests; (2) conventionally indirect requests and to (3) off-record requests, and in Situations 6 and 7, H has more power (higher status) than S. Therefore, the above results indicate that the bigger the power difference between S and H, when H has more power than S, the more indirect the requesting strategies become. Between the assessment of social distance between S and H and the choice of requesting strategies, there was a negative correlation in Situation 1.6 This result indicates that the bigger the social distance between S and H becomes, the more direct the requesting strategies become. Between the assessment of the degree of imposition of the requested act and the choice of requesting strategies, there were positive correlations in Situations 5, 6, 7 and 8.7 These results indicate that the higher the degree of imposition of the requested act, the more indirect the requesting strategies become.
Discussion As noted above it has been claimed in the literature that an implicit way of communication is valued in a collectivist culture, and Japanese culture is considered to be one of collectivist cultures, although it has gained more individualist ele-
TSL[v.20020404] Prn:3/09/2002; 12:22
F: PBNS33.tex / p.7 (269)
A cross-cultural study of requests
ments than before under the influence of economic growth. Even taking the growth of individualism in contemporary Japanese culture, it could be hypothesised that Japanese subjects would select more indirect requesting choices than British subjects, who are considered to be individualists. However, the results of this study showed that in four situations out of eight, Japanese subjects selected more direct requesting choices, as compared with their British counterparts. The same tendency for Japanese subjects to select more direct requesting choices than British subjects was also found in Fukushima (1996), although the situations used in that study were only among equals with different degrees of imposition, and linguistic realisations rather than strategy choices were investigated. In order to investigate what influenced the different choices of requesting strategies between British and Japanese subjects, the assessment of the three variables and the correlation between the assessment of those variables and requesting strategies were statistically analysed. From these results, the following can be concluded. The assessment of power differences and that of degree of imposition distinguished the requesting choices between British and Japanese subjects in Situation 6. Japanese subjects assessed power difference as bigger than did British subjects, while British subjects assessed the degree of imposition as higher than did their Japanese counterparts. There were also correlations between the choice of requesting strategies and the assessment of power differences, and between the choice of requesting strategies and the assessment of the degree of imposition. As a result of a bigger assessment of power differences by Japanese than British subjects and lower assessment of the degree of imposition by Japanese than British subjects, the Japanese chose more direct requesting strategies than British subjects. In Situation 4, it was not only from the results of the statistical analyses of this study which enabled identification of differences, but it may be conjectured that status difference was a possible factor which led to the significant difference in the choice of requesting strategies between the two groups of subjects. In this particular situation, a request was made from a higher status person to lower status person. LoCastro (1990: 269) notes the use of direct forms by higher status to lower status persons in Japanese, although the speech act of her example is not requesting but disagreement. This appears to be in the same line with Hassall’s (1999) results, which showed that Indonesian speakers selected imperatives when the speaker is of high status relative to the hearer. In Situations 5 and 8, there were positive correlations between the assessment of the degree of imposition and the choice of requesting strategies, but there were no significant differences between British and Japanese subjects in the assessment of the degree of imposition in either situation. This means that the degree of imposition was not a factor which differentiated the choice of requesting strategies by the two sets of subjects, so there must be some other factors which influenced strategy choice. In these situations, the power difference was small, from which it can
TSL[v.20020404] Prn:3/09/2002; 12:22
F: PBNS33.tex / p.8 (270)
Saeko Fukushima
be considered that S and H were equal in status. In such situations, Japanese subjects chose more direct requesting choices than British subjects, as was also found in Fukushima (1996). It is necessary to investigate further what kind of variables influence the different choices of requesting strategies between British and Japanese subjects, examining more variables. However, according to the results of this study, it was clear that whereas Japanese subjects differentiated their requesting strategies according to the situation, British subjects did so less, the most frequently chosen strategy by British subjects in all the situations being conventionally indirect requests. The reason why many British subjects chose conventionally indirect requesting strategies may be due to the fact that in English, conventionally indirect request strategies are “conventionalised,” because, as Brown and Levinson (1987: 248) explain, “In English, . . . , conventionalized indirect requests are so common that it is rare to hear a completely direct request even between equals. . . ”. Similarly, Levinson (1983: 264) notes that indirect speech acts are rather common in English, and that “the imperative is very rarely used to issue requests in English; instead we tend to employ sentences that only indirectly do requesting.” The results of this study show that Japanese subjects chose more direct requesting choices than British subjects, so challenging the stereotypes of communication styles in collectivist and individualist cultures as was described in the literature, collectivists being implicit and individualists being explicit. However, this does not mean that the results here are totally different from the features associated with collectivist/individualist cultures. In fact, the results here can also be described as Japanese subjects having exhibited more variability in the choice of requesting strategies than British subjects. This can be explained in terms of the differences in collectivist and individualist cultures. As noted above, people in collectivist cultures distinguish forms of discourse used to in-group members and out-group members, whereas people in individualist cultures use similar forms from situation to situation, i.e., they do not distinguish discourse forms and group membership as much as collectivists do. The results of this study, Japanese subjects having differentiated requesting choices more than British subjects, may be characterised as a phenomenon found in a collectivist culture. Likewise, the prevailing choice by British subjects, who preferred conventionally indirect requests most in all the situations, can be considered to be in line with the characteristics of an individualist culture. In this study, although the inside-outside group distinction was not clearly made, it may be a factor which influenced the requesting choices. Among the three variables investigated in this study, power differences and social distance will be closely related to the memberships–non-membership distinction. There may be some situations in which S regards H as an insider even when there is a power difference between S and H (e.g., after many supervisions, a supervisor may regard
TSL[v.20020404] Prn:3/09/2002; 12:22
F: PBNS33.tex / p.9 (271)
A cross-cultural study of requests
his/her supervisee as an insider), and also when there is no power difference between S and H (e.g., a student may regard his/her classmate as an insider at the end of the term, after having worked together). In similar relationships, there may be some situations in which S regards H as an outsider. For example, a supervisor may regard his/her supervisee as an outsider even after many meetings, and a student may regard his/her classmates as outsiders even at the end of the term. Although the distinction between insiders and outsiders is related to power differences and social distance, some other factors are also influential. For instance, sometimes emotional factors may also be related to the distinction. How people distinguish insiders and outsiders, whether there are any differences in differentiating group identity according to culture and how group distinction will influence requesting choices need further investigation. The requesters in this study may have regarded the requestees as insiders or outsiders depending on various factors, as discussed above, although they may not have consciously considered it. In this study, however, there was probably no situation in which a requester would surely regard a requestee as an outsider, because an interaction between total strangers, who are not normally regarded as insiders, was not included. The inclusion of strangers as one of the parties in request situations is probably important, because, as Lee-Wong (2000: 302) notes, “. . . formal politeness is evident in interaction with strangers. Subjects tend to perceive the use of conventionally indirect requests to strangers as polite and the use of impositives as rude.” Adding the situations of strangers as a category of requestee in future studies may reveal further differences in the choice of requesting strategies by British and Japanese subjects. There may be some differences between them in terms of interacting with strangers, because, as Triandis, et al. (1986: 266) point out, “. . . people in collectivist cultures compared with individualist cultures tend to deal with strangers differently than with ingroup members,” and these differences may lead to different choices of requesting strategies.
Conclusion The results of this study showed that there were differences in the choice of requesting strategies between British and Japanese subjects, Japanese having chosen more direct requesting strategies than British subjects. Some different assessments of the three variables suggested by Brown and Levinson, i.e., power difference, social distance between S and H and the degree of imposition of the requested act, were the factors differentiating the requesting choices. However, there were some situations in which it was not the three variables which differentiated the different choices of requesting strategies between Japanese and British subjects. As Turner (1996: 5)
TSL[v.20020404] Prn:3/09/2002; 12:22
F: PBNS33.tex / p.10 (272)
Saeko Fukushima
has suggested, these three variables are not enough, even though Brown and Levinson’s contribution to politeness theories by identifying these three variables was considerable. Indeed, most previous studies on requests were based on Brown and Levinson’s three variables (e.g., Bargiela-Chiappini & Harris 1996; Baxter 1984; Blum-Kulka & Olshtain 1984; Blum-Kulka, Danet, & Gherson 1985; Blum-Kulka & House 1989; Lee-Wong 1994; Lee-Wong 2000; Yeung 1997),8 although there are sometimes differences in terminology (see Spencer-Oatey 1996) and the components of one variable differ from study to study (e.g., affect in social distance) (see Slugoski & Turnbull 1988, for example). It is important to note that the variables are not independent, as Watts, Ide and Ehlich (1992: 9) argue. For example, the degree to which a social act is considered to be an imposition depends crucially on P and D. Brown and Levinson (1987: 78) also admit that the three factors interact with each other, saying that “. . . as S’s power over H increases, the weightiness of the FTA diminishes.” This feature of variables may also have to be considered in future studies, although how it is investigated empirically needs a further consideration. As is clear from this study, the variables which will influence the requesting choices need to be further investigated in future studies, together with the in- and out-group distinction, which may prove to be one of the variables which distinguishes requesting choices especially in collectivist and individualist cultures. Although this study is limited in having only three variables and in omitting some situations (e.g., interactions between strangers), it has shown that, contrary to expectations based on cultural stereotypes, Japanese subjects chose more direct requesting strategies than British subjects in half of the situations concerned. Furthermore, British subjects tended to use conventionally indirect requests across a wide range of situations. In this, they contrasted with the Japanese, who displayed more variability in their requesting choices. This indicates that such cultural assumptions on the surface as Japanese being indirect are dangerous.
Notes * This paper is based on my Ph.D. thesis (Fukushima 1999). I would like to extend my deep appreciation to Ron White and Ken Turner for their valuable comments and warm support. . Situation 1: MSB = 3.355, MSJ = 3.759, t = 3.458, p < .01; Situation 3: MSB = 3.669 MSJ = 4.308, t = 6.244, p < .001; Situation 6: MSB = 2.554, MSJ = 2.865, t = 2.449, p < .05; Situation 7: MSB = 3.107, MSJ = 3.917, t = 7.315, p < 0.1 (MSB = British subjects’ mean score; MSJ = Japanese subjects’ mean score) . Situation 1: MSB = 2.967, MSJ = 3.444, t = 3.730, p < .01; Situation 3: MSB = 3.223, MSJ = 3.947, t = 5.937, p < .001; Situation 5: MSB = 1.223, MSJ = 1.617, t = 4.387, p < .001; Situation 7: MSB = 3.157, MSJ = 3.579, t = 3.497, p < .01
TSL[v.20020404] Prn:3/09/2002; 12:22
F: PBNS33.tex / p.11 (273)
A cross-cultural study of requests . Situation 2: MSB = 2.116, MSJ = 2.699, t = 4.570, p < .001; Situation 3: MSB = 1.719, MSJ = 2.624, t = 7.898, p < .001; Situation 4: MSB = 2.165, MSJ = 1.789, t = 3.048, p < .01; Situation 6: MSB = 1.512, MSJ = 1.256, t = 3.100, p < .01 . Situation 4: z = 2.108, p < .05; Situation 5: z = 7.367, p < .001; Situation 6: z = 5.903, p < .001; Situation 8: z = 5.168, p < .01 . Situation 6: Spearman’s rank-order correlation coefficients (ρ s) = .125, p < .05; Situation 7: ρs = .187, p < .01 . Situation 1: ρs = –.233, p < .001 . Situation 5: ρs = .192, p < .01; Situation 6: ρs = .236, p < .001; Situation 7: ρs = .231, p < .001; Situation 8: ρs = .183, p < .01 . Blum-Kulka and Olshtain (1984) and Blum-Kulka, Danet, and Gherson (1985) investigated power and distance. The study by Blum-Kulka and House (1989), which uses six variables (1. the relative dominance of the request relative to the hearer; 2. the relative social distance between the interlocutors; 3. the hearer’s degree of obligation in carrying out the request; 4. the speaker’s degree of right in issuing the request; 5. the estimated degree of difficulty the speaker had in making the request, and 6. the estimated likelihood of compliance on the part of the hearer), can be considered to be based on Brown and Levinson’s three variables, the first variable being power, the second being social distance, and the third to the sixth being subsumed into imposition.
Appendix Situation 1. A tutor asks a student to deliver some gifts abroad. Situation 2. A student asks a colleague at part-time work to swop a working day. Situation 3. A student asks a tutor to write a letter of recommendation. Situation 4. A tutor asks a student for a lift home. Situation 5. A student asks a next door neighbour in a students’ hall for some salt. Situation 6. A student asks a senior member of the club to provide a phone number of the club captain. Situation 7. A student asks a tutor for a lift to an airport. Situation 8. A student asks his/her classmate to lend lunch money.
References Bargiela-Chiappini, F., & Harris, S. J. (1996). Requests and status in business correspondence. Journal of Pragmatics, 28, 635–662.
TSL[v.20020404] Prn:3/09/2002; 12:22
F: PBNS33.tex / p.12 (274)
Saeko Fukushima
Baxter, L. A. (1984). An Investigation of Compliance-Gaining as Politeness. Human Communication Research, 10(3), 427–456. Blum-Kulka, S., & Olshtain, E. (1984). Requests and Apologies: A Cross-Cultural Study of Speech Act Realization Patterns (CCSARP). Applied Linguistics, 5(3), 196–213. Blum-Kulka, S., Danet, B., & Gherson, R. (1985). The Language of Requesting in Israeli Society. In J. P. Forgas (Ed.), Language and Social Situations (pp. 113–139). Berlin: Springer-Verlag. Blum-Kulka, S., & House, J. (1989). Cross-Cultural and Situational Variation in Requesting Behavior. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-Cultural Pragmatics: Requests and Apologies (pp. 123–154). Norwood, NJ: Ablex. Blum-Kulka, S., House, J., & Kasper, G. (1989). Cross-Cultural Pragmatics: Requests and Apologies. Norwood, NJ: Ablex. Brown, P., & Levinson, S. C. (1987). Politeness: Some universals in language usage. Cambridge: Cambridge University Press. Clancy, P. M. (1986). The acquisition of communicative style in Japanese. In B. B. Shieffelin & E. Ochs (Eds.), Language socialization across cultures (pp. 213–250). Cambridge: Cambridge University Press. Fukushima, S. (1996). Request Strategies in British English and Japanese. In K. Jaszczolt & K. Turner (Eds.), Contrastive Semantics and Pragmatics Volume II: Discourse Strategies (pp. 671–688). Oxford: Pergamon Press. Fukushima, S. (1999). A Comparative Study of Pragmatic Strategies: Requests and Responses to Off-record Requests in British English and Japanese. Unpublished Ph.D. Thesis, The University of Reading. Fukushima, S. (2000). Requests and Culture: Politeness in British English and Japanese. Bern: Peter Lang. Hassall, T. (1999). Request Strategies in Indonesian. Pragmatics, 9, 585–606. Hatch, E., & Lazaraton, A. (1991). The Research Manual: Design and Statistics for Applied Linguistics. Boston: Heinle and Heinle Publishers. Hofstede, G. (1991). Cultures and Organizations: Software of the mind. London: McGrawHill Book Company. Kim, U., Triandis, H. C., Kagitcibasi, C., Choi, S., & Yoon, G. (1994). Introduction. In U. Kim et al. (Eds.), Individualism and Collectivism: Theory, Method, and Applications (pp. 1– 11). Thousand Oaks: Sage. Lebra, T. S. (1976). Japanese Patterns of Behavior. Honolulu: University Press of Hawaii. Lee-Wong, S. M. (1994). Imperatives in Requests: Direct or Impolite – Observations from Chinese. Pragmatics, 4(4), 491–515. Lee-Wong, S. M. (2000). Politeness and face in Chinese culture. Frankfurt am Main: Peter Lang. Levinson, S. C. (1983). Pragmatics. Cambridge: Cambridge University Press. LoCastro, V. (1990). Intercultural Pragmatics: A Japanese-American Case Study. Unpublished Ph.D. Thesis, The University of Lancaster. Marquez-Reiter, R. (2000). Conventional indirectness in English and Spanish: Same or different? Paper presented at Second International Conference in Contrastive Semantics and Pragmatics at Newnham College, Cambridge, 13 September, 2000.
TSL[v.20020404] Prn:3/09/2002; 12:22
F: PBNS33.tex / p.13 (275)
A cross-cultural study of requests
Miller, L. (1994). Japanese and American Indirectness. Journal of Asian Pacific Communication, 5(1–2), 37–55. Nakane, C. (1970). Japanese Society. Berkeley, CA: University of California Press. Okabe, R. (1983). Cultural assumptions of East and West: Japan and the United States. In W. Gudykunst (Ed.), Intercultural Communication Theory: Current Perspectives (pp. 21– 44). Beverly Hills, CA: Sage. Rinnert, C., & Kobayashi, H. (1999). Requestive hints in Japanese and English. Journal of Pragmatics, 31, 1173–1201. Rose, K. R. (1996). American English, Japanese, and Directness: More Than Stereotypes. JALT Journal, 18(1), 67–80. Scollon, R., & Scollon, S. W. (1995). Intercultural Communication. Oxford: Blackwell. Searle, J. (1969). Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Sifianou, M. (1995). Indirectness and politeness: The case of English and Greek. Reading Working Papers in Linguistics, Department of Linguistic Science, The University of Reading, 2, 241–253. Slugoski, B., & Turnbull, W. (1988). Cruel to be kind and kind to be cruel: Sarcasm, Banter and social relations. Journal of Language and Social Psychology, 7(2), 101–121. Spencer-Oatey, H. (1996). Reconsidering power and distance. Journal of Pragmatics, 26, 1– 24. Ting-Toomey, S. (1989). Identity and Interpersonal Bonding. In M. K. Asante & W. B. Gudykunst (Eds.), Handbook of International and Intercultural Communication (pp. 351–373). Newbury Park: Sage. Tracy, K., Craig, R. T., Smith, M., & Spisak, F. (1984). The Discourse of Requests: Assessment of a Compliance-Gaining Approach. Human Communication Research, 10(4), 513–538. Triandis, H. C. (1994). Culture and social behavior. New York: McGraw-Hill. Triandis, H., Bontempo, R., Betancourt, L., Bond, M., Leung, K., Brenes, A., Georgas, J., Hui, H., Martin, G., Setiadi, B., Sinha, J., Verma, J., Spangenberg, J., Touzard, H., & Montmollin, G. (1986). The measurement of the etic aspects of individualism and collectivism across cultures. Australian Journal of Psychology, 38, 257–267. Trosborg, A. (1995). Interlanguage Pragmatics: Requests, Complaints and Apologies. Berlin: Mouton de Gruyter. Turner, K. (1996). The principal principles of pragmatic inference: Politeness. Language Teaching, 29, 1–13. Watts, R. J., Ide, S., & Ehlich, K. (1992). Introduction. In R. J. Watts, S. Ide & K. Ehlich (Eds.), Politeness in Language: Studies in its History, Theory and Practice (pp. 1–17). Berlin: Mouton de Gruyter. Yamada, H. (1994). Talk-distancing in Japanese meetings. Journal of Asian Pacific Communication, 5, 19–36. Yeung, L. N. T. (1997). Polite requests in English and Chinese business correspondence in Hong Kong. Journal of Pragmatics, 27, 505–522.
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.1 (277)
Questions as indirect requests in Russian and Czech Michael Betsch Universität Tübingen
Introduction The usage of indirect constructions to convey directive speech acts is a widespread phenomenon. Here conventionalized indirect requests, as opposed to direct requests or non-conventional indirect speech acts like hints, play a very important role (Blum-Kulka 1989). In many European languages, conventionally indirect requests, which mostly take the form of questions about felicity conditions of the request, are the most important modes of expressing polite requests (for English, French and German see Blum-Kulka et al. 1989; for Spanish see le Pair 1996), even if the degree to which conventionalized indirect requests are used varies between the languages. In Russian, on the other hand, indirect requests exist, but the most frequent way of expressing requests is the imperative (Rathmayr 1994: 252). Indirect requests in Russian are also characterized by other features uncommon in West European languages, e.g. the use of negated questions (Rathmayr 1994; Comrie 1984: 42–45). In order to characterize the specific features of indirect requests in Russian, I want to compare it with another closely related language. We hope that the comparison between related languages can give results that do not depend on different morphosyntactic means used to express indirect requests. Russian and Czech are (as Slavonic languages) genetically closely related; they share also many grammatical means that can be exploited to construct indirect directives. In both languages, verbs have a conditional mood, which is formed analytically of the past tense of the verb and a conditional marker. In Russian this is an invariable particle by, in Czech the marker is a special auxiliary inflected for person and number (bych, bys etc.). In both languages a modal auxiliary moˇc’ (Russian) resp. moci (Czech) (“can”) is used for indirect requests. Both Russian and Czech make use of negation as a marker in indirect requests. In Russian the negation is
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.2 (278)
Michael Betsch
expressed by the particle ne, in Czech the negation marker is a verbal prefix. Both languages have a category of verbal aspect, where any verb belongs either to the imperfective or the perfective aspect; the future forms of perfective verbs are used to form indirect requests in Russian, but not in Czech. For some politeness phenomena, the distinction between distant and non-distant pronominal address is relevant: in both languages, non-distant address (T) is expressed by pronominal and verbal forms of the 2nd person singular, and distant (V) address is expressed by forms of the 2nd person plural. This paper will focus on types of questions that are used as conventionally indirect requests in Russian and Czech and on the linguistic means that are used to mark a (formal) question as a request (illocutionary marker) or to express a degree of politeness (politeness marker). Most grammars mention the usage of specific means to convey indirect requests to some extent (Barnetová et al. (1979: 844); Grepl & Karlík (1998: 455–458); Petr (1987: 272–273) on negation; Petr (1987: 340) on conditional and the auxiliary moci). The different realizations of linguistic politeness in Russian and Czech are specifically treated by the normative handbook of Formanovskaja and Tuˇcny (1986). I will try to characterize the usage and functions of central markers of indirect requests in both languages using data from corpora; the corpus data will be contrasted with data from interviews with native speakers.
Russian Several types of indirect requests can be expressed by questioning the addressee’s ability to fulfill the request (using the auxiliary moˇc’ “can”). A form that closely ressembles the English type “could you” is represented in example (1): this form consists of the negated conditional of the verb moˇc’ (can). This form is used either in a rather formal, written style (thus it is found in press interviews), or if an important request is to be expressed: the example cited here is taken from taped dialogues of a telephone information service, where the client is asking for information (the address) the service does not give – the request is declined. by skazat’? (1) A adres vy ne mogli and address pron2pl neg canPretPl conditional say “And what about the address, could you tell me the address?” (Služba 09) This type of request is rather new in Russian: it is almost missing in 19th century texts, but is more common in texts from the 20th century (Berger 1997). We can confirm that negation is almost obligatory; there is only one counterexample in the corpora used for this study. Its use is limited to distant (V) address: all ex-
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.3 (279)
Questions as indirect requests in Russian and Czech
amples in our corpora have distant address, and normative works, as for example Formanovskaja and Tuˇcny (1986: 163), give only examples with V address.1 Questions about the ability to fulfill a request can also be formed with the indicative mood of moˇc’ (can). Normative works mention such requests in negated as well as non-negated form (Formanovskaja & Tuˇcny 1986: 163). However, they are rare in most corpora, see Table (5); they are more frequent in a corpus of dialogues with a telephone information service, where however they come close to genuine requests about the ability to fulfill the request. Especially the rare nonnegated form, as in example (2), risks being understood as a genuine question instead of an indirect request (see Rathmayr 1994: 268). Some examples for the negated form are also attested in older works of fiction, like example (3), or in contemporary fiction (one example in the Uppsala Corpus, see (5)); they express important or pressing requests, but on the whole they are very rare. (2) Devuška, a vot Mytišˇci, spravoˇcnuju, vy možete dat’? – Girl, and now Mytišˇci, information, pron2pl canPres2Pl give? – Spravoˇcnuju Mytišˇc mogu vam dat’. – Da, dajte. information MytišˇciGen canPres1Sg pron2plDat give. – Yes, give. “Young woman, can you tell me the information bureau of Mytišˇci? – I can tell you the information bureau of Mytišˇci. – Well, then please tell me.” (Služba 09) (3) Radi boga, ne možete li vy menja oblegˇcit’? For God’s sake, neg canPres2Pl q pron2pl pron1sgAcc alleviate? “For God’s sake, could you take over some (duties) from me?” (Tolstoj, Vojna i mir) The most current form of indirect directives in Russian consists of the negated future tense form of a perfective verb, as in example (4). Such requests are generally viewed as questions about the addressee’s intention to fulfill the requests (Mills 1991: 71; Comrie 1984: 44), but they give the option to deny the possibility as well; the perfective future means also the ability to complete the action. This type of indirect request is specific to Russian; literal translations to e.g. Czech or Polish would not be understood as requests; in Czech, the negated conditional would be used instead (Mills 1992: 73). (4) Telefon sportkompleksa “Olimpijskij” ne podskažete? telephone sport complex “Olimpijskij” neg saypfFut2pl “Could you tell me the telephone of the sport complex ‘Olimpijskij’?” (Služba 09) Table (5) presents figures of occurrences of the types of indirect requests in the Russian corpora used for this study. These corpora are: The texts of the study
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.4 (280)
Michael Betsch
“Russkaja razgovornaja reˇc’ ” (Zemskaja 1973), abbrev. RR; a corpus of taped dialogues with the Moscow telephone information service (Služba 09, abbrev. Sl. 09); the Uppsala Corpus of modern Russian, which consists of fiction (UC/F) and nonfictional texts (the latter mainly from the press, abbrev. UC/NF); and a corpus of press interviews collected from Russian periodicals publicly available via the world wide web (IV). We see that the negated conditional of “can” occurs mainly in corpora consisting of written texts (IV and UC/NF). There are also several examples in Sl. 09; however, they represent only a very small fraction of the overall number of requests in this corpus. If neg + cond + “can” is used in spoken texts, as in Sl. 09, it expresses a much higher rank of politeness than the current negated perfective future. The negated future of perfective verbs, on the other hand, is represented in spoken texts (Sl. 09) or in fictional texts (UC/F), which are more close to a colloquial style than press texts. Negation of “can” (3) is attested almost exclusively in the dialogues of the telephone information service (Sl. 09). Here it applies to the specific situation where the question about the ability to fulfill the requests is always real. We could also confirm that negation is obligatory in the indirect requests mentioned: Without negation, requests such as (2) would be misunderstood as genuine questions; for the conditional of “can”, we found only one counterexample (without negation). All types of indirect requests however have a marginal position compared to directives formed with the imperative. Russian has a lexical politeness marker požalujsta that can only be used with imperative sentences.2 Thus this marker is an indicator of how often imperative sentences are used; the fourth line of table (5) shows the number of occurrences of požalujsta with imperative. Of course, the actual number of imperatives is still much higher than those combined with požalujsta. (5) Corpora3 RR Sl. 09 IV UC/NF UC/F neg + cond + “can” 1 10 7 7 – neg + “can” – 20 – – 1 neg + pf. future 4 179 – – 3 požalujsta + imp 41 908 26 7 6 No. of words 230 000 100 000 320 000 500 000 500 000
Native speakers’ judgments Six native speakers were asked how they would ask different hearers, varying by age and social status, to tell them the way to the railway station. The interviewees were
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.5 (281)
Questions as indirect requests in Russian and Czech
also asked to give more as well as less polite variants of their requests. The requests given use the constructions in (6) with the hierarchy of politeness indicated. (6)
neg + cond + “can” → neg + pf. future → imp + požalujsta
The interviewees did not mention the form neg + “can” (3); however, when they were asked specifically about possible usage of this form, most of them would also report situations where this could be used. This confirms the finding in the corpus data that such requests occur only marginally.
Summary for Russian Indirect requests occupy only a marginal position in Russian. Imperative forms are much more frequent, and the usage of indirect requests is subject to several restrictions: Indirect requests formed with conditional + “can” (1), a type very frequent in Western European languages, are very rare, their use is restricted to a formal style and they can only be used with distant (V) address. Indirect requests cannot be modified by the central lexical politeness marker požalujsta.4 This does not entail that indirect requests are perceived as less polite because they do not contain the lexical politeness marker; on the contrary, they are perceived as very polite by native speakers, even as too polite for many occasions (Mills 1992: 72; Rathmayr 1994: 266). In the types of indirect requests considered, negation is obligatory; if negation is omitted, these requests are not understood as requests, as in (2). Negation serves thus as a marker of illocutionary force; it cannot be used to differentiate between degrees of politeness, since it is always present. The restrictions on the use of indirect requests are linked with the marginal usage of indirect requests. We may assume that restrictions such as impossibility of the lexical politeness marker or the obligatory use of negation in indirect requests correspond to a situation where indirect requests do not (yet) have a strong position as requesting formulas. Usage of the same politeness markers as with the imperative indicates that in a given language indirect requests are to a great extent conventionalized as requests. This is not yet the case in Russian. Similar considerations apply to negation; the fact that negation is obligatory and that without it questions would not be understood as requests shows that questions are only conventionalized as requests to a low degree, and thus they need an illocutionary marker.
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.6 (282)
Michael Betsch
Czech In Czech, unlike Russian, requests formed with conditional + “can” are very frequent. They are used with distant ((9),(10)) as well as non-distant address, and they occur with ((8),(10)) as well as without negation. bys mi poradit? (7) . . . mohl canPastMSg conditional2Sg pron1sgDat advise “Could you give me some advice?” (8) Nemohl bys mi poslat ten neg + canPastMSg conditional2Sg pron1sgDat send dem ˇcasopis? journal “Could you send me your journal?”
váš your2pl
(9) Mohl byste rozvést vaše argumenty? canPastMSg conditional2Pl develop your2Pl arguments “Could you develop your arguments?” (10) . . . nemohl byste napsat nevelkou pˇredmluvu foreword . . . neg + canPastMSg conditional2Pl write little k ruskému vydání? to Russian edition “Could you write a little foreword to the Russian edition?” Table (11) shows figures of questions with the conditional of “can” in the Czech National Corpus. We see that except for a few dubious cases and some genuine hypothetical questions, most of these sentences are indirect requests.5 One type of situation where this type of directive is widely used is represented by interviews or discussions. Here there are numerous examples like (9) where the interlocutor is asked to give further information or to state his/her opinion. This type of directive (labeled “polite questions” in the following table) differs from other uses by the fact that here the question is almost never negated, thus it should be treated separately. Other directives occur more often without negation, but here negated forms are quite current in the corpus. (11) Directives Polite questions Genuine questions Dubious cases Sum
T 55 23 14 2 94
T + neg 36 – 2 1 39
V 104 259 18 – 381
V + neg 43 8 5 3 59
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.7 (283)
Questions as indirect requests in Russian and Czech
Another type of indirect request is formed by questions with the conditional. This corresponds to the questions about the addressee’s intention that are expressed by the perfective future in Russian (Mills 1992: 73). According to Petr (1987: 340), the conditional expresses the uncertainty whether the addressee is willing to fulfill the request. This type of request occurs with or without negation, see (12) and (13). (12) Pˇrinesla bys laskavˇe kabát? bringPastFSg conditional2Sg kindAdv coat “Could you please bring the coat?” (13) Nepˇreprala bys mi bílou košili? neg + washPastFSg conditional2Sg pron1sgDat white shirt “Would you mind washing the white shirt for me?” The following table (14) shows the distribution of negated and non-negated requests with some more common verbs. There are few examples, but possibly we may conclude that a tendency to use negation is stronger for V address; for jít, negation is clearly more common than the non-negated variant. (14) dát (give) rˇíci (say) jít (go/come)
T 2 2 4
T + neg – – 13
V 2 3 4
V + neg 3 1 9
It is possible to ask for an object by questioning whether the hearer has the object. Such requests are mentioned for Russian as well (e.g. Comrie 1984: 43; Mills 1991: 558), but there are only few examples in our corpora, so I excluded them from the comparison. These requests are questions with the verb mít “to have” and they occur with or without negation. (15) Nemáš kousek diazepamu? neg + havePres2Sg piece Diazepam (Valium) “Do you have some Valium?” (16) Babiˇcko, prosím tˇe, nemáš párky? grandmother, ask1Sg pron2sgAcc, neg + have2Sg sausages “Please, Grandmother, do you have some sausages?” (17) Máte prosím klíˇc pro místnost ˇcíslo padesát? have2Pl please key for room number fifty “Could you please give me the key for room No. 50?” Table (18) presents an overview of some requests formed with questions containing the verb mít “to have”. The first two lines represent figures of requests for one or more cigarettes (cigaretu, cigarety, cigáro) or change (drobné/drobný). The
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.8 (284)
Michael Betsch
third line represents requests modified by the lexical politeness marker prosím or prosím tˇe/vás. The requests for a cigarette are formed with or without negation. Here the non-negated form is used more frequently with non-distant address (T); with distant address (V), negation is preferred. Both requests for some change are negated. In requests with the politeness marker prosím negated questions are quite frequent as well. Here all non-negated examples are requests where the hearer has an obligation or professional duty to comply, as in example (17), where a porter is being asked for a key. The negated forms, on the other hand, are requests without such an obligation on the side of the hearer. (18) Cigarette Change “have” + prosím
T 4 – –
T + neg 3 1 3
V 1 – 3
V + neg 3 1 3
In questions used as requests for an object, negation varies according to several factors: for the “cigarette” question, negation is less frequent with small social distance (expressed by T address). Non-negated requests seem to be especially frequent if the object asked for has little value (as a cigarette). Negation is missing if the hearer has an obligation to comply, such as the porter in (17), or a salesperson. On the other hand, requests are more often negated if they are modified with the politeness marker prosím. In the types of indirect requests mentioned above, negation is frequent, but not obligatory, and may vary according to factors of the situation. However, there are also some types of indirect requests that are almost never attested with negation, e.g. (19): we found only one negated example, but 16 examples without negation. (19) Byl byste tak laskav a odložil bePastMSg conditional2Pl so gentle and lay downPastMSg tady aktovku? here briefcase “May I ask you to kindly deposit your briefcase?”
Native speakers’ judgments As for Russian, four interviewees were asked how they would ask different people for the way to the railway station. The results ((20)–(23)) present a greater variety of different answers. (20)
cond + “can” → neg + “know”
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.9 (285)
Questions as indirect requests in Russian and Czech
(21)
neg + cond (cf. (13)) → neg + “know”
(22)
cond + “can” neg + cond + “know”
(23)
neg + “know” → cond + “be nice” + “say” (cf. (19)6 )
→ neg + “know”
Generally here higher levels of politeness are marked by usage of the conditional, lower levels by simple negated questions for felicity conditions (neg + “know”). For requests formed with the conditional (13), the interviewees gave negated forms, but for the conditional of the modal auxiliary “can”, they gave non-negated forms, and even tended to reject the possibility of the negated form. The judgments of speakers thus differ from the corpus data, where negated requests of this type are still represented by a large minority of the examples (see (11)). This seems to indicate a change in progress: while Trávníˇcek’s grammar of 1951 only mentioned negated requests of this type (Trávníˇcek 1951: 663), contemporary grammars give non-negated (9) as well as negated (10) examples (e.g. Grepl & Karlík 1998: 597; Petr 1987: 340). Thus the descriptions of contemporary grammars correspond to the corpus data; from the interviewees’ judgments we may conclude that there is a tendency to abolish negation in this type of indirect request, and this tendency is more advanced in the speakers’ answers than in the corpus texts.7 On the other hand, negation seems to hold still firm where it is only combined with a verb denoting a felicity condition (“know”, “have”) or the requested action. Apparently negation is receding primarily in such types of indirect requests that contain several other markers of politeness or illocutionary force (here: modal verb + conditional).
Summary for Czech Indirect requests are more important in Czech than in Russian. They are not subject to similar restrictions of usage as in Russian, where one type is restricted to V address, and the central lexical politeness marker prosím (resp. prosím vás/tˇe) can be used with indirect as well as with direct requests, unlike Russian požalujsta, which is restricted to imperative sentences. Indirect requests can be formed with the conditional mood, the modal verb moci “can”, and other questions about felicity conditions. The conditional has a very important role in forming indirect requests, it is not a marginal form like in Russian. Unlike Russian, the future of perfective verbs is not used as a marker in indirect requests; instead, the conditional of the main verb can be used.8 Negation is not obligatory in indirect requests as in Russian. In some types of indirect directives, negation is almost missing; for types of indirect requests that
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.10 (286)
Michael Betsch
contain several other markers of politeness or illocution there is a tendency to restrict negation. In other types of indirect request negation serves as a politeness marker; its function seems to be to “give options” (“don’t coerce” in the sense of Brown & Levinson 1987: 172), and this could explain why it is not used in requests where the speaker would not expect a polite refusal, e.g. in requests to fulfill a professional duty (17).9 Negation is however also more common with distant address, and it correlates with the lexical politeness marker prosím.
Conclusion Russian has several types of conventionalized indirect requests; however, they are rarely used, they cannot be combined with the same lexical politeness marker as imperative forms and they are obligatorily negated, negation functions thus as an illocutionary marker. In this respect, Russian differs markedly from most West European languages. On the other hand, Czech has many types of conventionalized indirect requests that are widely used and that can be freely combined with the same politeness markers as direct requests. In this respect, Czech resembles West European languages. It differs from them by the use of negation in indirect requests; indirect requests are almost never negated in English or German (cf. House & Kasper 1987: 1263–1268). The use of negated indirect requests is a feature Czech shares with Russian, but in Czech negation in indirect requests has a function different from that in Russian: it is not obligatory, but it is used as a politeness marker in certain types of requests. From an areal-typological point of view, Czech is in an intermediate position on a scale between West European languages and Russian: indirect requests are more frequently used than in Russian, negation in indirect requests is used, but to a lesser extent than in Russian. Apparently the use of negation as a politeness marker (not obligatory), as in Czech, occupies an intermediate position between its usage as illocutionary marker (obligatory), as in Russian, and its non-usage, as in West European languages. Although negation in indirect requests is common in several Slavonic languages, it is not restricted to this language family, but it occurs e.g. in Danish (cf. House & Kasper 1987: 1263–1268 and Faerch & Kasper 1989: 227). The areal distribution of patterns of usage should be further studied; e.g. the usage of indirect requests in Polish is similar to that in Czech, cf. Marcjanik (1997: 161–166). In both languages considered here, some changes seem to indicate an expansion of features typical of West European languages. This concerns the decline of negation in certain constructions in Czech, and the recent emergence of indirect requests of the type neg + conditional + “can” in Russian (cf. (1); but the Russian
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.11 (287)
Questions as indirect requests in Russian and Czech
construction differs still by the use of negation). This is at least partly due to foreign influence (especially for the Russian construction, which is largely constrained to written texts). We may ask whether linguistic features of indirect requests spread historically from West Europe to the East, and whether the contemporary patterns in West European languages represent a later stage of an evolution that is also going on in Czech and Russian. The use of negation as a politeness marker might be an intermediate stage (in Czech, indirect requests apparently used to be much more frequent, see Trávníˇcek 1951: 663, so perhaps negation in earlier stages could not function as a politeness marker). I would propose the following hypothetical order of the evolution of indirect requests: first, in a language that mainly uses imperatives, simple questions about felicity conditions for the fulfillment of the request (ability, intention, availability of an object) are introduced. In later stages, the importance of indirect requests increases, and more complex constructions involving modal verbs and verbal mood (such as “can”, “could”) gradually replace the older simple constructions. Negation seems to be used primarily in the early stages of this process, and it tends to be gradually eliminated in the later stages. In order to verify this hypothesis, the evolution of conventionalized indirect requests should be further studied in different languages.
Corpora used The data on Russian are drawn from several corpora: The Uppsala Corpus of Modern Russian was created at the Department of Slavic Languages of Uppsala University. It consists of equal quantities of fictional (Russian prose from the years 1960– 1988) and non-fictional texts (drawn from the press from the years 1985–1988), of about 500,000 word forms each.10 Since pragmatic phenomena are represented differently in fictional and non-fictional texts (fictional texts are often closer to spoken language), figures are given for both parts of the corpus separately (UC/F resp. UC/NF). A corpus of press interviews (IV) from several Russian journals available via the World Wide Web is created and maintained by the Collaborative Research Centre “SFB 441 – Linguistic Data Structures” at Tübingen University.11 The texts from the study on Colloquial Russian “Russkaja razgovornaja reˇc’ ” (RR) have been made available to researchers in electronic form by the authors; they correspond partly to the materials in Zemskaja (1973). For a description of the corpus of taped dialogues from the Moscow telephone information service “Služba 09” (Sl. 09), see Kibrik (1992: 301–313). The Czech data are taken from the Czech National Corpus maintained by the Institute of the Czech National Corpus in Prague. The current research cor-
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.12 (288)
Michael Betsch
pus (SYN2000) consists of over 100 million word forms from various written texts published in the 1980s and 1990s.
Notes . There are some counterexamples to this rule to be found, but up to now they are extremely rare; cf. the following example from the journal “Ogonek” 29, 2000: Ne mog by ty snjat’sja na plakate našego neg canPretSgM conditional pron2sg photographInfRefl on poster ourGen magazina, . . . shopGen . . . “Could you (T) possibly have a picture of yourself taken for a poster of our shop . . . ” . In the “Sl. 09” corpus, of 154 questions containing požalujsta, all but one contain also an imperative. . Abbrev.: RR Russkaja razgovornaja reˇc’; Sl. 09 Služba 09 (Moscow Telephone Information Service); IV A corpus of newspaper interviews from the WWW; UC Uppsala Corpus of Modern Russian, with fictional and non-fiction texts: UC/NF non-fiction, UC/F fiction. . There are other lexical politeness markers in Russian, like bud’te dobry (resp. bud’ dobr/dobra), which can also be combined with indirect requests (Brehmer 2000: 53), but they are much less common; e.g., in the “Sl. 09” corpus, there are 1238 occurrences of požalujsta vs. 368 occurrences of bud’te dobry. . In this table, as in the following tables, numbers are given separately for distant address (V) and non-distant address (T) and for negated (T + neg, V + neg) and non-negated questions. . Unlike the construction in (19), here the addressee is a child, and a different adjective is used (hodný “nice”, “well-educated”). . It should also be taken into account that the Czech National Corpus contains some older texts, such as older works of fiction that were reprinted in the 1990s or 1980s, e.g. several ˇ works by Karel Capek (1890–1938) such as “War with the Newts” (Válka s mloky), “An Ordinary Life” (Obyˇcejný život) or “Stories from one pocket” (Povídky z jedné kapsy). . In Russian, directives formed with a question in the conditional are mentioned by normative works (Formanovskaja & Tuˇcny 1986: 164), but such directives are very rare in real usage (Berger 1999: 234). . That means the speaker would expect the hearer to decline the request only if he/she is really unable to fulfill it, but the hearer should not “opt out”. This is expected in interviews where the addressee has already consented to give answers to questions, or in the case of requests for cigarettes among people who are on T terms with each other, see table (18). . The Uppsala Corpus is available from the Department of Slavic Languages (Slaviska Institutionen) of Uppsala University at http://www.slaviska.uu.se/korpus.htm
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.13 (289)
Questions as indirect requests in Russian and Czech . The UC and IV corpora are available for online queries from the website of the Collaborative Research Centre “Linguistic Data Structures” at: http://www.sfb441.unituebingen.de/b1/en/korpora.html
References Barnetová, V., et al. (1979). Russkaja Grammatika, Vol. 2. Praha: Academia. Berger, T. (1997). Alte und neue Formen der Höflichkeit im Russischen – eine korpusbasierte Untersuchung höflicher Kommissiva und Direktiva. In P. Kosta, & E. Mann (Eds.), Slavistische Linguistik 1996. Slavistische Beiträge, Vol. 354 (pp. 9–29). München: Otto Sagner Verlag. Berger, T. (1999). Höfliche und unhöfliche Konditionale im Russischen. In K. Grünberg & W. Potthoff (Eds.), Ars Philologica. Festschrift für Baldur Panzer zum 65. Geburtstag (pp. 233–242). Frankfurt/Main: Lang. Blum-Kulka, S. (1989). Playing it safe: The role of conventionality in indirectness. In BlumKulka et al. (1989), Chap. 2 (pp. 37–70). Blum-Kulka, S., House, J., & Kasper, G. (Eds.). (1989). Cross-Cultural Pragmatics: Requests and Apologies. Advances in Discourse Processes, Vol. XXXI. Norwood, NJ: Ablex. Brehmer, B. (2000). Höfliche Imperative im Russischen. In K. Böttger, M. Giger, & B. Wiemer (Eds.), Beiträge zur Europäischen Slavistischen Linguistik (POLYSLAV) Band 3. Die Welt der Slaven. Sammelbände – Sborniki, Vol. 4 (pp. 47–57). München: Otto Sagner Verlag. Brown, P., & Levinson, S. C. (1987). Politeness: Some Universals in Language Usage. Studies in Interactional Sociolinguistics, Vol. 4. Cambridge, New York, Melbourne, New Rochelle, Sydney: Cambridge University Press. Comrie, B. (1984). Russian. In W. S. Chisholm, Jr., L. T. Milic, & J. A. C. Greppin (Eds.), Interrogativity. A Colloquium on the Grammar, Typology and Pragmatics of Questions in Seven Diverse Languages (pp. 7–46). Amsterdam, Philadelphia: John Benjamins. Faerch, C., & Kasper, G. (1989). Internal and external modification in interlanguage request realization. In Blum-Kulka et al. (1989), Chap. 9 (pp. 221–247). Formanovskaja, N., & Tuˇcny, P. (1986). Russkij reˇcevoj ˙etiket v zerkale ˇcešskogo. Moskva: Russkij jazyk. Grepl, M., & Karlík, P. (1998). Skladba ˇceštiny. Olomouc: Votobia. House, J. (1989). Politeness in English and German: The functions of Please and Bitte. In Blum-Kulka et al. (1989), Chap. 4 (pp. 96–119). House, J., & Kasper, G. (1987). Interlanguage pragmatics: Requesting in a foreign language. In W. Lörscher & R. Schulze (Eds.), Perspectives on Language in Performance. Studies in Linguistics, Literary Criticism, and Language Teaching and Learning. [To Honour Werner Hüllen on the Occasion of His Sixtieth Birthday] Tübinger Beiträge zur Linguistik, Vol. 317 (pp. 1250–1288). Tübingen: Gunter Narr Verlag. Kibrik, A. E. (1992). Oˇcerki po obšˇcim i prikladnym voprosam jazykoznanija (Universal’noe, tipovoe i specifiˇcnoe v jazyke). Moskva: Izdatel’stvo Moskovskogo Universiteta.
TSL[v.20020404] Prn:4/11/2002; 9:44
F: PBNS34.tex / p.14 (290)
Michael Betsch
le Pair, R. (1996). Spanish request strategies: A cross-cultural analysis from an intercultural perspective. In K. Jaszczołt & K. Turner (Eds.), Discourse Strategies. Contrastive Semantics and Pragmatics, Vol. II (pp. 651–670). Pergamon. Marcjanik, M. (1997). Polska grzeczno´s´c j˛ezykowa. Kielce. Mills, M. H. (1991). The performance force of the interrogative in colloquial Russian: From direct to indirect speech acts. Slavic and East European Journal, 35, 553–569. Mills, M. H. (1992). Conventionalized politeness in Russian requests: A pragmatic view of indirectness. Russian Linguistics, 16, 65–78. Petr, J. (Ed.). (1987). Mluvnice ˇceštiny, Vol. 3. Skladba. Praha: Academia. Rathmayr, R. (1994). Pragmatische und sprachlich konzeptualisierte Charakteristika russischer direktiver Sprechakte. In H. R. Mehlig (Ed.), Slavistische Linguistik 1993. Slavistische Beiträge, Vol. 319 (pp. 251–278). München: Otto Sagner Verlag. ˇ Trávníˇcek, F. (1951). Mluvnice spisovné ˇceštiny. Cást II. Skladba. Praha: Slovanské nakladatelství. Zemskaja, E. A. (Ed.). (1973). Russkaja razgovornaja reˇc’. Moskva: Nauka.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.1 (291)
The language of love in Melanesia A study of positive emotions Les Bruce Graduate Institute of Applied Linguistics & Summer Institute of Linguistics
.
Introduction
Anna Wierzbicka (1999: 292–294) proposes four universal types of concepts for emotions:1 ‘fear-like’, ‘anger-like’, ‘shame-like’ and ‘happy/love-like’ feelings. The more positive concepts of human languages, those referring to ‘happy/love-like’ feelings, are complex concepts involving components of feelings, desires, thoughts about objects, states and events, and sometimes a component of doing something. The first goal of this study was to describe concepts for positive emotions in English, Ancient Greek, and six languages2 in Melanesia. They are compared with each other and expressions in nine other languages are included to strengthen the motivation for the typology of concepts suggested herein. All eight languages had expressions in the positive emotions domain. This corroborates Wierzbicka’s hypothesis of the universality of the positive emotion domain. The second goal of the study was to investigate questions of the universality of positive emotions: The study corroborates the Natural Semantic Metalanguage (NSM) hypothesis that there are no universal positive emotion concepts in human language. The semantic descriptions herein will directly reveal the differences between concepts in different languages. Identical concepts in different languages are the exception rather than the rule and will be noted in the course of the report. Other issues pursued include: – –
Confirming that all languages have terms that can be included in the positive emotions domain. Identifying the semantic components that occur most commonly in positive emotion concepts.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.2 (292)
Les Bruce
Whereas there are no identical emotion concepts in common to all of the languages studied, there are commonly occurring component parts of the meanings of concepts in different languages. All of the languages have at least one expression that includes the defining component of the domain as part of its definition: “someone X feels something good about someone or something Y”. All of the languages have an expression that connects good feelings with the basic concept “want”, expressed in the component “someone X wants to be with someone Y”. The one component of positive emotion concepts that Wierzbicka (1999: 294) suggested may be universally found combines a desire with an action of doing something: “person X wants to do good things for person Y”. That component was also documented in expressions in all eight languages looked at here. The third goal of the research was to investigate the possibility of a typological analysis of concepts. The corpus of data was too small for a serious typological study. However, the findings show that attested concepts in a domain that is common cross-culturally occur with very limited variation compared to the theoretically possible combinations of semantic primitives and components comprised of them. The patterns of components that co-occur and components that are mutually exclusive indicate that it might be possible to construct a significant typology of positive emotions. The data are presented according to twelve common types of concepts. The most common themes for positive emotions are defined by the most commonly occurring semantic components. Additional findings include patterns of extension of meaning from the domains of Thinking, Feeling, and Wanting to the Positive Emotion domain. Finally the findings are discussed in relation to the psychology of emotions.
. Assumptions and theoretical framework This report uses the Natural Semantic Metalanguage framework to describe the meanings of concepts in the languages being studied. This framework has been developed by Wierzbicka and her students since the first publication of Semantic Primitives in 1972. NSM has been the continuation of the work of 17th century philosophers, particularly Gottfried Leibniz, and twentieth century linguist Andrzej Bogusławski.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.3 (293)
The language of love in Melanesia
There are several philosophical and procedural difficulties inherent in any project of comparative semantics. One of these is that meanings of lexemes can only be understood if they are ultimately explained in natural language expressions. This axiom is stated as the semiotic principle in NSM (Goddard et al. 1994: 7). Comparative semantics requires that concepts be described in the same terms if they are to be rigorously compared. If, however, meanings must eventually be explained in natural language definitions, what language should be used for cross language comparisons? Using a thoroughly described international language leads inevitably to ethno-linguistic skewing. Only a language independent semantic metalanguage can avoid the skewing problem. Another issue to deal with when attempting to create semantic descriptions is the serious problem of circularity. Since meanings of words can only be explained by using other words, then all of the words in a definition must be words that require explanation, and every explanation requires explanation ad infinitum. In order to break the infinite regress in definitions NSM proposes a semantic metalanguage which is self-explanatory. This is not merely an arbitrary, unmotivated, formal device. We assume human understanding, and “We can understand ourselves to the extent to which we can rely on some concepts which are selfexplanatory . . . ‘if nothing can be understood by itself nothing at all can ever be understood’ ”(Goddard et al. 1994: 17, translated from Leibniz 1980: 430). Leibniz sought to discover the basic alphabet of human thought by searching for semantic primitives of human language, basic, inherently understood concepts that are universal to human beings. NSM theory currently proposes from 57 (Goddard 1998) to 60 (Wierzbicka 1997) primitives3 that have been documented in a wide variety of languages. Whether or not these lists prove to be the final set of primitives these concepts have proven to be useful for defining more complex concepts in many languages. They provide a means to come close to describing concepts in different languages in the same terms. To the extent that the primitives are universal, semantic descriptions in different languages can be made in their own terms thus avoiding ethnocentricity. This study assumes that these 57 to 60 concepts are universal, primitive concepts of human language and thought. NSM is a formal theory that constructs semantic descriptions by means of reductive paraphrases that take the form of propositions that are themselves constructed using the semantic primitives. The paraphrases comprise natural language, primitive concepts that are related to each other by a natural language syntax. Semantic descriptions are not lists of abstract features or even primitive concepts that are not explicitly related to each other. The implication of this feature of the theory is that when concepts are compared they must be compared at different levels. Comparison can look at shared primitives, shared component propositions, and the similarity of how the components are related to each other.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.4 (294)
Les Bruce
. Hypotheses – –
–
– –
There are few universal, primitive concepts (57–60). While the set of primitives is small, the possible combinations of primitives into semantic components are very large. The set of possible concepts, defined as the possible combinations of components, is very large indeed. Actually occurring natural language concepts will be a small subset of all of the possible concepts that could be derived from the universal set of primitives formed into propositional components. If human concepts are a restricted set of possible concepts defined by the universal primitives, it would tend to suggest the possibility of a typology of concepts. Lexicalized concepts, morphemes with their associated sense of meaning, can be categorized as significant types of similar concepts. There are domains of experience that are universal in human language. That means that there are expressions in all human languages referring to some domains of experience. For example, all languages have an expression that refers to something like English love and angry.
. Methodology . Data collection The data used in this study comes from original research and published sources. The Alamblak research was conducted in Papua New Guinea over a fourteen-year period by the author and his wife. Kathleen Bruce (ms.) worked specifically on Alamblak emotion terms in 1983. She worked by elicitation with Jude Mengumari, a speaker of Alamblak as his first language. Recent research was conducted by the author during three training workshops in Vanuatu, Solomon Islands and Papua New Guinea from April through June 1999 under the auspices of the Summer Institute of Linguistics. In each workshop mother tongue (MT) speakers were trained in semantic analysis and lexicography over a period of three weeks. The schedule involved one to three hours of lectures intermingled with five hours of analytical work each day on the languages of the participants. In three cases MT speakers worked with trained linguists, one of whom was himself a MT speaker.
. Analysis of meaning Contrastive contexts for each lexeme were examined to determine the range of reference for each lexeme. Contexts were provided by discussions of case studies, nat-
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.5 (295)
The language of love in Melanesia
ural texts, and hypothetical situations that could be referred to by the lexeme being analyzed. Lexemes were then contrasted with similar lexemes in the same language, and then each semantic description was completed by describing the essence of the unifying features of the typical referents for each lexeme. Polysemous lexemes were identified and each sense was defined. Schemas were used to a limited degree to identify entailments and implications related to the lexemes being analyzed. This process often amounted to using probing questions such as “What are common causes of this emotion?” “What does this emotion typically lead to or result in?” The notion of schema is based on the notion as used by Artificial Intelligence researchers and cognitive linguists. According to DeBeaugrande and Dressler (1981: 90): ‘Schemas are global patterns of events and states in ordered sequences linked by time and proximity and causality.’ A schema is defined by the cultural patterns of behavior and worldview, i.e. conventional beliefs and knowledge, of the speech community of the language in question.
. Constructing definitions MT speakers either wrote definitions in their own language and then translated them into a language of wider communication, or they wrote them directly in a language of wider communication. In some cases a trained field linguist wrote the semantic descriptions, always in collaboration with the MT speakers. Definitions were tested for descriptive adequacy against the case studies, natural texts, and hypothetical situations that were prepared by the MT speakers. The research setting did not allow for the analyses and definitions to be checked with a wide variety of native speakers of the languages studied. Semantic descriptions generally followed the NSM format with reductive paraphrases comprised of semantic primitives. Workshop participants attempted to write one general definition before appealing to multiple meanings. Notions of prototypical scenario and comparison with such a scenario were used in the definitions. For example, the Maskelynes term l6mas3 compares how someone feels something good toward someone else to how a man feels when he wants a woman.
. Comparing concepts Concepts are similar to the degree that they share components of meaning. Sharing components entails sharing primitive concepts that comprise the components that are propositional statements. As was mentioned earlier, NSM theory dictates that primitive and propositional components of meaning must be related by a natural syntax. Therefore a complete comparison of concepts in different languages will
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.6 (296)
Les Bruce
include similarities in the ways propositional components are related to each other. The concepts in this study were compared primarily on the component level.
. Semantic typology Given the formal nature of semantic descriptions in NSM it is possible to precisely compare the similarities among concepts either within one language or among different languages. It is an empirical question whether human concepts fall into discrete groups revealing similar conceptualizations of experience or whether concepts in different languages form a continuum of concepts uniformly similar and different from each other. As a first step toward investigating a typology of positive emotions, this study will group concepts from different languages according to patterns of correlations of commonly occurring semantic components.
. Types of concepts It is impossible to completely overcome the subjectivity factor in the formulation of the definitions; the theory is not so formal as to be amenable to mechanical application in the precise wording of definitions. Of the semantic components that were used in the explications of the 46 lexemes in this study, thirteen components were observed in explications of lexemes in two or more languages. These components were accepted as definitive of cross-language types of concepts. These thirteen components are listed in Table 1. Table 1. Common semantic components Feeling
“someone X feels something good about someone or something Y”
Thought
“someone X thinks about Y”
Desires
“someone X wants to be with someone Y” “someone X wants to do something good for someone Y” “X wants something to happen” “X wants to do something with Y for X’s good”
Actions
“something good happened” “something bad happened”
Taking action
“someone X does something good for someone Y”
States
“X thinks about Y I am like a part of Y” “X is with someone Y” “something is good” “X has Y”
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.7 (297)
The language of love in Melanesia
Of the 4095 possible combinations of these 13 components,4 which define concepts, twelve combinations were documented in more than one language of the 17 languages looked at in this study. These 12 correlations of semantic components constitute the 12 semantic types of concepts used to present the data of this study. Much more field work is needed to construct a significant typology of positive emotion concepts of human language and several theoretical issues will need to be considered which are not incorporated here. The preliminary types of concepts presented here seem to be nontrivial, nonetheless, given the limited number of correlations of commonly occurring semantic components in the languages studied.
. Twelve common types of positive emotions The semantic components discussed in the preceding section comprise the basic nature of human concepts of positive emotions. Positive emotions involve good feelings coupled with related thoughts about the causes of those feelings. Additionally, positive emotions commonly involve a desire to be or have, a desire to do, or a desire for something to happen; less commonly they involve pure events, actions, or states without invoking a desire about them. – –
– –
The first seven types of concepts involve a desire to be with someone or something, or to do something, or simply to have something. Types 8 and 9 illustrate positive emotions related to Events (good or bad things happen) with or without an accompanying desire for something good to happen or to do something good. Type 10 illustrates two concepts related to Actions (doing something good). Types 11 and 12 illustrate positive emotions related to two States (something is good and someone has something).
Each definition is preceded by the name of the language with a location indicated in brackets, followed by the term set in a syntactic frame.
. Feel good, think about, want to be with, (and be with) The first type of positive emotion is a simple type of concept involving good feelings, thoughts about, and a desire to be with someone or something else.
English (U.S.) love1 In this study we are tentatively proposing four senses for English love in the domain of positive emotions. These senses cover the range of meanings of the nine
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.8 (298)
Les Bruce
emotion-type senses of the term enumerated in the Webster’s New World Collegiate Dictionary (WNWCD).5
English (U.S.), someone X loves1 someone Y someone or a thinking thing X feels something very good toward someone or a thinking thing Y like someone who is with someone he/she wants to be with because that person makes him/her feel good when that person is with him/her An example of this sense comes from an editorial. Sitting and hearing Bowden’s philosophies about discipline reminds you of an overly soft father. The kids love him, but they take advantage of him the minute he turns his back. (Jay Mariotti, “Warrick still can’t erase stigma”, SUN-TIMES, January 5, 2000)
This kind of emotion reflects an enjoyment of being with the object of one’s affection; there is no requirement or responsibility placed on the Subject who feels this emotion. The experiencer’s emotions do not move him/her to do things for their object of affection, but rather it is an emotion that responds to their own good pleasure. The explication of the meaning specifies that the experiencers and the objects of affection may be people or animals (thinking things). The hero-dog Buck is accredited with this kind of love in Jack London’s The Call of the Wild: But love that was feverish and burning, that was adoration, that was madness, it had taken John Thornton to arouse. This man had saved his life, which was something; but, further, he was the ideal master.
Maskelyne /Uliveo (Vanuatu), someone X l6]on3 Y someone X feels something good toward someone or something Y when X thinks about Y because of that X wants to be with Y This term may refer to someone’s feelings toward a person or an object. Its range of reference is best analyzed into three senses. Sense one includes feeling something, either good or bad; sense two refers to wanting to have something as in “Someone wants your radio. Would you give it to him?” The range of reference for sense three, discussed here, is still quite broad, overlapping in usage with English ‘like’, ‘love’, and ‘adore’ lacking specific components of each of them. An intensified expression based on this root morpheme is discussed in Type 2.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.9 (299)
The language of love in Melanesia
Maskelyne /Uliveo (Vanuatu), ri] n6lon 6lan (lit. “put your heart on him”) someone X ri] n6lon 6lan someone Y someone X feels something good toward someone Y when X thinks about Y X wants to be with Y because that person makes X feel good when that person is with him/her This Maskelyne idiom is used to refer to the attraction one person has for another. It resembles English love “but for your good not theirs”. (Healey, ms.)
Utku (Eskimo), someone X unga someone Y X feels something good toward Y X thinks something like this: I want to be with Y when I am with Y, I feel something good Unga expresses an emotion and desire to be with a loved one. The meaning is illustrated in the following example: “Inuttiaq’s children were said to unga him, to want to be with him, because he was never annoyed (urulu) with them” (Wierzbicka 1992: 161).
. Feel good, want to be with, and want to do good things for Most of the five concepts of this type, while sharing these three core components, differ in some way.
English (U.S.), someone X loves2 someone Y someone or a thinking thing X feels something very good toward someone or a thinking thing Y X wants to be with Y X wants to do good things for Y The following example is taken from The Pathfinder in which the hero felt concern for the safety of his niece; it is not a love of physical attraction: The old mariner now played his part manfully; for he was on his proper element, loved his niece sincerely, had a proper regard for his own person, and was not unused to fire, though his experience certainly lay in a very different species of warfare. A few strokes of the paddles were given, and the canoe shot into the bushes, Mabel was hurried to land by Jasper, and for the present all three of the fugitives were safe. (The Pathfinder, by James Fenimore Cooper)
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.10 (300)
Les Bruce
This dimension of devotion was also attributed to the dog, Buck in The Call of the Wild. “Because of his very great love, he could not steal from this man. . . Nothing was too great for Buck to do, when Thornton commanded.”
Maskelyne /Uliveo (Vanuatu), someone X l6]onb6n1 someone Y someone feels something good toward someone else Y when X thinks about Y X wants to be with that person and wants to do good things for that person This expression derives from l6]on3 ‘feeling good about someone and wanting to be with them’ by the addition of the intensifier suffix -b6n. The semantic effect of this suffix with emotion terms seems to be to increase the intensity of the feeling (cf. l6masb6n) or to increase the value of the ideal by shifting the good feelings and desires of the experiencer from a self-centered focus to an other-centered focus. L6]onb6n1 expresses the feelings of the experiencer who now seeks to do good things for the other person’s good. Note a similar effect in voib6nx6n in Type 6.
Maskelyne /Uliveo (Vanuatu), someone X l6masb6n1 someone Y someone X feels something good toward someone Y when X thinks about Y like someone who wants to be with that person and like married people who do not want to be apart and like someone who wants to do good things for that person This expression is used in contexts that focus on the desire of a person (people) to be together. It may refer to a woman and a man who are married. “They never run from each other.” That is, they always like to be together. L6masb6n is derived from l6mas2 (Type 4, an emotion and desires based on a kinship relationship). The prototypical use of this term is for describing the love of a couple in a close marriage relationship. It is not strictly limited to kinship, however, as the root l6mas2 is, a difference that is reflected in the explications of the two terms. Since the intensifier suffix -b6n is employed, we would expect that it would reflect an increase in the value of the emotion over that of the base l6mas2 as it does in l6]onb6n1 and voib6nx6n. L6masb6n is apparently the same type of emotion as that of a parent for his/her child, but intensified in the sense that the experiencer in the situation wants so much to be with the other person that he/she does not want to be apart from her/him.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.11 (301)
The language of love in Melanesia
Uare (PNG), someone X zamare2 someone Y someone X thinks that someone else Y is good because of that X feels good about Y and wants to be with that person and wants to do for that person something that person wants Zamare2 derives from the primary sense of the term, ‘think’. Alamblak dukay2 (Type 6) shows a similar extension of meaning, from the cognitive domain to the positive emotion domain. Zamare2 may refer to feelings and desires of human or supernatural experiencers most commonly for close friends or relatives. The feelings expressed by zamare may be caused by someone doing something good for another person, or something the first person enjoys, like telling him legends of the past. People feel this way toward someone who has returned from a long trip, or when something bad did not happen that was expected to happen such as when a very sick person recovers.
Koine Greek, agapa1 The ancient Greek concept agapa1 has been analyzed to include the component of “wanting to do what Y wants”. Koine Greek, someone X agapa1 someone Y someone X feels something good about someone Y and wants to be with Y and wants to do good things for Y and wants to do what Y wants This term is the other-centered expression of love in ancient Greek. This sense is used in contexts like: “If you agapa me you will obey what I say.” This sense contrasts with sense two, not discussed in this project, which refers to a desire to have something.
Yankunytjatjara (Australia), Mukuringanyi2 Mukuringanyi2 expresses the same desires to be with someone and to do something good for them. The concept differs from the others of this type by including an evaluative component. someone X Mukuringanyi2 someone Y Something like this is happening to X: X thinks Person Y is good
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.12 (302)
Les Bruce
I want to be with Y I want to do good things for Y Goddard (1990: 270) clearly includes this sense in the emotion domain as similar to English “feel fond toward”, although he does not include a feeling component in his explication.
. Feel good, (think about), want to be with, and want to do something good for Y, do something good for Y This type of concept combines the notion of a positive emotion with the resultant action that is motivated by that emotion together with attendant desires. This class of expressions shares the action “do something good for” with the expressions in Type 10.
Lunqa [lu]a] (S.I.), someone X and someone Y uluakako two people very much feel something good toward each other when they see each other because of that they want to be with each other and they do something that people can see and know that they feel and want this they do this because they want to This expression typically refers to the emotion felt between people of the opposite sex who are interested in marriage and/or desirous of a sexual relationship. The term is not restricted to this context, however. This emotion can be experienced by the best of friends of the same sex without a sexual association. Responses to this emotion could be smiling, laughing, hugging and jumping up and down in excitement. Note that “uluakako is more an emotion that results from X and Y seeing each other than from them thinking of each other,” (Alpheaus Zobule, personal correspondence). It may be that a cognitive component is entailed by the stated components of the Lunqa expression. That does not mean, however, that MT speakers conceptualize it that way. The concept expressed by uluakako depends upon a visual stimulus without reference to a cognitive aspect.
Maskelyne /Uliveo (Vanuatu), someone X l6masb6n2 Y Someone X feels something good toward someone Y when X thinks about Y and wants to be with that person
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.13 (303)
The language of love in Melanesia
and wants to do good things for that person because of that X does something good for Y L6masb6n2 is used in contexts in which the actor-experiencer expresses his/her feelings for the recipient in practical action. X’s desire to be with Y may not be greater than that expressed in l6mas2 (cf. in Type 4), but the emotion is intensified; more than wanting to be with someone or wanting to do something for Y, X actually does something for Y. It is possible to see both senses of l6masb6n as an intensification of l6mas2 (Type 4, an emotion and desires based on a kinship relationship). In the case of l6masb6n1 (Type 2) the focus is on the experiencer’s desire to be with the other person. In this case the focus is on the good that the actor does.
. Feel good, want to be with Y, X thinks about Y “I am like a part of Y” This type of concept includes the component that bases the feelings and desires on the relationship between the experiencer and the object of his or her emotion. Because of the distinctiveness of the relational component the self-oriented concept in Greek, with the component “X wants to have Y with him because Y makes X feel good”, is included here along with the other-oriented concepts of Sibe and Ontong Java that include the component “wants to do good things for Y”. The Sibe and Ontong Java lexemes are almost exact equivalents although the languages are unrelated. Sibe loma leinsi is more general, however, expressing good feelings toward people, animals or inanimate objects.
Sibe (PNG), someone X loma leinsi someone or something Y someone X feels something good toward someone or something Y and wants to be with Y and wants to do good things for Y like someone who thinks about Y I am like a part of Y The expression loma leinsi may be better described with two senses. On the other hand, the feelings and desires of this expression in all contexts share common defining principles. Differences between feelings for and associations with people versus inanimate things may be relegated to pragmatic, cultural factors. The concept, when referring to people, is based on the prototypical case of feelings and desires between relatives. In the case of feelings about inanimate objects the prototypical case is with one’s possessions that he/she wants to take proper care of. In both cases the experiencer has a special association with the object of his feelings and he/she wants to do good things for it.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.14 (304)
Les Bruce
The explication likens the feelings and desires of loma leinsi to someone who has a special association with Y. While the meaning of the concept is based on that prototypical scenario, the expression’s range of reference is wider than only feelings for relatives or one’s possessions. Any person may feel these emotions and desires for any other person of any age, social status whether or not that person is a relative. Situations that could give rise to this emotion are many: Y did something to please X, like giving X food or carrying X when X was sick or hurt, or finding X’s lost pig; X wants to marry Y and/or have sexual relations with Y; X found a gold nugget; something bad did not happen that was expected to happen such as when a very sick person recovers. Responses to these feelings could be smiling or laughing, wanting to give something to the other person, or wanting to keep an object (like a gold nugget).
Ontong Java (S.I.), someone X laoi someone Y someone X feels something good toward someone else and wants to be with that person Y and wants to do good things for that person because X thinks “Y is my kin”
This term commonly occurs in three collocations. Haikama laoi: Haikamana laoi: Haikina laoi:
‘parents’ love for their children’ ‘children’s love for their father’ ‘children’s love for their mother’
Koine Greek (ancient Indo-European), someone X phile someone or something Y someone feels good about something or someone Y when X has Y with him X wants to have Y with him because Y makes X feel good like someone who thinks about Y “I am close to Y like to a kin” The prototypical object of phile is a relative. The root freely compounds with other roots as in: φιλ´αδελφoς ‘love for one’s brother or sister’ φíλανδρoς ‘love for one’s husband’ φιλóτεκνoς ‘love for one’s own children’ φιλóτoργoς ‘love for those closely related to one, particularly members of
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.15 (305)
The language of love in Melanesia
one’s immediate family’ φιλανθρωπíα ‘love of mankind’
(Louw & Nida 1988: 293)
Maskelyne /Uliveo (Vanuatu), someone X l6mas2 Y someone feels something good toward his or her child Y when X thinks about Y X wants to be with Y and wants to do good things for Y This term refers exclusively to parental feelings toward children. This term typically evokes the image of a parent holding his/her child in his/her arms.
. Feel good, (think about), do something with Y that X wants to do This category comprises self-oriented concepts in which an emotion motivates and is motivated by a desire to satisfy oneself by doing something with the object of attraction.
English (U.S.), someone X loves3 someone or something Y Someone or a thinking thing X feels something very good when X thinks about something or someone Y Like someone does when that person is doing something with Y that that person wants to do. Love3 accounts for the three senses in WNWCD (1964: 869) “sexual passion. . . show love for by embracing, fondling, kissing, etc. . . . to delight in, take pleasure in an object or event.” In this third sense of love what the experiencer wants to do depends on the object of its love. “X loves his cologne,” implies X wants to smell it. “X loves Michelangelo,” can imply that X wants to look at his paintings, among other things. “X loves enchiladas,” implies that X wants to taste and eat them. “X loves the girl next door,” can imply that X wants to look at, talk to, or make physical contact with her. “X loves her catnip,” can imply that the cat wants to smell it. In the case of one person loving another, this sense of “love” can be paraphrased by “has a crush on.” The orientation of this emotion is toward the pleasure of the
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.16 (306)
Les Bruce
experiencer. This contrasts with sense one of this word that is oriented toward the good of the object of one’s love. Though love for the soil which had belonged to his ancestors kept the Sagamore of the Mohicans with a small band of followers who were serving at Edward, under the banners of the English king, by far the largest portion of his nation were known to be in the field as allies of Montcalm.
In this quote from The Last of the Mohicans, “The love for the soil” implies a desire to live in harmony with the earth and to derive one’s sustenance from the soil.
English (U.S.), someone X likes2 something or someone Y Someone or a thinking thing X feels good about something or someone Y when X thinks about something or someone Y like someone does when that person is doing something with Y that that person wants to do The difference between love3 and like2 is only a matter of intensity of the feelings. Either love or like could be used in the following context. a gentleman likes his delicacies, and a lady her feathers (The Pathfinder by James Fenimore Cooper)
Alamblak (PNG), nanho wafet The Alamblak idiom nanho wafet commonly expresses non-referentially something someone likes very much. Like the English sense of love above, what the Actor wants to do depends on the object of its love. someone says nanho wafet (lit. “it is my liver”) Someone X feels something very good about something Y because of that X wants to do something with Y that X very much wants to do Yirapam nanho wafet “fish meat is my liver” is a common Alamblak expression. It implies that X enjoys eating fish more than most things.
Alamblak (PNG), someone X yuhur2 something Y someone X feels good about something Y like someone does when he is doing something with Y that he wants to do because of that X wants to have Y Someone might have this feeling of liking and desire to have some new clothes so that he/she can wear them; he/she might want a child or more children.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.17 (307)
The language of love in Melanesia
A related idiom to this term is Tu-yuhur (lit. “something shoot-feel good about and want”) as in Rediot tu-yuhurw6tr (lit. “the radio it-shoots-he feels good about and wants (it)”) The experiencer is encoded as the Undergoer (Object) in this construction; the object that is desired is the surface Subject.
Maskelyne/Uliveo (Vanuatu), l6mas3 This expression includes the comparative component: “like a man feels when he wants a woman.” someone X l6mas3 someone Y someone feels something good toward someone Y when he thinks about Y like someone X feels when X wants to do something with Y so that X will feel something good This person feels like a man feels when he wants a woman For the prototypical situation for this sense of l6mas imagine a man who “has this feeling when his inner being steals a woman.” That is the Maskelyne description of a man burning with desire to have sexual relations with a woman. The component of “X wants to be with Y” has not been included in this definition. “Being with someone” is certainly entailed in “doing something with someone”, but in this concept the focus of attention is strongly on doing something with Y for X’s pleasure. This term is not used to refer to enjoying someone’s company in the sense of feeling content just to be with someone.
Utku (Eskimo), niviuq This Utku expression shares a specific manifestation of the component “want to do something with” with the other expressions in this concept type. someone X niviuq someone Y X thinks something like this looking at Y one feels something good like one feels looking at a baby because of this X wants to touch Y because of this X feels something good toward Y
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.18 (308)
Les Bruce
Yankunytjatjara (Australia), mukuringanyi3, and pukularinyi The Yankunytjatjara expressions mukuringanyi3 , and pukularinyi fit well in this type even though Goddard (1990) formulates the explications with different wording than that used in other terms in this class. The key similarity has to do with wanting to do something in relation to something Y. someone mukuringanyi3 something Something like this is happening to X: X thinks I want to do something, without Y I won’t be able to do it. someone pukularinyi something X is undergoing something like this X thinks Now I can do something I want to do
. Feel good, think about, want good things/not want bad things, want to do good for The examples of this subtype include feeling, thinking and wanting. The wanting component focuses on wanting good things for Y in the Maskelyne case and not wanting bad things for Y in the Eskimo case. Both components seem to be a part of the Alamblak example.
Alamblak (PNG), someone dukay2 something or someone Y someone feels good about someone or something Y like someone who thinks about something or someone else because that person wants to do good things for Y and does not want anything bad to happen to Y Dukay is the Alamblak word meaning ‘think’ in its primary sense (cf. Uare zamare in Type 2). This second sense derives an emotion term with a prima facie cognitive connection. Perhaps the English expression care or care for/about would be the closest to the feeling and motivational components of dukay. This expression could be used to reassure a child of his/her parent’s concern. Anurwahn, yifammaf dukewfn; Don’t cry, (your) parents care about you; yesëhrim hnehirahfn. they will bring some food to give to you.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.19 (309)
The language of love in Melanesia
Maskelyne /Uliveo (Vanuatu), someone X voib6nix6n someone Y someone feels good about someone else Y when he thinks about Y and wants to do good things for that person because X wants good things to happen to Y The Maskelyne morpheme Voi is the equivalent of ‘good’. With the transitivizer suffix, voix6n, the expression means “someone does something good for someone else”. Voi itself is an emotive term associated with good feelings. When someone does something good for someone else it is expected that the recipient will have good feelings (l6]on3 or ri]n6lon 6lan – literally ‘put your heart on him’) toward his benefactor. David Healey suggests (personal correspondence) that part of the motivation inherent in actions described by voix6n is less than altruistic, that “X wants Y to feel good toward him.” With the addition of the intensifier, -b6n, the value of the pseudo-emotive voix6n is augmented to an other-centered feeling on the part of the experiencer for the good of the recipient. Maskelyne speakers think of this expression as referring to an experiencer who wants “only good things to happen to a person.”
Utku Eskimo, naklik The Utku term naklik presents the motivation for this emotion in a partly negative frame. The experiencer does not want bad things to happen to Y. someone X naklik someone Y X feels something good toward Y X thinks something like this: I don’t want bad things to happen to Y X wants to do something good for Y Wierzbicka (1992: 161) reports the analysis done by Briggs (1970) that naklik is used to refer to “people who are hungry or for people who are cold, but not for people who are impaired physically. . . for taking care of the ill, for adopting orphans, and for marrying widows, all categories of people who are in need of physical assistance.” Clearly the prototypical schema for this emotion in Utku involves a response to people who are in bad circumstances, thus the perspective “I don’t want bad things to happen to Y.”
. Feel good, and want something Concepts in this subtype include the primitive “want”, but the specific complement of “want” is different in each case. This type of concept is at the boundary between
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.20 (310)
Les Bruce
emotions and desires simply because of the paucity of components to distinguish between them. Maskelyne l6mas, and l6]on are apparently polysemous with senses in both domains. Similar lexemes are found in other languages also. Uare urate is included in this type, but it also includes the additional component “want to be with”. It is included here rather than as a separate type because it is the only lexeme that combines just those three components. Yuhur2 (Alamblak) in Type 5 clearly includes the component “X wants Y”. It may be the case that Type 7 should be incorporated into Type 5 if “wanting something” proves to always imply that someone wants something in order to do something with it.
Lunqa (S.I.), someone nyorogua someone or something Y Someone X feels good about someone or something Y When X thinks about Y like someone who wants to have something Nyorogua is a general term of desire to have something and the associated positive emotion. It may refer to someone who likes someone else, or who likes and wants a coconut or potato; it may refer to the emotions and desires of domestic animals. The same term refers to how God or spirits feel toward people when they want to help or do something for them. The case of supernatural beings highlights the reason that the explication is not always “wanting to have something”, but an emotion that is “like someone who wants to have something,” the prototypical case. The emotional desire may be directed towards inanimate possessions and places, as well as supernatural beings, domesticated animals, and people (of the same or opposite sex and of any age or status). Note the difference between the component “wanting to have something” in the Lunqa term versus the expressions in Type 5, e.g., English like, “like someone does when that person is doing something with Y that that person wants to do.” The Uare expression urate includes the ubiquitous component “X wants to be with Y”.
Uare (PNG), someone X urate someone Y someone X feels something good about someone or something Y X wants something (usually for Y to do something for X) and X wants to be with Y The common component “thinking about” is not included in this explication. The reason for that exclusion is that it would be normal to express thinking separately from urate, as in “‘because he thought about (zamare) that person first, now he wants (urate) that person”’ (personal correspondence with Keiko Kikkawa). The
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.21 (311)
The language of love in Melanesia
emotional experience may include a cognitive component, without it being a part of the Uare concept. It is not likely that the meaning of urate entails the meaning of zamare since combining both expressions in one sentence would result in a tautology. This term refers to experiencers who are people, adult or child, or supernatural beings. They may feel this way toward people, domestic animals and natural or manmade possessions, but most commonly toward people. This emotion may be caused by being attracted to someone and wanting to marry that person and/or to have sexual relations with that person. It may also be caused by seeing something desirable or valuable. In response to the emotional desire the experiencer might want to buy an object, or ask someone to give it to him/her.
Hawaiian, aloha someone X feels something good toward someone Y X wants Y to feel something good This Hawaiian term is other-person-centered like types 2, 3, 4, 6, 9, and 10. It is placed here because what is wanted is very general. Wierzbicka (92: 153) analyzes aloha as implying “‘good feelings’ toward other people and ‘good wishes’ for other people, without any implication of personal bond, commitment, or active concern.” No one uses this expression because he/she “wants to be with someone”, or “wants to have someone”, or “wants to do something good for someone.”
. Feel good, (think about), something good happened / nothing bad is happening, (and someone wants that good thing to happen) Concepts of Type 8 are intransitive. They may or may not bring the cognitive component, “thinking about” into consciousness. The reason or grounds component, “something good happened”, may be expressed from a negative perspective, “nothing bad is happening now,” as in the Alamblak expression yindhor2 .
English (Australia), someone has joy X felt something because X thought something sometimes a person thinks: something very good is happening I want this to be happening when this person thinks this this person feels something very good
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.22 (312)
Les Bruce
X felt something like this because X thought something like this In her discussion of joy Wierzbicka (1999: 51–52) points out that joy is a short-term emotion based on something that is happening now.
English (Australia), someone is pleased Someone X felt something because X thought something sometimes a person thinks: something good happened I wanted this to happen when this person thinks this this person feels something good X felt something like this because X thought something like this Wierzbicka (1999: 55) illustrates pleased from a quote from Kemble: Mr. Butler, who is highly pleased with Mr. King’s past administration of his property, wished . . . to give him some token of his satisfaction.
English (U.S.), someone is happy someone X feels something good like people feel after something good happens that they wanted to happen and like people feel when they think that nothing bad is happening now The final component of happy is a way of expressing a desire for maintaining a preferred state, that nothing bad happens to change the status quo. We can’t give up our girls for a dozen fortunes. Rich or poor, we will keep together and be happy in one another. (Little Women, by Louisa May Alcott)
Ontong Java (S.I.), someone X ekauale’a1 someone X feels something good because something good happened and X is thinking about that This expression refers to the feeling people have when something good happens. This feature makes the term more like happy than the other English terms in this set.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.23 (313)
The language of love in Melanesia
The Ontong Java expression has the same explication as the following Alamblak expression. Data are not available for Ontong Java to determine whether the range of meaning of ekauale’a includes the situation “nothing bad is happening now” like yindhor2 . A second sense of ekauale’a parallels Alamblak yuhur and English like in Type 11.
Alamblak (PNG), someone yindhor1 someone X feels something good because something good happened and X is thinking about that The Alamblak expression yindhor has two senses according to K. Bruce’s analysis. The presenting situation in the second sense is “nothing bad is happening now” rather than “something good is happening.” The peace of tranquility, the lack of hardship or any other trouble brings this sense of good feeling.
Alamblak (PNG), someone yindhor2 someone feels something good because nothing bad is happening now and X is thinking about that It is no doubt that the good feeling of this Alamblak lexeme is the result of the experiencer’s perception of tranquility. Obviously the Alamblak speaker is not claiming that “nothing bad is happening now” anywhere, but by the pragmatic principle of minimal interpretation, nothing bad is happening within the immediate sphere of the experiencer.
Sibe (PNG), someone lotulotu Someone X feels very good X does something so that other people can see and know that X feels very good X does this because something good happened The experience of lotulotu involves doing something to express one’s feelings, such as some physical action with or without a vocalization like singing or exclaiming. Someone may experience this emotion upon hearing some good news, or otherwise learning that something good has happened, including receiving something that he or she needed.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.24 (314)
Les Bruce
Sibe (PNG), taba’lei’ This Sibe term overlaps in meaning with lotulotu. someone taba’lei’ someone X feels good because someone Y did something good to X Like lotulotu, taba’lei’ is an emotional response to something good happening to X. Taba’lei’ is restricted in range to a response to someone Y doing something good for X. This expression describes the more reserved response without the external expression of joy.
Ifaluk, ker and gachu Lutz describes these two terms with two common components: “something happens ‘that we want’ to happen”, and “Our insides feel good”. (Lutz 1982: 121) Ifaluk (Micronesia), someone ker1 Someone X feels something good like people feel when something good happens that they wanted to happen and like people A feel when other people say people A are very good people
Ifaluk, (Micronesia), someone X gachu2 Someone X feels something good like people feel when something good happens that they wanted to happen Gachu may be best analyzed with one, general sense. It has been tentatively analyzed into two senses here.
Japanese, ii kimochi The Japanese lexeme ii kimochi is mentioned in Wierzbicka (1999: 280) without reference to the experiencer thinking about the causing situation. The context given there of observing an event that elicits the good feeling suggests that thinking about it must be a part of the experience, but as noted earlier, thinking about it may not be a part of the emic conceptualization of the experience. someone ii kimochi someone X feels something good because something good happened
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.25 (315)
The language of love in Melanesia
Utku (Eskimo), someone quvia1 someone X feels something good because something good is happening Briggs (1970: 327) glosses this term as ‘happiness’. She describes three types of situations that elicit this emotion. Quvia is used to refer to an emotional response while doing something that one likes to do. The explication here is more general allowing for good things that happen that ego is not doing himself. From Briggs’ discussion the second component might be restricted to “because X is doing something that X wants to do (but does not have to do?).” We include this sense in this type in any case, because it can be construed that something good is happening, when ego is doing something ego wants to do. Other contexts for quvia suggest an emotion akin to contentment, like might result from being warm and fed. In other situations this emotion seems to relate to a mental state of happiness coupled with a sense of security. It may result, for example, from other people demonstrating their happiness, itself a sign of the lack of animosity, such as by smiling, laughing, telling stories and joking. In that type of context the additional component “that makes X think that nothing bad will happen now” may be warranted. More research would be required to postulate other senses of this term.
..
Feel good, think about, something good happens, and someone wants something to happen Concepts in subtype 5.8.1 are transitive; the complement of the emotion verb is an event. English (U.S.), someone or a living thing X loves4 something (EVENT) someone or a thinking thing X feels something very good when something good Z happens and when X thinks about that happening because of that X very much wants something Z to happen Examples for both human and animal experiencers follow: “I would love (for her) to win a gold medal in the Olympics.” Best of all, perhaps, he loved to lie near the fire, hind legs crouched under him, fore legs stretched out in front, head raised, and eyes blinking dreamily at the flames. (The Call of the Wild by Jack London)
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.26 (316)
Les Bruce
English (U.S.), like3 Someone or a thinking thing X feels something good when he thinks about something good Z happening because of that X wants something Z to happen As previously observed the difference between love and like in English has to do with the degree of intensity of the feelings and desires involved. The first example is referring to the feelings of the dog, Solleks. He did not like to be approached on his blind side. (The Call of the Wild by Jack London)
Alamblak (PNG), yuhur3 Someone X feels something good when X thinks about something good Z happening because of that X wants something Z to happen This Alamblak expression is virtually identical to the sense of like3 .
. Something bad happened, (feel good about), (want to do / should do / do something good) Concepts for positive emotions that are related to something bad happening also include either “feeling good”, and/or “wanting to do something good for someone” and/or “doing something good for someone.”
Sibe (PNG), someone X wikena someone Y someone X feels something good toward someone Y because something bad happened to Y X wants to be with Y and does something good for Y Wikena is attributed to people and supernatural beings that have feelings and desires toward other people and supernatural beings. This is common between all people regardless of age, sex, or social status. A typical scenario for this emotion is the sickness of a friend or family member. The typical response is included in the explication: to do something good for the person. Giving something to the person and crying and hugging may also be involved.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.27 (317)
The language of love in Melanesia
Maskelyne /Uliveo (Vanuatu), someone lolosa2 someone Y something bad happened to someone Y because of that X thinks Y feels something bad. Because of that X thinks something like I feel bad like Y feels X feels something good toward Y and he/she wants to do something to make that person feel good. This term closely resembles ‘to feel pity or compassion’ with a component of empathy in English. It is used to refer to the emotion someone feels upon observing someone else who is suffering emotional agony in difficult circumstances of life. The compassionate, empathetic person wishes a better life for the unfortunate one.
Alamblak (PNG), someone yohoaf someone or something Y something bad happened to someone or something Y someone X feels like someone who thinks this should not have happened and because of that this person feels something bad and wants to do something good for Y This term can be used to refer to states of affair, such as seeing a crippled person; it may refer to events like something bad happening to a loved one, the departure of a person who is significant to ego, witnessing a fight between brothers, or when something special to ego becomes broken. On such occasions ego is inclined to offer a gift, help the needy person or try to fix the broken object. These terms closely resemble the English compassion. Compare these with the following expressions discussed in Wierzbicka (1992: 142–174).
English, someone X has compassion for someone Y X thinks something like this something bad happened to Y when X thinks this, X feels something good toward Y if it happened to me, I would feel something bad because of this X feels something bad
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.28 (318)
Les Bruce
Russian, someone X žalost’ someone Y X thinks something like this: something bad is happening to Y because of this, X feels something good toward Y if X could, X would want to do something good for Y because of this, Y feels something bad I feel bad like I want that it had not happened
Samoan, someone X alofa someone Y someone thinks of person Y X thinks something like this: bad things can happen to a person X feels something good toward Y if something bad happens to a person, people should do something good for this person X wants to do something good for Y
Tahitian, someone X arofa someone Y someone feels something bad and feels something good toward someone, because of thinking that something bad could happen to that person and not wanting them to feel bad.
Ifaluk (Micronesia), someone X fago someone Y X thinks of someone Y X feels something good toward Y X thinks something like this: something bad can happen to a person when something bad happens to someone, some people should do something good for this person I don’t want bad things to happen to Y When X thinks that something bad can happen to Y X feels something bad because of this, X wants to do something good for Y Ifaluk fago is similar to English compassion, love, and sadness according to Lutz (1982). An example of its use would be, “If someone is sick or leaving the island, and I have nothing to give them, I feel fago and tang.” (Lutz 1982: 119)
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.29 (319)
The language of love in Melanesia
Pintupi (Australia), someone X yulatjarra someone Y X thinks of person Y X thinks something like this: something bad happened to Y X feels something good toward Y Y is like a part of me because of this, X feels something bad X doesn’t want not to feel this
. Feel good, think about, and do something good This type of concept entails taking action as a result of one’s feelings and thoughts.
Lunqa (S.I.), vaialalekanai The Lunqa lexeme has a reciprocal orientation. Someone X and someone Y vaialalekanai Someone X and someone Y very much feel something good toward each other when they think about each other because of that they do good things for each other For vaialalekanai people may like to be with each other, but “The important thing is they do good things for each other, whether or not they see or are with each other” (Zobule, personal correspondence). This emotion is expressed between people of any age, sex or social status, whether or not they are kin. The typical situations that cause this emotion have to do with a relationship that has developed to the point that the people involved give things to each other and help each other with their work. The typical responses of this emotion include wanting to give things or to help someone. People of the opposite sex can experience a relationship like this without having any sexual interest in one another.
Ontong Java (S.I.), lolosi Lolosi is similar to the Lunqa term, but adds the component “because of that Y knows that X feels something good toward him.” someone X lolosi someone Y someone X feels something good toward someone Y and
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.30 (320)
Les Bruce
X does something good for Y because of that Y knows that X feels something good toward him In this concept the good feeling of the experiencer moves him or her to do something good for another person. That act in turn brings to the recipient’s mind that his or her benefactor feels good toward him or her.
. Feel good, think about, something is good This type of concept involves a response to thinking about a state of being, with or without reference to a desire.
English (U.S.), someone X likes1 someone Y someone or a thinking thing X feels good about someone Y because X thinks something about Y is good The explication of like1 describes the sense of like used in sentences like: “I like Tony Gwen. He is loyal to San Diego and not controlled by the almighty dollar.” The distinction between the attitudinal sense of like and the emotional sense is difficult to make. Perhaps it is not so much qualitative as a matter of how strong the feeling is or how deeply a person thinks about the other person he/she likes; focused meditation over a period of time can feed back into the feeling level creating a clear emotional response out of what began as a thought about a disposition. When collocated with an adverb like really the emotive content is highlighted.
English (U.S.), adore Adore includes the additional component: “X thinks Y is better than anyone or anything else.” someone X adores Y When X thinks of things X knows about someone or something Y, X feels something very good about those things like someone who thinks, “Y is better than anyone or anything else I am thinking of now” and like someone who wants to be with Y The experiencer of this feeling may be a person or an animal that is thought to be capable of these thoughts, feelings and desires. For example, the term describes the hero of the story, John Thornton’s dog, in The Call of the Wild:
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.31 (321)
The language of love in Melanesia
For the most part, however, Buck’s love was expressed in adoration. While he went wild with happiness when Thornton touched him or spoke to him, he did not seek these tokens. (The Call of the Wild by Jack London)
This kind of response to his master is an admiration and devotion that does not require its own pleasure. It is not liking or loving that last only if there are rewards from his master. Adoration is certainly other-centered, but it is content to praise and be in the presence of the other person, not necessarily thinking about doing good things for her/him although that would not be out of character for the concept. In Little Women, Jo has decided that she did not truly love her boy friend. She tries to explain to him, “You’ll get over this after a while, and find some lovely accomplished girl, who will adore you. . . ” She denies feelings of adoration on his part also, by enumerating the differences between them that would cause embarrassment and dislike. Adoration seems to imply liking, even exalting another’s qualities, thus the definition includes “X thinks of things X knows about someone or something” and “feels something very good about those things”. Adoration sees the other as “better than anyone or anything else I am thinking of now”. This sentiment is expressed by another of Alcott’s characters. “Daisy made a galley slave of herself, and adored her brother as the one perfect being in the world”. (Project Gutenberg’s E-text of Little Women by Louisa May Alcott) The following quote from USA TODAY illustrates the use of adore to refer to an abstract object. “Both of you, however, adore romantic fiction.” (Languages for romance readers, by Deirdre Donahue, USA TODAY, 12/02/99)
Alamblak (PNG), someone X yuhur1 someone or something Y someone feels something good about someone Y because X thinks something about Y is good This term refers to an appreciation of someone or a good quality of someone. A common virtue that Alamblak people like in other people is their generosity.
Ontong Java (S.I.), someone X ekauale’a2 someone or something Y someone feels something good toward someone or something else because he/she thinks that Y is good This Ontong Java expression has two senses according to Father Edward Kolahai and Colin Nauhi. The first sense is a good feeling about something that happens (Type 8). This second sense is a good feeling about some person or thing that is good.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.32 (322)
Les Bruce
Ifaluk, (Micronesia), someone X gachu1 someone Y Someone X feels something good about someone Y because X thinks something about Y is good Lutz describes this term as most similar to English like, most commonly used to “describe one’s feelings toward peers.” (Lutz 1982: 123)
. Feel good, think about, someone X has something Y The final type is a happy emotion based on a state of having something in one’s possession.
Alamblak (PNG), nanho wafmkuyet (lit. “It is a piece of my liver”) someone Y nanho wafmkuyet someone X has something or a child Y and feels something very good about Y when X thinks about Y X does not want to not have Y This is the feeling a person has about one’s special possession or of one’s child, with which one does not want to part. The idiom does not collocate with one’s spouse except in jest. Other idioms with similar meanings are Nanho marbit korhw6t (lit. “My heart it sits (= is)”) as in Nanho marbit korhw6t rediotn (lit. “My heart it sits (= is) in/on the radio”) Pukw6 (lit. “cuts”) as in Rediot pukw6tr (lit. “The radio cuts him”)
Ifaluk (Micronesia), someone X ker2 something Y Someone X feels something good because X has Y
. Discussion . Universals Universal semantic domains hypothesis: Positive emotions All languages have terms that can be included in the positive emotions domain.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.33 (323)
The language of love in Melanesia
This “happy/love-like” feelings (Wierzbicka 1999: 292) domain is characterized by the core component: “someone X feels something good about someone or something Y”. This component is linked with one or more components that characterize concepts in this domain, including “someone X thinks about Y”, and “someone X wants something”. If these conclusions are accurate, human conceptualizations of positive emotional experiences always involve feeling good about something. Regardless of actual experience, it appears that concepts do not always acknowledge the cognitive component. Positive emotions are either associated with thinking about something or with desiring something, or both. Universal semantic component hypothesis for positive emotions Four semantic components universally occur either collectively or severally in lexicalized concepts in every human language, viz.: “someone feels something good about someone or something else” “someone X thinks about Y” “someone wants to be with someone Y” “someone wants to do good things for someone Y” These four semantic components were a part of the explication of at least one term in each of the eight languages analyzed in this study.
. Relativity The universals mentioned above are at the domain level and semantic component level. This study corroborates the hypothesis of Wierzbicka (1992) that at the concept level there are no universal emotion concepts in human language. Those emotion concepts which are dignified with a lexical form may have exact equivalents in some languages, but not in all languages.
. Typology Common themes of concepts for positive emotions While no single positive emotion concept can be documented in all of the world’s languages, many concepts share a common theme. The following components represent common aspects of concepts in human languages. “something good happened / is happening” Eleven lexemes in four languages in this study are based on something good happening. They are listed together with three concepts from other, published studies in Type 8 and subtype 8.1. Of these lexemes, six lexemes from this study and two
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.34 (324)
Les Bruce
from other sources include the notion of wanting that good thing to happen (“want something to happen”). “want to be with” and “(want to) do good for” Eleven lexemes in seven languages in Types 2–4, and one in Type 9 combine “want to be with”, and either “want to do good for”, or “do good for”. “X wants to do something with Y for X’s good” Five lexemes in this study of Type 5 are blatantly ego-oriented. In these concepts ego is attracted to someone or something because of the good that ego receives from the object of his/her attraction. “something bad happened”, and either “feel good toward Y” or “(want to) do good for” Four lexemes in this study (Type 9) refer to emotional responses to something bad that has happened. The five examples from other studies in Type 9 all include the notion of feeling good toward the unfortunate party. Other, less frequently observed components of meaning included: “X thinks about Y ‘I am like a part of Y’ ” “X is with someone Y” “something is good” “X has Y”
. Extensional relationships of positive emotion concepts Some positive emotion concepts in this study are rooted in the related domains of Thinking, Feeling, and Wanting. The following display shows the extensions of meaning from these domains to meanings in the Positive Emotion domain. Polysemy within the Positive Emotion domain is not discussed here; those relationships are presented in the body of the text. The discussion there also points out the close affinity between the domains of wanting and positive emotions that some terms exhibit.
. The psychology of emotions The purpose of this study is to analyze the meanings of conceptualizations of positive emotions in natural language. It has not been our purpose to analyze emotional experiences per se. Johnson-Laird and Oatley (1989: 84) proposed a theory of emotions as simple signals that “serve a communicative function both within
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.35 (325)
The language of love in Melanesia
Table 2. Extensional relationships of positive emotion concepts Alamblak: dukay Maskelyne: l6]on; l6mas Uare: zamare FEEL l6]on1 ‘feeling something, either good or bad’
WANT THINK
l6]on2 ‘want to have something’
dukay1 ‘think’
l6mas1 ‘want something very much like
zamare1 ‘think’
people want something when they ask someone for something and do not stop asking when that person does not say yes’
FEEL GOOD, THINKING ABOUT l6]on3 ‘feel something good toward Y, thinking about Y, wanting to be with Y ’ l6mas2 ‘someone feels something good toward his or her child Y when X thinks about Y, X wants to be with Y and wants to do good things for Y’ dukay2 ‘someone feels good about someone or something Y like someone who thinks about something or someone else; because that person wants to do good things for Y and does not want anything bad to happen to Y’ zamare2 ‘someone X thinks that someone else Y is good; because of that X feels good about Y and wants to be with that person and wants to do for that person something that person wants’
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.36 (326)
Les Bruce
the brain and within the social group.” For them basic emotional modes such as one represented by the English term happiness should be reflected in every human language by a semantically unanalyzable linguistic expression. Prototypical cognitive scenarios for all emotion concepts The results of this study challenge the idea that a psychologically basic emotion is necessarily conceptualized by a semantically primitive notion. Mother tongue speakers seem to think of their emotional experiences, those that are lexicalized, in terms of prototypical cognitive scenarios. The explications of emotions in an NSM format utilize proposed semantically primitive concepts like, feel, good, think, want, do, someone and something. Johnson-Laird and Oatley make a valid point that emotions cannot be understood by definitions without also experiencing those emotions. It may only be possible to capture the specification of what the feeling-good experience is in a positive emotional concept in terms of a prototypical cognitive scenario. Most of the concepts in this study are explicated in that way, using the primitive notion of similarity (like). Concepts without a cognitive component Current psychological views of emotions indicate emotions involve “cognitive evaluations” (Johnson-Laird & Oatley 1989: 85) together with “feelings”, not necessarily equated with bodily sensations (Wierzbicka 1999: 2–3). For Power and Dalgleish (1997: 100) “The concept of emotion includes an event, an interpretation, an appraisal, physiological change, a propensity for action, and conscious awareness.” They define the widely held notion of basic emotions, by appeal to common or perhaps culturally universal appraisal scenarios that distinguish those emotions.6 Johnson-Laird and Oatley note, however, that people may experience emotions without being consciously aware of the cognitive role in a particular experience. Research reported on here indicates that emotions are commonly conceptualized with a component of thinking about something in the causing situation. It is not a necessary feature, however. Emotions may be conceptualized without including the component of thinking about an event or an object of attraction, whether or not the cognitive component is a part of one’s experience. Such was Zobule’s intuition about his own language expression uluakako (Type 3).
Notes . Wierzbicka describes the English word emotion as a complex concept incorporating “a reference to ‘feeling’, a reference to ‘thinking’, and a reference to a person’s body.” Her view reflects current perspectives that emotions are cognitive based while recognizing the biological component also. Her notion of “feeling” is more abstract than a physical feeling like “itching”. “Feeling” for Wierzbicka includes cognitive-related feeling like that involved
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.37 (327)
The language of love in Melanesia
in “loneliness”, pure physical feeling like pain, and feeling that involves both, like fear and compassion. . The languages of the study and contributors of analyses and definitions are listed in the Appendix. . I, you, someone/person, something/thing[/anything], people, body think, know, want, feel, see, hear say, word, true this, the same, other one, two, some, many/much, all do, happen, move there is, have live/alive, die not, maybe, if, can, because, if. . . would (counterfactual) when/time, now, after, before, a long time, a short time, for some time where/place, here, under, above, far, near, side, inside part of, kind of good, bad, big, small very, more like . The combinatorial potential was figured on 12 variables since one of the components is present in every case. The factorial formula is C = (N factorial / K factorial × (N–K) factorial) + 1 where N = the number of variables and K = the number of positions for each number of possible combinations (C). One is added for the combination comprising all 13 components. . – strong affection for a person or persons – strong liking for or interest in something (event) – strong affection for a person of the opposite sex – sexual passion – in theology: a. God’s benevolent concern for mankind b. man’s devout attachment to God c. feeling of benevolence and brotherhood that people should have for each other – v.t. feel love for – v.t. show love for by embracing, fondling, kissing, etc. – v.t. to delight in, take pleasure in an object or event – v.i. to feel the emotion of love; be in love. . This cognitive position is a minority position in contrast to most psychologists who distinguish basic emotions by universal or physiological components.
References Alcott, L. M. Little Women. Project Gutenberg E-text.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.38 (328)
Les Bruce
de Beaugrande, R.-A., & Dressler, W. U. (1981). Introduction to Text Linguistics. London: Longman. Bogusławski, A. (1970). On semantic primitives and meaningfulness. In A. J. Greimas, R. Jakobson, & M. R. Mayenova (Eds.), Signs Language and Culture (pp. 143–152). The Hague: Mouton. Briggs, J. L. (1970). Never in anger: portrait of an Eskimo family. Cambridge, MA: Harvard Univ. Press. Bruce, K. Alamblak verbs of emotion. Ms. Bruce, L. P. (1984). The Alamblak language of Papua New Guinea. Pacific Linguistics, C–81. Canberra: The Linguistic Circle of Canberra and The Australian National University. Cooper, J. F. The Last of the Mohicans. Project Gutenberg E-text. ——— The Pathfinder. Project Gutenberg E-text. Goddard, Cliff (1990). The lexical semantics of “good feelings” in Yankunytjatjara. Australian Journal of Linguistics [Special issue on the semantics of emotions], 10(2), 257–92. ——— (1996). The ‘social emotions’ of Malay (Bahasa Melayu). Ethos, 24(3), 426–64. ——— (1997). Contrastive semantics and cultural psychology: “Surprise” in Malay and English. Culture and Psychology, 3(2), 153–81. ——— (1998). Semantic Analysis. Oxford Textbooks in Linguistics. Oxford: Oxford University Press. Goddard, C., & Wierzbicka, A. (Eds.). (1994). Semantic and lexical universals: Theory and empirical findings. Amsterdam: John Benjamins Pub. Co. Grimes, B. (Ed.). (1996–1999). Ethnologue: Languages of the world (13th ed.). Dallas: Summer Institute of Linguistics. Healey, D. Maskelyne-English dictionary. Ms. Johnson-Laird, P. N., & Oatley, K. (1989). The language of emotions: An analysis of a semantic field. Cognition and Emotion, 3, 81–123. Leibniz, G. W. (1980) [1705]. New essays concerning human understanding (P. Remnant & J. Bennett, Trans.). Cambridge: Cambridge University Press. Lewis, M., & Haviland-Jones, J. M. (2000). Handbook of Emotions (2nd ed.). New York: The Guilford Press. London, J. The Call of the Wild. Project Gutenberg E-text. Louw, J. P., & Nida, E. A. (1988). Greek-English Lexicon of the New Testament based on Semantic Domains, Volume 1: Introduction and Domains. New York: United Bible Societies. Lutz, C. (1982). The domain of emotion words on Ifaluk. American Ethnologist, 9.1, 113–28. Power, M., & Dalgleish, T. (1997). Cognition and Emotion. East Sussex, UK: Psychology Press. Webster’s New World Collegiate Dictionary. (1964). New York: The World Publishing Co. Wierzbicka, A. (1992). Semantics, Culture, and Cognition: universal human concepts in culture-specific configurations. New York: Oxford University Press. ——— (1997). Understanding cultures through their key words, English, Russian, Polish, German, and Japanese. Oxford: Oxford University Press. ——— (1999). Emotions across languages and cultures: diversity and universals. Cambridge: Cambridge University Press.
TSL[v.20020404] Prn:4/11/2002; 9:47
F: PBNS35.tex / p.39 (329)
The language of love in Melanesia
Appendix: Languages and contributors
Language
Location
Classification
Analysts/source
Alamblak
Papua New Guinea Papua New Guinea Papua New Guinea Solomon Islands
Sepik-Ramu
Kathleen Bruce, Les Bruce
East-Papuan
Mark Sipaala
Sibe Uare/Kwale Lunqa/Lungga
Ontong Solomon Islands Java/Luaniua Maskelyne/Uliveo Vanuatu English
United States United Kingdom Australia No longer in use
Ancient Greek, Koine Hawaiian
United States
Ifaluk
Micronesia
Samoan
Samoa
Tahitian
Tahiti
Utku
North-central Canada Japan Russia Australia Australia
Japanese Russian Pintupi Yankunytjatjara
Trans-New Guinea Inara Gogina, Keiko Kikkawa (linguistic advisor) MalayoMarconi Podokolo, Rosella Papa, Polynesian Alphaeus Zobule, MA (linguistic advisor and mother tongue speaker of Lunqa) MalayoF. Edward Kolahai, Colin Nauhi Polynesian MalayoOtel Nagof, Edly Wakon, David Polynesian Healey (linguistic advisor) Indo-European Les Bruce Wierzbicka (’92, ’99) Indo-European
Les Bruce
MalayoPolynesian MalayoPolynesian MalayoPolynesian MalayoPolynesian Eskimo, Inuit
Wierzbicka (’92)
Wierzbicka (’92)
Japanese Indo-European Pama-Nyungan Pama-Nyungan
Wierzbicka (’92, ’99) Wierzbicka (’92) Wierzbicka (’92) Goddard (’90)
Lutz (’82) Wierzbicka (’92) Wierzbicka (’92) Wierzbicka (’92)
TSL[v.20020404] Prn:4/11/2002; 11:46
F: PBNS36.tex / p.1 (331)
Everyday rituals in Polish and English Ewa Jakubowska University of Silesia, Katowice, Poland
.
Introduction
Everyday interaction involves ritual, convention and routine (Coulmas 1981). The recurrence of certain communicative goals in interpersonal communication results in some communicative strategies being turned into “interaction rituals”, as Goffman (1967) calls them. He compares these “little ceremonies of everyday life” to religious rituals. Ritual is a kind of symbolic action. This is an act, either individual or collective, which always, even when involving a certain amount of improvisation, is performed in accordance with certain rules (Cazaneuve 1971; Rothenbuhler 1998; cf. Leach 1989). The Longman Dictionary of Contemporary English (1987: 902) defines ritual as “one or more ceremonies or customary acts which are often repeated in the same form: Christian ritual(s) (= the form of church services)”. But rituals are not restricted entirely to religious contexts. As was mentioned above, the recurrence of certain communicative goals in everyday life “has led to the evolution of standardised strategies for their accomplishment” (Coulmas 1981: 3; Malinowski 1923). They are used automatically, and their occurrence is highly predictable (Griffin & Mehan 1981). In some cases the strategies have turned into full rituals, in others the wording differs from one performance to another, but the sequential organisation is more or less constant (Coulmas 1981). The “ritual” character of exchange in social interaction has been noticed by many. Huxley (1966) saw the role of ritualization in improving the signal and therefore communication. Goffman (1967, 1971, 1981), Ferguson (1981), Laver (1981), and Brown and Levinson (1987) stressed the ritual character of politeness. Ritual prepatterned behaviour was investigated by Malinowski (1972, 1980). “Interaction rituals” (also called “interpersonal rituals” (Ferguson 1981) and “rituals of exchange” (Brown & Levinson 1987), they will here be called everyday
TSL[v.20020404] Prn:4/11/2002; 11:46
F: PBNS36.tex / p.2 (332)
Ewa Jakubowska
rituals) have a social function. They are acts “through whose symbolic component the actor shows how worthy he is of respect or how worthy he feels others are of it” (Goffman 1972: 328). In other words, the function of these rituals is to establish and/or maintain a state of “ritual equilibrium” in social interaction. The state of “ritual equilibrium” is necessary to sustain one’s face, “the positive social value a person effectively claims for himself ”, or “an image of self delineated in terms of approved social attributes” (Goffman 1967: 5). To secure their image, people engage in what Goffman calls “face-work”, performing action to make whatever they are doing consistent with their own face and with the face of their interlocutor. Our everyday behaviour is subject to ritual constraints which have to do with “how each individual ought to handle himself with respect to each of the others, so that he does not discredit his own tacit claim to good character or the tacit claim of the others that they are persons of social worth whose various forms of territoriality are to be respected” (Goffman 1976: 266). Thus, the performative nature of ritual implies two independent characteristics. First, ritual is always action according to pre-existing conceptions shared by the members of the scene in which a performance occurs. Second, ritual is always performance for someone, the interlocutor (Rothenbuhler 1998). In various social situations people behave in a conventionalised way by performing fossilized social rituals. To maintain a state of ritual equilibrium people have to address each other properly with respect to the context of the situation, the relationship between the interlocutors and their social status. Greetings and farewells are used as “access rituals” (Goffman 1971: 79). “Greetings mark the transition to a condition of increased access and farewells to a state of decreased access” (ibid.: 47). They have three main functions: attention-production, identification, and reduction of anxiety in social contacts (Firth 1972; Malinowski 1923; cf. Laver 1981). There are two kinds of ritual interchanges: “supportive rituals”, which are performed for the sake of mutual support (e.g. thanks, congratulations, condolences), and “remedial rituals”, performed when the Speaker (S) tries to remedy an offence he has committed and thus re-establish a state of ritual equilibrium (e.g. apologies) (Goffman 1971). It is often said that in some situations thanks and apologies are merely ritual, i.e. that S is simply doing what is expected of him (Fraser 1981; Aijmer 1996). Some of these rituals can be performed verbally or nonverbally, others only verbally with the use of certain routine formulae (called also polite formulae) (cf. O˙zóg 1990, 1997). Thus, to perform these rituals people use: – – –
terms of address formulae beginning a conversation – greetings formulae ending a conversation – farewells
TSL[v.20020404] Prn:4/11/2002; 11:46
F: PBNS36.tex / p.3 (333)
Everyday rituals in Polish and English
– – –
formulae expressing gratitude – thanks formulae expressing apology other polite formulae (e.g. compliments, congratulations, good wishes, toasts, and condolences).
. Aims of the study Everyday rituals, which are the subject of my research, are conventional acts of a communicative but arbitrary kind, performed verbally, having the effect of establishing and/or maintaining ritual equilibrium in social interaction. Their use is traditionally motivated by politeness. Everyday rituals encode cultural beliefs and reflect community social organisation, and as such are language- and culture-specific. The aims of my study are: – –
to analyse everyday rituals as a complex cognitive category to provide a pragmatic contrastive analysis of greetings and farewells in Polish and English, stressing the most striking differences and similarities between the two languages.
The data for the study was obtained from different sources: participant observation, introspection, and elicitation tests. The informants, aged 20–50, were mainly of middle class or academic background.
. Methodology . A cognitive approach to everyday rituals Everyday rituals, and politeness in general, are not commonly analysed in the framework of cognitive linguistics. However, this is the model of analysis I have chosen for my study of such rituals. Greetings and farewells, like other speech acts, are fuzzy categories based on ´ 1993). prototypical properties (Kalisz & Kubinski The cognitive network model (Langacker 1987) can be a useful tool in analysing greetings, farewells and other everyday rituals. The use of this model can help explain the fuzziness of these categories and show the role of context in eval´ uating them (cf. Lakoff 1987; Kalisz & Kubinski 1993). This is particularly vital because the appropriateness of social behaviour can be evaluated only with respect to the context of the situation.
TSL[v.20020404] Prn:4/11/2002; 11:46
F: PBNS36.tex / p.4 (334)
Ewa Jakubowska
Non-prototypical members of the two categories can be represented as extensions of the schema by virtue of a difference in behaviour with respect to the cluster of properties of a given category (see examples (1) and (2)). (1) Thanks for the book.
(non-prototypical greetings)
(2) Don’t forget about our meeting on Tuesday.
(non-prototypical farewells)
All speech acts can be either direct or indirect. Directness and indirectness should not be treated as clear-cut categories, but rather as a spectrum from the most literal expressions to expressions requiring some contextual information. Indirectness is handled in cognitive linguistics as metaphorization, mapping one cognitive ´ 1993). domain onto another (Kalisz & Kubinski The role of ritual constraints is to help us to secure our self-image (face) and the self-image of our interlocutor. Certain kinds of actions are intrinsically facethreatening. Such actions can threaten the Hearer’s (H) negative face, i.e. indicate that S wants to impede H’s freedom of action (e.g. requests, suggestions). Other kinds of actions are acts threatening H’s positive face wants, e.g. acts that indicate that S does not care about H’s feelings and wants, or does not approve of some aspects of H’s positive face (e.g. expressions of disapproval, mention of taboo topics, use of address terms and other status-marked indications in initial encounters in an offensive or embarrassing way). All these acts are called face-threatening acts. They can threaten both H’s face, as in the case of requests, and S’s face, as in the case of promises, thanks, and excuses (Brown & Levinson 1987; cf. Dirven & Verspoor 1998). Some acts are more face-threatening than others. The value of the face threat can be measured on a gradient scale.
. Contrastive tools The pragmatic contrastive analysis mentioned above will be conducted in terms of prototype theory and gestalts. Prototypes are said to be culture specific. But the prototype analysis can be also a good contrastive tool. Such contrastive analysis should include (Kalisz 1988): – – – – –
establishing a prototype and family resembling forms in L1 analysing corresponding L2 forms on the basis of family resemblance with respect to a prototype established for L1 establishing a prototype and family resembling forms in L2 analysing corresponding L1 forms on the basis of family resemblance with respect to a prototype established for L2 showing properties shared by the two prototypes.
TSL[v.20020404] Prn:4/11/2002; 11:46
F: PBNS36.tex / p.5 (335)
Everyday rituals in Polish and English
If two linguistic forms in two different languages belong to one category within the domain of contrastive studies, the properties crucial in determining category membership are of unequal status as category determinants (Krzeszowski 1990). The properties which are relevant in determining pragmatic equivalence constitute a gestalt (Lakoff 1977). The three important parameters are (Krzeszowski 1990): – – –
type of speech act modality anticipated perlocutionary effects.
In determining pragmatic equivalence one also cannot forget about the extralinguistic context in which a given speech act is performed.
. A cognitive analysis of greeting and farewell rituals Everyday rituals constitute a complex cognitive category. It has the structure of a schematic network, an assembly of overlapping units. The unit which is naturally most salient, most often thought of, and most likely to be chosen as representative of the category is called the prototype. In the case of greetings and farewells the prototypes are the greeting and farewell formulae most often resorted to in everyday conversations (see Tables 1 and 2). The structure of the network is given by two categorising relationships between nodes: specialisation (A———> B) and extension (A– – – > B) (Langacker 1987). Thus, besides prototypical expressions, there are other unproblematic instantiations of the category defined by the schemas [GREETING] and [FAREWELL], respectively. These are formulae which are less representative of the respective catTable 1. GREETINGS Prototypical expressions: – nominal formulae (indirect)
Table 2. FAREWELLS Prototypical expressions: – nominal formulae (indirect) – next-meeting promises (indirect)
Specialisation: 1. direct greetings 2. indirect greetings – phatic exclamations – how-are-you-type questions – miscellaneous phatic expressions
Specialisation: 1. direct farewells 2. indirect farewells – farewell exclamations – closing expressions – invitation-like formulae
Extension: – thanks – compliments – congratulations
Extension: – apologies – good wishes – compliments
TSL[v.20020404] Prn:4/11/2002; 11:46
F: PBNS36.tex / p.6 (336)
Ewa Jakubowska
Table 3. The category of EVERYDAY RITUALS described as a schematic network
egory and less frequently used than the prototypical expressions, because of contextual restrictions. Other category units, like thanks, compliments and congratulations, in the case of greetings, are assimilated to this category only through extension. These formulae can be used as greetings only in some specific contexts.
TSL[v.20020404] Prn:4/11/2002; 11:46
F: PBNS36.tex / p.7 (337)
Everyday rituals in Polish and English
The same can be said about apologies, good wishes and compliments, in the case of farewells. The relationship between categorising units within one category can be very complex. A schematic network, even though it resembles standard taxonomic hierarchies, can serve three important functions: categorisation, the capture of generalisations (expressed by schemas), and the sanction of novel structures (Langacker 1987). The category of everyday rituals also can be presented as a schematic network (see Table 3).
. A pragmatic contrastive analysis of greeting and farewell rituals in Polish and English As was mentioned above, a pragmatic contrastive analysis of greeting and farewell rituals will be conducted in terms of prototype theory and gestalts.
. A contrastive analysis of greetings The first step of the prototype analysis, as proposed by Kalisz (1988), is to establish a prototype and family resembling formulae in Polish (3). The second step is to analyse corresponding English forms on the basis of family resemblance with respect to the prototype established for Polish (4). (3) Polish greetings a. Prototypical expressions Dzie´n dobry (“Good day”) Dobry wieczór (“Good evening”) b. Specialisation Witam Pana bardzo serdecznie (“I greet (you) Sir very heartily.”) Witaj stary! (inf.) (“Greet (you) old chap!”) Cze´s´c! (inf.) (“(I regard you with) reverence.” – now semantically empty) Serwus! (inf.) (“(Your) slave!” (Lat) – now semantically empty)
(4) English greetings Good morning, Good afternoon Good evening Welcome to our hotel!
Hi! (inf.)
TSL[v.20020404] Prn:4/11/2002; 11:46
F: PBNS36.tex / p.8 (338)
Ewa Jakubowska
Jak si˛e masz? (“How are you?”) Moje uszanowanie (form.) (“My respect (towards you)”) Kop˛e lat! (inf.) (“Many years!”) c. Extension Dzi˛ekuj˛e za kartk˛e z Rzymu. (“Thank you for the postcard from Rome.”) Bombowo wygladasz! ˛ (“You look great?”) Gratuluj˛e udanego wystapienia! ˛ (“I congratulate you on your good presentation!”)
How are you?
Long time no see! (inf.)
Thanks for the book!
Great! You look terrific! Congratulations!
Next a prototype and family resembling forms are established in English, and the corresponding forms in Polish are analysed on the basis of family resemblance with respect to the prototype established for English (see (5) and (6)). (5) English greetings a. Prototypical expressions Good morning Good afternoon Good evening b. Specialisation Welcome to our new home! Hi! (inf.) Hello! How are you? Nice to see you! Lovely day, isn’t it? Long time no see! (inf.) Claudia! Just the person I wanted to see. (inf.) c. Extension Thanks for the note! You look gorgeous!
(6) Polish greetings Dzie´n dobry Dzie´n dobry Dobry wieczór Witamy szanownych go´sci! (“Welcome (our) honourable guests!”) Cze´s´c! (inf.) Jak si˛e masz? Co za dzie´n! (“What a day!”) Kop˛e lat! (inf.) Wła´snie o tobie my´slałam. (“I was just thinking of you.”) Dzi˛ekuj˛e za wiadomo´s´c. Wygladasz ˛ wspaniale!
TSL[v.20020404] Prn:4/11/2002; 11:46
F: PBNS36.tex / p.9 (339)
Everyday rituals in Polish and English
Well done, Mary! (inf.)
Dobra robota! (inf.) (“Good job!”)
Both Polish and English prototypical greetings are nominal formulae, in origin “welfare wishes”, now mainly treated as “affirmations” (Ferguson 1981). Although they are fully equivalent formally and semantically, they are used in different temporal contexts (see example (7)). (7) Polish English Dzie´n dobry (during the day) Good morning (before noon) Good afternoon (after 12 noon or after lunch to the end of the working day) Dobry wieczór (after dark) Good evening (after work or from 6 pm onwards)
. A contrastive analysis of farewells A contrastive analysis of farewells can be carried out in the same way. It will not be presented step by step, but only its final results will be discussed (see examples (8) and (9)). (8) Polish farewells a. Prototypical expressions Do widzenia/zobaczenia (“(Goodbye) till (we) see (each other again)”) Dobranoc (“Good night”) b. Specialisation ˙ Zegnaj! (form.) (“I bid you goodbye.”) Cze´s´c! (inf.) Pa-pa! (inf.) (“Bye-bye!”) No, na mnie ju˙z czas/pora. (“Well, it’s time for me to leave.”) Musz˛e lecie´c. (inf.) (“I must fly.”) No to b˛edziemy w kontakcie. (“Well, we’ll be in contact.”)
(9) English farewells a. Prototypical expressions Goodbye See you (later/soon/tomorrow) Good night b. Specialisation Farewell! (form. or old)
Bye-bye! (inf.) Well, I’d better be going. (inf.)
I must rush/go. (inf.) I’ll be in touch.
TSL[v.20020404] Prn:4/11/2002; 11:46
F: PBNS36.tex / p.10 (340)
Ewa Jakubowska
c. Extension Przepraszam, ale musz˛e lecie´c. (inf.) (“Sorry, but I must fly.”) ˙ e) miłego dnia! (Zycz˛ (“(I wish you a) good day.”) Trzymaj si˛e! (inf.) (“Keep (in good health)!”) Uwa˙zaj (na siebie)! (inf.) (“Look after (yourself)!”) Szcz˛e´sliwej podró˙zy! (“(I wish you a) happy journey!”) Miło si˛e z Panem rozmawiało, ale. . . (“It’s been nice talking to you Sir, but. . . ”)
c. Extension Sorry, I’ve got to go now. (inf.)
(Have a) good day! Take care! (inf.) Look after yourself! (inf.) Have a good journey/trip!
It’s been nice talking to you, but. . .
Both Polish and English prototypical farewells are nominal formulae, except for the Polish formula Do widzenia, which has the form of a prepositional phrase and is closest in meaning to the English formula See you. The other three formulae, like Polish and English prototypical greetings, are in origin “welfare wishes”, now treated as “affirmations”. Some farewell formulae, especially those having the form of a good wish, even though classified here as extensions of the schema, are very frequently used and the frequency of their occurrence is comparable to that of the prototypical farewells.
. The gestalt analysis A complementary tool helping to determine pragmatic equivalence is the gestalt analysis. Only a sample of such an analysis will be presented here. Consider the following example: (10) Cze´s´c! (inf.) (as a greeting) (“(I regard you with) reverence.”) (11) Hi! (inf.) (12) Hello! In the case of example (10) the most important parameters constituting the relevant gestalt are:
TSL[v.20020404] Prn:4/11/2002; 11:46
F: PBNS36.tex / p.11 (341)
Everyday rituals in Polish and English
semantically empty, polite greeting, only with a phatic function, used in informal situations, by young people or friends. Example (11) is the most accurate equivalent of example (10), in semantic and pragmatic aspects. So, it can be said that example (11) is the most prototypical equivalent of example (10), while example (12) is less prototypical, because the formula Hello! is not limited to the use “in informal situations, by young people and friends”. The gestalt analysis, taking into consideration all important properties more privileged than others in assigning two formulae in the two languages to one category within the domain of contrastive studies, helps the researcher to investigate in greater detail the objects to be compared, and thus makes it easier to determine their pragmatic equivalence.
. Conclusions The aims of the study were: – –
to analyse everyday rituals as a complex cognitive category to provide a pragmatic contrastive analysis of greeting and farewell rituals in Polish and English, conducted in terms of prototype theory and gestalts.
The cognitive analysis of the category of everyday rituals conducted in terms of the cognitive network model provided an exhaustive description of the relationships between categorising units within the category. The contrastive tools, complementing each other, allowed the researcher to carry out an adequate pragmatic contrastive analysis of greetings and farewells in the two languages. On the basis of the contrastive analysis the following typology of greeting and farewell formulae can be proposed: a. formulae which are fully equivalent b. formulae which are partially equivalent: – formally different – semantically different – used in different contexts.
TSL[v.20020404] Prn:4/11/2002; 11:46
F: PBNS36.tex / p.12 (342)
Ewa Jakubowska
Abbreviations form. — formal H — hearer inf. — informal L1 — first language L2 — second language Lat — Latin S — speaker
References Aijmer, K. (1996). Conversational Routines in English. London, New York: Longman. Brown, P., & Levinson, S. C. (1987). Politeness. Some Universals in Language Usage. Cambridge: Cambridge University Press. Cazaneuve, J. (1971). Sociologie du Rite (Tabou, Magie, Sacré). Paris. Coulmas, F. (1981). Introduction: Conversational routine. In F. Coulmas (Ed.), Conversational Routine (pp. 1–17). The Hague: Mouton. Coulmas, F. (Ed.). (1981). Conversational Routine. The Hague: Mouton. Dirven, R., & Verspoor, M. (1998). Cognitive Exploration of Language and Linguistics. Amsterdam/Philadelphia: John Benjamins Publishing Company. Ferguson, Ch. A. (1981). The structure and use of politeness formulas. In F. Coulmas (Ed.), Conversational Routine (pp. 21–35). The Hague: Mouton. Firth, J. R. (1972). Verbal and bodily rituals of greeting and parting. In J. S. La Fontaine (Ed.), The Interpretation of Ritual. Essays in Honour of A. J. Richards (pp. 1–38). London: Tavistock. Fraser, B. (1981). On apologizing. In F. Coulmas (Ed.), Conversational Routine (pp. 259– 271). The Hague: Mouton. Goffman, E. (1967). Interaction Ritual: Essays in Face-to-Face Behavior. New York: Doubleday, Garden City. Goffman, E. (1971). Relations in Public. New York: Harper. Goffman, E. (1972). On face-work: An analysis of ritual elements in social interaction. In J. Laver, & S. Hutcheson (Eds.), Communication in Face to Face Interaction (pp. 319– 346). Harmondsworth: Penguin Books Ltd. Goffman, E. (1976). Responses and replies. Language in Society, 5(3), 257–313. Goffman, E. (1981). Forms of Talk. Philadelphia: University of Pennsylvania Press. Górska, E. (Ed.). (1993). Images from the Cognitive Scene. Kraków: Universitas. Griffin, P., & Mehan, H. (1981). Sense and ritual in classroom discourse. In F. Coulmas (Ed.), Conversational Routine (pp. 189–213). The Hague: Mouton. Huxley, J. S. (Ed.). (1966). Ritualization of Behaviour in Animals and Men. [Philosophical Transactions of The Royal Society of London, Series B, 251 (772)] (pp. 247–526). Kalisz, R. (1988). The problems of directionality in contrastive studies based on cognitive linguistics. Papers and Studies in Contrastive Linguistics, 23, 37–45.
TSL[v.20020404] Prn:4/11/2002; 11:46
F: PBNS36.tex / p.13 (343)
Everyday rituals in Polish and English
´ Kalisz, R., & Kubinski, W. (1993). Speech act as a radial category. In E. Górska (Ed.), Images from the Cognitive Scene (pp. 73–85). Kraków: Universitas. Krzeszowski, T. (1990). Contrasting Languages. Berlin, New York: Mouton de Gruyter. La Fontaine, J. S. (Ed.). (1972). The Interpretation of Ritual. Essays in Honour of A. J. Richards. London: Tavistock. Lakoff, G. (1977). Linguistic gestalts. CLS, 13, 236–287. Lakoff, G. (1987). Women, Fire and Dangerous Things: What Categories Reveal About the Mind. Chicago: Chicago University Press. Langacker, R. W. (1987). Foundations of Cognitive Grammar. Vol. 1. Theoretical Prerequisites. Stanford: Stanford University Press. Laver, J. (1981). Linguistic routines and politeness in greeting and parting. In F. Coulmas (Ed.), Conversational Routine (pp. 289–304). The Hague: Mouton. Laver, J., & Hutcheson, S. (Eds.). (1972). Communication in Face to Face Interaction. Harmondsworth: Penguin Books Ltd. Leach, E. (1989). Kultura i komunikowanie. In E. Leach & A. J. Greimas, Rytuał i narracja ´ (pp. 5–98). Warszawa: Panstwowe Wydawnictwo Naukowe. Leach, E., & Greimas, A. J. (1989). Rytuał i narracja. (E. Leach, Culture and Communication, and A. J. Greimas, La Mythologie, and Pour une L’Interprétation du Récit Mythic ´ ´ (M. Buchowski, A. Grzegorczyk, E. Uminska-Plisenko, Trans.)) Warszawa: Panstwowe Wydawnictwo Naukowe. Longman Dictionary of Contemporary English. (1987). Harlow: Longman. Malinowski, B. (1923). The problem of meaning in primitive languages. Reprinted in C. K. Ogden & I. A. Richards (1949) The Meaning of Meaning (pp. 296–336). London: Kegan Paul. Malinowski, B. (1972). Phatic communion. In J. Laver & S. Hutcheson (Eds.), Communication in Face to Face Interaction (pp. 146–152). Harmondsworth: Penguin Books Ltd. Malinowski, B. (1980). Wierzenia Pierwotne i Formy Ustroju Społecznego. Dzieła, Vol. 1. Warszawa. O˙zóg, K. (1990). Zwroty grzeczno´sciowe współczesnej polszczyzny mówionej. Zeszyty Naukowe Uniwersytetu Jagiello´nskiego CMXIII, 98. Kraków. O˙zóg, K. (1997). J˛ezykowe rytualizmy w interakcji. Socjolingwistyka, 15, 63–70. Rothenbuhler, E. W. (1998). Ritual Communication. London, New Delhi: Sage Publications, Thousand Oaks.
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.1 (345)
A question of time? Question types and speech act shifts from a historical-contrastive perspective Some examples from Old Spanish and Middle English Verena Jung and Angela Schrott University of Leeds, UK / Ruhr-University, Bochum, Germany
Introduction Our paper on question types and speech act shifts has a twofold aim. Firstly, we want to combine the perspective of contrastive pragmatics with a historical dimension in order to analyse the speech act of asking a question and speech act shifts from a cross-linguistic historical perspective. The premise is that speech acts change and that speech acts described in texts are therefore understood differently at different times. Secondly, in order to document illocutionary changes and shifts we make use of translation studies and determine possible changes via a comparison between the original text and its translations. Our study has the following structure. After introducing the concept of historical pragmatics and discussing some crossroads between historical and contrastive pragmatics (1) we link both disciplines with the field of translation studies (2). Based on this theoretical framework we demonstrate the relevance of historical pragmatics for translation practice by looking at two examples of historical texts – one from Old Spanish and one from Middle English – and their various modern translations. In the Old Spanish Cantar de mio Cid (3) we focus on question types and ask whether the medieval text presents question types which are not asked any longer. In our second text, Chaucer’s Wife of Bath’s Prologue and Tale (4), we analyse two examples of speech act shift portrayed in the various translations of this historical text and discuss possible motivations for the illocutionary shift in two of the translations given.
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.2 (346)
Verena Jung and Angela Schrott
.
Contrastive pragmatics and historical pragmatics: Crossroads in the field of pragmatics
. Historical pragmatics As the title of our paper suggests, we want to introduce a historical dimension into the contrastive study of verbal interaction. Before we turn to the interrelations of contrastive and historical pragmatics we want to sketch our understanding of historical pragmatics. As pragmatics is concerned with the study of language use as social interaction it is a genuinely historical discipline in the sense that it always considers the sociohistorical situation in which the interaction takes place (Cherubim 1980: 4, 6f., 7– 9, 19). This historical dimension of the pragmatic approach becomes particularly clear when we study verbal interactions in speaker communities of the past. It is this branch of pragmatics that focuses on speaker interactions in the past that is usually referred to by the notion of historical pragmatics (Jucker & Jacobs 1995: 4–5). The central aim of historical pragmatics is to describe verbal interactions as historically determined ways of interacting and to study linguistic means and their use across different historical stages of a language (Jucker & Jacobs 1995: 4–6, 11–14, 19–21; Fritz 1995: 469, 1997: 47f.). As a subdiscipline of historical pragmatics, historical dialogue analysis considers dialogues as social behaviour regulated by language structures and historical contexts of acting and communicating (Cherubim 1980: 7–9). Historical dialogue analysis interprets dialogue forms as tools for solving communicative tasks and describes the ways in which elements of dialogue interaction are subject to historical changes.1 Very prominent elements of dialogue interaction are of course speech acts and speech act sequences. Recent studies in historical dialogue analysis (cf. the articles in Jucker, Fritz & Lebsanft 1999) show that changes affect dialogues in many ways: changes include not only maxims of interaction and politeness rules (Ehlich 1992) but also speech act patterns and speech acts themselves. In the perspective of historical pragmatics speech acts are seen as historically determined types of verbal interaction whose realisation is influenced by a number of social parameters and cultural traditions (Schlieben-Lange & Weydt 1979: 66– 67; Schlieben-Lange 1983: 13; Arnovick 2000: 2, 39, 142f.). Changes can concern the speech act itself, the rules for its performance and the illocutionary patterns in which it is embedded. As types of verbal interaction speech acts are texts, and as texts they can form traditions of their own which are handed down from one speaker-generation to the next. If we consider speech acts as texts that have their own traditions, it follows that speech acts like asking a question or making a complaint do not belong to the history of a specific language but to the history of a cultural community that uses this speech act as a communicative tool (Coseriu in Schlieben-Lange & Weydt 1979: 74–76, 77). Speech acts therefore cannot simply be
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.3 (347)
Question types and speech act shifts
linked to the history of a language but are part of the cultural history of a specific community. These cultural traditions have to be taken into consideration when we want to understand dialogue interactions in medieval texts and render them adequately when translating. Cultural traditions can concern politeness rules and patterns of interaction as well as the roles a cultural community attributes e.g. to men and women. In most cases, changes in speech acts and dialogue patterns are not limited to the linguistic level but go hand in hand with changes in the realia of the extra-linguistic world (Lebsanft 1999: 270f., 284f.).
. Historical and contrastive pragmatics After this brief sketch of some implications of historical pragmatics, we now turn to the question of how historical pragmatics is related to contrastive pragmatics. The common ground of both disciplines has been pointed out by various studies in historical pragmatics (e.g. Jucker & Jacobs 1995: 3f.). Both disciplines compare the realisation of components of interaction in different cultures – be it cultures that follow each other on the time line or cultures that exist at the same time in different environments. Like historical pragmatics, contrastive pragmatics understands verbal interaction as social behaviour shaped by cultural and sociohistorical parameters and focuses on the differences and the variety of elements and patterns of interaction. Contrastive pragmatics aims at describing and explaining the variety of interactional styles found in different cultures and wants to interrelate the performance of speech acts with the cultural and sociohistorical parameters that may affect their performance (Blum-Kulka & House 1989: 5). Here, differences in the performance of speech acts lie not only in the linguistic expressions but also in a different system of cultural values (Wierzbicka 1991: 25ff., 40, 44, 50ff., 59f., 61). There are not only institutionalised speech acts like baptizing that are culturebound but all speech acts are culture specific (Wierzbicka 1991: 149f., 156). Therefore, we should not ask how members of different cultures perform e.g. apologies but how they behave e.g. in a situation when they have done wrong to a person and want to remedy the situation. Thus, both disciplines are concerned with the realisation of elements of verbal interaction in different languages and cultures and both disciplines concentrate on comparing the ways in which a specific speech act is realised in different communicative and cultural contexts (Jucker & Jacobs 1995: 3f.; Jucker & Taavitsainen 2000: 68). But let us go back to the interrelationship of both disciplines. How ‘contrastive’ is historical pragmatics and how ‘historical’ is contrastive pragmatics? The contrastive nature of historical pragmatics becomes evident in the blueprint for the domain of historical dialogue analysis given by Fritz (1995: 469–470, 1997: 47f.). Fritz distinguishes three phases of historical dialogue analysis: the interpretation of individual texts, the contrastive study of historical
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.4 (348)
Verena Jung and Angela Schrott
texts and, finally, the systematic analysis of the evolution of dialogue forms. Yet the study of a medieval text is in itself already contrastive as we implicitly compare the text to the texts of our own discourse universe and use our own communicative experience to understand the text. This contrastive view is always inherent in historical pragmatics, even if we concentrate on a single medieval text (Fritz 1995: 469; Lebsanft 1999: 269f.). However, whereas historical pragmatics always offers a contrastive perspective, contrastive pragmatics does not necessarily embrace a historical dimension: contrastive pragmatics compares a certain speech act in different speaker communities in a synchronic cross-section and does not aim at sketching the diachronic evolution of the speech act (Jucker & Taavitsainen 2000: 68). Nevertheless, contrastive pragmatics may have a diachronic dimension, e.g. when we realise that a speech act is differently performed in a specific speaker community and we want to know the historical reason for these differences (cf. Stein 1985). The conceptual affinity implies that both disciplines have to face similar problems concerning the variability and historicity of speech acts. In order to describe the verbal realisations of a speech act, both disciplines take this speech act as an invariable tertium comparationis (Jucker & Taavitsainen 2000: 69) for the description of differences in speech act performance, be it synchronically or diachronically. Recently, linguists in historical pragmatics as well as in contrastive pragmatics have become aware of the fact that we cannot presuppose the invariability of a certain speech act. Not every culture has the same set of speech acts. A speech act like apologizing does not necessarily have an identical function in different cultures: the difference may not only concern the verbal realisation but also the function speakers assign to a certain action (Wolfson 1989: 179). Another problem both disciplines share is whether changes merely concern the linguistic representation of a speech act or whether the speech act itself changes (Lebsanft 1999: 270f., 284f.) – a vital distinction which we will have to consider in our comparison of the medieval texts with their modern translations.
. Translation studies and pragmatics After giving an outline of the relations between contrastive pragmatics and historical pragmatics, we want to introduce the ways in which both fields and our text analysis are to be linked with translation studies. Before we take a closer look at the historical perspective, some remarks on the relationship between translation theory and pragmatics in general. While linguistics has come to be seen as an important tool both in the description of the two languages a translator deals with and in the description of the translation process (Catford 1965; Albrecht 1973; Diller & Kornelius 1978),
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.5 (349)
Question types and speech act shifts
the newer disciplines of linguistics such as pragmatics have entered translation studies much more slowly. Studies in contrastive linguistics tended to be more descriptions of the linguistic system than of language in action (Hawkins 1986) and translation methodology books also tended to concentrate more on the system side than on pragmatic considerations. While text typology and the influential skopos theory (Reiss 1982; Reiss & Vermeer 1984: 214) became increasingly a factor seen as determining different translation strategies for different text types, the level of translational strategy of the individual sentence was interpreted to be the semantic rather than the pragmatic level of the individual phrase. However, in the so-called functionalist approach, represented by Hönig and Kussmaul (1982), Wilss (1982, 1996) and Nord (1997), pragmatics becomes relevant not only in terms of the determination of the text-type but also for the translation of the individual element. Hönig and Kussmaul have integrated speech act theory into translation theory and teaching methodology in depth and explain the relevance of seeing utterances as speech acts in their 1982 didactic handbook on translation. Kussmaul (1980) also has translated and edited some of the major writings on speech act theory. The functionalist approach postulates that the illocution itself remains unchanged, but this often means a rather substantial change concerning the choice of linguistic means. Illocutionary invariance often implies that syntactic structures and sentence types cannot remain unchanged and have to undergo variance. As shown by House (1997: 82f.) German and English differences in formulating apologies or offers prove that the linguistic realisation of the basic forms of speech acts can vary considerably from one speaker community to the other. Thus, the polite offer “Why don’t you take your coat off?” of an English host might be interpreted as an almost philosophical or personal question by a German guest, and might not produce the expected “Thank you” and a handing over of the coat but rather a reaction like “Well, I usually wear my coat indoors” or “Why do you want to know?”. The German structural equivalent “Warum ziehst du deinen (ziehen Sie Ihren) Mantel nicht aus?” – as this English phrase is often rendered in translations of American or English TV series – creates a different and very strange effect in the German version. Wierzbicka (1991: 212) analyses this form as a “whimperative construction” that should be taught as such to learners of the English language (1991: 212, 216, 218).2 Because of the disastrous consequences of ignoring the pragmatic level some translation methodologists and practitioners, such as Hönig and Kussmaul (1982) and also functionalist Christiane Nord (1997) advocate a top-to-bottom process of translation, i.e. the pragmatic level must be seen as the first level from which all other translation decisions must be derived. Thus, the identification of the speech act would precede vocabulary questions and decisions. Like Hönig and Kussmaul (1982), Christiane Nord (1997) gives pragmatics top priority in the process of translation:
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.6 (350)
Verena Jung and Angela Schrott
In functional translation, problems should therefore be dealt with in a topdown way. This means that a functional translation process should start on the pragmatic level by deciding on the intended function of the translation (documentary vs. instrumental). A distinction is then made between those functional elements of the source text that will have to be reproduced “as such” and the ones that must be adapted to the addressee’s background knowledge, expectations and communicative needs [. . . ]. (1997: 68)
In a recent article, Hönig (1995) develops the idea that conference interpreters need to be aware of the speech act intended by the speaker in order to be able to translate the illocution correctly. In his article he analyses examples of political criticism formulated as questions in Prime Minister’s Question Time at the House of Commons. Thus, when the opposition party in a discussion on police action against domestic violence asks the question “Why is it that with a problem of such importance and when so many women throughout the country are at risk no action has been taken since January?” the illocution is not that the opposition party is asking to be supplied with information. The interrogative structure here functions as a rhetorical question with assertive value and expresses the conviction that there is no valuable reason for not taking action against that dangerous situation. As a consequence, the German rendering of the interpreter, whether in question form or not, must render this as a critical statement and not as a mere thirst for more information, otherwise the communicative intention has not been rendered despite achieving “invariance on the content level” (Reiss 1982: 20). The functionalist approach also demands that the translator bears in mind the text function as a whole, i.e. the text type and the traditions that mark the text. This can best be explained by using Reiss and Vermeer’s (1984) text typology to describe how genre considerations influence the translator’s decision. Reiss and Vermeer postulate that advertising texts, being examples of operative texts (Reiss & Vermeer 1984: 214), demand extreme faithfulness on the message level, i.e. what is the aim of the text, but are not particularly concerned with rendering invariance on the semantic content level, while literary texts and technical texts tend to demand more invariance on the semantic level. When dealing with literary texts, the translator has to consider especially the poetic traditions that mark the text on a formal and functional level. Now let us return from those more general remarks on the pragmatic perspective in translation to our idea of using translation studies in the field of historical pragmatics. One link is already hinted at by Jucker and Jacobs (1995). Among the branches of historical pragmatics Jucker and Jacobs (1995: 11) count the approach of “pragmaphilology”. Pragmaphilology describes the contextual aspects of historical texts, the conditions of their production and reception. If we want to know how historical texts were received through time, translations are important witnesses
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.7 (351)
Question types and speech act shifts
for the reception of a text and the way it was understood or misunderstood. The way literary texts like the Cantar de mio Cid and the Wife of Bath’s Prologue and Tale have been received through time can be described via an inventory of their translations. In our paper, however, we do not embrace a complex pragmaphilological approach that would seek to write the history of a literary text reflected in its translations. Although our translations were chosen in order to compare different ‘styles’ of translation we cannot pretend to give a well-balanced and representative choice of translation styles or schools of translation. The primary concern of our paper is not how the translation generally reflects or modifies the literary text but how it renders certain types of illocutions we find described in the medieval text. To some extent, our approach could be called pragmaphilological on a speech act level in the sense that we study the way certain speech acts are received and rendered in the translations. Translation comparison is one of the most frequent exercises used in translation studies. It can be used for various purposes, didactic – distinguishing between different types of translation – or as a quality assessment exercise. In order to be used as a quality assessment, criteria must first be established and it must be explained in what way they are relevant to the making of a good translation (House 1997: 9ff.). But any translation comparison, whether for prescriptive or descriptive purposes, should concentrate on predefined elements in order to be conclusive (Frank 1988), just as any linguistic study would. In this study, the speech act level will be the level of comparison. For our analyses we have chosen an Old Spanish text, the Cantar de mio Cid from the 12th century and a Middle English text from the 15th century, the Wife of Bath’s Prologue and Tale from Chaucer’s Canterbury Tales. The temporal distance implies that the texts are deeply marked by what Jauß (1977: 10–16) called the ‘alterity’ of the medieval text – a strangeness of the text that demands special efforts of understanding on the part of the reader. This strangeness characterises the text not only on the level of literary aesthetics but also on the level of the interactions we find depicted in the text: those interactions are mimetic representations of the way people acted or were supposed to act many centuries ago (Cerquiglini 1981: 247). As speech acts and patterns of interaction are exposed to historical changes, it is possible that a reader or translator is confronted with an illocutionary type or pattern that is no longer part of his culture. Here, the modern reader or translator has to reconstruct the illocution, making use of a hermeneutic circle and referring to the historical background as well as to the literary traditions of the text. When translators have grasped the pragmatic profiles of the interactions they render them with the linguistic means of modern English, Spanish or German. Thus, the linguistic transfer follows the cultural transfer. The translator takes the pragmatic function as a starting point and looks for the linguistic means which are apt to render the illocution. In order to achieve this aim, the translator often
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.8 (352)
Verena Jung and Angela Schrott
not only uses different linguistic means but may also add explanatory passages to the text and remodel the pragmatic structure of the text. A pragmatic remodelling of a text is given when the semantic and syntactic modifications in the translated text imply a change of the pragmatic pattern in the original. The translator can e.g. change the sequence of speech acts, insert an additional speech act or render a certain speech act with a speech act type that does not exactly correspond to the original. We can think of different reasons for such a change in the pragmatic pattern: it may be due to a simple misunderstanding of the medieval text or it may result from the specific strategies chosen by the translator. However, in the perspective of the historicity of speech acts highlighted by historical pragmatics, a third possibility comes up. If we keep in mind the variability of illocutions over time, a pragmatic remodelling can also indicate that the speech act or speech act sequence described in the medieval text no longer exist in the translator’s culture (Wierzbicka 1991). In that case, a faithful rendering of the pragmatic pattern could easily cause misunderstandings or produce a speech act sequence whose meaning the modern reader cannot grasp. Therefore, in order to keep intact the basic communicative value of a string of interactions, the translator has to remodel the pragmatic pattern more or less profoundly – the illocutionary shift then reflects a pragmatic evolution and enables the linguist to trace diachronic changes in the speech act set. In that perspective, pragmatic remodellings are possible indicators that illocutions or illocutionary sequences have undergone changes. Based on these assumptions, contemporary translations of historical texts can make good material for tracking down evolutions in diachronic pragmatics. The description of what the translator needs to do in order to render a passage in such a way that the communicative needs of the addressee are met corresponds in some way to the task of historical pragmatics when trying to determine if a speech act has changed either in form or content from one period to the other. This shows that using the knowledge of historical pragmatics may be an essential tool to top-down translation methodology for a translator of medieval texts. We will see this illustrated in the text examples from the Cantar de mio Cid.
. Questions not asked any longer? Question types in the Cantar de mio Cid and their translations . The act of asking a question If dialogues and speech acts have a historical dimension we have to consider the possibility that an elementary illocution like the act of asking a question can also be exposed to changes over time – a strong possibility as we analyse interrogative acts described in a text of the late 12th century. On the functional level processes
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.9 (353)
Question types and speech act shifts
of change can affect questions in two ways. Firstly, the pragmatic profile of a specific type of question like “Where are you, my brave vassal?” can change. Secondly, the way a question is performed in certain dialogue contexts can change, i.e. the conversational presentation of the interrogative act as well as its embedding in the illocutionary context can undergo variance. In order to describe the specific functions of the questions in the Old Spanish text and their translations we have to define the pragmatic function of the act of asking a question. Against the background of this definition we can characterise the function in the text and describe possible shifts in pragmatic meaning.3 Questions are marked by a missing propositional parameter in the set of information represented in the utterance. In wh-questions of the type “Where did the brave vassal go?” the lacking information is marked by the interrogative adverb; in yes/no-questions like “Did the brave vassal go to Castile?” the proposition is complete but the truth value of the proposition as a whole is not assured. The speaker’s knowledge about a relevant situation is fragmentary and the speaker lays open a knowledge deficit in need for specification.4 However, the knowledge deficit exposed by the question does not always function as a request to the interlocutor to deliver the lacking information – the speaker can present the deficit without expecting the addressee to fill the gap. Therefore, questions do not automatically function as requests for information.5 But even if a question is not aimed at eliciting information, the evocation of an information deficit has a highly activating potential: by pointing at the knowledge gap, the speaker can call upon the addressee for a response or reaction (Liedtke 1998: 165–167; Bucher 1994: 242). The highly activating power of questions to call up responses manifests itself in the fact that questions are often bound to the pattern of the adjacency pair of question and answer where the question as first element draws the answer after it.6 The wide definition given here is to be understood as a prototype for the act of asking questions (Schrott 1999: 338; cf. also Arnovick 2000: 16) – a definition that is wide enough to embrace not only typical but also ‘untypical’ question acts like the ones we will find in the Old Spanish text. As examples for the question types used in the Cantar we have deliberately chosen two types of wh-questions which are clearly marked by their syntactic structure as interrogative sentences in Old Spanish. This makes sure that in all the examples we analyse, the translators were confronted with an utterance that is syntactically marked as an interrogative structure in the Old Spanish original. If the translator chooses to render this structure with an illocution other than a question, we can be sure that this modification cannot be due to a syntactic ambiguity of the Old Spanish sentence structure – the translator’s decision must be motivated differently.
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.10 (354)
Verena Jung and Angela Schrott
. Some remarks on the text and its translations As a textual basis for this analysis we refer to the Cantar de mio Cid, an Old Spanish epic poem or cantar de gesta, probably written at the end of the 12th century.7 With a medieval text like the Cantar de mio Cid the historical understanding of the literary genre is especially important.8 The relationship between the Cid and his vassals e.g. can only be understood on the background of the feudal system in general and of the Cid’s sacrosanct authority in particular. It is also important that the Cantar as an early vernacular text spreads from an oral tradition and is influenced by a performance culture (Walsh 1990: 1). The chosen translations cannot pretend to represent the spectrum of translations that have been made of the Cantar de mio Cid nor do we intend to depict the reception of the medieval text via its translations.9 The translations represent different ‘styles’ of translation but as we have not been able to screen all translations of the Cantar into modern English or Spanish the versions presented constitute in a certain way a ‘random sample’.10 For our comparison we have chosen two translations into modern Spanish and two translations into modern English. In the case of the Spanish versions, we have intralingual translations from Old Spanish (castellano medieval), a period that ends around 1450 (Eberenz 1991), into modern Spanish whose beginnings are located around 1650. Here, the Spanish translations can also be seen as ‘modernisations’ of the text that remain within the diachronic evolution of the Spanish language. As we pointed out already, our translation comparison focuses on the rendering of speech acts and on the pragmatic remodelling of illocutions. For historical pragmatics as a discipline that wants to trace changes in the speech act culture, it is crucial to be able to decide whether a remodelled pattern may reflect a change or whether the modification is conditioned by translational strategies. Therefore, it is very helpful to throw a glance at the guidelines and aims the translators evoke in their prefaces. As translations into modern Spanish we have chosen the well-known translation by the poet and essayist Pedro Salinas, written in 1926, and a more recent translation by Francisco Marcos Marín (1997). With his translation Salinas11 wants to tear down the linguistic barriers that prevent the modern Spanish reader to know and love the Cantar de mio Cid – Salinas does not write for specialists but for a reader who is generally interested in literature without special historical or literary knowledge (Salinas 1926: 10).12 With his verse translation, Salinas means to preserve the poetic qualities of the text and recreate the first masterpiece of Spanish poetry for a modern public (Salinas 1926: 11). The translation shall fill the text with life, communicate the “fuerte virtud poética y humana” of the Cantar and win the epic poem the hearts of today’s audience (Salinas 1926: 11–12). For the translation, Salinas points out the following guidelines: the translation is supposed to be clear
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.11 (355)
Question types and speech act shifts
and absolutely faithful to the medieval text (“fiel y claro”). He wants to translate “punto por punto” and consequently retains e.g. the repetitions in which the text abounds.13 Salinas wants to recreate the Cantar in a modern version that is a poetic text of its own – the translator seems to take the part of a modern juglar or singer of tales and act as a mediator between the medieval text and modern readers. This ‘autonomy’ of the translation as a poetic recreation becomes also apparent in the first edition of the translation (1926) where we find only the translation and no reprint of the Old Spanish text is given. Like Salinas, Francisco Marcos Marín points out that the maxim of his translation is faithfulness to the text. Marcos Marín’s version modernises the text so that a modern reader with a certain culture can understand most of it.14 However, this modernisation is limited by Marcos Marín’s maxim to work out a translation that is as close as possible to the original in order to respect what he calls the ‘coherence of the text’.15 Marcos Marin’s verse translation follows this guideline even at the cost of keeping some expressions that may have an archaic character for the modern reader. The faithfulness to the original implies that the translation carefully preserves the formulaic style of the Cantar (1997: 108f.). Marcos Marín modernises the text only when the ‘faithful’ translation could lead to a misunderstanding of the text (1997: 110). In a way, the translation is a part of the text edition like the foot notes in which Marcos Marín comments the Old Spanish text. The translation is fully integrated in the edition and is not meant to be a poetic recreation of the text.16 As English translations we have chosen two quite recent translations of the Cantar, one by Peter Such and John Hodgkinson (1987), and the other one by Rita Hamilton and Janet Perry (1975). Similar to Salinas and Marcos Marín, Peter Such and John Hodgkinson (1987) want to give a translation that is faithful to the text and keeps its poetic style. Such and Hodgkinson favour a “literal rendering of the text” and it is their aim “to preserve details of form as far as to do so would produce an acceptable English reading” (1987: 35). In their English version they intend to “reproduce something of the style and flavour of the original” as far as possible in modern English – the translators are in general reluctant to depart from the text and modify the text only “in order that our translation should work as a convincing English version” (1987: 35). By following these guidelines they hope to retain the spirit of the text as well as the “formality of idiom” that characterises the poetic work.17 Different guidelines are advocated by Rita Hamilton and Janet Perry in their prose translation (1975). In contrast to Such and Hodgkinson, they are not primarily concerned with the maxim of faithfulness to the text but highlight that they intend to give a translation whose style is more “current” for today’s readers than the epic style of the original text (1975: VII). The main aim of Hamilton and Perry is to support the modern reader with a text that presents “current English usage”. To achieve this, they abandon the formulae of the Old Spanish text and sometimes
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.12 (356)
Verena Jung and Angela Schrott
suppress the repetitions of the epic poem. Their translation, which is meant to be read independently from the original, does not aim at retaining the poetic character of the text but wants to adapt it to the communicative experiences of the modern reader. The translation comparison will show in what respects the guidelines may have influenced the way the Old Spanish text is modelled on the pragmatic level.
. A special use of the where-question When working on the question types in the Cantar, the linguist realises that some uses of questions or other speech act types do not seem to cause any problems of understanding even for a modern reader without special knowledge of this text tradition: the every-day communicative experience of the modern reader is sufficient to understand the speech act. Those illocution types are used more or less in the same way we use them today and their pragmatic functions have remained unchanged. However, we find other question types whose use puzzles even the medievalist. It is those puzzling uses that indicate that the performance of a speech act type may have changed. Here, problems of understanding arise which can be reflected in the translations of the text and shifts in pragmatic meaning may indicate that the interaction represented in the text had to be rendered differently because it no longer forms a part of the modern reader’s communicative experiences. Our first example is such a ‘puzzling’ question type, a special use of the wherequestion. In the Cantar de mio Cid, we find several very similar situations where the Cid, surrounded by his most faithful liegemen, calls upon one vassal to perform a mission. He begins his briefing by asking a where-question (V.2618–2623; Ed. Montaner, English translation by Such and Hodgkinson):
¿Ó eres, mio sobrino, tú, Félez Muñoz? Primo eres de mis fijas amas d’alma e de coraçón, mándot’ que vayas con ellas fata dentro en Carrión, verás las heredades que a mis fijas dadas son. Con aquestas nuevas vernás al Campeador. Dixo Félez Muñoz: Plazme d’alma e de coraçón.
Where are you, my nephew, Félez Muñoz? You are the cousin of my daughters, whom I love with heart and soul. I command you to accompany them as far as Carrión, and see the estates which have been given to my daughters as an inheritance. With a report of this you shall come to the Battler. Félez Muñoz said: In my heart and soul I am pleased to do this.
As the vassal is present before the Cid’s eyes, the question is obviously not supposed to elicit information about the whereabouts of the vassal. By asking the question
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.13 (357)
Question types and speech act shifts
the Cid does not want to know where the vassal is, he wants to know whether the vassal is there for him, prepared to perform his orders. Here, the where-question belongs to a ritualised strategy which expresses the ideal of a well-functioning feudal system. In the Cantar we find four similar scenes in which the where-question is used to focus on one candidate and serves as preliminary to a command – the interrogative structure functions as question act but it fulfils a purpose no longer familiar to the modern reader. For our translation comparison, two leading questions arise. Firstly, we have to ask whether the translation preserves the communicative value the speech act or speech act sequence has in the Old Spanish text. Secondly, influenced by the approach of historical pragmatics, we want to find out to what extent the translation retains the cultural characteristics of the interaction sequence, i.e. to what extent a possible alterity and strangeness of the interaction are still present in the translation. As to the translation of the interrogative structure, the versions form two groups. Marcos Marín and Such and Hodgkinson keep the interrogative structure of the where-question and preserve the question act as such. On the level of the speech act sequence, they do not modify the illocutionary pattern of the Cantar.
Cantar (Ed. Montaner)
Salinas
Marcos Marín
Such/Hodgkinson Hamilton/Perry
(1) V. 102–104 Llegó Martín Antolínez, a guisa de membrado:
(1a) Les saluda el burgalés, muy atento y muy taimado:
(1b) Llegó Martín Antolínez, como hombre avisado:
(1c) Martín Antolínez arrived, a clever man:
¿Ó sodes, Rachel ¿Cómo estáis, e Vidas, los mios Raquel y Vidas, amigos caros? amigos míos tan caros? En poridad En secreto yo fablar querría querría hablar con amos. con los dos un rato. (2) V. 1804, 1808 (2a) ¿Dó sodes, ¿En dónde caboso? ¡Venid estáis, grande aca, Minaya! hombre? Venid para acá, Minaya.
¿Dónde estáis, Ragel y Vidas, mis amigos amados? En secreto querría hablar con ambos.
Where are you, Raquel and Vidas, my dear friends? I would like to speak to both of you in confidence. (2b) (2c) ¿Dónde estáis, Where are you, honrado?, my noble venid acá, Minaya? Come Minaya; [. . . ] here.
(1d) Martín Antolínez, like the shrewd fellow he was, came up and greeted them: Is that you, my good friends, Rachel and Vidas? I should like a word with you two in private. (2d) Come here, my gallant Minaya. [. . . ]
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.14 (358)
Verena Jung and Angela Schrott
(continued) [. . . ] e cras a la mañana irvos hedes sin falla [. . . ]
[. . . ] Mañana al romper el día habéis de marchar sin falta. (3) V. 2618, 2620 (3a) ¿Ó eres, mio ¿Dónde estás, sobrino mío, sobrino, tú, Félez Muñoz? dónde estás, [. . . ] Félez Muñoz? [. . . ] Mándot’ que mando que vayas con ellas vayas con ellas fata dentro en allá hasta el Carrión [. . . ] mismo Carrión, [. . . ] (4) V. 2901, 2903 (4a) ¿Ó eres, Muño ¿Aquí estás, Gustioz, mio Muño Gustioz, vasallo de pro? tú mi vasallo [. . . ] de pro? [. . . ] Lieves el mandado a Castiella al rey Alfonso, [. . . ]
Mañana por la [. . . ] and mañana os iréis tomorrow sin falta [. . . ] morning you will leave for certain, [. . . ] (3b) (3c) ¿Dónde estás Where are you, mi sobrino, tú, my nephew, Félez Félez Muñoz? Muñoz? [. . . ] [. . . ] Te mando vayas con ellas hasta dentro de Carrión,
I command you to accompany them as far as Carrión, [. . . ]
(4b) ¿Dónde estás Muño Gustioz, mi vasallo de pro? [. . . ]
(4c) Where are you, Muño Gustioz, my worthy vassal? [. . . ]
Tomorrow morning you must set out, [. . . ]
(3d) (He called Félez Muñoz to him and said) Are you there, my nephew, Félez Muñoz? [. . . ] Go with them as far as Carrión [. . . ]
(4d) (He called Muño Gustioz and said to him) Muño Gustioz, my brave vassal, A Alfonso, rey Lleva el recado Carry this [. . . ]. You shall de Castilla, irás a Castilla, al rey message to Castile carry this message con esta Alfonso; to King Alfonso; to King Alfonso in misión: [. . . ] Castile.
A different choice has been made in the translations by Salinas and Hamilton and Perry. Whereas Hamilton and Perry never translate with a where-question, Salinas keeps the type of the where-question in example (2a) and (3a) where he even repeats the interrogative structure. The repetition in (3a) introduces a touch of impatience as if the Cid was already waiting for the vassal to come and fulfil his orders: “Where are you, where are you? I want you to leave for Carrión.” Although this modification slightly changes the pragmatic value it is still very close to the original and it can remind the reader of a pattern he is familiar with. In the examples (1a) and (4a) Salinas does not keep the where-question. In (1a) the changes are rather complex. Firstly, Salinas changes the context of the illocution. Whereas the original only mentions the arrival of the clever Martín Antolínez, Salinas inserts a passage that qualifies the following enunciation as a greeting. What is more, he changes the where-question into the question ¿Cómo estáis? (“How are you?”) which is clearly marked as a greeting formula. Here, the pragmatic pattern undergoes a deeper modification than in (3a). However, in spite of the modification,
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.15 (359)
Question types and speech act shifts
the basic illocutionary value is preserved as the greeting formula picks out the addressee, focuses on him and functions as a preliminary to what the speaker has to say.18 The modification Salinas chooses in (4a) has the same effect. Here, Salinas replaces the where-question with a yes/no-question that has only weak interrogative power. The speaker knows that the addressee is there and asks only for confirmation or some other signal of co-operation. A corresponding translation is also chosen by Hamilton and Perry in example (3d). Like in the greeting example, Salinas here (4a) succeeds in keeping intact the communicative value of the pattern. In the translation by Hamilton and Perry we find similar techniques. Like Salinas, Hamilton and Perry insert a short narrative passage that presents the question as a greeting (1d). However, they do not go so far as to replace the wherequestion by a genuine greeting like “How are you?” but opt for a yes/no-question as if the speaker was not sure about the addressee’s identity. As questions like “Is that you, my good friends. . . ?” are frequently used for establishing contact with the addressee, the translation can render the focusing effect the where-question has in the original. We find a similar effect in (3d) where the yes/no-question functions analogously to Salinas’ version (4a). In two examples, Hamilton and Perry suppress the question act as a whole. In (2d) where the original presents a sequence of question and imperative they suppress the question, keep only the imperative and underline the element of command. In (4d) the vassal is simply addressed with his name before the briefing starts. In this example, and likewise in (3d), the translators feel obliged to insert a short stage direction in brackets in order to illustrate the communicative situation. The stage instructions are slightly ambiguous: they imply that the addressee has to be called before the Cid whereas the original makes clear that the vassals are present before the Cid’s eyes. Whereas Hamilton and Perry retain the communicative value of the pragmatic pattern in examples (1d) and (3d), this can be doubted for their versions (2d) and (4d) in which they suppress the question. In (4d) the stage direction and the address of the vassal still render in some way the idea of a focusing speech act that prepares the briefing; in (2d) however, this pattern is not given any longer and is replaced by a simple order to the liegeman. The translations compared illustrate two different strategies on the pragmatic level. That each strategy is represented by a Spanish and an English translation points to the fact that the different translational styles are not due to the different language systems. Marcos Marín’s intralingual version and Such and Hodgkinson’s interlingual translation retain the pragmatic pattern of the original. As this pattern is not familiar to the reader, he may easily understand Where are you, my nephew, Félez Muñoz? (3c) as a question asking for the whereabouts of the knight – in that case, keeping the sentence type would produce a different speech act. As we showed in our discussion of this verse, a philological interpretation is necessary to reconstruct the pragmatic profile: in modern Spanish as well as in modern English the
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.16 (360)
Verena Jung and Angela Schrott
illocutionary pattern retains the strangeness and alterity that characterises the interaction sequence in the Old Spanish text. The translation is careful not to touch the aesthetic signifiant of the text and produces a transparent translation behind which the original text in its alterity is still visible – in a way the translation functions as a ‘paraphrase’ of the original speech act sequence. The difficulty for the reader does not lie in the Old Spanish question type ¿Ó eres. . . ? whose pragmatic potential overlaps with where-questions in modern Spanish and English. What has changed is the part the question plays in the illocutionary sequence whose function is closely linked to the feudal system. Thus, modern readers find themselves confronted with a pragmatic pattern whose meaning they may find difficult to grasp – they are in the same position as the medievalist who seeks to reconstruct the interaction in the Old Spanish text. In contrast to the ‘paraphrasing’ translations that leave part of the cultural transfer to the reader, Salinas’ Spanish version and Hamilton and Perry’s English translation remodel the pragmatic pattern in order to create a speech act sequence that belongs to the cultural repertoire of the reader while keeping intact the original’s communicative values. As we have seen, this attempt to find functional equivalences in today’s culture is successful in most cases. However, what is lost to a certain extent, is the alterity of the interactional culture we find in the Cantar de mio Cid.
. Rhetorical questions The second example we want to discuss is concerned with the use of rhetorical questions. We define rhetorical questions as a special use of interrogative sentences where the interrogative structure serves to perform an assertion (Meibauer 1986: 76). Thus, rhetorical questions can be seen as indirect assertions (Meibauer 1986: 32–42; Schrott 2000: 270–271) that can be inferred from the interrogative structure. As a rhetorical device, rhetorical questions are neither bound to a specific language nor to specific linguistic structures – whether a question is rhetorical or not depends on the context. In the Cantar we find four passages where the fictional narrator uses rhetorical questions in a very specific way. We limit the analysis to one rhetorical question of that type and compare the way it is rendered in the modern translations.19 In the example given the interrogative structure “Who could count the gold and silver?” implies the assertion that nobody is able to count those riches and evokes a notion of infiniteness (Meibauer 1986: 132). With this rhetorical question the fictional narrator highlights that the riches are so enormous that they surpass the imagination of the audience as well as his own imagination. Whereas Marcos Marín and Such and Hodgkinson in both examples keep the rhetorical question, Salinas drops the second rhetorical question and paraphrases the concept of infiniteness
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.17 (361)
Question types and speech act shifts
Cantar (Ed. Montaner)
Salinas
Marcos Marín
V.1214, 1218 el oro e la plata, ¿quién vos lo podrié contar? [. . . ]
(a) y el oro y plata ganados ¿quién los podría contar? [. . . ]
(b) el oro y la plata ¿quién os lo podría contar? [. . . ]
Such/Hodgkinson Hamilton/Perry
(c) and who could reckon the value of the silver and the gold they seized? [. . . ] y las otras who could reckon e los otros averes y además le ganancias the value of the ¿quién los podrié tocan bienes que no se ¿quién las other goods he contar? pueden contar. podría contar? gained?
(d) [. . . ] there were untold qualities of gold and silver [. . . ]
the value of the rest in kind was beyond reckoning.
(“he got riches that cannot be counted”). In Hamilton and Perry’s version no trace of the rhetorical device is left – they replace both rhetorical questions by paraphrases that keep the idea of infiniteness on the content level.20 However, in the Cantar the idea of infiniteness is not simply stated, with the rhetorical questions the fictional narrator addresses the audience – an interaction that functions as a fictional recreation of the interaction between the singer of tales, the juglar, and his audience. The translation by Hamilton and Perry fades out the figure of the narrator and effaces those artfully elaborated traces of oral performance. The pragmatic pattern of the original is not only remodelled but simply left out so that the aesthetic signifiant of the text and its alterity undergo a profound change. Thinking of the communicative experiences of the modern reader who is well-acquainted with rhetorical questions it is in our opinion a modification that does not seem necessary in order to adapt the text to today’s communicative culture.
. Translation comparison and historical pragmatics: An example of speech act shifts in Chaucer translations . Understanding Chaucer – shifted focus in the speech acts portrayed in the various translations – statement of status versus lament As in the translations of the Cantar, the following analysis of shifted focus in Chaucer’s Wife of Bath’s Prologue and Tale highlights the question of consistency of the communicative value of the original, therefore, the first step is to look at the original and determine the type of illocution contained in one text passage from the original, the historically determined text. We will then look at the translations, in this case, two so-called intralingual translations, translations into modern English, and two into modern German and attempt to determine their illocutionary value.
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.18 (362)
Verena Jung and Angela Schrott
As the first two translations are in fact translations into a more modern stage of the same language, they are particularly interesting in terms of historical pragmatics, as the language system will not be responsible for most of the necessary adaptations. Therefore, a change in either form or function of the speech act is likely to be the reason for a translational decision. As has been shown in the case of the where-questions in the Cantar, keeping a particular sentence type can cause a pragmatic shift in the understanding of the modern reader. This phenomenon was demonstrated even more clearly by the example of “Why don’t you take your coat off?” and its German non-equivalent “Warum ziehst du deinen (ziehen Sie Ihren) Mantel nicht aus?” – here, keeping a particular surface structure creates a different speech act in the new community. Only after differences in speech acts have been analysed does the question of motivation for these shifts come into play. For the translation practitioner, determining which version is closer to the original communicative purpose is crucial – for the purpose of historical pragmatics this question is also relevant, but it is only one aspect in the tracing of changes in speech acts through the passage in time. However, for a shift in emphasis to be exclusively determined by historical change, it would have to be used by several or all translators of one passage. If the shift occurs only in some versions but not in others, this suggests rather that a different translational strategy was used by the different translators. This can be seen in the following example from Chaucer:
Chaucer Text (Riverside Ed.) III, 1–3 Experience, though non auctoritee/ were in this world, is right ynogh for me/ to speke of wo that is in marriage.
Coghill
Lumiansky
Kemmler
Lehnert
If there were no authority on earth/ Except experience, mine, for what it’s worth, and that’s enough, for me, all goes to show/ That marriage is a misery and a woe.
My experience gives me sufficient right to speak of the trouble there is in marriage, even if there were no other authority in the world.
Erfahrung ist für mich, selbst dann, wenn es keine Autoritäten auf dieser Welt gäbe, völlig ausreichend, um über das Leid im Ehestand sprechen zu können.
Erfahrung kann hinlänglich mich belehren, wenn nicht dafür Autoritäten wären, Wie doch der Ehestand is voller Leid.
In this famous example, the very beginning of the Wife of Bath’s Prologue to her tale, all five versions, the original, the intralingual and the interlingual translations have the form of a statement. On the surface, the differences between each of the two
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.19 (363)
Question types and speech act shifts
versions in the same language are syntactic in nature. However, in view of a topdown process, the syntactic restructuring would of course only be a consequence of a perceived communicative function for which a restructuring may or may not be necessary in order to create the same function, depending on the closeness of the two speaker communities. The sentence focus of the original and of the versions by Lumiansky and Kemmler is on experience, whereas the focus of Coghill and Lehnert is on woe. In more traditional linguistics, this might have been seen as a stylistic difference. But changing the focus of a sentence has implications for the speech act it contains. If the focus of the sentence is experience, then the illocution of the passage would be to state authority, to claim a certain status. If the focus of the sentence is woe in marriage, then the passage is a lament. Both are appealing to the listener, but in the one case for respect and attention and in the other version more for pity. That the versions of Lumiansky and Kemmler are closer to the speech act pattern of the original is our interpretation, as there are slight differences in focusing elements between English and German and there could have been different focus rules in Middle English. The question of motivation for the shift in speech act from the original in Coghill’s and Lehnert’s versions can have several interpretations. One would be to suggest that the copying of the syntactic structure of the original was expected to automatically create the same illocution (which is not the case, due to differences in focus structure between the two languages, cf. Doherty 1990: 282; Jung 2002: 140ff.), the other would be that the modern reader or two of the modern readers did interpret the speech act differently due to a change of cultural interactional patterns. One reason for this could be a different perception of a woman’s role in the world, the idea that the interpretation of the speech act as ‘lament’ would fit better into the cultural picture of this time or the perceived traditional view of the historic time in which the text was written. Kussmaul (1995: 32f.) cites an interesting example of motivation for what is clearly a translation mistake, very clearly even a misreading of a text, but a misreading which, in his view, is supported or even created by a shift in thought patterns and attitudes of a society. In this case, the time lapse between the text and the translation is not, as in our case here, more than 500 years, but only 40 years. In a passage from the book Male and Female written by anthropologist Margaret Mead in 1949 [. . . ] both men and women share the same images of what makes a marriageable or an unmarriageable woman, a good husband, a fascinating lover whom any woman would be a fool to marry, or a born old bachelor [. . . ] (Margaret Mead. 1962 [1949]. Male and Female. Penguin Books, Harmondworth, p. 271)
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.20 (364)
Verena Jung and Angela Schrott
the sentence a fascinating lover whom any woman would be a fool to marry was translated in the sense of “whom any woman would be a fool not to marry” by a surprising number of students. While overlooking a negation or putting one in is not uncommon in translators, especially untrained ones, the fact that this happened to so many students with this particular sentence was seen by Kussmaul as an indication that general beliefs act as semantic frames when interpreting a text. This would explain the fact that most women in 1990 immediately assumed that a good lover must make good husband material and therefore inserted a not here. Thus, changes in our verbal interaction expectations might make us misunderstand historical texts for that reason more than or as much as problems with meaning changes of certain vocabulary or changes in the rules of syntax structure. Therefore, one possible interpretation for the different versions of this passage could be that the two translators who have a different illocution to the one in the Chaucer text came to that interpretation not only through failure to recognise syntactic differences between German and Middle English or modern English and Middle English, but also, because the lamenting Wife of Bath fitted better into their general image of medieval women than the self-assertive version. This is of course not a claim that could be maintained on one passage alone, the fact is simply that the portrait created by means of the verbal portrayal of this Wife of Bath, in Coghill and Lehnert, clearly differs from that of Lumiansky and Kemmler.
. Different contextualisation and explicitness create different assumptions: Self-reproach versus boast Chaucer Text (Riverside Ed.)
Coghill
Lumiansky
III, 149–151 In wyfhod I wol use myn instrument/ As frely as my Makere hath it sent./ If I be If I turn If I am dangerous, God difficult, God standoffish, yeve me sorwe. give me sorrow. may God send me sorrow.
Kemmler
Lehnert
Sollte ich damit Bin ich gefügig geizen, möge nicht, geb Gott mir Gott mir Sorgen. Kummer bereiten.
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.21 (365)
Question types and speech act shifts
The second example from Chaucer’s Wife of Bath’s Prologue is concerned with a different structural level, this time, the five sentences all look completely alike apart from the translation of one word. Again, this could be seen simply as a question of the linguistic discipline of lexicology, not of pragmatics. Clearly, the word in question, dangerous, has changed in meaning from 1450 to today, therefore none of the two English translations use the etymologically closest equivalent in modern English. The Middle English glossary (Chaucer Riverside Edition glossary) lists the following meanings: 1. domineering, disdainful 2. disdainful, standoffish, unaccommodating 3. niggardly, grudging (the latter is the suggested meaning for this passage in the Chaucer Riverside Edition glossary) But of the many meanings of dangerous offered in the Middle English glossary Coghill and Lehnert come up with a translation that seems removed from the context. The translations as difficult and nicht gefügig (which means ‘not obedient’) suggest meaning 1, “domineering” rather than 3, “grudging”. Again, we are presented with two very different interpretations of the meaning of this passage. We see Coghill’s and Lehnert’s versions as separate from the context, as the question of obedience was not discussed. This is the end of a passage in which the Wife of Bath discusses the merits of being either chaste or a wife and is trying to favour the virtues of wifehood and active sexuality versus holiness and chastity, at least aiming at giving them an equal status – again a question of gaining status and respect. Again, the same two translators, Lumianksy and Kemmler, go for the version “I will fulfil my status”, interpreting dangerous as “not fulfilling marital duties”, while Coghill’s and Lehnert’s versions (“If I should not be obedient. . . ”) create a meaning of self-reproach, which is not relevant to the context, but probably relevant to the overall frame evoked by these two translators, that of a lamenting woman, reproaching both herself and her husband for the woe in marriage. That a different cultural concept may be to blame here more than a misunderstanding of this word is further supported by the fact that there is a passage in the Chaucer text (III, 514) where all four translators translated dangerous with the meaning chosen by Lumiansky and Kemmler here, “standoffish, non-compliant”, rather than “domineering”, but in this case the adjective applies to the husband’s duties. As was demonstrated in the examples from the Cantar de mio Cid, the fact that the translator took a certain decision indicates that some changes in general interactional behaviour are perceived. This particular passage also points at the problem of greater explicitness of the translation: The translator does not really produce the same illocution, he produces the same illocution but in a more forceful manner. Lumiansky and Kemmler, while in our view closer to the interpretation of the original, still had to make their illocution more forceful, losing the multiplicity of
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.22 (366)
Verena Jung and Angela Schrott
meanings and having to opt for one specific, sexual interpretation of dangerous here. But Coghill and Lehnert are equally explicit, though in our view focusing on an illocution not relevant to the original author. However, the general statement that translations by nature have to be more explicit than originals (Blum-Kulka 1986: 9) is only relevant if we assume that the illocution of the translated text could ideally correspond exactly to that of the original, yet it is only through the translation variants that we find out about the possible range of illocutions contained in the original (Frank 1988). It is the task of historical pragmatics to then point out which one best fits the general cultural values of the time.
. Conclusion In our paper we have tried to sketch a framework that could serve as a basis for combined approaches in historical pragmatics and translation studies. As has been shown, both historical pragmatics and translation studies can use each other as indicators for their research. For historical pragmatics, translation comparison on the pragmatic level constitutes valuable material for tracing changes in speech acts and speech act patterns. On the other hand, the historicity of speech acts which is at the very centre of historical pragmatics can help translation studies to determine the particular challenges and problems that are faced by the translator when translating historically determined texts into the modern idiom and the modern frame of speech acts. This is especially true for translation scholars who deal with historical texts of a by-gone culture and who are often torn between the necessities of pragmatic remodelling and the wish to retain the alterity of the text – be it out of philological traditions or out of respect for a text considered as a poetic masterpiece. Through our text analyses and their theoretical frame we hope to have shown that the co-operation between an inherently contrastive historical approach in pragmatics and translation studies can be rewarding and stimulating for both disciplines. For historical pragmatics the method of translation comparison not only provides useful theoretical and methodological insights but also valuable text material for studies on the evolution of speech acts and their traditions through time. When analysing historical texts like the Cantar de mio Cid or the Wife of Bath’s Prologue and Tale, one could e.g. analyse how a specific speech act pattern was translated not only in the 20th century but through a much longer period of time – however, this is a chapter in translation studies and historical pragmatics that we will not open in this paper.
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.23 (367)
Question types and speech act shifts
Notes . For a discussion on the historicity of dialogues see Fritz 1995: 469, 470f. and 1997: 47f. . Wierzbicka (1991: 212) gives the following paraphrase for this conversational routine: Why don’t you do X? a. b. c. d.
I say: I want you to say, if you can, why you/we shouldn’t do X. I think you can’t (say why). I think it will be a good thing if you/we do it. I say this because I would want you to do it.
. For a more explicit discussion of the speech act of asking a question see also Schrott 1999 and 2000. . Cf. for example Schlieben-Lange 1983: 96; Kerbrat-Orecchioni 1991: 18. . Cf. Searle’s definition of the question as a special type of directive, Searle 1969: 66, 69 and 1979: 44–47, 48–51. For more literature on questions see Schrott 1999 and 2000. . On the question-answer sequence as an adjacency pair see Sacks, Schegloff, & Jefferson 1974: 711, 716f., 728; Kerbrat-Orecchioni 1991: 9; Bucher 1994: 239. . The period of Old Spanish extends from the 12th to the 15th century, for a closer look on the periodisation of the Spanish language see Eberenz 1991. . Cf. the concept of “alterity” that characterises medieval texts according to Jauß 1977: 10, 13–14. . For translations as receptions of medieval texts see also Grosse 1986; for information on the international reception of the story of the Cid see Rodiek 1990. . López Estrada gives an overview of translations into modern Spanish (1961: LXIV– LXV). . Salinas also wrote several essays on the Cantar (Salinas 1961). . Salinas 1926: 10: “la antigüedad del lenguaje tiene confinada esta obra de arte en un círculo de lectores forzosamente reducido.” . Salinas 1926: 11: “Fidelidad absoluta al texto del poema, sin desviaciones en busca de ornato, sin amplificaciones ni desarrollos casi nunca.” . Marcos Marín 1997: 108: “Actualizamos el texto del Cantar de manera que sea inmediatamente comprensible para un lector moderno, con un cierto grado de cultura, dentro de algunes límites.” . Marcos Marín 1997: 108: “Para mantener una coherencia textual hemos procurado que nuestro texto no se apartase mucho del original, aunque ello implicase mantener alguna expresión más arcaizante, [. . . ].” . On the techniques with which the translation was elaborated see Marcos Marín 1997: 110–111. . Here Such and Hodgkinson (1987: 35) quote a passage that Tobias Smollett includes in the preface of his translation of Don Quijote (1755): “He has endeavoured to retain the spirit and ideas, without servilely adhering to the literal expression of the original; from which,
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.24 (368)
Verena Jung and Angela Schrott
however, he has not so far deviated, as to destroy that formality of idiom, so peculiar to the Spaniards, and so essential to the character of the work.” . For a study on linguistic structures and functions of greetings see Lebsanft 1988. . Rhetorical questions of the type discussed here are also to be found in the following passages of the Cantar: V. 698–700; V. 1966–1971. For a more thorough analysis of this question type see Schrott 2000: 273–276. . If we look at all four examples in the Cantar, we see that Hamilton and Perry always suppress the rhetorical question whereas Salinas in general keeps all the rhetorical questions and only leaves out the one in the example we quote.
References Primary texts Cantar de mio Cid. (1993). Edición, prólogo y notas de Alberto Montaner (2. edición corregida). Barcelona: Crítica. Translations Poema de mio Cid. (1926). Puesto en romance vulgar y lenguaje moderno por Pedro Salinas. Madrid: Revista de Occidente. Cantar de mio Cid. (1997). Edición de Francisco Marcos Marín. Madrid: Editorial Biblioteca Nueva. The Poem of the Cid. (1975). A Bilingual Edition with Parallel Text. Translated by Rita Hamilton and Janet Perry with an Introduction and Notes by Ian Michael. London: Penguin. The Poem of My Cid. (1991). Translated with an Introduction & Commentary by Peter Such & John Hodgkinson. Second corrected impression. Warminster: Aris & Phillips.
The Riverside Chaucer. (1988). General editor: Larry D. Benson. Oxford: Oxford University Press. Translations Chaucer, Geoffrey: The Canterbury Tales. (1977). Translated into modern English by Nevill Coghill. London: Penguin. Chaucer, Geoffrey: The Canterbury Tales. (1977). A new modern English prose translation by R. M. Lumiansky. New York: Pocket Books. Chaucer, Geoffrey: Canterbury-Erzählungen. (1989). In deutsche Prosa übertragen von Fritz Kemmler. München: Goldmann. Chaucer, Geoffrey: Canterbury-Erzählungen. (1963). Herausgegeben und übersetzt von Martin Lehnert. Ausstattung u. Illustrationen: Werner Klemke. Berlin: Rütten & Loening.
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.25 (369)
Question types and speech act shifts
Works cited Albrecht, J. (1973). Linguistik und Übersetzung. Tübingen: Niemeyer. Arnovick, L. K. (2000). Diachronic Pragmatics. Seven Case Studies in English Illocutionary Development. Amsterdam/Philadelphia: John Benjamins. Blum-Kulka, S. (1986). Shifts of Cohesion and Coherence in Translation. In J. House & S. Blum-Kulka (Eds.), Interlingual and Intercultural Communication (pp. 7–35). Tübingen: Narr. Blum-Kulka, S., & House, J. (1989). Investigating Cross-Cultural Pragmatics: An Introductory Overview. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-Cultural Pragmatics: Requests and Apologies (pp. 1–34). Norwood: Ablex. Bucher, H. J. (1994). Frage-Antwort-Dialoge. In G. Fritz & F. Hundsnurscher (Eds.), Handbuch der Dialoganalyse (pp. 239–258). Tübingen: Niemeyer. Catford, J. C. (1965). A Linguistic Theory of Translation. London: OUP. Cerquiglini, B. (1981). La parole médiévale. Paris: Minuit. Cherubim, D. (1980). Zum Programm der historischen Sprachpragmatik. In H. Sitta (Ed.), Ansätze zu einer pragmatischen Sprachgeschichte (pp. 3–21). Tübingen: Niemeyer. Doherty, M. (1990). Focus Hierarchies. Acta Linguistica Hungarica, 40, 275–284. Diller, H. J., & Kornelius, J. (1978). Linguistische Probleme der Übersetzung. Tübingen: Niemeyer. Eberenz, R. (1991). Castellano antiguo y español moderno: Reflexiones sobre la periodización en la historia de la lengua. Revista de Filología española, 71, 79–106. Ehlich, K. (1992). On the historicity of politeness. In R. Watts, S. Ide & K. Ehlich (Eds.), Politeness in Language. Studies in its History, Theory and Practice (pp. 71–107). Berlin/New York: Mouton de Gruyter. Frank, A. P. (1988). Understanding Literature via translation: The translation as hermeneutic device. Literatur in Wissenschaft und Unterricht, 21–1, 33–37. Fritz, G. (1995). Topics in the History of Dialogue Forms. In A. H. Jucker (Ed.), Historical Pragmatics. Pragmatic Developments in the History of English (pp. 469–498). Amsterdam/Philadelphia: John Benjamins. Fritz, G. (1997). Remarks on the History of Dialogue Forms. In E. Petri, D. Laroche-Bouvy, & S. Stati (Eds.), Dialoganalyse V. Referate der 5. Arbeitstagung Paris 1994 (pp. 47–55). Tübingen: Niemeyer. Grosse, S. (1986). Überblick über die Rezeption der deutschen Literatur des Mittelalters im 19. Jahrhundert. In Peter Wapnewski (Ed.), Mittelalter-Rezeption. Ein Symposion (pp. 377–391). Stuttgart: Metzlersche Verlagsbuchhandlung. Hawkins, J. (1986). A Comparative Typology of English and German. London & Sydney: Croom Helm. Hönig, H. G., & Kussmaul, P. (1982). Strategien der Übersetzung. Ein Lehr- und Arbeitsbuch. Tübingen: Narr. Hönig, H. G. (1995). Wieviel Sprechakttheorie braucht ein Dolmetscher? In N. Salnikow (Ed.), Sprachtransfer – Kulturtransfer (pp. 23–30). Frankfurt a. M.: Peter Lang. House, J. (1997). Translation Quality Assessment – A Model Revisited. Tübingen: Narr. Jauß, H. R. (1977). Alterität und Modernität in der mittelalterlichen Literatur. Gesammelte Aufsätze 1955–1976. München: Fink.
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.26 (370)
Verena Jung and Angela Schrott
Jucker, A. H., & Jacobs, A. (1995). The Historical Perspective in Pragmatics. In A. H. Jucker (Ed.), Historical Pragmatics. Pragmatic Developments in the History of English (pp. 3– 33). Amsterdam/Philadelphia: John Benjamins. Jucker, A. H., & Taavitsainen, I. (2000). Diachronic speech act analysis. Insults from flyting to flaming. Journal of Historical Pragmatics, 1, 67–95. Jucker, A. H., Fritz, G., & Lebsanft, F. (Eds.). (1999). Historical Dialogue Analysis. Amsterdam/Philadelphia: John Benjamins. Jung, V. (2002). English-German Self-Translation and its Relevance for Translation Theory and Practice. Frankfurt a.M.: Peter Lang. Kerbrat-Orecchioni, C. (1991). Introduction. In C. Kerbrat-Oreccioni (Ed.), La question (pp. 5–37). Lyon: Presses universitaires. Kussmaul, P. (1980). Sprechakttheorie. Ein Reader. Wiesbaden: Athenaion. Kussmaul, P. (1995). Zur Relevanz einiger Semantikmodelle für die Übersetzung. In N. Salnikow (Ed.), Sprachtransfer – Kulturtransfer (pp. 31–47). Frankfurt a. M.: Peter Lang. Lebsanft, F. (1988). Studien zu einer Linguistik des Grußes. Sprache und Funktion der altfranzösischen Grußformeln. Tübingen: Niemeyer. Lebsanft, F. (1999). A Late Medieval Bargain Dialogue (‘Pathelin’ II). Or: Further Remarks on the History of Dialogue Forms. In A. H. Jucker, G. Fritz, & F. Lebsanft (Eds.), Historical Dialogue Analysis (pp. 269–292). Amsterdam/Philadelphia: John Benjamins. Liedtke, F. (1998). Grammatik der Illokution. Über Sprechhandlungen und ihre Realisierungsformen im Deutschen. Tübingen: Narr. López Estrada, F. (1961). Prólogo. In F. López Estrada (Ed.), Poema del Cid (pp. XI–LXVIII). Madrid: Castalia. Meibauer, J. (1986). Rhetorische Fragen. Tübingen: Niemeyer. Nord, C. (1997). Translation as a Purposeful Activity – Functional Approaches Explained. Manchester: St. Jerome. Reiss, K. (1982). Texttyp und Übersetzungsmethode – Der operative Text. Heidelberg: Gross. Reiss, K., & Vermeer, H. (1984). Grundlegung einer allgemeinen Translationstheorie. Tübingen: Niemeyer. Rodiek, C. (1990). Sujet – Kontext – Gattung. Die internationale Cid-Rezeption. Berlin/New York: Walter de Gruyter. Sacks, H., Schegloff, E. A., & Jefferson, G. (1974). A Simplest Systematics for the Organization of Turn-Taking for Conversation. Language, 50, 696–735. Salinas, P. (1961 [1945]). El “Cantar de mio Cid”. Poema de la honra. In J. Marichal (Ed.), Pedro Salinas - Ensayos de literatura española. Del ‘Cantar de mio Cid’ a García Lorca (pp. 27–43). Madrid: Aguilar. Schlieben-Lange, B. (1983). Traditionen des Sprechens. Elemente einer pragmatischen Sprachgeschichtsschreibung. Stuttgart/Berlin etc.: Kohlhammer. Schlieben-Lange, B., & Weydt, H. (mit Beiträgen von E. Coseriu und H. U. Gumbrecht) (1979). Streitgespräch zur Historizität von Sprechakten. Linguistische Berichte, 60, 65–78. Schrott, A. (1999). Que fais, Adam? Questions and Seduction in the ‘Jeu d’Adam’. In A. H. Jucker, G. Fritz, & F. Lebsanft (Eds.), Historical Dialogue Analysis (pp. 331–370). Amsterdam/Philadelphia: John Benjamins.
TSL[v.20020404] Prn:4/09/2002; 15:12
F: PBNS37.tex / p.27 (371)
Question types and speech act shifts
Schrott, A. (2000). ¿Quí los podrié contar? Interrogative Acts in the ‘Cantar de mio Cid’. Some Examples from Old Spanish on Asking Questions. Journal of Historical Pragmatics, 1, 263–300. Searle, J. R. (1969). Speech Acts. An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Searle, J. R. (1979). Expression and Meaning. Studies in the Theory of Speech Acts. Cambridge: Cambridge University Press. Stein, D. (1985). Perspectives on Historical pragmatics. Papers from the Workshop on SocioHistorical Linguistics. Folia Linguistica Historica, 6, 347–352. Walsh, J. K. (1990). Performance in the ‘Poema de mio Cid’. Romance Philology, 44, 1–25. Wierzbicka, A. (1991). Cross-Cultural Pragmatics. The Semantics of Human Interaction. Berlin/New York: Mouton de Gruyter. Wilss, W. (1982). The Science of Translation: Problems and Methods. Tübingen: Narr. Wilss, W. (1996). Übersetzungsunterricht. Tübingen: Narr. Wolfson, N. (1989). Problems in the Comparison of Speech Acts Across Cultures. In S. Blum-Kulka, J. House, & G. Kasper (Eds.), Cross-Cultural Pragmatics: Requests and Apologies (pp. 174–196). Norwood: Ablex.
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.1 (373)
The contrasts between contrasters What discussion groups can tell us about discourse pragmatics* Piibi-Kai Kivik and Krista Vogelberg University of Tartu
The paper presents the results of a contrastive study of multi-party talk in Estonian and American English. The data come from two formally comparable speech events: extended sessions of student group discussions geared towards preparation for a task set by the experimenter (where the task consists in giving advice on problems of Estonian-American intercultural communication). The general theoretical framework encompasses contrastive conversation/discourse analysis and the pragmatics of conversation structure. The central analytical concepts are those of frame and framing, where framing is understood as a further level of dynamic interpretation and production of a situation broadly delimited by frame as a relatively static given. The three main ‘given’ frames – those of experiment, task and informal conversation – are found to be not only differentially foregrounded by the two groups but also subjected to differential further interpretation/framing oriented to different goals. In particular, the American group frames informal conversation as personalised story-telling with the goal of yielding good individual performances, whereas the Estonian group enters into a non-competitive in-group debate where personal experiences become generalised and abstracted. While the characteristic features of the American group’s talk are typical narrative structures, the Estonian group engages in extensive co-construction of argument. Analysis of conversation-structural and pragmatic correlates of the framings does not uphold straightforward American-Estonian cross-cultural generalisations based on earlier research. Rather, frames and framings in their complex interplay emerge as the central determining factors for choosing interactional strategies, with the picture further compounded by bidirectionality between choice of strategies and framing. However, the wider context of culture can still be shown to be operant in determin-
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.2 (374)
Piibi-Kai Kivik and Krista Vogelberg
ing parameters of talk, though indirectly, via its formative influence on processes of framing.
.
Background and objectives
The approach employed in the current paper derives from Contrastive Conversation Analysis (Maynard 1990, 1997) and cross-cultural studies of framing of speech events, e.g., in discussion groups (Watanabe 1993) and oral proficiency interviews (e.g., Ross 1998). All of these projects focussed on features of American and Japanese conversation (conversation is used here in the wider meaning of talkin-interaction), and traced the observed differences in framing to a wider cultural context. While Maynard’s basis for contrasts are aspects of conversation management, such as listener feedback, Watanabe detects culturally conditioned frames of a speech event (discussion group), as evidenced by more global discourse features, such as opening or closing discussion, and framing of particular communicative acts (Watanabe 1993: 181). The notion of frame has been defined in different traditions and with a variable degree of precision (see e.g., Tannen & Wallat 1993; Tannen 1993).1 Here, ‘frame’ is used to mean the participants’ interpretation of the ongoing speech event as evident in linguistic and contextual cues, following Bateson’s (1972) identification of the “metacommunicative level of communication”. However, we find it necessary to further distinguish between the broad idea of a frame in the socio-cultural, speech-event interpretation sense, which (even in Watanabe’s culture-dependent sense) is relatively static and based on expectations, on the one hand, and the idea of a frame as a ‘superordinate message’ about what is really meant, how the broad frame is further interpreted in interaction, on the other (i.e., to distinguish between frame as in ‘interview frame’ and as in ‘joking frame’). Shea (1994: 362) suggests shifting “the understanding of participation structure from the analysis of speaker positions toward a more dynamic and interactive understanding of speaker positioning, from the analysis of structure toward an understanding of structuring”. In a similar manner, we are going to use ‘framing’ as a term denoting a further level of interpretation and signalling within a more general frame of a speech situation and expectations about it. Goodwin and Duranti (1992: 5) name as one of the aspects of context dynamics “the ability of participants to rapidly invoke within the talk of the moment alternative contextual frames”. Still another aspect of “situated discourse” (cf. Gu 1999) which we consider relevant to framing is that participants collaboratively realise their goals.
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.3 (375)
The contrasts between contrasters
The present study of conversation in Estonian and (American) English is undertaken against the backdrop of recent work on contrastive aspects of Estonian and American communicative style, politeness and realisation of speech acts as well as conversational routines. Up to now, anecdotal information from intercultural encounters has mainly been supplemented by observational, questionnaire and interview data. Vogelberg (1997) analysed the message-construction strategies of Estonians as interviewees in gate-keeping interviews, and found a strong preference for negative politeness and adherence to the Principle of Modesty, conflicting with the positive politeness of the North American gate-keepers. Kivik’s (1998) questionnaire and interview study of conversational style suggested that Estonians’ beliefs about conversational behaviour differ from those of their North American peers, e.g., they are more tolerant of pauses and silence in conversation and less likely to initiate interaction. The results again indicate Estonian preference for negative/independence interactional strategies. Several student diploma theses supervised by the authors have employed questionnaire and DCT data to compare conversational routines as well as realisation of speech acts in Estonian and English. Preliminary results from those studies have pointed to culture/situation interaction in the distribution and frequencies of speech act strategies, less routinisation and more variation in the Estonian samples (see also Vogelberg 2000). At this point, there is a pronounced need for analysis of naturally occurring conversations in both languages. While English-language conversation has been studied in detail, these data are often not suitable for comparative purposes, since comparable data should be obtained from at least formally comparable speechevents (cf. also Maynard 1990). The present study is the first attempt to elicit and analyse such data. Systematic study of spoken Estonian, based on a corpus of naturally occurring interactions, is still at an early stage (e.g., Hennoste 2000). Thus a database for comparative and cross-cultural study is only starting to emerge. In cross-cultural and interlanguage pragmatics research, Boxer (1995), Bardovi-Harlig (1999), a.o., have recommended ethnographic interviews to validate the existing, or help to set up new, hypotheses, as well as for a triangulation of conversational and/or questionnaire results. Interview data regarding communication between North Americans and Estonians could reveal some of the more salient contrasts in conversational style and interpersonal pragmatics. Those contrasts would also help complement the description of Estonian as used in native-speaker interaction. Therefore, our project set out to obtain both ethnographic ‘interview’ data and natural conversation data in the form of audio-recorded group-discussion. The setup was designed with a view to recording natural conversation within the framework of a specific speech event. This paper will not be concerned with the interview data, and will focus on the structural aspects of discussion group talk. However, the two parts of the project interact. In particular, the speech event de-
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.4 (376)
Piibi-Kai Kivik and Krista Vogelberg
signed to elicit comments on cross-cultural communication also enabled specific, culture-based aspects of talk construction to emerge. The original objectives were to obtain contrastive data in English and Estonian on the construction of multi-party talk, with respect to pragmatic strategies and conversation structure and to analyse contrastively the frames and participant roles in discussion-group talk, determining the culture-specific discourse parameters of student discussion groups as a speech event. Meanwhile, the analysis of data led to results that in a way transcended the original objectives and suggested a crucial role of the process of framing in determining all the other parameters under study. The presence of cultural factors was found to be less straightforward than expected, revealed through the differential framing of the ‘given’ speech situation.
. Method and data . The design The research design benefited from Watanabe’s (1993) study of framing in discussion groups, Sun’s (1995) study of peer-review discussion groups, and Myers’s (1998) study of conversational aspects of focus-group discussion. Bardovi-Harlig and Hartford’s (1996) research into advising sessions was followed in choosing a speech-event that the participants would be familiar with. Two student discussion-groups (Estonians N = 7, North Americans N = 6)2 were set up, one at a Midwestern university at the U.S., the other at the University of Tartu, Estonia. The Estonian participants were all native speakers of Estonian and the Americans native speakers of North American English, the discussions took place in the native languages of the participants. Two researchers (the two authors) gave identical instructions to the participants, maximal equivalence in translation was aimed at. The participants were asked to meet for an informal discussion session after classes. The meetings took place in university offices; however, the atmosphere was casual, with coffee and snacks provided. The purpose of the session as described to the participants was to put together a panel discussion/instruction session for a group of visitors from the target culture (Estonian and the U.S. respectively) for a successful sojourn at their university. The talk was tape-recorded, the researchers were not present. The participants knew they were being recorded for the purpose of a cross-cultural study but did not know what aspect of their talk would be studied. They also knew the panel/instruction session was hypothetical, although they were told such a session (if successfully prepared) could well be organised, if not to meet its original purpose of informing visitors, then as a presentation at some event focussing on cross-cultural communication.
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.5 (377)
The contrasts between contrasters
After the meeting, informal interviews were conducted with the participants, and they were asked about their impressions of the discussion session.
. The participants The makeup of the groups was comparable, including graduate as well as senior undergraduate students. The participants were recruited based on their contacts with the target culture (they had all spent some time in Estonia and/or Finland3 or the U.S. respectively) and came from different university departments. Both groups included one ‘informed participant’. These were linguistics students, who had a very general idea of the researchers’ interests but were not aware of the exact purpose of the experiment. Both groups included people who were familiar with each other, and those that met for the first time. A significant difference in the otherwise comparable situations was the fact that the researchers were both Estonians, which could have made the task less facethreatening for the Estonian group, while the Americans were requested to reveal their perception of another speech community to a representative of that speech community. However, the face-threat for the Americans was minimised on the one hand, by the fact that the researcher was not present at the recording (which, we believe, resulted in more direct expression of opinions, as evidenced, in particular, in remarks potentially face-threatening to the researcher, and not likely to have surfaced in her presence). On the other hand, the relative status of the researchers is believed to have counterbalanced the adverse effect of the Americans working with a researcher who was not their in-group member. The researcher collecting data in Estonia was a professor from among the most senior teaching staff the students have contact with while the Americans interacted with a graduate assistant of roughly the same age and status as the participants. Thus while the Estonian group was most likely more relaxed due to the absence of face-threat to an outgroupmember, they may have been more constrained by the relative power-distance. In addition, considering the length of the discussion sessions (70 minutes each) the contaminating influence of recording or research situation on conversational behaviour is not likely to be significant. There was another discrepancy in the participant structure: all participants were female except for one male in the American group. However, it was not possible to determine any gender-effects in this study, as the male participant was also the one with most topic knowledge.
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.6 (378)
Piibi-Kai Kivik and Krista Vogelberg
. Results The instructions given to the participants were formally identical. On the whole, the event was presented by the researchers as a group-discussion in preparation for a specific task. Still, the exact framing of the talk had to be negotiated by the participants. The three broad frames that can be identified in the talk of both groups are those of task, free conversation and experiment-participation. Those frames directly derive from the speech situation, and are thus ‘given’. The speakers move between those three frames, and the frames have different proportions in the two groups. Yet a more detailed analysis reveals that the frames are differently realised in different groups, enabling us to speak of further framing of the frames. The presentation of results will be structured here according to the three given frames as they can be traced in the transcripts of the discussion groups as well as the follow-up interviews with the participants.
. The experiment frame The experiment frame, i.e., the speakers’ understanding of the event as an experiment situation and their perception of themselves as participants in an experiment/research project influenced the temporal characteristics of speech, especially in the Estonian group. As reported later in the follow-ups, the participants felt more pressed to talk because of the recording. The American group was more concerned with the outcome of the experiment, whereas for the Estonians the experiment-situation elicited jokes and self-ironic comments. As could be expected, the experiment frame surfaces more in the opening and closing sequences of the discussion. Excerpt 1.
Estonian
Opening of the discussion. 1 Inge:
ma arvan / et meil läheb ikka mitu kassetti I think it will sure take us several tapes 2 Mall: arvad või? vaevalt küll You think so? not likely 3 Liisa: saaks kohvi äkki? could I get coffee perhaps? 4 Mall: jah / palun yes please 5 Ellu: kohe saab in a minute (sounds of pouring coffee)
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.7 (379)
The contrasts between contrasters
6 Inge: pärast on ainult kohvijoomine seal afterwards it will be only coffee-drinking there 7 Piia: terve lindi täis kohvijoomist a whole tape full of coffee-drinking 8 Mall: no-nii/ hakkab peale so-well / it starts /. . . / 15 Ellu: läheks nüüd asja juurde ka või/ millal te siis nagu Ameerikas käisite? what about getting to the point or what/ when did you then like go to America? 16 Inga: mida me ameeriklastest üldse arvame what do we think about the Americans in general (. . . ) 17 Ellu: xxxx nüüd algab see tüüpiline eestlaste vaikus now this typical Estonian silence is starting 18 Mall: kas on juba tüüpiline aeg eestlaste vaikust või võibis it already a typical time of Estonian silence or can we 19 Jane: = minu meelest = I think 20 Kai: enam-vähem tüüpiline more or less typical The pre-discussion talk includes getting coffee, and joking about being recorded (1–2), (6–7). The opening to discussion proper (15) is emphatically low-key, Ellu employs ka või (‘or what?’) and nagu (a common hedge, Hennoste 2000), both characteristic of colloquial speech. The opening statement would sound impolite if used outside the peer group. Right from the opening sequence, positive politeness is employed for in-group marking, expressed in low key, use of constructions typical to the colloquial speech of students, self-irony in (17–20). Already in the opening sequence we see repetitions (in 17–20), which are typical of this group’s talk. The American group’s opening also includes a pre-discussion preparatory sequence, but they are more explicitly interested in the purpose and the outcome of the experiment already in the opening phase. Excerpt 2.
American
1 Ann: I ‘m gonna / I kind of /cause I like work with xxxxx/ and I need to ask you for one second only/ because I know it’s really hard to hear this transcript 2 Dee: oh
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.8 (380)
Piibi-Kai Kivik and Krista Vogelberg
3 Ann: so it’d be better /if you try to talk uhm /try not to overlap or talk really loudly/ otherwise like this ain’t gonna do like anything good huhuh 4 Dee: hmmm 5 Ann: soo: and then like / do it again or something huhuh (. . . ) /. . . / 11 May: So // does anyone have any ideas // of how to organise a presentation? 12 Dee: Ann / do you focus on that in your studies?/ do you focus on communication studies? In (1), Ann, the informed participant, explicitly comments on recording, instructing others to speak clearly. This reinforces the idea that recording the content of the talk seems to be the most relevant aspect of the experiment, and that the participants see themselves as responsible for providing the necessary content information. Such responsibility and concern for the result is expressed by the Estonian group as well, yet in a joking manner: ‘a whole tapeful of coffee-drinking’ (7). The Americans’ opening proper here focuses narrowly on the task, ‘how to organise the presentation’ (11). They are considerably more realistic and serious about the task. Closings of the discussion (initiated by the researcher’s return to the room) mirrored the openings. Upon the arrival of the researcher and her question how the group was doing, the Estonian group jokingly replied that they were ready but there was still room on the tape (parallel to the opening discussion where they joke that they will need more tapes, 1: 1). The Americans said that they still had no structure for the presentation, i.e. they had not been able to ‘complete the task’, parallel to their opening concern with ‘how to organise the presentation’. We could say that the American experiment-frame actually constituted a task-frame.
. The task frame This frame is expressed first, in ‘task preparation frame’, i.e., explicit negotiation of the task (preparation of the instructions/recommendations) and second, in ‘taskrealisation’, i.e., ‘self-reported’ or ‘projected’ speech, the potential future utterances while carrying out the task giving instructions or suggestions, often introduced by a ‘reporting’ clause such as: “they should be told that”, “we should let them know that”. Task-preparation is present in both groups. It is realised early in the session as ‘negotiation of the task’, where participants clarify what they will have to do. All through the discussion, task-preparation is cued by change of register, as formal/academic register is used to sum up ideas colloquially expressed in previous turns. The register is characterised by careful choice of lexis, well-formed sentences, lack of overlaps or interruptions.
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.9 (381)
The contrasts between contrasters
The American group spends more time explicitly negotiating the task. Excerpt 3.
American
1 Tom: there’s something I experienced /that might not have anything to do with it / but when I was in Peace Corps they did some / they acted out some/ things/ to show like the differences/ they got the Estonian instructors to act like the/ Americans/ and they got the Americans to act like Estonians/ and they did a skit/ so // I don’t know if that/ sparks anybody’s ideas 2 Dee: oo:h 3 Tom: like what sorts of ideas would you show 4 Ann: stereotypes 5 Lyn: would you want to write that up/ to do a series of workshops and lectures? like / maybe we want to have /like someone acting in such a lecture 6 Tom: well if it’s getting to be a really elaborate thing /we might do that/ though I guess she really didn’t specify/ whether it was some weeklong immersion training / or was it a thirty-minute workshop 7 Fran: it says here a panel discussion /which would lead me to think of one hour/ one hour 8 Ann: yes This example of task discussion involves first, typical negative politeness strategies: in (1) Tom’s ‘pessimism’ about his contribution, in (5) Lyn’s, (6) Tom’s and (7) Fran’s indirectness and hesitation; and second, appeal to group-membership (“we”). The following excerpt demonstrates a contrast between the registers. Excerpt 4.
American
1 Tom: I mean/ when she said this/ I wrote some things down/like I was immediately reminded of/ immediately when I came back/after being /being in Estonia for two years /and going going in the store/ and coming like this close to somebody /and they say excuse me!/and I was like/ I’m in their way or something/ and nobody ever said excuse me in Estonia/ and here/ it has to do with / Americans have this really xxxx idea of personal space/ and that’s one thing/ and like if an Estonian comes here / they might be warned to stand further away from people/ if they talk to you /and also when you walk excuse me excuse me /I’m in your way / and I’m not even in your way 2 Fran: or / you’re in my way huhuhuh 3 Tom: well yeah / and there’s other stuff too /
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.10 (382)
Piibi-Kai Kivik and Krista Vogelberg
4 May: so uhm what you’re saying/ is that/in Estonia / um people /erm stand close to each other/ or pass closer to each other /without feeling the need to apologize = 5 Tom: = right 6 May: whereas here [they apologize /more/ OK 7 Tom: [coming to twelve feet they’re like ohh May’s lines (4) and (6) summarise Tom’s turn. While Tom’s style (1, 3, 7) is colloquial, May carefully re-words and sums up his point. This would not be usual in a natural relaxed conversation, but is a feature of task-oriented discussion. A similar feature can be observed in the Estonian group. Excerpt 5.
Estonian
1 Mall: nii et / okei välimus aga-/ mis sa tahtsid lisada? so well/ OK/ appearance/ but-/what did you want to add? (Did you want to elaborate on this?) The question in (5: 1), although informal (as the informal second person singular pronoun is used), is not typical of a conversation among friends, but of a discussion, because of the word lisada – a formal word for ‘add’. In the following, a confusion about the nature of the task is clarified. Note Mall’s careful formulation of the task in (9). Excerpt 6. 1 Ellu: 2 Mall: 3 Ellu: 7 Mall: 8 Inga: 9 Mall:
Estonian. kas me peame rääkima/ et millised eestlased on pigem/ või do we have to talk about / what the Estonians are like rather/or jah yes mmhmh jah yes no teeme siis niimoodi et eestlased = well let’s do it this way then that Estonians = = ja millised on eestlaste-ameeriklaste erinevused / mille põhjal võiksid and what the differences are between the Estonians and Americans/on tekkida probleemid/ selles mõttes / /no nii/ välimus// sul on? the basis of which problems could emerge/ in this sense/ so-well/ appearance// you have?
More significant differences between the two groups emerge in the second aspect of task-frame, which can be described as envisaged task-realisation. Both groups
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.11 (383)
The contrasts between contrasters
use ‘projected speech’ for purpose, however it is further framed completely differently. ‘Projected speech’ of the American group is often realised as bald-onrecord instructions, even bald imperatives, while unlike in the Estonian group, bald-on-record strategies are never used in talk directed at other participants. In Excerpt 7 Dee is using direct imperatives, her intonation contour is unmarked, just presenting a list of successive recommendations or instructions. She can be seen as offering some mitigation with the joke at the end, but at the same time she is presenting her behaviour as a potential model to follow. Excerpt 7.
American
1 Dee: well maybe/ well I was just thinking that /of saying that/ err on the more formal side/ or if you know you think something/ might be offensive don’t say it//that brings also up the issue also well/ if you think something may be inappropriate/ or would be better to say/ ask the person//well that’s what I do/ I mean when I’m in a country/ usually I’m not swearing all over the places/ 2 All: (laughter) To word the potential advice, positive and negative politeness strategies were used in reporting clauses (task preparation, addressing other group-members) and absent from ‘reported’ direct speech (task realisation), e.g., “if an Estonian comes here they might be warned to stand further away from people”, “that is something Americans do all the time /. . . / it is your obligation to say. . . ” (appeal to norms), “I would advise Estonians/ that if you need directions / then it is OK to ask a policeman/ as they probably know”. For the Estonians, quite differently, this aspect of the task frame was further framed as a ‘mock4 task frame’, references to the task were made jokingly, using exaggerated and stylised imagery, intonation, and lexis. Excerpt 8.
Estonian
1 Mall: ja-jah/ okei/ no et välimus on nagu siis selge / et muidu samasugused nagu teie/ ainult riietuvad palju stiilsemalt ja paremini MAR/ OK /FIL, appearance is then like clear/ that otherwise the same (nom:pl) as you (nom:pl) / only (they) dress (3pl:pres) much more stylishly and better 2 Jane: ikka sure 3 Kai: see oleks meiepoolne sõnum neile that would be our message to them In Excerpt 8 the second person plural deictic teie ‘you’ is used to denote Americans as if in a direct address, and third person to denote Estonians. This ‘direct quote’
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.12 (384)
Piibi-Kai Kivik and Krista Vogelberg
is confusing even to the participants, and clarification is provided by Kai in (3), where meie ‘our’ denotes the discussion group and neile ‘to them’ the potential American instructee. Person deixis may be seen here as corresponding to the abovementioned aspects of the task frame: task realisation (1, Mall) and preparation (3, Kai). Mall (1) is clearly self-ironic, following the segment of discussion where the Estonian group notes that Americans are more relaxed about clothing than Estonians, and may be looked down upon in Estonia for that reason. The following example is another self-critical remark following the observation that American small-talk and politeness routines may be taken literally by Estonians, and thus result in miscommunication. Excerpt 9.
Estonian
1 Mall: nii et/ ütleks neile kohe nagu ära et/ “ärge tehke tühje lubadusi!” (mock chastising intonation) SO THAT SAY (cond) TO THEM RIGHT (MAR) AWAY THAT DO NOT MAKE (2:pl) EMPTY PROMISES so let’s tell them like right away “do not make empty promises!” Note the hesitations (use of conditional verb form and hesitation marker nagu) in the ‘task preparation’ and direct imperative in the ‘task realisation’. However, the intonation cue is crucial here, voice quality in the suggested direct ‘instruction’ is changed and emphasis exaggerated with a humorous intent to make the utterance sound like an admonition. While the ‘task realisation’ utterances have similar structure in both groups, e.g., bald imperatives, their meaning and function are very different. The Americans’ instructions are realistic and could be well uttered in the same form at an instruction session. The Estonians’ instructions are joking, and are clearly not meant to be used in the real situation, as they would constitute a serious face threat to the hearers. A typical exchange pattern in the Estonian group is an introduction of some idea by the participants, followed by a more formal summary or re-phrasal (usually by the informed participant), which is immediately followed by another joking remark as ‘downtoner’. As a rule, in such cases the preceding discourse has portrayed Americans as somehow not conforming to the Estonian norms, and sometimes outright criticism has emerged. In the next excerpt (10) Kai and Mall are jointly summarising the idea that Americans may dress inappropriately for the Estonian eye, and that they should be warned about possible reactions. Mall in (2) is formal, her word choice is not that of a free conversation. This is immediately ‘challenged’ or counterbalanced by Kai’s colloquial completion of Mall’s utterance in (3), before Mall can finish her utterance. Just the suggestion that Estonians might not accept Americans in
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.13 (385)
The contrasts between contrasters
society because of the way they are dressed is obviously a little too much for Kai to be taken seriously, and she cuts in with a mildly ironic comment, her agreement marker ja-jah corresponds here to the American ‘yeah, right’. Excerpt 10.
Estonian
1 Kai:
et see ameeriklane saaks siis aru/ et ei ole mitte miskit muud kui riietus that the American would then understand/ that it is nothing else but clothing 2 Mall: et mitte et me ei aktsepteeri teda ühiskonnas/ vaidthat it is not that we are not accepting her in the society/ but3 Kai: = ja-jah, et puhtalt miniseelik MAR: AGR that it is all about the mini-skirt Thus, in the Estonian group the task-realisation frame is without an exception realised in a ‘mock task’ frame, and unlike the Americans they do not produce a single realistic instruction.
. Free conversation The third frame, which is actually the main frame in the discussion, can be called free conversation, as this is the part of talk where no reference is made to either the task or the experiment, and the participants engage in relatively unconstrained conversational exchange of ideas. This given frame was realised very differently by the two groups, whose further framings structured the talk as ‘individual telling of stories’ and ‘in-group debate’ by the American and Estonian groups respectively. The differences are evident in numerous conversation-structural signals. For the Estonian group, the discussion takes the form of collaborative message construction to the extent that one complete fluent utterance is built up from turns by three different participants, and in places syntactic parallelism can be observed. There is almost no competition for floor, smooth transitions are characteristic. There appear to be a number of transition-relevant places within the turns, though not always overtly signalled by pauses. The following is an example of co-constructed talk typical of the Estonian discussion. Excerpt 11.
Estonian
1 Inga: tähendab põhiline on see nagu/ et kui sa ikka inimest tunned/ ja ja noh/ ütleme ta ei pruugi su sõber olla, sa temast mööda ei jaluta niisama it means the main thing is this like/ that when you really know the person /and and FIL/ let’s say he does not have to be your friend/ you will not just walk by him just so. . .
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.14 (386)
Piibi-Kai Kivik and Krista Vogelberg
2 Ellu: 3 Inga: 4 Mall:
5 Jane: 6 Kai:
7 All: 8 Mall:
9 All: 10 Mall: 11 All:
jah/ täpselt niimoodi yes/ exactly like this sa lehvitad ja naeratad / ja midagi ikka teed you wink and smile/ and do something still = tervitad /nagu selles mõttes see ei ole nagu jah mõtteline küsimus/ et “how are you” (you) greet/ like in this sense that it is not really an information question/ that “how are you” jah yeah aga see on mingi selline pealiskaudne viisakus / mis tegelikult eestlastel BUT THIS IS SOME SUCH SUPERFICIAL POLITENESS/ THAT ACTUALLY ESTONIANS (ade) iseenesest puudub, meil on pigem see /et prill ette ja keeran kähku nurga tahaIN ITSELF LACK (3sg:pres) WE (ade) IS RATHER THIS /THAT GLASS(ES) (nom:sg) FRONT (ill) AND TURN (1sg:pres) QUICKLY CORNER (gen:sg) BEHIND (ill) but this is a sort of such superficial politeness/ which actually Estonians do not have/ we have rather this / that [I’ll put] glasses on and quickly turn behind the corner [ (laughter) [ = või siis pilk maha OR THEN GLANCE (nom:sg) DOWN (ill) or then look down (laughter) või siis pilk maha/ kui ei anna ära joosta or then look down /if there is no way of running away (laughter)
Turns in (1), (4) and (6) and (8), and (11) can be connected to form a single utterance thanks to the parallel verb-forms used, and the connectives aga ‘but’ in (6) and või ‘or’ in (8). The verb form used by different participants in consecutive turns is second person singular (you), which is only one of the several structures that can be used to express a generalised situation impersonally, e.g., in (6) Kai switches to first person singular in the same function, and Mall (8) and (10) are elliptical with no verb. While the expressions in (6, Kai) and (8, 10, Mall): ‘prill ette’, ‘pilk maha’ are semantically related – both have to do with avoiding eye contact – they are also structurally similar: singular noun in the nominative case + locative particle in the illative case. Furthermore, there is even alliteration and assonance in ‘prill’ and ‘pilk’.
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.15 (387)
The contrasts between contrasters
‘Prill ette’ is actually a marked colloquial variant of ‘prillid ette’ ‘[put] on glasses’, here used for humorous effect, especially without a verb. ‘Pilk maha’ is more common as an expression, though it has a plural alternative and would normally be used with a verb. The ‘absolute construction’ as well as the particular form can be suggested to have been conditioned by the structure used in the previous turn. In co-constructing the message, speakers are tuned to each other’s talk to the extent of synchronising their linguistic choices. Another characteristic feature of the Estonian group’s talk is abundance of verbatim repetitions. These can be supportive moves or listener back-channels. Excerpt 12.
Estonian
1 Inga: kõik oleneb inimesest all depends on the person 2 Kai: kõik oleneb inimesest jah all depends on the person MAR:ACK Here Inga is mitigating the previous speaker’s bald on record disagreement with a point, and Kai joins her with supportive echo plus acknowledger. Excerpt 13.
Estonian
1 Piia:
oh mis need vanad asjad oh what these old things 2 Inga: Eestis on kõigil sellised in Estonia everyone has such ones 3 Piia: jaa Eestis on kõigil sellised yeah in Estonia everyone has such ones Here, Piia is supplying a prototypical Estonian compliment response (rejection), then Inga complements it with a possible continuation or an alternative rejection, and then Piia is providing what looks like a back-channel, acknowledging Inga’s example. In Excerpt 14 Kai is verbatim repeating the key phrase from Mall’s turn and adding endorsement + agreement markers. Excerpt 14.
Estonian
1 Mall: see vist oleneb sell-/ nagu noh ka et sellest/ et kui me räägime nüüd ameerikast/ siis kogemused vist “regionaalselt erinevad”(emphatic and very clear intonation mocking “academic” talk) It probably depends on thi-/ like FIL also that this/ that when we talk now about America/ then the experiences are probably regionally different
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.16 (388)
Piibi-Kai Kivik and Krista Vogelberg
2 Kai:
regionaalselt erinevad/ täiesti õige jah / ei vaidle üldse regionally different/ absolutely right MAR /(I) do not argue at all
Here Mall is displaying a self-ironic attitude to her position as an ‘informed participant’ by intonationally exaggerating her use of ‘academic’ vocabulary. Kai’s repetition can be seen as a true ‘cognitive repetition’, as she is concentrating on the meaning, or it can serve to enhance Mall’s face, as if saying ‘you need not worry, I would say the same thing’. The explicit performative negated (“I do not argue”) does not constitute a denial of presumed argument, but functions just as an agreement marker. For the Estonian group, this frame is also characterised by a number of baldon-record disagreements as well as a general scarcity of mitigating devices, which is not perceived as impoliteness but in this frame serves as a means for in-groupbuilding. The disagreements do not lead to conflicts, but rather to elucidation of problems discussed, no participant returns to her contested argument to defend it. Excerpt 15.
Estonian
1 Mall: ei tegelikult/ kui sa teed midagi nende heaks teed/ hindavad / väike asi aga aega ja panust iseenesest/ meil nagu võtad iseeenestmõistetavana/ kui aitad no really/ if you do something for them/ they appreciate it/ a small thing/ but the time and effort/ here you like often take it as selfevident/ if you help 2 Liisa: ega nad seal ei ole ikka midagi nii abivalmis /eestlased on hulka enam minu meelest they are really not at all that willing to help there/ Estonians are much more I think 3 Mall: võib-olla perhaps Excerpt 16.
Estonian
1 Mall: et kui nad tulevad/ siis hoiatame neid/ et huhuh politsei poole ärge pöörduge that when they come/we will warn them/that huhuh do not turn to police 2 Inga: võib-olla perhaps 3 Kai: mina ei tea/ minul pole meie politseile küll midagi ette heita I don’t know/ I have nothing bad at all to say about our police
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.17 (389)
The contrasts between contrasters
4 Piia: 5 All:
ei jah/ ega nad küll eriti efektiivsed ei ole/ aga noh MAR/ they are not especially efficient /but well (laughter)
Excerpt 17.
Estonian
1 Kai:
elementaarne viisakus on hakanud tulema/avalikus kohas ikka piipu ette ei pane elementary politeness has started to appear/ in public places you really do not light your pipe 2 Liisa: mina vaidlen vastu /kui silti ei ole/ nagu Pronto on suitsusaun/ kui mul ikka kõrvallauas ja taga suitsetatakse/ tuhatoos on laua peal/ miski ei ütle mulle /et minu kohus oleks mitte suitsetada / ka minu elementaarne viisakus mitte I disagree [contradict, lit: I argue against (your point)]/ if there is no sign/ like Pronto is a smoke-sauna/ well when they are smoking in the next table and behind me/ ash-tray is on the table/ nothing tells me / that I should be obliged not to smoke/ even not my elementary politeness 3 Kai: oleneb inimesest it depends on the person 4 Mall: ja kohast and place
These excerpts exemplify a typical pattern of bald-on-record disagreements, in Excerpt (17: 2) there is even an explicit performative (mina vaidlen vastu ‘I oppose/argue against your point’), which are then followed by a mitigation offered by the next speaker. The mitigation can be either a hesitation, in (15) (‘perhaps’), an agreement/ acknowledgement, (16), or a disclaimer (17). There is never a further disagreement, nor any arguments developed. In contrast, the characteristic strategy of the Americans in free-conversation frame was story-telling to illustrate their point. Although they also engaged in more abstract discussion and brought brief examples like the Estonians did, there was a total of nine personal experience narratives of an average length of 160 words (with the characteristic features outlined in Polanyi 1985) as compared to only one similar structure (71 words) in the Estonian group. In addition to being longer, the American narratives were also richer in details and more ‘performative’, i.e. the story-tellers employed direct speech and different voice-quality to ‘perform’ the roles of the characters in the stories (cf. Couper-Kuhlen 1999), while the Estonians were using figurative language, colourful metaphors or comparisons to render the situation.
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.18 (390)
Piibi-Kai Kivik and Krista Vogelberg
Feedback in the American group was mostly in the form of acknowledgement and continuer markers, no repetitions were observed. There was a notable competition for floor, testified to by overlaps and latchings. Often this resulted in simultaneous development of two related, but still separate topics. Excerpt 18.
American
1 Tom: yeah that might not be really everyone’s daily experience/ but yeah -/ [well it never happens in Estonia huhuhhh 2 Ann: [well I think that Finns3 Tom: = rarely people would say hi to you no4 Ann: = yeah, Finns freak out if strangers say anything to them/ I remember. . . Here Tom is still in the middle of telling about Estonians when Ann in (2) cuts in and starts talking about Finns, but Tom is reluctant to give up the floor and finishes his utterance in (3), whereas Ann goes on with hers in (4), where yeah seems to be a continuer for her own line in (2). Excerpt 19.
American
1 Lyn: 2 Tom:
there’s some kind of a weird= well /yeah/ I had a similar experience when I was in in/ Seville Spain with a Finnish girlfriend and we were in a bar/ and it was almost the size of this room/ and there were two Finnish women/ in their flamenco dresses/ which was also strange3 Ann: tchhh 4 Tom: [and she said 5 Fran: [yeah big trend 6 Tom: yeah/ and she said oh those women are speaking Finnish/ and I said /oh really? are you goin’ to talk to them/ and she said no: are you kidding/and I said why not /for chrissake you know7 others: [laughter 8 Ann: = [‘coz an American would 9 Tom: = yeah that’s10 Ann: = where you fro:::m/ I’m from Texa::s huhuh 11 Tom: well /she’s like no!/ and then later she did though/ but it was not the first thought/ that she would have to go over and talk to them This is an example of a story told by Tom in (2) , (6) and (11) as he grabs the floor from Lyn. Ann and Fran in (3) and (5) is a side-sequence commenting on flamenco in Finland, and Ann’s (8) and (10) is an attempt to insert a comment + example into Tom’s story, but apart from a brief response in (9), Tom returns to his story as it was cut off in (6) and finishes it in (11) undisturbed.
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.19 (391)
The contrasts between contrasters
The Americans used negative politeness strategies to introduce a topic or to suggest an idea. They typically displayed pessimism about their contribution. Although group-membership was enhanced with positive politeness in the conversation frame, the negative politeness strategies used (hesitations, pessimism) also served the face of other group-members, not imposing one’s own views. Excerpt 20.
American
1 Dee: that’s something that happened last year in Estonia/ really anyone/ I mean // I don’t know if this is true/ and if anyone else has experienced this/ but I would come into a group situation/ and I wouldn’t be introduced/ and it really bothered me/and //I was like/ why //am I not being introduced?/ I mean I didn’t know anyone’s name Excerpt 21.
American
(. . . 3.9) 1 Tom:
there’s / kind of change of subject ( under breath)/ there was mention earlier of handshakes/ and maybe it’s just from a male perspective/ I don’t know but/ again/ me in the US/ maybe I’m just strange 2 others: (laughter) 3 Tom: but what I’ve noticed is In the conversation frame, face-threat was obviously greater in the American group, as any disagreements are definitely avoided. There is only one disagreement during the whole session, and this is introduced very carefully and backed up with a lengthy argument to justify the conflicting opinion. Excerpt 22.
American
(Here Dee is disagreeing with Tom’s suggestion that in the U. S. it is always safe to approach police-officers for help.) 1 Dee: well, you know what/ I have something to say about that /my // both my grandfather and my great-grandfather were policemen/ typical Irish cops/ they had REALLY bad things to say/ my grandfather told me / from when I was in sixth grade /like never ever ever stop for a cop /and when you get pulled over/ and you’re alone/ in a car/ keep on driving till you get to a public/ you get to a public spot and/ then get out/ this is like/ you know/ someone2 Tom: = this is what they tell you in Argentina too 3 Lyn: really? 4 Dee: well this is / he was a New York City cop /and I grew up in New Jersey /and he said you can never trust any cops/ so that was someone/ who was actually/ you know/ also a policeman for a while/ saying that too/ only pull over in public spots
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.20 (392)
Piibi-Kai Kivik and Krista Vogelberg
The Americans, like Estonians also display positive politeness strategies, such as joking and explicit in-group marking, mostly by the use of person deictics. However, an interesting feature associated with storytelling that is present in the American group but not in the Estonian one, is compliments, which reveal that stories told by other participants are evaluated as “performances”. For example: Excerpt 23.
American
(Following a personal experience narrative by Tom) 9 Dee: ohhhh OK 10 Tom: that/ I wouldn’t have believed this /if I had not seen it 11 Ann: that is fun 12 all: (overlapping/laughter) scary The Estonian group would agree or disagree with statements by others, but would not evaluate others explicitly. Jokes are responded to with laughter, but no comment is made. The American group also employs characteristic ‘voices’, marked intonation to report speech by others (cf. Couper-Kuhlen 1999). While Estonians mostly vary their voice quality and stress patterns in the task frame to denote ‘projected’ speech by themselves, reported ‘voices’ play an important role in the Americans’ narratives, most of which portray some event where speech is crucial.
. Discussion This project was originally designed as an initial stage of a cross-cultural study, contrasting pragmatic and structural aspects of Estonian and North American conversation. Based on earlier research by the authors, the results were envisaged to testify to several contrasts between the two (roughly outlined) cultures with respect to politeness strategies, listener response, turn-taking, etc. Estonians have been portrayed as generally negative-politeness oriented, reluctant to reveal their opinions, not too skilled in small talk but sometimes too direct. The identically presented speech event was hoped to draw forcefully the expected contrasts in natural talk. Although no cross-cultural generalisations could be suggested based on this limited comparison only, we hoped to identify guidelines for a more extensive cross-cultural study. However, during the analysis of data it became obvious no such direct clearcut comparisons can be made, and probably should not be made. What we believe emerged from our data as the central determining force for most pragmatic choices seems to be how participants frame the situation for themselves. Framing is seen as the participants’ interpretation of situation, foregrounding the aspects of situa-
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.21 (393)
The contrasts between contrasters
tion and context that are most relevant for achieving the goals in interaction at a particular moment. While as analysts we can identify frames on the basis of linguistic features, the determining forces are bi-directional in real-life interaction as well. Once certain linguistic features have been selected and used, these establish a frame, which may in the end define the whole event for all participants. The group differences we noted consisted in both frame proportion and frame realisation in further framing. While all three broad or ‘given’ frames were present during the whole speech event, they were differently foregrounded. The experiment frame is more marked for the Estonian group than the American one, since being recorded is much less common in Estonia. Also, while the settings mirrored each other, it is not usual in Estonia to serve snacks and coffee to students in the context of an academic discussion or meeting, while it is common practice in the U.S. The physical context of the experiment clearly disturbed the Estonian participants more. The Estonian group resolved the self-consciousness and concern for the success of the experiment with joking. At first, their attitude may appear casual, but the self-ironical comments and joking reveal uneasiness about the situation. Its function is face-saving by anticipatory self-criticism, and the source of face-threat in the experiment frame is the very situation of having to participate. It is suggested that the joking and emphatically casual style of the Estonians served both to save their face in the experiment situation from the face-threat originating from the situation (researcher) and to create in-group feeling to neutralise potential face-threat from other group-members. Note, however, that once joking had been chosen as a strategy, this created a conversation-internal factor that ‘set the tone’ for the rest of the discussion. In general, for the American group, the task frame was more significant than for the Estonian group. The reason could lie in the fact that student group-work is far more common and thus more institutionalised in the U.S. education system than it is in Estonia, at least in the form of assessed group-work. This might have activated the frame of being assessed for performance as a group-member for the American group participants even if they were not concentrating on their task explicitly, in other words, the ‘study’ frame may have been transferred to this ‘after classes discussion’. Both the task and the experiment frame were more institutionalised for the American group, and performance in these frames was important. Therefore, the group were concerned with the exact nature of their task, and also discussed the purpose of the experiment, which the Estonian group did not touch upon. Both the openings and closings of the discussions reflect the importance of the task frame for the Americans and the joking but slightly self-conscious perception of the experiment frame of the Estonians. Tannen (1993: 23) reports a similar observation in her framing study of Greeks and Americans retelling a film, where American par-
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.22 (394)
Piibi-Kai Kivik and Krista Vogelberg
ticipants, but not the Greek, were concerned with the purpose of the experiment in order to be able to ‘perform’ better. In the task-preparation frame we observed negotiation of the task and explicit discussion of its structure in the American group. They employed negative politeness strategies to put forward their ideas one at a time (cf. Watanabe’s (1993: 199) observation of the ‘single account’ strategy of the American discussion groups), in a way that students doing group-work would all contribute, but also leave room for the others’ contributions. At times this may have brought about competition for floor. As this group was task-oriented, they perceived face-threat in making an irrelevant contribution or a claim without enough data or facts to back up their point. At least some of the participants in this group had defined their goals as providing ‘insider’ views and ‘expert’ information on communicating with the target group. Thus there may have been a fear of not performing well as participants in the experiment, which was then addressed with hesitation and pessimism. The use of these strategies was made necessary by the underlying competitiveness present in the frame. As it can be suggested that for the Americans the task-preparation frame was associated with assessed group-work, this could be one of the factors that divided the group into separate individuals, precluded co-construction of talk and also accounts for the explicit evaluations of others’ performances. Indeed, the very presence of compliments signals that the participants are at some level assessing one another. The lack of such evaluation in the Estonian group where assessments pertain to ideas rather than performances corresponds to the general lack of competitiveness, and their participation as a group, rather than separate individuals. Additionally, the intercultural communication situation in the task preparation frame could have affected the discussion. In both groups negative politeness strategies were employed in the task frame, where they served to protect the face of the speaker, and the recipient in this case was the potential audience of representatives of the ‘target culture’ group. In both groups, when criticism of the target group surged, it was mitigated with self-critical or ironical comments, to save one’s own and the in-group’s positive and the recipients’ negative face. In the experiment frame the potential for face-threat stemmed from the researcher, who, although not physically present, was the recipient of the talk (Goffman 1981). The threat may have been greater in the American group as the researcher was from the target culture. The Americans were apprehensive of directly criticising Estonians; when they presented a potentially face-threatening opinion, they emphasised that it might have been just an isolated personal experience and displayed pessimism about their contribution. The second aspect of the task frame, that of ‘task realisation’, conjured up a more concrete, though imagined, intercultural communication situation, as the
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.23 (395)
The contrasts between contrasters
discussants pictured themselves as advice-givers in an ‘instruction session’ to representatives from the target culture. Here strategies differed notably. The American group did not seem to have any problems with giving instructions in the taskrealisation frame, and indeed, both in the U.S. and abroad such ‘training-sessions’ are fairly common. Their ‘projected speech’ included bald imperatives as well as redressed suggestions or then less direct hints, such as “this is what people do here” or “this is not done there” implying those are the norms visitors should follow. The Estonians, however, framed ‘task realisation’ jokingly, as a ‘mock task’ frame. The reasons suggested for the joking and ‘mock task’ frame include the very nature of the task. Here the perspective of Estonians and Americans differed. To give instructions and advice to Americans was somewhat unnatural for the Estonians, as in the real life the role-distribution would normally be different because of the common consultant (advisor, gate-keeper) role of the American. Thus advicegiving was perceived as face-threatening due to the implicit power-dimension, and was constantly downplayed through humour and irony, resulting in ‘instructions’ that would have been rude and hostile if uttered outside the framing. Often Estonians opted for avoidance, and topics of a particular cultural difference were concluded without deciding upon any suggestion to be made in the task frame. Therefore, it would actually be impossible to put together the imaginary advising session in Estonia, as practically no realistic advice was put forward. In the free conversation frame, in the Estonian group the task triggered discussion, in the form of debate. While this was not ‘argument as war’ type of heated debate, it was not just small talk to pass the time either. The participants were collaboratively developing, indeed co-constructing, topics, often abstract in the sense of dealing with contrasts between the two cultures on ideological or values level rather than just observed behaviour. They were not so much participating as individuals, but as ‘carriers of certain ideas’, the disagreements and competition happened on the level of ideas rather than individuals. Therefore, there was no need for redressive or mitigating strategies when disagreeing, as the whole group was working together to develop the debate, and the participants were relatively ‘faceless’, i.e. their faces were protected as a group-face by positive politeness (humour, irony, grotesque) employed to build in-group feeling, and negative politeness in the intercultural communication context. The main goal of this group, as they reported in the follow-up, was to enjoy a good intellectual conversation. Disagreements here seem to be accepted as a natural part of the process of ‘search for truth’, as a generative force for topic development, which does not threaten group harmony. Competition is not a feature of the Estonian discussion, and there is no need to prove one’s point at any cost, therefore we do not see the disagreeing participants return to their argument or make an effort to bolster their position. Mitigations that are supplied by the next speaker or disclaimers following a disagreement serve less the
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.24 (396)
Piibi-Kai Kivik and Krista Vogelberg
purpose of attending to face-needs than to develop or conclude the topic under discussion. The American group, in contrast, for most of the time took turns in telling personal experience stories of mismatch in norms for social interaction between the two cultures. These realisations of the conversation frame go back to the framing of the task (as an assessed groupwork) and that of the experiment. The goal for several American participants was to perform well in the task by contributing real facts, and producing useful advice. The analysis also revealed bidirectionality between choice of conversational strategies and the frame. Once several American group-members had started telling their experiences as sources of information on contrasts that they need to contribute in the experiment frame, and that serve as a basis for advice in the task frame, they had set the frame, and it was difficult to change it. Meanwhile, the telling of situated narratives, where context is not shared by the listeners, is a very ‘individualised’ activity. The story-teller presents him/herself as an individual and his/her contribution as a performance. Each group-member is participating as an individual with their own contribution. Naturally, their individual faces are vulnerable and need protection, as there is competition, and the contributions are assessed – either explicitly by the listeners or then implicitly in the evoked frame of group-work assignment. This, in particular, accounts for the separate threads of argument that are developed in parallel, but not jointly, and also for the abundance of face-work, both positive (compliments, group-markers) and negative (hesitations, pessimism). However, in the task realisation frame, the Americans perceived themselves as a more powerful group, thus less redressive action was needed. There is only one instance of disagreeing in the whole American discussion, where repetitive argumentation seems intended to forestall a potential refutation. Disagreement is very clearly a dispreferred strategy in the American group, as it constitutes a face-threat to the individuals presenting their personal experiences and emotions. A much used device for discourse cohesion for the Estonian group is repetition, often verbatim reproduction of a chunk from the previous turn. In the American group, indirect repetition is observable in the form of paraphrasing, e.g., in a move between frames. Paraphrasing is employed by the Estonian group as well, but then with a strong emphasis on different frame, with mock ‘academic’ register or ‘authoritative’ tone of voice. Sunakawa (2000) observed a similar contrast in the use of repetitions in Japanese and English advising discourse, with the possible cause of differential attitude towards the listener. The Japanese advice-givers were moving on a more interactional level, thus employing verbatim repetitions for involvement. whereas the Americans focused more on the message-level, providing explications and their view-points of what the advice-seeker had uttered. Another related reason we suggest for these repetitions is cognitive, an advice-giver who
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.25 (397)
The contrasts between contrasters
puts more effort into actually considering the problem before voicing a response ‘internalises‘ the speech she hears and then ‘externalises’ it again in the course of her own intensive thought process. In a sense, there is “a continual movement back and forth from thought to word, and from word to thought” (Vygotsky 1986: 218) not only intrapersonally but also interpersonally. Similarly, in our data the abundance of repetition points not necessarily to the fact that repetition would be a common Estonian back-channel, but rather to the framing of the situation by this particular group as demanding attention and mentally challenging, also, to the groupmembers being attuned to each other and connecting their speech very closely to the previous discourse. Myers (1998: 89) points out the homogeneity and a “sense of something shared” among the focus-group participants, which is created by the selection procedure, and enhanced by the moderator. In our discussion groups, this sense was initially created by the voluntary participation (the group was selected on the basis of shared experience) and constantly re-created by the participants. However, the means for signalling group-identity were differently employed by the Estonian and American groups. They both used deictic markers (‘we’, ‘we here’, ‘none of us’, ‘here’), back-channelling, expressions of agreement (‘I completely agree’, ‘yeah that’s right’, ‘that’s true’, ‘the same thing happened to me’) and assessments (‘yeah that’s a good point’, ‘that’s awesome’, ‘scary’; ‘exactly’), yet the American groupidentity was realised by acknowledging each other’s role as an individual contributor towards the task, the Estonian group emphasised group-harmony in the joint production of talk. A threat to group-harmony could spring from one participant being perceived as a group-leader. In the Estonian group, the leader-role was constantly downplayed by the ‘informed participant’, who used humour, self-irony and negative redressive strategies whenever she had to display her domain knowledge or assume a ‘moderator’ role. The latter aspect can be viewed as a conflict between the frames of experiment (the informed participant’s concern for the success of the experiment), task (concern for fulfilling the task) and conversation (maintaining good interpersonal relations through group harmony). Finally, we can look for reasons behind the particular framings observed in more general cultural differences. The educational system mirrors a wider cultural context. The Estonian higher education system – as indeed, the whole traditional culture of the country – although changing fast, is still not as competitive as that in the U.S. Student competition is not too much encouraged, neither in class nor in the system itself (e.g., there are no tuition fees in state institutions, no percentile assessments, competition for scholarships and in the job market are relatively new). Students, as a rule, do not perceive each other as competitors, while this is common in the U.S.
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.26 (398)
Piibi-Kai Kivik and Krista Vogelberg
. Conclusion The main conclusion of the study was that no direct clear-cut generalised crosscultural comparisons between conversation-structural and pragmatic choices can be made. Thus, contrary to predictions, our Estonian group did not use negative politeness in interaction with each other, instead, they used avoidance in the talk framed as giving instructions to the out-group. The Americans’ choice of strategies was almost the opposite in the two frames. The face-threat that calls for strategies originates not only from the P and D factors related to the immediate interlocutors, but also other potential recipients and originators of talk in the frame, as well as the goals set in the frame. In general, throughout our analysis frames and framings in their complex interplay emerged as the central determining factor for choosing interactional strategies. In this study, we attempted a distinction between the broader and more static notion of frame as ‘given’ by a speech event and its further framing as a more dynamic and interactional notion. The American and Estonian groups compared differed clearly both in the proportions of foregrounding the ‘given’ frames, and in their further framing as evidenced in the contrasts observed in the linguistic features of the group discussions. The reasons for the differences were partly conversation-internal, yet can also be traced to cultural differences. Among the latter, one can distinguish more particular ones, such as the differential presence of assessed group-work in American and Estonian higher education, and possibly more general cultural factors such as higher achievement-orientation and competitiveness in American mainstream society. Thus culture may still be operant in determining parameters of talk, yet indirectly, in the form of culture-situation interaction, mediated by the dynamic processes of framing. We hope the study contributes towards finding ways of combining the cross-cultural and contrastive linguistic approaches to discourse pragmatics.
Transcription and glossing conventions Italics Regular Underlined CAPS Bold / // (. . . )
original Estonian speech/sounds English translation or English original emphasis extra stress structure highlighted for analysis pause marking “chunks” within a turn a noticable pause within a turn a noticable pause between turns
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.27 (399)
The contrasts between contrasters
(laughter) [ – = ? ! e: xxxx MAR FIL
comments on intonation, non-verbal behaviour, etc. overlapping talk point of being cut short latching (no time between turns) intonation of a question exclamation lengthened vowel not clearly audible marker (AGR – agreement, ACK – acknowledge) filler
Morphological glossing (supplied only when directly relevant to the analysis) 1, 2, 3 person ADE adessive GEN genitive COND conditional ILL illative NOM nominative PL plural PRES present tense SG singular
Notes * The authors are listed alphabetically, their contribution is equal. The authors would like to extend their warmest thanks to all the participants in the study. The first-named author is also grateful to Dr. Beverly Hartford for reading an earlier discussion of the U.S. data, now incorporated into this paper. . More specifically, Tannen and Wallat (1993: 60) use an interactive notion of frame, i.e., “a definition of what is going on in interaction, without which no utterance (or movement or gesture) could be interpreted.” Frame here refers to a sense of what activity is being engaged in, how speakers mean what they say. Watanabe (1993: 178) proceeds from Goffman’s definition of frames as “principles of organization which govern events – at least social ones – and our subjective involvement in them”. Gumperz (1992: 232) considers framing, “understood in Goffman’s sense as signalling what is expected in the interaction at any stage, the third, most global of the three levels of which contextualization cues aid interpretation”. . From here on, ‘Estonians’ and ‘Americans’ in this paper denote solely the participants in these two groups. No generalisations to the respective populations are intended.
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.28 (400)
Piibi-Kai Kivik and Krista Vogelberg . For the American group, the target cultures were both Estonian and Finnish to increase the number of participants. As obtaining content information came only second in this experiment, the actual differences between Estonian and Finnish communicative styles were overlooked, and the two lumped together into a ‘cultural area’ for the purpose of this discussion. Research on Estonian and Finnish conversational strategies, e.g. Pajupuu (1997) has suggested several contrasts. However, we think the sharpness of these contrasts is levelled by a further comparison with North Americans. . ‘mock’ here used in the literal sense of ‘mockery’.
References Bardovi-Harlig, K. (1999). Researching method. Pragmatics and Language Learning, 9, 237– 264. Bardovi-Harlig, K., & Hartford, B. (1996). Input in an institutional setting. Studies in Second Language Acquisition, 18, 171–188. Bateson, G. (1972). Steps to an Ecology of Mind. New York: Ballantine Books. Boxer, D. (1995). Ethnographic interviewing as a research tool in speech act analysis: The case of complaints. In S. M. Gass & J. Neu (Eds.), Speech acts across cultures: challenges to communication in a second language (pp. 217–240). Berlin: Mouton de Gruyter. Couper-Kuhlen, E. (1999). Coherent voicing: On prosody in conversational reported speech. In W. Bublitz, U. Lenk, & E. Ventola (Eds.), Coherence In Spoken and Written Discourse. Amsterdam: John Benjamins. Goodwin, Ch., & Duranti, A. (1992). Rethinking context: Introduction. In A. Duranti & Ch. Goodwin (Eds.), Rethinking context: Language as an interactive phenomenon (pp. 1–42). Cambridge: Cambridge University Press. Goffman, E. (1981). Forms of Talk. Philadelphia: University of Pennsylvania Press. Gu, Y. (1999). Towards a model of situated discourse analysis. In K. Turner (Ed.), The semantics/pragmatics interface from different points of view (pp. 149–178). Oxford etc.: Elsevier. Gumperz, J. J. (1992). Contextualization and understanding. In A. Duranti & Ch. Goodwin (Eds.), Rethinking Context: Language as an interactive phenomenon (pp. 229–258). Cambridge: Cambridge University Press. Hennoste, T. (2000). Sissejuhatus suulisesse eesti keelde. [An Introduction to Spoken Estonian]. Akadeemia, 8, 1771–1806. Kivik, P.-K. (1998). What silence says: Communicative style and identity. Trames, 1, 66–90. Maynard, S. K. (1990). Conversation management in contrast. Journal of Pragmatics, 14, 397–412. Maynard, S. K. (1997). Analyzing interactional management in native/non-native English conversation: A case of listener response. International Review of Applied Linguistics: language teaching, Vol.XXXV/1, 37–59. Myers, G. (1998). Displaying opinions: Topics and disagreement in focus groups. Language in Society, 27, 85–111.
TSL[v.20020404] Prn:4/11/2002; 10:10
F: PBNS38.tex / p.29 (401)
The contrasts between contrasters
Pajupuu, H. (1997). Eestlased ja soomlased-probleemitud suhtlejad. [Estonians and Finnsunproblematic interactants] Keel ja Kirjandus, 8, 547–550. Polanyi, L. (1985). Conversational storytelling. In T. A. van Dijk (Ed.), Handbook of discourse analysis, Vol. 3. Ross, S. (1998). Divergent frame interpretations in language proficiency interview interaction. In R. Young & A. W. He (Eds.), Talking and testing: discourse approaches to the assessment of oral proficiency (pp. 333–353). Philadelphia: John Benjamins. Shea, D. (1994). Perspective and production: structuring conversational participation across cultural borders. Pragmatics, 4, 357–389. Sun, H. (1995). Pause and co-construction in Chinese peer review discussions. Pragmatics and Language Learning, 6, 121–142. Sunakawa, Chiho (2000). Repetition in advising discourse in Japanese and English. Paper presented at the Second International Conference of Semantics and Pragmatics, Cambridge. Tannen, D. (1993). Introduction. In D. Tannen (Ed.), Framing in discourse (pp. 3–14). Oxford: Oxford University Press. Tannen, D., & Wallat, C. (1993). Interactive frames and knowledge schemas in interaction: examples from a medical examination interview. In D. Tannen (Ed.), Framing in discourse (pp. 57–76). Oxford: Oxford University Press. Vogelberg, K. (1997). Message-construction strategies in the interlanguage of Estonian learners of English: the example of gate-keeping encounters. In L. Diaz & C. Perez (Eds.), Views on the acquisition and use of a second language. Barcelona: Universitat Pompeu Fabra. Vogelberg, K. (2000). Contrastive and interlanguage studies of the realisation of speech acts and their relevance to second language learning. In Kiira Allikmets (Ed.), Languages at Universities Today and Tomorrow (pp. 124–131). Tartu: Tartu University Press. Vygotsky, L. (1986). Thought and Language. Cambridge, MA: The MIT Press. Watanabe, S. (1993). Cultural differences in framing: American and Japanese group discussions. In D. Tannen (Ed.), Framing in discourse (pp. 176–209). Oxford: Oxford University Press.
TSL[v.20020404] Prn:8/04/2002; 15:26
F: PBNSP04.tex / p.1 (1)
P IV
The semantics/pragmatics boundary: Theory and applications
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.1 (405)
Cross-linguistic implementations of specificity* Klaus von Heusinger University of Konstanz
.
Introduction
The semantic-pragmatic category specificity, which was introduced in the late 60ies, has recently received a new interest (e.g. Enç 1991; Abbott 1994; Farkas 1995; Mathewson 1999; Geurts 1999 among others). However, there is no agreement on the range of phenomena that is related to specificity nor on the very subtle judgements of constructions involving specific NPs. For building a feasible theory of specificity we need (i) a better definition of the referential properties of the linguistic contexts that determine specific readings, and (ii) a list of grammatical implementations of specificity in those contexts. In particular, I will show that accusative case in Turkish is a good indicator for specificity. However, we find more instances of accusative case marking in Turkish than is predicted by the general approach to specificity, which assumes that specific indefinites are “scopeless”, i.e. show always widest scope with respect to other operators. I develop a different approach to specificity according to which specificity indicates that the referent is referentially anchored to another expression, which could be the speaker but also some other expression in the sentence itself. The concept of specificity was introduced in the late 60ies by transferring the de re-de dicto distinction of definite NPs to indefinite NPs. The contrast is illustrated by example (1), which can be assigned two readings: the specific reading of a monk is motivated by the continuation (1a), while the non-specific reading can be continued by (1b): (1) Umberto Eco: “I desired to poison a monk.” a. He lived in the famous monastery Bobbio in the year 1347. b. Therefore, Eco started to write a novel about a monastery.
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.2 (406)
Klaus von Heusinger
A specific reading of an indefinite NP is pretheoretically characterized by the “certainty of the speaker about the identity of the referent”, “the speaker has the referent in mind”, “the speaker can identify the referent”, etc. A weaker version of this characterization is that the referent of a specific NP is fixed and that it matters which referent we select out of the set of entities that fulfill the description. It is also generally assumed that specific indefinites are “scopeless” like proper names or demonstratives, i.e. they always show widest scope. Furthermore, the insertion of a certain is assumed to indicate specificity. (2) Pretheoretical and informal characterization of specificity i. certainty of the speaker about the identity of the referent ii. the referent is fixed iii. specific indefinite NPs are “scopeless”, i.e. they behave as if they always have widest scope iv. specific indefinite NPs can be paraphrased by a certain I will show that this characterization captures only certain cases of specific readings. In the course of this paper I argue that specificity is not based on the characterization (2i) and (2iii), but is rather to be characterized as the property of an NP being referentially anchored: The characterization (2ii) and (2iv) are vindicated in this view. There are two goals which are closely interconnected. First, I will show that the range of the phenomena that are related to specificity is far broader than generally assumed. Second, I will sketch a new semantic analysis for specific NPs based on a variety of cross-linguistic data. The general method I employ in this paper differs from the established semantic methodology of testing sentences for possible ambiguities. In this paper, I will fix the referential properties of the background and check how the linguistic expressions mirror these properties. In order to control the referential properties, I use the novel The Name of the Rose by Umberto Eco as the background for the sentences under investigation. In comparing the translations of one and the same sentence (in one and the same referential setting), we can illustrate the cross-linguistic implementation of one and the same semantic-pragmatic category (see example (4) below for illustration). Although using translations may cause other problems such as a more poetic style or deviations from unmarked forms in order to match the original text, they are used here to ensure that the examples from different languages have a similar referential background. The paper is organized in the following way: In Section 2, I illustrate the comparative semantic method of the paper using the semantic category of “genericity”. Furthermore, it will be shown that (i) referential categories encompass a great variety of semantic phenomena, and (ii) that one and the same referential property can be cross-linguistically expressed by different means. The point will be illus-
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.3 (407)
Cross-linguistic implementations of specificity
trated by quotations from The Name of the Rose. In Section 3, I indicate the range of phenomena associated with specificity by mentioning a few examples from different languages. Section 4 gives a summary of the different implementation and illustrates the different grammatical possibilities by comparing translations of a fragment of the novel. In Section 5, I sketch the two classical semantic approaches to specificity, the scope theory and the lexical ambiguity theory. The two theories do not cover all the cases of specificity, in particular they cannot describe cases of “relative specificity”, which I present in Section 6. I develop an indexical theory of specificity which accounts for the various phenomena discussed so far. The basic idea of being anchored is elaborated and discussed by illustration of some examples from the novel. In Section 7, I give a short summary.
. Referential properties of noun phrases NPs can express different referential properties in that they have various ways to refer to their referents. The type of referent can vary in multiple ways. Generally, we may distinguish the following semantic-pragmatic categories of the way expressions can refer: (3) Referential contrasts of nouns i. ii. iii. vi.
singular vs. plural generic vs. particular vs. predicative definite vs. indefinite specific vs. non-specific
Each of these referential properties can be implemented in different ways, and there are considerable contrasts between languages – some of the properties are overtly marked by morphology or syntax, some will only show up in the interaction with other expressions, and others are not marked at all. This will be illustrated by the referential property of genericity1 in (4) and (5). In (4) the first phrase ai semplici, its English translation the simple-minded, and the German translation die Laien expresses genericity by using definite NPs (in the plural), while the corresponding phrase un monaco, a monk and ein Mönch expresses genericity by using an indefinite expression in the singular. (4) a.
“. . . talora gli ordini dati ai semplici vanno rinforzati con qualche minaccia, come il presagio che a chi disubbidisce possa accadere qualcosa di terribile, e per forza soprannaturale. Un monaco invece. . . ” (42)
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.4 (408)
Klaus von Heusinger
b. “. . . sometimes orders given to the simple-minded have to be reinforced with a threat, a suggestion that something terrible will happen to the disobedient, perforce something supernatural. A monk, on the contrary. . . ” (33) c. “. . . Manchmal müssen Verbote für die Laien mit einer gewissen Drohung unterstrichen werden, etwa mit der Voraussage, daß dem Ungehorsamen etwas Schreckliches widerfahren könnte, etwas Übernatürliches selbstverständlich. Ein Mönch dagegen. . . ”(50) Languages also differ from each other in the means to express generic readings, as illustrated in (5). Given the context and background of the story, the phrase a monk is to be read generically. This generic reference is implemented in Italian by the definite article in the singular il monaco in (5a), in English by the indefinite article in the singular a monk in (5b), while German uses the indefinite in the plural Mönche in (5c) (note that there is no indefinite article in the plural). We also see similar differences in the implementation of the predicative function: un uomo (indefinite article and singular), human (no article), and Menschen (no article and plural).2 (5) a.
“Anche il monaco è un uomo,” sentenziò Aymaro. Poi aggiunse: “Ma qui sono meno uomini che altrove.” (132) b. “A monk is also human,” Aymaro declared. Then he added, “But here they are less human than elsewhere.” (125) c. “Auch Mönche sind Menschen”, sagte Aymarus sentenziös. Dann fügte er hinzu: “Aber hier sind sie es weniger als woanders.” (167)
In the following example, all three languages start with an indefinite article plus the noun in the singular, expressing some general contrast to the behavior of Benno; but then genericity is implemented by the plural definite in Italian (6a), by a plural indefinite (without an article) in English (6b), and by the singular with an indefinite article in German (6c). Again this differs from the examples in (5). (6) a.
Un monaco dovrebbe certo amare i suoi libri con umiltà, volendo il ben loro e non la gloria della propria curiosità: ma quello che per i laici è la tentazione dell’adulterio e per gli ecclesiastici regolari è la brama di ricchezze, questa per i monaci è la seduzione della conoscenza. b. (The day before, Benno had said he would be prepared to sin in order to procure a rare book. He was not lying and not joking.) A monk should surely love his books with humility, wishing their good and not the glory of his own curiosity; but what the temptation of adultery is for laymen and the yearning for riches is for secular ecclesiastics, the seduction of knowledge is for monks. (183)
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.5 (409)
Cross-linguistic implementations of specificity
c.
Gewiß, ein Mönch sollte seine Bücher in Demut lieben, sich lediglich ihrer Erhaltung widmen und nicht der Befriedigung seiner Neugier. Doch was für den Laien die Verlockung des Ehebruchs ist und für den städtischen Priester der Zauber des Reichtums, das ist für den Mönch die Versuchung des Wissens und der Erkenntnis. (241f.)
The preceding examples have illustrated that the encodings of genericity differ from language to language: il monaco, a monk, and Mönche in (5a–c). Moreover, the examples illustrate that a referential category like genericity is implemented in even one language by different grammatical means, such as un monaco in (4a), il monaco in (5a), and i monaci in (6a). This shows that there is no one-to-one correlation between a grammatical contrast and a semantic-pragmatic category that expresses a referential property. Therefore, we must consider the following three aspects in order to understand semantic-pragmatic categories, such as genericity or specificity. First, we need enough contextual background, second we need a wide range of linguistic observations, and third we must design a feasible theory. The first ingredient for my analysis of specificity is given by the novel. Second, I present cross-linguistic observations that are related to the pretheoretical concept of specificity in the next two sections. Third, I develop a feasible theory of specificity that is based on the observations presented.
. Cross-linguistic observations As opposed to definiteness, specificity is not morphologically marked in most Indo-European languages. However, there are other kinds of grammatical contrasts that can be related to the referential property of specificity. There are: (i) quasiuniversal contrasts, which can be found in most languages; (ii) common contrasts that can be found in many languages; and (iii) grammatical phenomena that are language specific. In the following I will present an example from each group. Quasi-universal specificity contrasts include the tendency for an indefinite to be (more) specific as more descriptive material is inserted into it, as illustrated in (7): (7) Everything turns on the theft and possession of a book, which was concealed in the Finis Africae, and which is now there again thanks to Malachi’s intervention. (446) Another quasi-universal property of specific indefinites is that they tend to have wide scope, and therefore it is easier to establish anaphoric links to them, while non-specific indefinites depend much more on discourse domains and they interact with the scope of operators like negation. This is illustrated by (8) where
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.6 (410)
Klaus von Heusinger
the non-specific indefinite (in the plural) ceci, chickpeas, Kirchererbsen in the question (i) and in the negated sentence (ii–iii) cannot serve as antecedents. Therefore, the expression has to be repeated in (iv), while in the second sentence of the contribution in (iv–v) they can be referred to by anaphoric pronouns li, them, sie.3 (8) a.
i. ii. iii. iv. v.
b. i. ii. iii. iv. v. c.
i. ii. iii. iv. v.
“Hai ceci?” La domanda, diretta a me, mi sorprese. “No, non ho ceci”, dissi confuso. “La prossima volta portami dei ceci. Li tengo in bocca vedi la mia povera bocca senza denti, sinché non si ammollano tutti.” “Do you have any chickpeas?” The question, addressed to me, surprised me. “No, I have no chickpeas,” I said, confused. “Next time, bring me some chickpeas. I hold them in my mouth – you see my poor toothless mouth? – until they are soft.” “Hast du Kichererbsen?” Die Frage galt mir, und ich antwortete verwirrt: “Nein, ehrwürdiger Vater, ich habe keine Kichererbsen.” “Das nächste Mal bring mir Kichererbsen mit. Ich nehme sie in den Mund, sieh meinen armen zahnlosen Mund, und kaue sie weich.”
It is quite common that modality may co-occur with specificity. For example, the subjunctive or conjunctive of an embedded clause strongly suggests that an indefinite is to be understood as non-specific. This is illustrated in example (9): (9) a.
Peraltro parlava sempre di cose così buone e sagge che era come se un monaco ci leggesse le vite dei santi. (35) b. For that matter, he spoke always of things so good and wise that it was as if a monk were reading to us the lives of the saints. (27) c. Doch er [William] sprach stets so klug und erbaulich, daß es war, als läse ein Mönch aus den Viten der Heiligen vor. (41)
A language specific implementation of specificity is found in Turkish (Kornfilt 1997: 219ff). Turkish does not have a definite article, while it has the indefinite article bir, which is derived from the numeral bir, but which differs in distribution. The direct object can be realized by the absolut(ive) without case endings or by the accusative with the case ending -I. Thus the definite reading is generally expressed by the accusative case ending, while the indefinite reading is realized by the
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.7 (411)
Cross-linguistic implementations of specificity
indefinite article plus the absolutive. However, the combination of the markers for definiteness and indefiniteness in (10c) expresses an indefinite specific NP.4 (10) a.
(ben) kitab-ı oku-du-m I book-acc read-past-1sg “I read the book.” b. (ben) bir kitap oku-du-m I a book read-past-1sg “I read a book.” c. (ben) bir kitab-ı oku-du-m I a book-acc read-past-1sg “I read a certain book.”
A similar contrast exists for the subject of embedded sentences. The predicate of an embedded sentence in Turkish is a nominalized form that shows agreement with the subject, realized by the possessive marker -I. The subject is realized either as a genitive with the case ending -In, or as the absolut. The combination of the indefinite article bir with the genitive case marks a specific subject (Kornfilt 1997: 219ff, ex. (762)=(11a)). Note that the non-specific subject tends to be closer to the predicate, while the specific one stands rather at the first position in the sentence. (11) a.
[köy-ü haydut bas-tı˘g-ın]-ı duy-du-m [village-acc robber raid-nom-poss.3sg]-acc hear-past-1sg “I heard that robbers raided the village.” b. [bir haydut-un köy-ü bas-tı˘g-ın]-ı duy-du-m [a robber-gen village-acc raid-nom-poss.3sg]-acc hear-past-1sg “I heard that a certain robber raided the village.”
. Grammatical means of marking specificity In the last section, I have given examples of quasi-universal, common and language-particular grammatical means of marking specificity. There are many more ways to indicate specificity, which cannot be listed here for reasons of space. For the time being, we can summarize the observations as follows:5
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.8 (412)
Klaus von Heusinger
i.
quasi-universal contrasts
Specific indefinites contain more descriptive material (see (7)), they are assumed to be “scopeless”, i.e. they behave as if they had widest scope (see discussion below in section 5 for a different view); due to their scope behavior it is easier to link anaphoric pronouns to them6 (see (1) and (12b.iii)), they can be lexically marked by indefinite pronouns (a certain, certi, ein bestimmter) as in (12b.v). ii.
common contrasts
Specific indefinites are more often found in assertive sentences, while non-specific ones can be found in interrogative or imperative ones (see (8)). Modality is also related to specificity (see (9)). A subjunctive in a relative clause promotes a nonspecific reading of its indefinite head in Italian (and other Romance languages), while the indicative motivates a specific reading, as in (12a.i + ii). Specific indefinites are more often found with past than with future tense. iii.
grammatical phenomena that are language specific
German shows a contrast in negating indefinite NPs: nicht + ein vs. kein. The latter can approximately be used if the indefinite is non-specific. Some languages show double marking, either combination of article plus case endings (such as Turkish, see (10) and (11)), or a combination of two articles (Plank 2002). Givón (1978) reports on two indefinite articles in Bemba (Bantu) expressing the contrast between specific and non-specific. Turkish gives us better diagnostics for specificity than English, Italian or German, since specificity is morphologically marked by case endings. These diagnostics will be used to show that the assumed referential properties (given by the background) do show up in the grammatical implementation. This is illustrated by the following example (12). The context of the novel is that one monk indicates to William of Baskerville (the medieval Sherlock Holmes) that he knows something (specific!), but that he is not ready to disclose it: “But in the abbey there are rumors, . . . strange rumors . . . ” – “Of what sort?” (12a) i. “Strane. Diciamo, di un monaco che nottetempo ha voluto avventurarsi in ii. biblioteca, per cercare qualcosa che Malachia non aveva voluto dargli, e ha iii. visto serpenti, uomini senza testa, e uomini con due teste.” (12b) i. ii. iii. iv. v. vi.
“Strange. Let us say, rumors about a monk who decided to venture into the library during the night, to look for something Malachi had refused to give him, and he saw serpents, headless men, and men with two heads. He was nearly crazy when he emerged from the labyrinth. . . ” [. . . ] And besides, if it is evil to handle certain books, why would the Devil distract a monk from committing evil?” (89)
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.9 (413)
Cross-linguistic implementations of specificity
(12c) i.
“Garip söylenti-ler örne˘gin, [bir rahib-in geceyarısı, [[M’nin kendine strange rumor-pl for example, [a monk-gen midnight [[M.-gen himself-dat ii. ver-mek iste-me-di˘g-i] bir kitab-ı bul-mak için] gizlice give-inf want-neg-nom-poss.3sg] a book-acc find-inf to] secretly iii. kitaplı˘g-a girmey-e kalkı¸s-tı˘g-ı] (. . . ) dair söylenti-ler library-dat enter-to venture-nom-poss.3sg] about rumor-pl A somewhat “literal” translation would be: “There are strange rumors, for example rumors about [a monk midnights secretly into the library venturing [to find a book [that Malachi did not want to give him]]]
In the English translation the anaphoric pronoun he in (12b.iii) shows that the indefinite NP a monk is at least not embedded under the NP rumors, but allows anaphoric linkage. This indicates that it has a specific reading. In the Italian original the indicative mood of the relative clause (ha voluto) indicates that the head noun un monaco is specific. This is confirmed by the Turkish translation, where the subject bir rahib-in of the embedded sentence that ends in kalkı¸stı˘gı shows double marking (indefinite article plus case ending). We can further note that the Turkish translation bir kitabı for the Italian qualcosa or English something in line (ii) is marked as specific. The specificity of this NP is confirmed by the setting of the novel: Malachi (the librarian) can only refuse to give something to the monk if the monk had asked for a specific thing. In Italian, the predicate aveva voluto in the relative clause is in the indicative, and thus indicating that the head noun qualcosa is specific. In English, the relative clause modifying something contains the proper name Malachi, which again is a good indication that the indefinite pronoun is linked to the referent of that proper name. In comparing the three languages, it is obvious that Turkish marks specificity obviously, while we have to look for subtle indicators in English or Italian.
. Reconstructing specificity by constructing theories The classical reconstruction of the contrast between specific and non-specific indefinite NPs was formulated in terms of a scope interaction between the indefinite NP and other operators. In this view, the two readings of (1), repeated as (13), are paraphrased as (13a) where the existential quantifier has scope over desire, yielding the specific reading; and as (13b) with narrow scope for the indefinite yielding the non-specific reading: (13) Umberto Eco: “I desired to poison a monk” a. There is a monk and Umberto Eco desires to poison him. [spec] b. Umberto Eco desires that there is a monk such that he poisons him. [non-spec]
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.10 (414)
Klaus von Heusinger
In the following I present two of the problems with this configurational analysis of specificity. The first observation concerns island restrictions for quantifier raising, and the second the independence of scope and specificity. Quantifiers in an embedded sentence cannot take scope over a definite NP in the matrix clause. (14) has only the reading (14a) that there is one rumor such that each student had been called before the dean, but the reading (14b) “for each student there is a rumor such that. . . ” is not available. This is explained by the assumption that quantifiers cannot be raised out of islands like the embedded sentence in (14). However, Fodor and Sag (1982) observe that indefinite NPs can leave such islands, as illustrated in (15). Example (15) can receive a reading (15a) with wide scope of the rumor, but also a reading (15b), in which the indefinite takes scope over the rumor. (14) John overheard the rumor [that each student of mine had been called before the dean]. a. the rumor . . . each student b. *each student . . . the rumor (15) John overheard the rumor [that a student of mine had been called before the dean]. a. the rumor . . . there is a student . . . b. a certain student . . . the rumor . . . Fodor and Sag (1982) come to the conclusion that indefinite NPs can receive two different readings via two indefinite articles with the same surface form: an existential or non-specific article ae , and a referential or specific article ar . The referential article applied to the common noun forms a referential indefinite NP, which is “scopeless” like proper names or deictic expressions. Such expressions behave as if they had always widest scope. The existential article is represented as the existential quantifier, which can be raised like the universal quantifier. The existential reading, however, does not allow for quantifier raising out of islands, providing only the narrow scope reading (15a). Fodor and Sag’s lexical ambiguity theory of indefinites makes a clear prediction about cases with more than one operator besides an indefinite that is embedded in a scope island, as in (16). (16) is predicted to have only two readings: the reading (16a), where the indefinite takes scope over the two other operators since it receives a referential or specific reading, and the reading (16b) with narrow scope with respect to both other operators licensed by the existential article. However, there should be no reading like (16c), the so called “intermediate” reading, since it would be neither specific (“only widest scope”) nor existential (since the indefinite cannot leave the island). The prediction is borne out by the intuition to example (16).
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.11 (415)
Cross-linguistic implementations of specificity
(16) Each teacher overheard the rumor that a student of mine had been called before the dean. a. a student . . . each teacher . . . the rumor (= spec/referential reading) b. each teacher . . . the rumor . . . a student (= non-spec/existential reading) c. *each teacher . . . a student . . . the rumor (“intermediate reading”) Contrary to this observation, Farkas (1981) and later others (e.g. Geurts 1999) have observed that there are examples with the same structure as (16) that allows for an intermediate reading, as in (17). The indefinite NP some condition can take scope over three arguments and under each student. That is, the sentence has a reading where for each student, there is some condition such that the student should find three arguments against this condition. Thus the lexical ambiguity theory of specificity faces serious obstacles, which are not easy to overcome.7 (17) Each student has to come up with three arguments that show that some condition proposed by Chomsky is wrong. a. each student . . . some condition . . . three arguments (“intermediate reading”)
. Relative specificity The second problem for the scope (as well as the lexical ambiguity) theory of specificity are cases in which the specificity seems not to relate to the speaker or the context of utterance, but to some other element in the sentence. Higginbotham (1987: 64) illustrates this by the examples (18) and (19): “In typical cases specific uses are said to involve a referent that the speaker ‘has in mind.’ But this condition seems much too strong. Suppose my friend George says to me, ‘I met with a certain student of mine today.’ Then I can report the encounter to a third party by saying, ‘George said that he met with a certain student of his today,’ and the ‘specificity’ effect is felt, although I am in no position to say which student George met with.” (18) George: “I met a certain student of mine.” (19) James: “George met a certain student of his.” Hintikka (1986) had made a similar observation in his discussion of the expression a certain. He shows in (20) that the specific indefinite a certain woman can receive narrow scope with respect to the universal quantifier and still be specific: it is a specific woman for each man. Hintikka suggests that the specific indefinite NP is to be represented by a Skolem-function that assigns to each man the woman who
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.12 (416)
Klaus von Heusinger
is his mother. With Farkas (1997) we can describe the dependency of the specific NP a certain woman from the universal quantifier every man by the concept of “co-variation”: The referent for the specific indefinite covaries with the referent for man. In other words, once the reference for man is fixed (during the process of interpreting the universal quantifier), the reference for the specific indefinite is simultaneously fixed. (20) According to Freud, every man unconsciously wants to marry a certain woman – his mother. (Hintikka 1986) These observations motivate a revision of the pre-theoretical description of specificity as the “certainty of the speaker about the referent”. It was shown that a specific indefinite NP need not depend on the speaker or the context of utterance, it can also depend on other linguistic entities like the universal quantifier every man in (20) or on the proper name George in (19). In the following I assume that specificity marks that the specific expression is referentially anchored to another expression, rather than “absolutely” related to the speaker. With this view, we can now go back to some of the earlier examples and account for their specificity in terms of referential anchoring. In example (12), repeated as (21), we had the impression that the indefinite pronoun something is specific, which was confirmed by the accusative case markings in the Turkish translation. At that stage of description, we had no means to account for this, and the generally given description of specificity did not allow for this since the referent is not “known by the speaker” or “identified by the speaker” (if we take the writer of the book as the speaker). However, in the new approach we can referentially anchor the indefinite pronoun to the expression a monk. The Turkish translation (21b) double-marks the specific indefinite bir kitabı by the indefinite article and the accusative case. Example (22) shows that the anchor bir rahip for the specific indefinite bir kitabı can be itself non-specific, as is indicated by the conditional form of the verb isterse. (21) a.
(Strange. Let us say, rumors about a monk who decided to venture into the library during the night), [to look for something [Malachi had refused to give him]] b. [[ver-mek iste-me-di˘g-i] bir kitab-ı [[give-inf want-neg-nom-poss.3sg] a book-acc bul-mak için] (. . . ) find-inf to] (. . . )
(22) bir rahip bir kitabı al-mak isterse a monk a book-acc take-inf want-cond “If a monk wants to take a book [spec]”8
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.13 (417)
Cross-linguistic implementations of specificity
Specificity expresses the dependency of the reference of one term on the referent of another term. This view generalizes the pre-theoretical idea of “the certainty of the speaker towards the referent” in at least two dimensions: first it is not only the speaker who can act as anchor for a specific expression, and second the relation between the anchor and the specific expression is not described by any mental relation like “certainty” or “ability to identify” etc. I assume that these notions are not part of linguistic description or theory proper. The anchoring relation rather expresses that the reference of the specific expression is determined by the reference of another expression. One instance was example (20), where the referent of the specific a woman covaries with the referent of man. In (17), repeated as (23), the referent of the specific indefinite some condition can either be anchored to the speaker yielding wide scope, or to the student yielding intermediate scope, or to three arguments yielding narrow scope. (23) Each student has to come up with three arguments that show that some conditioni proposed by Chomsky is wrong. a. i = speaker widest scope b. i = student intermediate scope c. i = three arguments narrow scope Often the speaker is the anchor for a specific expression. In that case the specific NP receives the same scope as the speaker, namely widest scope with respect to any other operator, which is equivalent to the pre-theoretical view. However, in cases like (20)–(22) the specific expression depends on a different expression, but not on the speaker, and therefore it does not have widest scope. For example, on one reading the specific indefinite bir kitabı in (22) is specific with respect to the monk, but does not receive scope over the conditional. Example (24)9 involves several possible anchors: The speaker William, the hearer Jorge, the person they talk about, Malachi, and the two persons mentioned in the sentence. Thus, in principle the indefinite a book from the Finis Africae could be anchored to each of them. However, the setting of the novel strongly suggests that Malachi had not known the identity of the book for most of the time. So it seems more likely that the specific indefinite is anchored to Berengar (or Severin). (24) [William to Jorge de Burgos about Malachi:] You probably told him Berengar had been intimate with Severinus, and as a reward Severinus had given him a book from the Finis Africae. We reconstruct the anchoring by assuming that any NP introduces an anchor and that an indefinite NP receives an additional index that must be linked to an accessible anchor. Now we can represent sentence (24) as (24a). (24b) shows one way of linking the indefinite to an anchor, here: Berengar:
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.14 (418)
Klaus von Heusinger
(24) a.
<William, Jorge>[You probably told him [= Malachias] <Malachias>[Berengar had been intimate with Severinus, and as a reward [Severinus had given him [a book from the Finis Africae]i ]]] b. <William, Jorge>[You probably told him [= Malachias] <Malachias>[Berengar had been intimate with Severinus, and as a reward [Severinus had given him [a book from the Finis Africae]Berengar ]]]
. Summary I argued that the pretheoretical characterization of specificity as (i) certainty of the speaker about the identity of the referent, (ii) the referent is fixed, (iii) specific indefinite NP is “scopeless”, and (iv) specific indefinite NPs can be paraphrased by a certain, can only describe a restricted set of specific expressions. I presented cross-linguistic implementations of specificity that clearly extended the range described by the mentioned characterization. In particular, the Turkish accusative case ending marking specificity is used in a broader way. In explaining this use and by accounting for the scope behavior of specific indefinites, the so-called cases of “relative specificity”, I developed a new theory of specificity. Specificity expresses a referential relationship between two referential expressions: the reference of a specific expression depends on the “anchor” expression. Once the reference for the anchored is determined, the reference for the specific term is also determined. Thus the scope behavior of specific expressions can explained on this line.
Notes * This article is the revised version of a talk at the Second International Conference in Contrastive Semantics and Pragmatics in Cambridge in September 2000. First of all I would like to thank the organizers, Katarzyna M. Jaszczolt and Ken Turner, for the organization and for helpful comments on the paper. I would also like to thank the audience at the conference, as well as the participants of the workshop on Sentence Type and Specificity at the Zentrum für Allgemeine Sprachwissenschaft in Berlin for comments, and in particular Barbara Abbott, Nick Asher, Werner Frey, Bart Geurts, Michael Grabski, Jaklin Kornfilt, Manfred Krifka, Albert Ortmann, Kerstin Schwabe, and Carla Umbach for very inspiring discussions; Paul Portner gave very detailed and constructive comments on an earlier version of this paper. All
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.15 (419)
Cross-linguistic implementations of specificity
remaining shortcomings are mine. The research was supported by a Heisenberg-Fellowship of the German Science Foundation. . I cannot give an account of the whole phenomena of genericity (see Carlson & Pelletier 1995) or its cross-linguistic implementation (see Gerstner-Link 1995). . (5) illustrates the method of controlling the referential background via translation and testing the linguistic expression against this background. All three sentences (5a–c) stand in the same referential setting. Even though they refer in the first part by the quite different expressions il monaco, a monk, and Mönche, they continue the second sentence with the same anaphoric plural pronoun they and sie (the Pro-drop language Italian only shows plural agreement on the verb sono). I do not claim that the pronoun them expresses a close anaphoric relation based on binding – it seems to be rather a “pragmatic” pronoun which can refer to some pragmatically salient set of individuals. However, the point is that it is the same referential setting which is reflected in the anaphoric link. . Nick Asher and Paul Portner pointed out to me that the anaphoric relation between chickpeas in the imperative (iv) and the pronoun they in the following declarative sentence is quite complex: While the indefinite chickpeas is existential, the pronoun depends on a generic or habitual operator and therefore cannot be (dynamically) bound by the antecedent. . See Dede (1986) for futher conditions on the use of the accusative case marker. In particular, she shows that any direct object must be marked by the accusative case if it is moved out of its preverbal base position. . Fodor and Sag (1982: 358ff) list “a number of factors which favor either a quantificational [= non-specific, KvH] or a referential [= specific, KvH] understanding of an indefinite”. Among others they mention the following: (i) The content of a noun phrase: the more content the more referential. (ii) Topicalization and Left Dislocation strongly favor the referential reading. (iii) The use of non demonstrative this strongly, perhaps uniquely, favors a referential reading. (iv) There-insertion favors existential readings. (v) Relative clauses add material, thus induce a referential reading; this holds even more for non-restrictive relative clauses. (vi) A certain favors a referential reading (but the semantics of a certain is “completely obscure”). . Werner Frey (pc) pointed out to me that anaphoricity is no indicator for specificity, contrary to what some people believe. The point I wanted to make here is that specific indefinites behave differently with respect to negation, conditionals and other domain creating operators. This is reflected in their potential to act as antecedents for anaphoric pronouns. . Recently Kratzer (1998) has revived the lexical ambiguity theory by using choice functions for the referential article, and the existential quantifier for the existential article of Fodor and Sag. See also von Heusinger (2000) for a short discussion and criticism of that approach. . Jaklin Kornfilt (pc) pointed out to me that she thinks that (22) is an ungrammatical, or at least very marked, Turkish sentence. . The sentence is uttered during the show-down at the end of the novel (night of the seventh day). The two main protagonists, William of Baskerville and Jorge de Burgos, are meeting in the secret room of the library, the famous “Finis Africae”, and William gets hold of the book the story is about. The two men report to each other their reconstruction of the events
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.16 (420)
Klaus von Heusinger
in the monastry and try to show that each of them was intellectually ahead of the other. In this context they utter the sentence (i) which I simplified to (24): (i)
[William to Jorge de Burgos about Malachi]: “You probably told him Berengar had been intimate with Severinus, and as a reward Severinus had given him a book from the Finis Africae.” (465)
References Abbott, B. (1994). Referentiality, Specificity, Strength, & Individual Concepts. In E. Duncan, D. Farkas, & P. Spaelti (Eds.), The Proceedings of the Twelfth West Coast Conference on Formal Linguistics (pp. 473–484). Stanford: CSLI. Carlson, G., & Pelletier, F. (1995). The Generic Book. Chicago: University of Chicago Press. Dede, M. (1986). Definiteness and Referentiality in Turkish Verbal Sentences. In D. Slobin & K. Zimmer (Eds.), Studies in Turkish Linguistics (pp. 147–164). Amsterdam: Benjamins. Eco, U. (1980). Il nome della rosa. Milano: Bompiani. Eco, U. (1994). The name of the rose. [With a postscript by the author.] (Trans. by W. Weaver). San Diego; New York; London: Harcourt Brace & Company. Eco, U. (1986). Der Name der Rose. (Trans. by B. Kroeber) München: Deutscher Taschenbuch Verlag. Eco, U. (1986). Gulun Adi. (Trans. by S. Karadeniz). Istanbul: Can Yayinlari. Enç, M. (1991). The Semantics of Specificity. Linguistic Inquiry, 22, 1–25. Farkas, D. [1981] (1985). Intensional Description and the Romance Subjunctive Mood. New York: Garland. Farkas, D. (1995). Specificity and Scope. Linguistic Research Center. UC Santa Cruz. LRC– 95–01. Farkas, D. (1997). Dependent Indefinites. In F. Corblin, D. Godard, & J.-M. Maradin (Eds.), Empirical Issues in Formal Syntax and Semantics. Selected Papers from the Colloque de Syntaxe et de Sémantique de Paris (CSSP 1995) (pp. 242–267). Bern: Lang. Fodor, J., & Sag, I. (1982). Referential and Quantificational Indefinites. Linguistics and Philosophy, 5, 355–398. Gerstner-Link, C. (1995). Über Generizität. Generische Nominalausdrücke in singulären und generellen Aussagen. München: Fink. Geurts, B. (1999). Specifics. In B. Geurts, M. Krifka, & R. van der Sandt (Eds.), Focus and Presupposition in Multi-Speaker Discourse (pp. 99–129). ESSLLI 99, Utrecht. Givón, T. (1978). Definiteness and Referentiality. In J. Greenberg, C. Ferguson, & E. Moravcsik (Eds.), Universals of Human Language (4 vols.), Vol. 4 (pp. 291–330). Stanford: Stanford University Press. von Heusinger, K. (2000). The Reference of Indefinites. In K. von Heusinger, & U. Egli (Eds.), Reference and Anaphoric Relations (pp. 247–265). Dordrecht: Kluwer. Higginbotham, J. (1987). Indefinites and Predication. In E. Reuland, & A. ter Meulen (Eds.), The Representation of (In)definiteness (pp. 43–70). Cambridge, MA: MIT Press. Hintikka, J. (1986). The Semantics of ‘a certain’. Linguistic Inquiry, 17, 331–336. Kornfilt, J. (1997). Turkish. London: Routledge. (Descriptive Grammars Series).
TSL[v.20020404] Prn:4/11/2002; 10:13
F: PBNS39.tex / p.17 (421)
Cross-linguistic implementations of specificity
Kratzer, A. (1998). Scope or Pseudoscope? Are there Wide-Scope Indefinites. In S. Rothstein (Ed.), Events and Grammar (pp. 163–196). Dordrecht: Kluwer. Matthewson, L. (1999). On the Interpretation of Wide-Scope Indefinites. Natural Language Semantics, 7, 79–134. Plank, F. (2002). Double Articulation. In F. Plank (Ed.), Noun Phrase Structure in the Languages of Europe. Berlin: Mouton de Gruyter.
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.1 (423)
The semantics–pragmatics interface The case of grounding Esam N. Khalil University of Nijmegen, the Netherlands
.
Introduction
This paper is a contribution to the current debate on the semantics– pragmatics interface. This debate reflects divergent perspectives on the nature of the relation between semantics and pragmatics and the role they play in language (production). Therefore, interface may be taken to imply the presence of a dividing line between the two disciplines and hence a division of labour between them. It may also suggest that semantics and pragmatics enjoy a common boundary and, as a result, some interaction, mutual influence, or complementarity. In this paper I will explore some features of the semantics– pragmatics interface by focussing on a fundamental and elusive property of discourse, namely grounding (viz. the foreground–background structure). As a discourse phenomenon, grounding is appropriate for the investigation of the semantics– pragmatics differentiation – or perhaps their mutual influence – since it reveals various strategies by writers in the process of textualization. The vast literature on grounding has shown a great deal of conceptual confusion; definitions have been imprecise and various levels of description conflated (for details, see Khalil 2000). In previous work (Khalil 2000), I examined grounding in discourse-semantic terms and distinguished it from several other structures. I therefore distinguished grounding from other semantic properties such as coherence, and from the cognitive non-textual level of information structure. I also distinguished grounding from the more or less prominent ways in which it is signaled in surface structure. My approach to grounding is that it is a distinct dimension within text semantics that organizes propositions on a continuum of grounding values. I considered grounding in terms of the proposition and not the single concept or lexical item.
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.2 (424)
Esam N. Khalil
‘A proposition is the unit of meaning that identifies the subject matter of a statement; it describes some state of affairs, and takes the form of a declarative sentence, e.g. Mary loves Michael’ (Crystal 1997: 107). My analysis is empirically based; data are taken from short news texts. These texts are appropriate for the present study because as van Dijk (1999: 125) points out: 1. Besides everyday conversation and professional discourse, news is the kind of discourse that most of us are confronted with most frequently. 2. It is undoubtedly the kind of discourse from which we learn most about the world. It should by now be apparent that I am not concerned with speech acts (see e.g. Austin 1962; Searle 1969; Grice 1975) as applied to spoken discourse or to dialogue situations and short face-to-face communication. It should however be added that the term ‘discourse’ does not refer only to dialogue and conversation, and that a written text may also be viewed as a ‘monologue discourse’ (van Dijk 1977: 8). In fact, ‘text takes on more of the interactive qualities of discourse’ (Garrett & Bell 1998: 2). I will focus on news communication, where writers perform certain global and local informative acts, assuming that they are an authority on the subject matter and that they have evidence for the factual information they impart. Readers usually believe that news writers have something newsworthy to communicate and that they are sincere in their intention to communicate it. Since communication (in spoken/written discourse) typically involves more than one proposition or sentence, and given the complexity of communicating meaning (beyond sentence-boundaries), my assumption is that not all facets of the grounding phenomenon can be adequately accounted for in exclusively semantic or pragmatic terms. I will therefore examine certain features of the semantics– pragmatics interface by attempting to gain more insight into the nature of propositional/semantic grounding and its surface structure expression. Specifically, I will examine some properties of the content, the context, and the surface structure. In so doing, my approach to grounding will be also in text pragmatics terms.1 As will become apparent, a central notion in text pragmatics, and in the present study, is intentions. The intentions of the writer determine the form the message takes, and, in its turn, the form provides readers with clues to interpret the message. That the notion of intentions has been associated with pragmatics is apparent in descriptions such as the following, which maintains that pragmatics is ‘the study of action deliberately undertaken with the intention of causing the intended interpreter to re-assess his model of how things are, including his system of values and his model of the speaker’s beliefs, attitudes, and intentions’ (Green 1989: 5).
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.3 (425)
The semantics –pragmatics interface
Some other descriptions have nevertheless suggested that semantics is not devoid of intentions: Pragmatics and semantics both take into account such notions as the intentions of the speaker, the effects of an utterance on listeners, the implications that follow from expressing something in a certain way, and the knowledge, beliefs, and presuppositions about the world upon which speakers and listeners rely when they interact. (Crystal 1997: 120)
Although this characterization also suggests that semantics and pragmatics are two complementary and interrelated fields (see e.g. Leech 1983), several discerning features have been proposed. Therefore, semantics has been viewed as the study of meaning in natural language (word meaning, sentence meaning), and pragmatics as the study of meaning in interaction ‘which includes speaker meaning and consideration of the wider context’ (Jucker 1998: 830). The type of knowledge has also been suggested as a discerning feature: semantics comprises those aspects of the interpretation process that rely on linguistic knowledge only while pragmatics comprises those ‘which must have recourse to non-linguistic information (world knowledge or common ground between the speaker and addressee)’ (Jucker 1998: 830). I would like to suggest that pragmatics is a discipline that is assumed to focus on the relation(s) between the way utterances or texts are organized and the contribution of the extra-linguistic context, including the intentions of the writer. This view involves a broad interpretation of pragmatics, which includes ‘the interpretation of acts assumed to be undertaken in order to accomplish some purpose’ (Green 1989: 3).2
. Different intentions, semantics, and context As a starting point in the examination of the semantics– pragmatics differentiation, I suggest that although intentions are sometimes ascribed to both semantics and pragmatics, the nature and the extent of intentions in each may vary. As we will see in the case of grounding, intentions, indeed, ‘come in various types and strengths’ (Jaszczolt 1999: 200) and utterances ‘come with various types of intentions’ (Jaszczolt 2000b: 173). It should however be made clear that I am not dealing with different intentions within the same utterance where an utterance has a primary or basic intention, namely to secure the referent of the speaker’s utterance (for that see Jaszczolt 1999). I am also not dealing with intentions in terms of the analysis of sentence meaning or the question of ambiguity in – and the different interpretations of – utterances. In discourse grounding there is a scale of ‘degrees of intentionality’ (Jaszczolt 2000b: 170) towards the organization of meaning/propositions. This re-
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.4 (426)
Esam N. Khalil
sults in different grounding-structures. Differences in intentions account for the distinction between what I refer to as semantics prone (or semantically motivated) grounding-structures and pragmatics prone (or pragmatically motivated) grounding-structures. In other words, grounding-structures allow – or are subject to – different degrees of pragmatic intervention/influence. Semantically motivated grounding-structures manifest a low degree of writer’s intentions. The writer may have a default type of intention that has a default value that ‘triggers the default semantics for an utterance’ (Jaszczolt 1999: 200).3 The writer is primarily concerned with the linguistic context, or rather with a type of context that is much more restricted in scope and limited in role, that is to say, restricted to determining content (see Bach 1999: 72) in terms of default propositional grounding. In this regard, semantically motivated grounding-structures are usually severed from contextual factors.4 Of course, conventional world knowledge is required for the understanding of semantic structures. But this knowledge is restricted to text-type rules, and there is no – or a minimal degree of – context accompanying meaning. I would like to suggest that semantically motivated grounding-structures are one manifestation of ‘narrow context’ (Bach 1999: 93) which is relevant to semantics. A narrow context, thus, may be associated with ‘default semantics’ or semantic intentions. Pragmatically motivated grounding-structures, on the other hand, exhibit a high degree of writer’s intentions, determined by the extra-linguistic context.5 This results in the (re)organization of grounding-structures and brings about whatever changes are necessary for propositions to become more or less contextually informative.6 Contextual constraints include shared knowledge by writers and readers, relevance, and the interestingness of information in a given situation. In other words, what is of relevance is a ‘broad context’ that is relevant to pragmatics (see Bach 1999: 73).7
. Discourse grounding as a dynamic process Viewed from the perspective of language production, discourse grounding is a dynamic process (Givón 1987) that is part of dynamic semantics.8 It involves the dynamic unfolding of the meaning of a text for readers by organizing propositions on a foreground–background continuum, and distinguishing between different grounding values. In this regard, discourse grounding differs from another semantic property of discourse, namely coherence. Coherence has to do with the organization of meaning in text in terms of a macro-micro hierarchy. The hierarchy reflects how facts, events or states of affairs
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.5 (427)
The semantics –pragmatics interface
are related, and how language users understand them or how they might have represented them in their mental models, independent of communicative intentions, interestingness, or newsworthiness. Therefore, a proposition at the highest level of generalization need not be a foreground proposition. A foreground proposition does not necessarily summarize a discourse topic. Any information may in principle be mapped as a foreground proposition. While coherence may tend to cluster concepts in propositions in order to establish conceptual closeness in a hierarchical network of relations between facts and objects, grounding may be insensitive to conceptual closeness, and propositions may be expressed in non-contiguous sentences. For example, a proposition that is conceptually very close to the top macroproposition – because it is, for instance, a sub-topic or a specification of the proposition that refers to the main event – may, for grounding considerations, be realized at the end of the text. It is clear that grounding is a further contribution to textuality, and that the organization of meaning in terms of a foreground–background gradient is a decision that may be taken after the construction of macro- and micro-propositions. In fact, the writer may manipulate macro and micro propositions in the process of distributing meaning along a continuum of grounding values. Grounding would therefore be closer to actual expression than macro-and micro-structures.
. The foreground– background structure Grounding is one of the organizing principles of discourse meaning. It organizes (viz. distributes) propositions in terms of a scale that shows a gradual decrease in propositional importance as is typical of news discourse.9 The foreground–background structure may be represented as a three-level scale: foreground (FG), midground (MG), and background (BG). Foreground refers to all meaning that is high on the scale. A proposition that is assigned a foreground value denotes the main, most recent, current, or ongoing (speech) event. The event is often unknown, unpredictable, or inaccessible to readers and it is usually expressed at the beginning of the news text. Midground refers to all meaning that is lower on the scale than FG meaning. A midground proposition denotes specific characteristics of one or more properties of the same/main (speech) event. Midground provides explanation and is usually expressed after expressing foreground. Background refers to all meaning that is lower on the scale than midground. A background proposition may be assigned to the spatio-temporal setting of the event, its circumstance, or occasion. It may map information that is old, known, or accessible (e.g. common knowledge) to readers and writers as members of a speech
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.6 (428)
Esam N. Khalil
community. The information may also be recalled, and hence is a reminder of other information. Background provides orientation and is usually expressed towards the end of the news text.
. Assignment of grounding values The assignment of grounding values to propositions is based on the ways in which information about events is organized in mental models. Information therefore is not a property of text but of the cognitive structure and the knowledge of language users, e.g. as represented in their mental models of events (Johnson-Laird 1983; van Dijk & Kintsch 1983) underlying the text. It is mapped onto semantic representations, i.e. text meaning, as propositions. Information in mental models may be organized in terms of: 1. Greater or lesser importance. Importance is taken as a measure of information in terms of the size of its knowledge implications. It is a property of what language users know about (the structure of) events. Importance implies a hierarchy of facts or events (e.g. with respect to the structure of the world, causes and consequences) as represented in the mental models of language users. For example, information about death is generally more important (i.e. has more news value) than information about sickness, since it has more serious consequences. 2. More or less relevance for language users in the present communicative situation. Information relevance is defined in terms of how consequential facts are in the life of writers and readers. For example, writers may attach more or less relevance to information according to their knowledge, interests, and what they consider to be the possible usefulness or effect of this information for (some) people as well as what is known about their current information needs (see van Dijk 1995). Contextually or pragmatically relevant information is different from – and may be independent of – both the importance of information in event models and the inherent importance of facts, or relevance somewhere in the world, that is, in absolute, context-free terms (e.g. a war is considered more important than the occurrence of a minor border incident). It should be noted that relevance is a key discourse pragmatic notion in the communicative situation and that relevant information fulfils a pragmatic function of informing readers about certain properties of events and states of affairs.10
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.7 (429)
The semantics –pragmatics interface
. Surface structure expression of grounding Surface structure – the last phase in the process of textualization – does not express semantic meaning only, but also properties of the context, such as the knowledge and intentions of language users (see Section 6). It also provides readers with clues as to the intended grounding structure and helps them retrieve and interpret this structure. Since readers interact with the text as they read it (Coulthard 1977), they may capture the various signals that surface structure provides, and successfully interpret the intent of the writer in performing certain linguistic acts (see Green 1989).11 One important surface structure property of discourse is prominence.12 Prominence characterizes the way in which sentences and their constituents stand out in text and hence shows their relative conspicuousness. Since prominence is by definition a gradual phenomenon (Ungerer & Schmid 1996: 182), certain sentences – whether they express propositions about actions, verbal reactions, or states of affairs – may catch the reader’s attention more than others do. In news discourse, what is prominent is expressed in the beginning; and what is non-prominent is expressed at the end. In the linear order of the text, prominence serves a communicative function of influencing the reader’s perception of what is reported. In order to communicate the FG–BG structure and to distinguish gradual and relative prominence in text, writers make use of staging.13 They manipulate sentence structure and sentence order in order to signal different groundingstructures. In the default grounding-structure, the slot for sentences that are prominent (what I will call news upstage) tends to subsume typical FG propositions, and the slot for sentences that are non-prominent (what I will call news downstage) tends to subsume typical BG propositions. In other words, news upstage and news downstage encompass a hierarchy of text segments whose prominence corresponds in a decreasing fashion to the grounding values of the underlying propositions. On the other hand, signaling the non-default grounding-structure (i.e. pragmatically motivated) would require staging operations where news upstage would be the slot for sentences that express BG propositions, and news downstage would be the slot for sentences that express MG propositions.
. Signaling different patterns of grounding In the process of communicating grounding, news writers assign variant grounding values to propositions and express these propositions prominently or nonprominently. These staging operations are governed by different degrees of intentions: whether to express a default or a non-default pattern of grounding.
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.8 (430)
Esam N. Khalil
. Signaling the default FG–BG structure In the default FG–BG structure, information importance is the primary basis for the assignment of grounding values to propositions and for the expression of these values in surface structure. A proposition would therefore be selected for prominent expression on the basis of high importance (semantic ‘importance’) of information. Semantic considerations play the most important role, and the grounding structure is predominantly content-directed. The writer’s intention is to encode grounding values so as to have the main event first in a foreground proposition followed by midground referring to sub-events or details, and background at the end.14 This means that events, objects and states of affairs are presented without giving special significance to certain (background) propositions. Thus in the default FG– BG structure there are textual constraints, with a minimum degree of intervention from the extra-linguistic context. It may therefore be said that texttype canonical FG– BG structure manifests semantics (viz. propositional grounding) under neutral context constraints.15 An essential part of a neutral context is knowledge of text-type conventions and text-type specific FG– BG structure.16 Consider the following text.17 (1) Dutch Government Rejects New Airport AMSTERDAM (AP)— S1 1a The Dutch government rejected plans Friday for a new international airport to be built in the North Sea, 1b calling the 40 billion guilder ($ 18.6 billion) price tag too high. S2 Plans to build runways in the sea were drawn up last year because Schiphol International Airport outside Amsterdam can no longer accomodate explosive growth. (IHT 18–19/12/1999) The text shows a gradual downward progression from FG to BG. S1 has two constituents: 1a, expressing meaning about the main event that the writer has selected as FG, and 1b, which gives details of this event, hence it has a MG interpretation. The meaning expressed in S2 serves a different grounding function, namely BG, since the underlying information is about an earlier event that provides the context (viz. what prompted the plan to build runways in the sea) that is interpreted as less important than the information about the decision by the Dutch government. Surface structure shows the writer’s intention to express this general pattern of canonical grounding-structure where BG propositions tend to map less important information and to be expressed less conspicuously, i.e. later in the text. The default grounding-structure is semantically motivated. Textual constraints propel text-type rules of the FG– BG structure.
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.9 (431)
The semantics –pragmatics interface
. Signaling the non-default FG– BG structure In the non-default FG– BG structure, information relevance is the primary basis for the assignment of grounding values to propositions and for the expression of these values in surface structure. A proposition would therefore be selected for prominent expression on the basis of high relevance (pragmatic ‘importance’) of information. The grounding-structure is governed by strong – or a high degree of – intentions. In other words, contextual or pragmatic relevance constraints (e.g. interests, values, beliefs, ideological preferences by writers) may override the relative importance of information about specific facts and properties of events in the models of language users. In this case, writers may opt to emphasize a certain feature of an event not because it is an important piece of information, but because they interpret it as being more relevant – for them and the lives (and the interests) of readers (as a result of political, socio-cultural, (inter)national or human factors) – than (more) important information. In this case, the FG–BG structure is pragmatically motivated and predominantly context-directed.18 It manifests semantics (viz. propositional grounding) under pragmatic/contextual constraints.19 The intention of the writer is to make readers aware that certain propositions are more relevant at a given point in text and context than other propositions. That they are of immediate concern, is made apparent by prominent expression in the text (e.g. a BG proposition that refers to context). Here the relation between writers and readers is a crucial factor in determining the grounding structure. In other words, the semantic structure interacts with contextual information. Writers’ knowledge and their subjective interpretation of contextual factors determine their communicative behaviour in terms of grounding.
. The influence of pragmatics on grounding Pragmatically motivated grounding-structures – based on the pragmatic importance (i.e. the relevance) of certain meanings to writers – are usually apparent in text surface structure. In what follows I will briefly discuss and illustrate two of the major effects of pragmatic influence on semantic grounding. They are: 1. Foregrounding and backgrounding. 2. Extra expressions: sentence-initial markers in Arabic. The above are characterized by a high degree of intention to influence readers’ interpretation of text meaning (e.g. events and states of affairs) in terms of grounding. They also represent extra dimensions in communicating grounding, since they are governed by the perspective of writers, where the pragmatic relevance of the
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.10 (432)
Esam N. Khalil
facts (see van Dijk 1985b) plays an important role and may take precedence over semantic hierarchy and grounding value (see also van Dijk 1988).
. Foregrounding and backgrounding Foregrounding and backgrounding mean a pragmatic interpretation of the distribution of meanings in terms of grounding values, or rather a pragmatic assignment of higher or lower grounding values in text. This means that foregrounded and backgrounded propositions retain the (original) grounding values that are assigned to them, yet they become more or less informative in a given context. This is realized by expressing propositions respectively as news upstage, i.e. earlier, and news downstage, i.e. later, thus assigning certain facts and events the maximum or the minimum contextual informativeness in the linear ordering strategies. This suggests that foregrounded and backgrounded propositions are not necessarily the same as foreground and background propositions. In fact, they may coexist in surface structure with expressed foreground or background propositions. For example, a background proposition may be foregrounded and prominently expressed as part of the lead sentence that customarily expresses foreground meaning, while some component parts of midground meaning may be backgrounded and non-prominently expressed in the vicinity of other background propositions later in the text. I will illustrate three features of foregrounding/backgrounding operations.
A. Foregrounding in the lead sentence The lead sentence may be the locus for foregrounding operations. Consider, for example, the following text. (2) Heavy Rains Bring Floods to Botswana GABORONE, Botswana— S1 Rain, normally a blessing in this African desert country, came down in torrents Friday, virtually cutting off the capital, washing out a railroad and flooding homes and a prison. S2 Five straight days of rain have killed at least 74 people in southern Africa, including Mozambique, where helicopters were deployed to pluck stranded villagers to safety. S3 Thousands of motorists remained stranded by washed-out roads and bridges in South Africa, Mozambique and Botswana. S4 The flooding was the worst in memory in Botswana, a country of 1.5 million people, where rain is so precious that the local currency is called the “pula”, which means rain. (AP) (IHT 12–13/2/2000)
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.11 (433)
The semantics –pragmatics interface
The writer of this text introduces background meaning about the significance of rain in Botswana – a welcome natural phenomenon – immediately after the introduction of the first component of foreground in the lead sentence about floods and damage to humans and infrastructure. In the situation described, the background proposition acquires a high communicative value irrespective of any possible inherent information unimportance and low grounding value. The writer’s mental model structure of facts includes information inferred from parts of scripts such as those about the country and the climate. Under the influence of communicative context constraints, this information is assumed to be relevant for expression. The inclusion of this information reflects the knowledge of the writer and his/her intention to communicate this to readers as early as possible in the text. Readers (are assumed to) recognize the writer’s intention as regards grounding-structure and realize that some manipulation has occurred in the propositional content, namely that a proposition having a low grounding value has been assigned more than usual prominence, i.e. it has been foregrounded.
B. Foregrounding in sentence-initial position Communicative context constraints may lead to the expression of BG meaning before MG through the use of constituents such as participles or adverbials in sentence-initial position. Fronted constituents retain their background status or function, serving as the ‘ground’ or ‘scene-setting’ for what follows (Quirk et al. 1985: 491) as well as for what has gone before. Sometimes a constituent in sentenceinitial position expresses what is called ‘a complete setting’ (Geis 1987: 107) such as when a complex adverbial denotes features of an event, time, and place. Complex adverbials in sentence-initial position occur often in news reporting (Geis 1987). The constituent 2a in text (3) illustrates the grounding-signaling function of the participle. (3) Japan’s Obuchi Visits Thailand BANGKOK— S1 Prime Minister Keizo Obuchi of Japan began a two-day visit to Thailand on Thursday aimed at enhancing a close bilateral relationship and putting Asia’s economic recovery on a more solid footing, officials said. S2 2a Arriving in Bangkok for the last stop on a tour of Southeast Asia that has taken him to Cambodia and Laos, 2b Mr. Obuchi met with Prime Minister Chuan Leekpai, who thanked him for support in dealing with the economic crisis. S3 Mr. Chuan added that Thailand still needs help from Japan – its second-largest trading partner – to complete its recovery, A Thai government spokesman said. (Reuters, AP) (IHT 14/1/2000)
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.12 (434)
Esam N. Khalil
Initiated by the -ing participle, 2a encodes meaning about the context of the visit of the Japanese premier to Bangkok as well as about previous visits in the same tour. Because of its position early in the text, background in 2a is relatively prominent. It shows the writer’s communicative intention to signal as prominent – or assign more prominence to – a BG meaning that maps non-important information and to provide readers with this information at that point in text and prior to providing further details of the main event (i.e. midground) in 2b. In this position (i.e. early in the sentence, that is, in initial position, and early in the text, that is, as part of news upstage), the information acquires/reflects greater contextual or pragmatic relevance, and hence it is interpreted as of immediate concern. A background proposition, of course, may be expressed in a main clause or an independent sentence. Thus 2a may be expressed as follows: S2 Mr. Obuchi arrived in Bangkok for the last stop on a tour of Southeast Asia that has taken him to Cambodia and Laos. In a typical grounding structure, this sentence may also be expressed later in the text, as is customary for BG propositions with less importance.
C. Backgrounding in sentence order Pragmatic motivation may be apparent in backgrounding that involves the order of whole sentences. Consider text (4). (4) Britain and Iran Strengthen Ties LONDON— S1 Couching differences over human rights issues in diplomatic niceties, Britain and Iran signed a joint declaration Tuesday to promote trade, drug trafficking controls and political and academic contacts. S2 The declaration at the end of a two-day visit to Britain by the Iranian foreign minister, Kamal Kharrazi, marked a giant leap in relations between two countries that only exchanged ambassadors in May after a 20-year break. S3 Mr. Kharrazi, who was dogged by demonstrators both days, and the British foreign secretary, Robin Cook, also agreed in the declaration on the need to “fight against all forms of terrorism wherever it occurs.” (AP) (IHT 12/1/2000) S1 expresses FG meaning about the joint declaration that Britain and Iran signed. S2 expresses BG meaning about the context of the declaration: when it has been signed and what its significance is. Background meaning has been assigned greater relevance, demonstrating the writer’s intention to make this information more informative than it usually is. Further details of the declaration that have midground
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.13 (435)
The semantics –pragmatics interface
value are expressed in the last sentence. It has been backgrounded and expressed less prominently, based on the assignment of a lower degree of relevance that may not (necessarily) coincide with a similar degree of importance.
. Extra expressions: Sentence-initial markers in Arabic The second major effect of pragmatic influence on grounding is the occurrence in some languages of extra expressions in sentence-initial position. A Semitic nonIndo-European language like Arabic employs several prefatory expressions that serve a variety of discourse functions, including the function of communicating grounding. Empirically-based observations about differences between English and Arabic in ways of signaling grounding at sentence-initial position demonstrate that Arabic news texts are more explicit than their English counterparts in marking at that position the relative grounding value of text meaning (for details, see Khalil 2000). Unlike other connectives (in other types of text and other languages), these markers are obligatory in Arabic news texts. These extra markers are context sensitive. Their use is subject to the writer’s interpretation of the grounding value of the proposition in the sentence that the marker prefaces. Their use also makes explicit the writer’s communicative intentions with regard to the way the FG–BG structure should be understood and interpreted by readers. In this sense, these extra expressions may be considered as speech acts that characterize the grounding status of propositions and provide instructions to readers and help them interpret – successfully – the grounding value intended by the writer. One way in which Arabic news texts differ from English news texts in the explicitness of signaling background at sentence-initial position is apparent in text (2) cited earlier. S4 encodes BG meaning about the nature of the floods in Botswana and about how rain is viewed there. In order to express that meaning in Arabic, it would be necessary for the writer of the Arabic text to use an initial marker such as the following: Mimm¯a yudkaru anna20 (among things to be mentioned is that) By stating that what follows is among things to be mentioned within the framework of what has already been reported, the writer makes explicit his/her intention towards the grounding value of the upcoming proposition: that the underlying information is incidental, but also relevant and hence recalled.21 By using this marker, the writer performs an illocutionary act of informing readers of certain facts and events, assuming their lack of knowledge of them. This feature shows that contextual factors where writers and readers are present govern the way grounding values are encoded. It also shows that there
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.14 (436)
Esam N. Khalil
are also semantic constraints on the expression of grounding: that groundingstructures determine the selection of the appropriate linguistic marker in sentenceinitial position. Furthermore, the use of initial markers in Arabic news texts may determine – or even modify – the (interpretation of the) grounding value of the proposition.
. Conclusion Taking a text-level property of discourse, namely grounding, the paper examined the semantics– pragmatics interaction and suggested that while certain grounding patterns are usually accounted for in terms of semantics, pragmatics may account for certain other patterns. That grounding does not fall exclusively within either semantics or pragmatics supports the view that ‘both pragmatics and semantics turn into inter-disciplinary endeavors’ (Jucker 1998: 830). In this regard, the paper has shown that dynamic semantics conveys some pragmatic meaning and that pragmatics contributes some semantic interpretation. The paper highlighted two key sources for the assignment of grounding values to propositions in text (i.e. information importance and information relevance) and related them to the semantics– pragmatics interface as well as to the degree of pragmatic influence. It also suggested that propositions on the FG– BG scale are selected for prominent expression in text on the basis of high importance (semantic importance) of information and high relevance (pragmatic importance) of information. The terminological distinction that has been made between foreground and foregrounded meaning and between background and backgrounded meaning suggests that event model importance of information is the source for the assignment of grounding values to propositions, and that context model relevance of information is the source for foregrounded and backgrounded meanings. The paper also examined the pragmatic notion of intentions and suggested that different grounding-structures are governed by different types of context as well as by varying degrees of intentions. The paper discussed and illustrated a number of effects of pragmatic – or communicative context – constraints on semantic grounding. It showed that pragmatic influence on grounding is apparent in text organization and in the use of linguistic markers such as the use in Arabic texts of extra markers in sentence-initial position. Governed by contextual factors, these expressions are text acts that have the function of maximizing the signaling of the FG– BG distinction.
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.15 (437)
The semantics –pragmatics interface
Notes . ‘In text pragmatics, a ‘text’ is defined as a communicative event in which the communicator transfers to the audience, by means of language, and with social consequences, some propositional content’ (Togeby 1998: 1008). Text pragmatics ‘is not a well-established discipline within linguistics or pragmatics’ (ibid.). It is ‘a name for certain non mainstream theories about the functions of language in communication’ (ibid.) . In its narrowest interpretation, pragmatics involves the ‘interpretation of indexical expressions (words or phrases like me, here, then whose REFERENCE cannot be determined without taking into account the context of the utterance of a linguistic expression)’ (Green 1989: 2). . In default semantics (Jaszczolt 1999), semantic representations are established with the help of a default intention in communication. It should be noted that my use – and interpretation – of (the meaning of) the notion of default semantics applies to discourse and discourse grounding. . ‘. . . it is in the context of communication that the question arises of where to draw the line between semantics and pragmatics’ (Bach 1999: 67). . Intentions are considered constitutive parts of context (Jaszczolt 2000b: 170). . ‘Contextualism allows for the contextual information to contribute to the propositional form of an utterance. Anti-contextualism regards contextual information as implicatures that function in addition to the propositional form in the process of utterance interpretation. . . ’ (Jaszczolt 2000a: 5). . In the broad context the extra-linguistic context includes any information that the hearer takes into account to determine the speaker’s communicative intention (Bach 1999: 72). . It has been suggested that semantics ‘is no longer always restricted to decontextualized analyses of word or sentence meanings but includes, for instance, the intended social function of an utterance as part of its meaning. . . ’ (Jucker 1998: 830). . Depending on text-type, other interpretations of the significance of sentence-order might be possible. In certain narratives, a background proposition may be realized in an early position and a foreground proposition about the point of the story may be expressed later, where that position is construed as having more prominence. In the narrative text-type, the ordering strategy may correspond to a natural or chronological text-world ordering pattern. . Although relevance has not been investigated thoroughly in extended written discourse (in fact the analysis of relevance has been limited to the sentence or utterance level (see e.g. Allwood 1985)), some work has been made on written communication such as the discussion of relevance in news discourse (see van Dijk 1984, 1985a: 79). . The hearer seeks to identify/recognize the speaker’s intention in making the utterance (Bach 1999: 74). ‘In effect the hearer seeks to explain the fact that the speaker said what he said, in the way he said it’ (ibid.). . Prominence is ‘a formal, surface structure notion, viz. defined in terms of the (set of) textual devices that express importance or relevance of information’ (van Dijk 1995: 263).
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.16 (438)
Esam N. Khalil . The stage metaphor has been used to refer to the particular perspective from which speakers present what they want to say (Grimes 1975), and to the ways in which the sender communicates relative importance (Grimes 1982). According to the stage metaphor ‘the setup of the constituents in a sentence is similar to what happens in a play on the stage’ (Ungerer & Schmid 1996: 182). . It should be noted that a FG proposition – representing a main event – that is high in the semantic hierarchy may coincide with a macrotopic/macroproposition, and a BG proposition – representing a sub-event – that is low in the semantic hierarchy may coincide with a microtopic/microproposition. . If we look at meaning in terms of levels (see Thomas 1995: 2ff), what is relevant is not the first level of abstract (i.e. decontextualized) meaning but the second level of contextual meaning (ibid.). At that level of meaning, the default foreground – background structure is constructed. . Grounding-structures are assumed to be part of readers’ passive knowledge. ‘Readers are required to use their knowledge of text structures and the world to construct parallel or similar interpretations of the text to that intended by the writer’ (Meyer 1985: 66). It is also suggested that ‘discourse structure is an essential component in discourse interpretation and results from integrating pragmatic and semantic information together’ (Asher 1999: 20). . Illustrative examples are taken from the International Herald Tribune (IHT). Sentences and sentence segments are numbered for the purposes of the analysis. . What is relevant is a ‘broad context’ (see Bach 1999) and hence the grounding-structure is relevant to pragmatics. . This is perhaps similar to the third level of meaning: when the speaker’s or writer’s communicative intention, that is the force of an utterance (Thomas 1995: 2), is of relevance. . The first word in the expression mimm¯a yudkaru anna is an assimilation of min (from/among) and m¯a (what). It denotes partition or division. Yudkaru (is mentioned) is the imperfective passive form of verb dakara (mention). The particle anna (that) introduces the subject. . Writers may include information they want readers to be reminded of (see Searle 1969).
References Allwood, J. (1985). On relevance in spoken interaction. In S. Bäckman & G. Kjellmer (Eds.), Papers on Language and Literature (pp. 18–35). Göteborg: Acta Universitatis Gothoburgensis. Asher, N. (1999). Discourse structure and the logic of conversation. In K. Turner (Ed.), The Semantics/Pragmatics Interface from Different Points of View (pp. 19–48). Oxford: Elsevier. Austin, J. L. (1962). How to Do Things with Words. London: Oxford University Press.
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.17 (439)
The semantics –pragmatics interface
Bach, K. (1999). The semanitcs-pragmatics distinction: What it is and why it matters. In K. Turner (Ed.), The Semantics/Pragmatics Interface from Different Points of View (pp. 65–84). Oxford: Elsevier. Coulthard, M. (1977). An Introduction to Discourse Analysis. London: Longman. Crystal, D. (1997). The Cambridge Encyclopedia of Language (2nd ed.). Cambridge: Cambridge University Press. van Dijk, T. A. (1977). Text and Context: Explorations in the Semantics and Pragmatics of Discourse. London: Longman. van Dijk, T. A. (1984). Structures of International News: A Case Study of World’s Press. University of Amsterdam, Department of General Literary Studies, Section of Discourse Studies. van Dijk, T. A. (1985a). Structures of News in the Press. In T. A. van Dijk (Ed.), Discourse and Communication: New Approaches to the Analysis of Mass Media Discourse and Communication (pp. 69–93). Berlin: Walter de Gruyter. van Dijk, T. A. (1985b). Semantic Discourse Analysis. In T. A. van Dijk (Ed.), Handbook of Discourse Analysis, Vol. 2: Dimensions of Discourse (pp. 103–136). London: Academic Press. van Dijk, T. A. (1988). News Analysis. Hillsdale, NJ: Lawrence Erlbaum. van Dijk, T. A. (1995). Discourse semantics and ideology. Discourse & Society, 6, 243–289. van Dijk, T.A. (1999). Context models in discourse processing. In H. van Oostendorp & S. R. Goldman (Eds.), The Construction of Mental Representations During Reading (pp. 123–148). Mahwah, NJ: Lawrence Erlbaum Associates. van Dijk, T. A., & Kintsch, W. (1983). Strategies of Discourse Comprehension. New York: Academic Press. Garrett, P., & Bell, A. (1998). Media and discourse: A critical overview. In A. Bell & P. Garrett (Eds.), Approaches to Media Discourse (pp. 1–20). Oxford: Basil Blackwell. Geis, M. L. (1987). The Language of Politics. New York: Springer-Verlag. Givón, T. (1987). Beyond foreground and background. In R. S. Tomlin (Ed.), Coherence and Grounding in Discourse (pp. 175–188). Amsterdam: John Benjamins. Green, G. M. (1989). Pragmatics and Natural Language Understanding. Hillsdale, NJ: Lawrence Erlbaum. Grice, P. H. (1975). Logic and Conversation. In P. Cole & J. L. Morgan (Eds.), Syntax and Semantics, Vol. 3: Speech Acts (pp. 41–58). New York: Academic Press. Grimes, J. E. (1975). The Thread of Discourse. The Hague: Mouton. Grimes, J. E. (1982). Reference spaces in text. In S. Allén (Ed.), Text Processing. Proceedings of Nobel Symposium 51 (pp. 381–414). Stockholm: Almqvist & Wiksell International. Jaszczolt, K. M. (1999). Default semantics, pragmatics, and intentions. In K. Turner (Ed.), The Semantics/Pragmatics Interface from Different Points of View (pp. 199–232). Oxford: Elsevier. Jaszczolt, K. M. (2000a). Introduction: Belief reports and pragmatic theory: The state of the art. In K. M. Jaszczolt (Ed.), The Pragmatics of Propositional Attitude Reports (pp. 1–12). Oxford: Elsevier. Jaszczolt, K. M. (2000b). The default-based context-dependence of belief reports. In K. M. Jaszczolt (Ed.), The Pragmatics of Propositional Attitude Reports (pp. 169–185). Oxford: Elsevier.
TSL[v.20020404] Prn:3/09/2002; 12:57
F: PBNS40.tex / p.18 (440)
Esam N. Khalil
Johnson-Laird, P. N. (1983). Mental Models: Toward a Cognitive Science of Language, Inference, and Consciousness. Cambridge: Cambridge University Press. Jucker, A. H. (1998). Semantics and pragmatics. In J. L. Mey (Ed.), Concise Encyclopedia of Pragmatics (pp. 830–831). Amsterdam: Elsevier. Khalil, E. N. (2000). Grounding in English and Arabic News Discourse. Amsterdam: John Benjamins. Leech, G. N. (1983). Principles of Pragmatics. London: Longman. Meyer, B. J. F. (1985). Signaling the structure of text. In D. H. Jonassen (Ed.), The Technology of Text, Vol. 2 (pp. 64–89). Englewood Cliffs, NJ: Educational Technology Publications. Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J. (1985). A Comprehensive Grammar of the English Language. London: Longman. Searle, J. R. (1969). Speech Acts: An Essay in the Philosophy of Language. Cambridge: Cambridge University Press. Thomas, J. (1995). Meaning in Interaction: An Introduction to Pragmatics. London: Longman. Togeby, D. (1998). Text pragmatics. In J. L. Mey (Ed.), Concise Encyclopedia of Pragmatics (pp. 1008–1010). Amsterdam: Elsevier. Ungerer, F., & Schmid, H.-J. (1996). An Introduction to Cognitive Linguistics. London: Longman.
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.1 (441)
On translating ‘what is said’* Tertium comparationis in contrastive semantics and pragmatics K. M. Jaszczolt University of Cambridge
.
Levels of equivalence
This paper consists of the following parts: it (i) introduces one of the main problems of theoretical contrastive linguistics, namely the issue of the platform of comparison, (ii) postulates an adequate unit of comparison for contrastive semantics and pragmatics, (iii) discusses the controversy surrounding the semanticspragmatics boundary and, finally, (iv) assesses the impact of this dispute on translation theory. First, some terminological remarks are needed. It has been widely acknowledged that contrastive analysis has to comprise theoretical linguistic research on all of the levels of linguistic study, as well as psycholinguistic and sociolinguistic studies (cf. Sajavaara 1984; James 1980). The term ‘contrastive analysis’ stands for applied contrastive studies (henceforth: ACS) which deal with the practical consequences of differences between contrasted languages for teaching purposes, bilingual analysis or translation. The theoretical linguistic component of ACS is very important and it gave rise to theoretical contrastive studies (henceforth: TCS). TCS can be performed on the level of phonology, lexicon, syntax, semantics, pragmatics or text linguistics. Such theoretical studies deal with a universal category and the ways this universal category is realized in contrasted languages. In phonology, for example, the functioning of phonological features is contrasted (cf. Fisiak et al. 1978). The main problem of TCS is how to contrast languages, i.e. against what criterion of measurement. Depending on the platform of reference called by Krzeszowski (1990: 15) ‘tertium comparationis’, two objects of analysis may appear either similar or different. Also, it is obvious that each level of linguistic analysis will
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.2 (442)
K. M. Jaszczolt
have its own criterion of comparison. For pragmatics, it is common to contrast the illocutionary force. For example, sentences (1) and (2) can be contrasted as responses to a speech act of complimenting on one’s appearance in English and in Polish. (1) Thank you. (2) To tylko stara sukienka. (It’s only an old dress.) Semantically, syntactically and lexically they hardly exhibit any similarities which would justify a comparison. But they may both constitute answers to a speech act of complimenting: thanking in English, and self-denigrating in Polish (cf. Jaszczolt 1995a, b). So, when we talk about an equivalence for example in translation, it is necessary to distinguish syntactic, semantic and pragmatic equivalence. It may well be the case that there is also an overall equivalence which is a function of these (cf. Kalisz 1981), but in this paper I intend to concentrate on semantic and pragmatic equivalence and on the difficulty with defining the unit of analysis for each level. As I shall demonstrate, since semantics does not have a clear boundary with pragmatics but rather they overlap, the translator’s decision as to whether to adopt semantic or pragmatic equivalence will not always be possible. It has been acknowledged at least since the 1960s and Nida (1964), that ‘faithful’ translation means translating the author’s intentions, assumptions, rather than structures and style (see Gentzler 1993: 58). So, since “. . . the equivalence between a text and its translation can be neither in form nor lexical meanings, but only in the experience of text receivers” (De Beaugrande 1980: 291), we want to know how this identity of experience is to be achieved in the multi-layered process of translation. The semantic layer alone won’t do; since the meaning is often culture-bound, we need the socio-cultural layer. For instance, a Russian idiomatic greeting ‘with easy steam’ can only reasonably be translated as a comment ‘I hope you had a good bath’, and even for non-idiomatic expressions we have to resolve what situation (state, event) the sentence refers to. Superficially, contrastive pragmatics may seem unproblematic: conversational effects should be kept constant and one should look at the contrast between the ways languages achieve these effects. Here the obvious areas of study are illocutionary forces of exclamations, differences in levels and meaning of self-assertion, differences in terms of address, and many other culture-bound phenomena. Now, in order to contrast languages on the pragmatic level, one has to decide what the equivalence of contrasted structures on the pragmatic level means. The following definition was proposed by Oleksy (1984): pragmatic equivalence holds between two expressions in L1 and L2 if they can be used to perform the same speech act in these two languages. All we have to do now is to keep speech acts steady and look at the sets of strategies used in L1 and L2 to perform these acts. The problem arises
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.3 (443)
On translating ‘what is said’
as to whether to admit indirect speech acts. As has been frequently pointed out in the pragmatic literature, the direct/indirect distinction for speech acts is untenable. Sperber and Wilson (1995: 245) give the following example: (3) The weather will be warmer tomorrow. (4) The speaker is predicting that the weather will be warmer tomorrow. Sentence (3) can successfully function as a prediction without the speaker’s intending to communicate the information in (4). Indeed, (4) need not be recovered by the hearer at all for (3) to function as a prediction. On the other hand, in (5), the act of bidding has to be communicated, either directly or by inference. (5) (I bid) two no trumps. Speech acts are either institutional, social, like bidding in bridge or thanking, or performed without being recognized as such, e.g. warning, threatening, or their category can be recognized and universal, as in the case of saying, telling and asking. Also, as is well known from the collapse of the performative hypothesis and the literal force hypothesis (Levinson 1983), there is no reliable correlation between the sentence type and speech act type on the one hand, and the meaning of the performative verb and the type of speech act on the other (see also Sperber & Wilson 1995: 246–247). Hence, instead of relying on speech acts, it may be more adequate to follow Sperber and Wilson and talk about the recovery of the propositional form of the speaker’s utterance which is an interpretation of a mental representation of the speaker’s and which is entertained with an appropriate attitude to render assertions, questions, requests, and advice (cf. Sperber & Wilson’s diagram in 1995: 232). Also, since there is no clear-cut definition of directness, there is no one-to-one correlation between sentences and acts, even within one language. Cross-linguistically, the situation becomes more complicated due to the fact that an illocution in one culture can be a perlocution in another (cf. Wierzbicka 1991). So, the speech act is not an adequate tertium comparationis, we need a different unit of pragmatic equivalence.
. Semantic and pragmatic equivalence We can say that expressions are pragmatically equivalent if they communicate the same content. They are not necessarily also semantically equivalent. In a language in which metaphors are a common means of expression, one may use a metaphor instead of speaking literally. Here I would like to concentrate on cases of literal meaning. Grice proposed a well-known distinction between what is said and what is implicated, distinguishing truth-conditional aspects of meaning as what is said,
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.4 (444)
K. M. Jaszczolt
and conventional and conversational implicatures as what is implicated (cf. the diagram in Horn 1988: 121). In this distinction, semantics and pragmatics overlap: there is no clear-cut boundary. Conventional implicatures, such as the meaning of contrast in ‘but’, the conclusion to premises in ‘therefore’, or overcome difficulty in ‘manage’, do not contribute to the truth-conditional content of sentences and yet are part of word meaning. Generalized conversational implicatures, such as enrichment from ‘three’ to ‘exactly three’, do not require context for their occurrence and are default interpretations, regarded by some as (i) semantic, e.g. in the framework of Discourse Representation Theory (Kamp & Reyle 1993) or Default Semantics (Jaszczolt 1999a, b), by others as (ii) belonging to the middle level of meaning (Levinson 1995, 2000), and (iii) are denied the default status and classified as context-dependent nonce-inference by relevance theorists. So, they can be regarded as either semantic or pragmatic, unlike particularized conversational implicatures which are context-dependent and certainly arrived through pragmatic processes of inference. Bearing this difficulty in mind, I suggest trying the following hypotheses: A1
Semantic equivalence is the equivalence of what is said.
A2
Pragmatic equivalence is the equivalence of what is implicitly communicated.
These definitions are not very informative as they stand. Since the problem of the fuzzy boundary between semantics and pragmatics is unresolved, we are only pushing the terminological difficulty one step on to the equally problematic notions of ‘what is said’ and ‘what is implicated’. But there is an advantage to be gained. What is said and what is implicated have been subject to extensive studies and heated debates in the last decade (e.g. Carston 1988, 1998a; Récanati 1989, 1993; Bach 1994a, b; Levinson 1995, 2000; Jaszczolt 1999a, b). The starting point to the debate is the observation that Grice seriously underestimated the role of pragmatic processes in establishing the representation of the utterance of the sentence which can be subject to the provision of truth conditions, i.e. the propositional form. In addition to reference assignment and disambiguation which he acknowledged, there are many processes of enrichment of the proposition, or patching up of the incomplete propositional form, which have to be performed in order to arrive at the relevant, truth-evaluable representation. For example, the sentential connective ‘and’ can be enriched to include the indication of temporal sequence or causal consequence and this enrichment is relevant to the truth conditions of the proposition, which can be tested, for example, by placing the sentences in the scope of logical operators such as negation or implication. I shall return to the tests for contributions to truth conditions in Section 6. Suffice it to say that what is said and what is implicated seem to constitute a promising departure for the improvements on the definitions of semantic and pragmatic equivalence in translation.
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.5 (445)
On translating ‘what is said’
To sum up, there have been various proposals of how to draw the boundary between what is said and what is communicated. The main standpoints can be summarized as follows: i.
Some sentences are semantically ambiguous. This traditional view was advocated, among others, by Russell. ii. There is no semantic ambiguity. The differences in meaning between the two (or more) readings can be attributed to implicated information. This is Grice’s postulate of unitary semantics complemented with conversational implicatures. iii. There is no semantic ambiguity and the differences in meaning belong, explicitly or implicitly, to what is said rather than what is implicated. Semantics is underspecified as to these aspects of meaning. The process of supplementing the semantic form with pragmatic information is called completion (saturation) and expansion (strengthening; cf. Bach 1994a, b; Récanati 1989). According to Bach, these aspects constitute a level of what is implicit in what is said (impliciture). iv. There is no semantic ambiguity, semantics is underspecified, but the differences in meaning belong to the middle level located between semantics and pragmatics. The main proponent of this view is Levinson (e.g. 1995, 2000) and the idea is that there are default meanings arrived at through the three heuristics, Quantity, Informativeness and Manner, corresponding to his rearrangement of Grice’s maxims of conversation. On this view, what is said and what is implicated are not disjoint. In addition, in my other work (e.g. Jaszczolt 1997, 1998a, b, c, 1999a, b, 2000 and especially 2002), I presented arguments against both the underspecification and ambiguity positions by demonstrating that within the frameworks that espouse the dynamic approach to meaning construction, such as Discourse Representation Theory (henceforth: DRT, Kamp & Reyle 1993) or File Change Semantics (Heim 1988), a non-ambiguous interpretation ensues without the need for the level of underspecified semantics. In other words, I demonstrated that the term ‘underspecification’ has been grossly overused and its application should be confined to the context of ambiguous logic (see van Deemter 1998).1 Instead, there are default representations arrived at through the interaction of the propositional form with the intentions in communication. Hence, the fifth option is v.
Default Semantics.
In this paper, I shall put the differences between (iii), (iv) and (v) aside and examine how the fact that pragmatic processes contribute to the propositional representation affects the definitions of semantic and pragmatic equivalence.
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.6 (446)
K. M. Jaszczolt
As a result of this multitude of standpoints, instead of talking directly of the semantics/pragmatics boundary, I shall now attempt to delineate what is said and what is implicitly communicated. What is said is a proposition, an entity which is either true or false. Hence, we can try the following: A1
Two sentences are semantically equivalent if they correspond to the same proposition.
A2
Two sentences are pragmatically equivalent if they render the same set of implicatures, triggered by rules of conversational inference.2
‘Corresponding to’ the same proposition can be investigated by comparing the propositional form and, usually, the truth conditions.3 A2 is obviously contextdependent. But, so is A1 . Firstly, there are intensional contexts such as belief contexts (‘A believes that B φs.’) in which the mode of presentation of the referent contributes to the truth conditions (cf. Schiffer 1992; Jaszczolt 1998a). Secondly, not all sentences express full propositions, and not all propositions expressed by sentences are sufficiently informative to be taken as what is said without enrichment, to quote only (6) and (7). (6) Steel isn’t strong enough [for what?]. (7) I haven’t eaten
[since when?].
(Bach 1994a: 268)
Sentence (6) does not express a complete proposition, although syntactically it is complete, it has a complete logical form (which is the output of grammatical analysis). In order to provide truth conditions, an essential pre-condition has to be fulfilled: we have to have a proposition, a unit to which we can assign truth conditions. Hence, we have to complete (6) by stating what it is that steel is not strong enough for. Sentence (7) is different: it expresses a complete proposition, but it is a proposition which is blatantly false and dubious as a candidate for what the speaker intended to communicate. Assigning truth conditions to (7) is unproblematic: the proposition is true just in case there is no time prior to the time of the utterance when the speaker (here we need reference assignment to the indexical expression ‘I’) ate. But just as it is easy, so it is pointless to do so. Rather, we should infer the speaker’s assumptions first, expand the proposition to make it include all contextually salient information pertaining to this particular proposition, and only then assess the expanded proposition as to what would make it true or false. Logical form as the output of syntactic rules will not suffice to draw the boundary between semantic and pragmatic equivalence. So, all in all, I shall talk about what is said instead and its pragmatic constituents (or: ‘what is implicitly stated’) – completing and expanding, respectively, the logical form of the sentence.4 In order to preserve the definition of semantic
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.7 (447)
On translating ‘what is said’
equivalence in A1 , one has to resolve these difficulties of enrichment and completion. A1 is worth preserving because there does not seem to be a more adequate definition of sentence meaning than that provided by truth-conditional semantics. So, what is said is partly semantic and partly pragmatic, as is what is implicated: implicatures have their own logical forms and hence their own semantics, but they are the product of pragmatic processes of inference.
. Cultural rationalism Now, in order to accept A1 and A2 we need to specify one further theoretical commitment. Namely, we accept that propositions are recognizable cross-culturally, that is, that it is not the case that people of different cultures live in different worlds. In other words, we accept cultural rationalism (and realism) against cultural relativism and in this way give pragmatic equivalence access to a universally, crossculturally valid theory of implicature of any post-Gricean flavour. But this commitment is not driven by the need to ease the task of defining pragmatic equivalence, this move would be circular. The justification comes from the strength of the independent arguments for realism. Presenting them is the task to which I shall now turn. At first glance, it may seem that relativity is the dominant orientation in linguistics, brought to the fore by cognitivists (e.g. Lakoff 1987), standing in direct opposition to universalism. Sperber (1985, 1996, 1997) takes universalism and cultural relativism further. He approaches culture as follows: Ideas can be transmitted, and, by being transmitted from one person to another, they may even propagate. Some ideas – religious beliefs, cooking recipes, or scientific hypotheses, for instance – propagate so effectively that, in different versions, they may end up durably invading whole populations. Culture is made up, first and foremost, of such contagious ideas. (Sperber 1996: 1)
So, in order to explain culture, one has to explain how it happens that some ideas are ‘contagious’. This enterprise is dubbed by Sperber (1996: 1) an ‘epidemiology of representations’. During the process of transmission, the original information is affected by the receiver’s memory and set of assumptions. In other words, communication normally slightly transforms the message. Now, explaining cultural beliefs boils down to explaining the distribution of ideas in a population, explaining how individual processes of adopting an idea contribute to its dissemination. It is essentially a Darwinian approach, linked to those of Cavalli-Sforza and Dawkins, where the Darwinian idea of selection is applied to culture. Sperber calls it a ‘naturalistic
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.8 (448)
K. M. Jaszczolt
approach’ because social issues are approached through the study of cognition, or through psychological processes of individuals who combine to form a society. Ideas propagate and change, sometimes they are understood, sometimes misunderstood or half-grasped. Since memory and communication transform original ideas, the ideas are stored as representations of the world or of the speaker’s representations.5 In other words, Social-cultural phenomena are (. . . ) ecological patterns of psychological phenomena. Sociological facts are defined in terms of psychological facts, but do not reduce to them. (Sperber 1996: 31)
The idea is that people have representations of situations in the form of memory, belief, or intention. Representations can also be public – for example written texts whose function is communication are such public representations. Now, some of these representations are repeated, communicated more often than others and become cultural representations of a social group. These are shared beliefs, norms, myths, techniques, or classifications. Such representations are conveyed by producing another representation, which is either a description or an interpretation of the original representation (cf. the diagram in Sperber & Wilson 1985: 232). Interpretation occurs when, for example, the receiver has to expand on the content of some tribal belief in order to make it understandable. In other words, we make partial and often speculative interpretations in order to understand others, especially representatives of different cultures. Then we attribute beliefs, desires and intentions to people or communities in a way which makes them look rational. And that is why we need interpretations: The anthropologist must (. . . ) go beyond mere translation: only then can she hope to understand what she hears, and thus be genuinely able to translate it. She must speculate, synthesize, reconceptualize. (Sperber 1996: 39)
Putting aside the status of the modal verbs in the quotation, we can summarize Sperber’s proposal as follows. To understand cultural phenomena, the observer/hearer can either (i) generalize over them through making an interpretation of one phenomenon wider and wider, applicable to other phenomena; (ii) look for common patterns, themes or relations between these patterns in a structuralist manner; (iii) show how a cultural phenomenon is beneficial for the group by way of a functional explanation, or (iv) answer the question why some representations are more successful than others and become shared, cultural representations, i.e. undergo epidemiology of representations (cf. Sperber 1996: 50). What is of interest for the search for a translation equivalence is the idea that representations are not reproductions of what was said but rather construction by the hearer of his/her own thoughts which are normally related rather closely to the thoughts represented by what is said (cf. Sperber 1996: 58). So, if a tribesman strives
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.9 (449)
On translating ‘what is said’
to understand a representative of a different culture, he/she will store some of the representations in a semi-understood way, hoping to fill them in with the intended meaning at a later stage. For example, if a tribesman says that there are dragons in the woods, we take this proposition literally and assume that it is stored in the native’s memory not in a fully decomposed way but rather semi-propositionally, as a semi-understood idea, as a belief about some representation, rather than about a perceptually verifiable fact. We do that rather than jumping to the conclusion that the worlds of the two cultures are incompatible. It is a metarepresentation, a representation of someone else’s representation. Such rationalism allows for comparing propositions cross-culturally. Sperber (1985: 35) gives the following example. A tribesman of Southern Ethiopia comes to Sperber and asks him to kill a dragon whose heart is made of gold, who has a horn on his neck, is golden all over and lives somewhere nearby. Sperber believes the tribesman to be a sensible, respectable person, but how would a sensible person believe that there was a dragon living within walking distance? This is an example of an apparently irrational belief. Many anthropologists would say that people of different cultures live in different worlds, that beliefs in dragons may appear rational to some cultures; others advocate symbolism. Sperber rejects these views. Instead, he believes in rationalism. He says that people can have various types and degrees of commitment to beliefs – beliefs are not all held in the same way, and so the criteria of rationality can also differ. Rationalism acquires additional support from Sperber’s account of concept formation which is compatible with epidemiology: . . . we have an innate disposition to develop concepts according to certain schemas. We have different schemas for different domains: our concepts of living kinds tend to be taxonomic; our concepts of artefacts tend to be characterized in terms of function; our concepts of colour tend to be centred on focal hues; and so on. Concepts which conform to these schemas are easily internalized and remembered. Let us call them basic concepts. A large body of basic concepts is found in every language. Of course, basic concepts differ from one language to another, but they do not differ very much. The basic concepts of another language tend to be comparatively easy to grasp, learn and translate. (Sperber 1996: 69)
So, there are universal schemas for basic concepts, rules for forming representations and universal patterns of their spread. The point of difficulty for delimiting pragmatic equivalence seems to be located in assessing the type of representation formed by the hearer. Now, we can summarize Sperber’s account by saying that humans can represent not only the sensory input and experience, but also someone else’s and their own representations and mental states. This is called metarepresenting and allows
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.10 (450)
K. M. Jaszczolt
us to have, for example, an attitude of disbelief or doubt towards a representation. Metarepresenting also allows us to store representations which are not fully understood. Some such half-understood ideas are not steps towards full understanding but rather create mysteries which spread in the community. Representations which fit best with other mental representations, and which cannot be interpreted, are most successful. They lead to what we see as irrational cultural beliefs. But, when we consider them to be semi-understood ideas, then they are only apparently irrational (cf. Sperber 1996: 73). Belief has normally been regarded, at least since Russell, as a type of propositional attitude, an attitude to an entity which is either true or false. But: . . . the objects of our ‘propositional attitudes’, the ideas we hold or otherwise entertain, are not always strictly propositional in character. Just as it would be mistaken to define ‘speaking’ as ‘uttering sentences’, it is mistaken, I suggest, to define thinking in terms of attitudes to propositions: many of our utterances do not match sentences but semi-grammatical strings; similarly, many of our thoughts are what we might call semi-propositional, they approximate but do not achieve propositionality. (. . . ) [I]f it were true that the objects of belief necessarily were propositions, then we could only believe ideas which we fully understand. I am arguing that we can also hold as beliefs incompletely understood ideas. (Sperber 1985: 51)
So, there are propositional representations which correspond to fully understood ideas, and semi-propositional representations which correspond to ideas which are not fully understood. These are conceptual representations which do not identify a unique proposition – some concepts are missing. Such semi-propositional representations are useful for storing information which is not fully understood. These metarepresentations are not factual beliefs, but representational beliefs. Hence, there are four potential classes of beliefs: factual beliefs with propositional content which are rational, observation-based; factual beliefs with semi-propositional content which do not normally occur; representational beliefs with propositional content (metarepresentations, including also scientific assumptions not fully believed); and finally representational beliefs with semi-propositional content, such as religious beliefs, mysteries (cf. Sperber 1985: 58). Cultural beliefs are such representational beliefs. Including its recent developments, Sperber’s explanation of cross-cultural differences and similarities is as follows: there are two types of beliefs, intuitive (factual) and reflective (representational).6 Intuitive beliefs are based on innate, universal mechanisms, based on perception and inference, and hence are similar across cultures. Reflective beliefs are not factual, not basic, they are, so to speak, ‘in quotes’: they are believed because they are embedded in intuitive beliefs. They may be half-understood and leading either to scientific knowledge or to myster-
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.11 (451)
On translating ‘what is said’
ies, often religious. These mysterious beliefs are the ones that vary cross-culturally and may even appear irrational from another culture’s point of view (cf. Sperber 1996: 91–92). In other words, . . . explaining cultural beliefs, whether intuitive or reflective, and if reflective, whether half-understood or fully understood, involves looking at two things: how they are cognized by individuals and how they are communicated within a group; or to put it in the form of a slogan: Culture is the precipitate of cognition and communication in a human population. (Sperber 1996: 97)
This discussion is particularly relevant for our study of semantic and pragmatic equivalences as it demonstrates clearly the need to look at human behaviour, including verbal behaviour, in the context of an anthropological culture to which it belongs. It demonstrates that there are types of beliefs which, when externalized, cannot be easily interpreted by members of other communities (here also linguistic communities) because they are metarepresentations. Although it does not seem to be true that people of different cultures live in different worlds, we have to know something about the culture to be able to classify and represent beliefs. We can suggest tentatively at this stage that when a belief is semi-propositional, and a fortiori the sentence uttered does not have the backing of a proposition, the translator’s task is not to elaborate on the degree of awareness of the proposition; this is the anthropologist’s task. The translator has to substitute a sentence in the target language which is a successful representative of this proposition, whether it is understood fully or partially by the speaker7 – unless culture-specific implicatures intervene. This is the next problem to be discussed.
. What is said As can be seen from the discussion in Section 3, in order to arrive at what is said, world knowledge and the assumption of cultural rationalism are required more than a detailed insight into the anthropological culture of the speaker: what is said is available cross-culturally. Semi-propositional beliefs can immediately strike the hearer as problematic, as in the case of the belief in golden dragons, but they are equally puzzling for the holder of the belief and the hearer of the expression of belief. So, the best strategy seems to be to preserve what is said unless the rule A2 dictates otherwise. The next step is to delimit what is said. There is no consensus regarding its scope. The first difficulty is engendered by the fact that people often speak loosely, non-literally, because it is more efficient to do so; such utterances require less processing effort on the part of the hearer. For example, a mother assures a child who cut his finger:
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.12 (452)
K. M. Jaszczolt
(8) You are not going to die, Peter.
(Bach 1994a: 267)
The meaning of this proposition is that there is no time following the time of utterance at which the addressee will die. This proposition is obviously false. What was communicated instead was an expanded proposition (9). (9) You are not going to die from this cut, Peter. The mother did not say (9) because it was obvious. So, did what she said differ from what she meant, or did she say that Peter was not going to die from this wound? Here opinions differ. While some pragmaticists claim that what was said was the expanded proposition with its pragmatic constituents (cf. Récanati 1989), others maintain that what was said was the minimal proposition and the mother was speaking loosely, non-literally (Bach 1994a, b). Similarly, (7) is used non-literally. (7) and (8) have literal meaning and express propositions with clear truth conditions: that the speaker has never eaten and that the boy is immortal, respectively. This is the minimal proposition, and from its obvious falsehood it is clear that the speaker meant instead an expanded proposition. Similarly, when a proposition is very obviously true, the speaker is taken to mean an expansion of it, as in (10) and (11). (10) The park is some distance from where I live. (11) It will take us some time to get there.
(Carston 1988: 164)
The expansions are, respectively, ‘a rather long way away’ and ‘a rather long time’. Now, all this additional information is not in the form of additional propositions but rather it is an expansion of the original proposition, and this information contributes to the truth conditions. It is then of the level of semantic equivalence. Examples (7) and (8) have to be kept apart from (6). Sentence (6) is different. This sentence does not have clear truth conditions, it does not express a complete proposition. It is semantically general, underdetermined. The meaning of this sentence is only a so-called propositional radical which requires completion to become a full proposition: ‘strong enough for what?’. Here opinions vary again as to what the speaker said.8 Some pragmaticists claim that what is said includes the pragmatic completion (cf. Récanati 1989). For example, (6) expresses (12). (12) Steel isn’t strong enough for spare parts for this machine. Others maintain that what is said is the propositional radical, and the completion is meant, implicit in what is said (cf. Bach 1994a, b). Grice did not talk about these processes of expansion and completion, although they require precisely the same pragmatic processes as these involved in implicatures. Bach (1994a: 270) observes that “. . . there is no line to be drawn between what is said and what is implicated. Instead, there is considerable middle ground be-
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.13 (453)
On translating ‘what is said’
tween them.” The middle ground consists of implicitly stating the completions and expansions (a.k.a. saturation and strengthening, Récanati 1989). Relevance theorists subsume this middle ground under explicature – the proposition expressed, the elaboration of the logical form of the sentence.9 But this decision is only theoretical, dictated by the fact that all this information contributes to the truth conditions of the original proposition. This enriched proposition is called by Bach impliciture: Implicitures go beyond what is said, but unlike implicatures, which are additional propositions external to what is said, implicitures are built out of what is said. (Bach 1994a: 273)
Bach’s argument against calling implicitures part of what is said is as follows. (7) uttered by a robot in a science-fiction story can mean that there is no time prior to the time of utterance at which the speaker ate. So, the minimal, not the expanded proposition is here what is said. Hence, what is said would have to differ from one situation to another, which is methodologically unsatisfactory. If we accept that what is said is uniformly the expanded proposition, then what is meant would have to be less than what is said, which is counterintuitive. So, the remaining option is that there are pragmatic aspects of what is said, but these are implicit in what is said. Be that as it may, what is of interest here is the coarsely-grained, broad notion of what is said. Whatever contributes to the proposition expressed rather than constituting a separate proposition (implicature) is the domain of what is said and of truth-conditional semantics. Naturally, such embellishments of the original proposition can be endless. When the intuitively plausible proposition is reached, we say that this truth-conditionally relevant embellishment is completed.10 Let us take stock. The problems with semantic equivalence based on sentence meaning and the notion of what is said are three-fold: (i) the proposition expressed by the speaker’s utterance of a sentence is obviously uninformative and obviously true or blatantly false, which triggers the search for an enriched proposition that would be a better candidate for what the speaker intended to convey; (ii) the sentence uttered by the speaker does not correspond to a full proposition and hence has to be completed; and (iii) the sentence uttered by the speaker, although informative, does not correspond to an attributable proposition in either the literal or non-literal interpretation. The hearer concludes that the speaker holds a semi-propositional, representational belief (cf. Sperber 1985, table on p. 58). It seems that sentences falling under (iii) have to be rendered in translation by adhering to semantic equivalence (as usual, unless A2 intervenes). It will be equally diaphanous to the reader of the source text as to the reader of the target text that the corresponding belief is semi-propositional: this is the gain from accepting cultural rationalism.
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.14 (454)
K. M. Jaszczolt
. Which equivalence? Coming back to the definition of semantic equivalence in translation, we want to ensure that what is said is preserved. Hence, if it is not obvious to the addressee of the translation that (8) communicates an expanded proposition, this expansion should be explicitly stated (and this applies both to spoken and written discourse). The test can be the obvious falsehood or the blatant truth of the literal meaning. Similarly with propositional radicals, if the completion is not obvious in the translation whereas it is obvious in the original discourse, it has to be explicitly stated. Semi-propositional beliefs have propositional form of the corresponding sentences which translates across cultures and hence can be left unattended to: a speaker who holds a semi-propositional belief communicates a proposition willingly, knowing that it does not correspond to a propositional belief but rather to a belief in something being the case where this ‘something’ is stored for possible future understanding. The solution would be simple indeed if not for implicatures. By saying (8) the speaker communicates a range of implicatures, the most obvious ones being “There is no need to worry”, “The wound is not serious”, “The wound will heal quickly”, etc. Attaining pragmatic equivalence requires preserving these obvious implicatures. And in order to do so, the semantic equivalence may have to be compromised: sentence (8) may have to be replaced with (13) or another set phrase dictated by particular socio-cultural, linguistic conventions. (13) It’s not a big deal, Peter. So, the translator has to juggle these two levels of equivalence. Gutt (1991), who advocates the view that translation is part of communication, observes that translation in communication theory has normally been treated through the ‘code model’ of communication, where communication is a matter of encoding and decoding messages. But the subsequent suggestion of Sperber and Wilson’s (1995) that inference is a more essential element of communication should, according to Gutt, percolate to translation theory (cf. Gutt 1991: 21–22). The idea is, however, not completely new. We can trace it back to Nida (1964; Nida & Taber 1969) who developed the notion of dynamic equivalence: Dynamic equivalence is (. . . ) to be defined in terms of the degree to which the receptors of the message in the receptor language respond to it in substantially the same manner as the receptors in the source language. This response can never be identical, for the cultural and historical settings are too different, but there should be a high degree of equivalence of response, or the translation will have failed to accomplish its purpose. (Nida & Taber 1969: 24)
This was applied to Bible translation but seems to be widely applicable.
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.15 (455)
On translating ‘what is said’
Coming back to the standpoints on the what is said/what is communicated debate, it seems that the Default Semantics (option v) gains some support from the theory and practice of translation in that when the inferences are obvious, they need not be translated. And being obvious includes being natural, primary, unmarked. It has to be borne in mind in the translating exercise that not all inferences in conversation (or written text) are pragmatic. There are default interpretations, the level of so-called utterance-type meaning, located between sentence meaning and utterance-token meaning: This third layer is a level of systematic pragmatic inference based not on direct computations about speaker-intentions, but rather on general expectations about how language is normally used. These expectations give rise to presumptions, default inferences, about both content and force. . . (Levinson 1995: 93)
Hence, some expansions and completions may proceed through a default interpretation, guaranteed by socio-cultural and linguistic conventions and the phenomenon of pragmatic constituents is not as uniform as it would appear to be: sometimes we need pragmatic processes, sometimes conventions and defaults, and sometimes merely semantics to arrive at what is said. As I argue elsewhere (Jaszczolt 1997, 1999a, b), the propositional representation can combine with the default senses guaranteed by the default presence of various types of intentions in communication. Now, the problem arises as to whether the translator should supply the contextual assumptions if it is obvious that the reader will lack them. Nida and Taber (1969: 205) claim that only linguistically implicit information from the original can be made explicit. As we would say now, it is probably the information which can be recovered through maxims of conversation. Hence, contextual assumptions and implications are largely excluded when they pertain to cultural information. Other theorists allow for the inclusion of cultural information (see Gutt 1991: 81). As Gutt (1991: 94) says, following relevance theory, a translation “. . . should convey to the receptors all and only those explicatures and implicatures that the original was intended to convey.” But this is untenable: conveying the same message will depend on the context and inferences, and these cannot be kept steady (cf. Gutt 1991: 99). At best, we can aim at an approximation. All in all, achieving complete interpretive resemblance may not be attainable, or at least there may not be a steadfast rule as to how to achieve it. Devices may have to be changed even within one text (cf. Gutt 1991: 185–186). In general, if translation aims at successful communication, we may not be able to distinguish semantic and pragmatic equivalence but only communicative and cognitive. The boundary may only be constituted by the distinction between these ‘implicatures’ that can be called expansions and completions of the original proposition, and
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.16 (456)
K. M. Jaszczolt
genuine implicatures which have their own logical forms. Since implicatures force a decision as to whether to make them explicit or hope they will be derived, there may be a boundary, if implicatures that pertain to expansion and completion are treated as a separate phenomenon of impliciture, explicature, or pragmatic constituents of what is said, in other words: as the domain of semantics. And this is the idea I followed in this paper. If translation is communication, then a theory of communication will do for defining equivalence. That is why we have to distinguish explicit from implicit communication: what is said and what is implicated. This is how we can assess motivations for choices e.g. between direct and indirect translation. ‘Direct’ means here literal, preserving not only meaning, but also the way it was expressed (cf. Gutt 1991: 125); ‘indirect’ means free translation. The idea that translation is communication and includes what is said and what is implicated has very respectable roots in the philosophy of language, notably in the work of Rudolf Carnap, dating back to the 1940s. Carnap analysed semantic meaning as follows: intralingually, two sentences are equivalent if both are true or both are not true, and the individual expressions must stand for the same thing to be equivalent (Carnap 1942, 1947). Equivalence can be either material, where there is equivalence of objects and facts, or logical, where the truth of the sentence can be deduced from the semantic rules. We also need the equivalence of intensions, i.e. the way of thinking about objects and events, not only the equivalence of events. For this stronger equivalence the identity of structure is required, i.e. intensional isomorphism. For example, ‘three’ and ‘the square root of nine’ are L-equivalent but do not have the same intensional structure. Carnap’s approach failed because it was behaviouristic: he defined mental attitudes such as belief that p as dispositions to assent to sentences intensionally isomorphic with p, which is obviously inadequate. But this is how intensions gave rise to possible worlds and to contemporary semantics: there are intensional contexts in which the identity of truth conditions will not guarantee semantic equivalence, such as belief and other propositional attitude contexts. But this is a topic for a different occasion (see Jaszczolt 1997, 1998a, c, 1999b).
. Intuitions and ‘what is said’ Let us come back to the fuzziness of the boundary between semantics and pragmatics. The category of semantic underdetermination, sentences expressing propositional radicals, is somewhat eclectic. Examples (14) and (15) quoted after Bach (1994a) do not seem to be clearly underdeterminate, as opposed to (16) which is very obviously incomplete. (14) ?She is leaving
[from where?]
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.17 (457)
On translating ‘what is said’
(15) ?She wants a taxi [to do what with?] (16) I am too tired
[to do what?]
The matter would be easier to handle if all propositional radicals had slots in their syntactic representation to be filled, as in the case of ellipsis in (17). (17) Bill wants pie for dessert and Bob ø pudding ø.
(Bach 1994a: 281)
Also, the elided material in (17) can be easily recovered, unlike in the case of underdetermination in (14)–(16). Similarly, in the case of indexicals, there is a slot for reference assignment as in (18). (18) I am happy. Propositional radicals do not have such syntactic gaps. The syntactic forms are complete and the logical forms are incomplete (cf. Bach 1994a: 283). Hence, completion is largely arbitrary. Coming back to the Gricean distinction between semantics and pragmatics (see Horn’s diagram in 1988: 121), its oversimplification is obvious. According to Grice, the propositional form results from the conventional meaning of the sentence, and disambiguation and reference assignment to referring expressions. But as we can see from Bach’s, Récanati’s and Carston’s examples, propositions may also require expanding or completing. And this resulting proposition differs from implicatures which are independent from it functionally and logically. It remains to ask about the scope of this contribution of pragmatic processes to what is said. Even the meaning of the connective ‘and’ can be regarded as partly arrived at through such an enrichment where ‘and’ is temporal or consequential, as demonstrated in the contrast between (19) and (20) on the one hand, and (21) and (22) on the other. (19) They got married and had a baby. (20) They had a baby and got married. (21) 2 + 2 = 4 (22) I like cheese and wine. The order of conjuncts is free in (21) and (22), but it is fixed in (19) and (20) where ‘and’ reads as ‘and then’. Changing the order changes the truth conditions, as can be tested by (i) the functional independence principle mentioned above or by (ii) embedding the sentences in the scope of logical operators such as negation and implication (‘if. . . then’):
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.18 (458)
K. M. Jaszczolt
(23) If the old king died of a heart attack and a republic was declared Sam will be happy, but if a republic was declared and the old king died of a heart attack, Sam will be unhappy. (cf. Cohen 1971: 54; Carston 1988: 172) These tests of functional independence and scope do not always work. It has been proposed (Récanati 1989) that what is said should be delimited on an intuitive basis instead. Récanati agrees with Grice that postulating semantic ambiguity is methodologically unsatisfactory and instead we should allow for pragmatic processes to contribute to the semantic representation of the utterance. This is what came to be known as the principle of Modified Occam’s Razor. But he says that we can only tell whether pragmatic information is necessary for what is said if we already know what is said. And this is circular. Instead, what is said is determined intuitively: In deciding whether a pragmatically determined aspect of utterance meaning is part of what is said, that is, in making a decision concerning what is said, we should always try to preserve our pre-theoretic intuitions on the matter. (Récanati 1989: 106)
He calls this suggestion the Availability Principle. What is said is consciously accessible, we appeal to common sense to determine it. Generalized conversational implicatures are subsumed under what is said: numerals are strengthened from the meaning ‘at least n’ to ‘exactly n’ and disjunction is strengthened from inclusive to exclusive ‘or’. All in all, there are various ways of delimiting what is said and hence delimiting the unit of semantic equivalence in translation. Now, the choice as to whether to follow semantic, pragmatic, syntactic or stylistic equivalence can perhaps never be resolved; it is widely acknowledged that the preferences may change even within one text. But delimiting tertium comparationis, the platform of reference, is a necessary pre-condition for making such choices.
. Concluding remarks As is well known, some philosophers of language claim that objective translation is impossible. Quine (1960) makes translation dependent on a ‘translation scheme’ or a theory.11 But it has been widely accepted that propositions, units of meaning, are language-independent. So, they constitute, in our terms, tertia comparationis, a universal semantic category, realized as sentences of particular languages. Sentences have their own logical forms which may require further development into a full propositional form. This propositional form is the unit of semantic equivalence in translation. When the aim is to preserve communicative equivalence, the decisions
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.19 (459)
On translating ‘what is said’
may have to be made as to whether it is best done by preserving the semantic or pragmatic equivalence, the first relying on the syntactic equivalence and sameness of logical form. Semantic equivalence is not attainable for the comparison of grammatical constructions. It has to be text-bound rather than systematic (involving the comparison of grammatical and lexical systems, cf. Krzeszowski 1989; Jaszczolt 1995a: 2), just as semantics has to incorporate the pragmatic constituents of what is said. It was also suggested that preserving semantic and pragmatic equivalence may not be compatible: the decision has to be made as translation progresses. This approach is taken to be applicable both to spoken and written discourse.
Notes * I would like to thank the audiences of my talks at the International Conference on Germanic and Baltic Language Studies, Teaching and Translation at the University of Vilnius and the Second International Conference in Contrastive Semantics and Pragmatics at the University of Cambridge for their comments on the previous drafts of this paper. . Ambiguous logic is a logic for underspecified representations, used for the purpose of demonstrating to what extent the system has to disambiguate the representation for logical reasoning to proceed. Disambiguation stops as soon as the representation becomes sufficiently specific to act in further reasoning. . Be it Gricean, neo-Gricean, or relevance-theoretic. . ‘Correspond to’ is left deliberately vague: the proposition expressed by the sentence may have to be enriched to arrive at what is said. ‘Usually’ is intended as a disclaimer to account for tautologies: analytic sentences (in the wide sense) are always true. . The logical form is derived through linguistic semantics. See Récanati 1989, Carston 1988, 1998a, b. . So, in discussing culture we have to access cognitive psychology. . For the distinction into intuitive and reflective beliefs see Sperber 1996, 1997. . For criticism see Récanati (1997: 93): “Sperber’s claim concerning semi-propositionality can (. . . ) be constructed as a claim concerning the epistemical state of the user, rather than a claim about semantic content.” In other words, underdetermination is epistemic rather than semantic. . Experiments have been performed concerning speakers’ intuitions as to what is said but so far they have been inconclusive: two sets of experiments led to opposite results. See Gibbs ant Moise 1997; Nicolle and Clark 1999. . Récanati calls it ‘pragmatic constituents of what is said’. . Here Carston (1988) refers to the principle of relevance in establishing the upper boundary on the development of the logical form of the sentence.
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.20 (460)
K. M. Jaszczolt . “. . . manuals for translating one language into another can be set up in divergent ways, all compatible with the totality of speech dispositions, yet incompatible with one another.” Quine (1960: 27).
References Bach, K. (1994a). Semantic slack: What is said and more. In S. L. Tsohatzidis (Ed.), Foundations of Speech Act Theory (pp. 267–291). London: Routledge. Bach, K. (1994b). Conversational impliciture. Mind & Language, 9, 124–162. Beaugrande, R. de (1980). Text, Discourse, and Process: Toward a Multidisciplinary Science of Texts. Norwood, NJ: Ablex. Carnap, R. (1942). Introduction to Semantics. Cambridge, MA: Harvard University Press. Carnap, R. (1947). Meaning and Necessity: A Study in Semantics and Modal Logic. Chicago: University of Chicago Press. Carston, R. (1988). Implicature, explicature, and truth-theoretic semantics. In R. M. Kempson (Ed.), Mental Representations (pp. 155–181). Cambridge: CUP. Carston, R. (1998a). Postscript (1995). In A. Kasher (Ed.), Pragmatics: Critical Concepts, Vol. 4 (pp. 464–479). London: Routledge. (postscript to Carston 1988). Carston, R. (1998b). Informativeness, relevance and scalar implicature. In R. Carston & S. Uchida (Eds.), Relevance Theory: Applications and Implications (pp. 179–236). Amsterdam: John Benjamins. Cohen, L. J. (1971). Some remarks on Grice’s views about the logical particles of natural language. In Y. Bar-Hillel (Ed.), Pragmatics of Natural Languages (pp. 50–68). Dordrecht: D. Reidel. ´ Fisiak, J., Lipinska-Grzegorek, M., & Zabrocki, T. (1978). An Introductory English-Polish ´ Contrastive Grammar. Warszawa: Panstwowe Wydawnictwo Naukowe. Gentzler, E. (1993). Contemporary Translation Theories. London: Routledge. Gibbs, R. W., & Moise, J. F. (1997). Pragmatics in understanding what is said. Cognition, 62, 51–74. Gutt, E. A. (1991). Translation and Relevance: Cognition and Context. Oxford: Blackwell. Heim, I. (1988). The Semantics of Definite and Indefinite Noun Phrases. New York: Garland Press. Horn, L. R. (1988). Pragmatic theory. In F. J. Newmeyer (Ed.), Linguistics: The Cambridge Survey, Vol. 1, Linguistic Theory: Foundations (pp. 113–145). Cambridge: CUP. James, C. (1980). Contrastive Analysis. London: Longman. Jaszczolt, K. (1995a). Typology of contrastive studies: Specialisation, progress and applications. Language Teaching, 28, 1–15. Jaszczolt, K. (1995b). Contrastive analysis. In J. Verschueren, J.-O. Östman, & J. Blommaert (Eds.), Handbook of Pragmatics: Manual (pp. 561–565). Amsterdam: J. Benjamins. Jaszczolt, K. (1997). The default de re principle for the interpretation of belief utterances. Journal of Pragmatics, 28, 315–336. Jaszczolt, K. M. (1998a). Reports on beliefs: Default interpretations and default intentions. Journal of Literary Semantics, 27, 31–42.
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.21 (461)
On translating ‘what is said’
Jaszczolt, K. M. (1998b). Referring in discourse: Referential intention and the “taking for granted” principle. Journal of Literary Semantics, 27, 96–109. Jaszczolt, K. M. (1998c). Discourse about beliefs. Theoretical Linguistics, 24, 1–28. Jaszczolt, K. M. (1999a). Default semantics, pragmatics, and intentions. In K. Turner (Ed.), The Semantics/Pragmatics Interface From Different Points of View (pp. 199–232). Oxford: Elsevier Science. Jaszczolt, K. M. (1999b). Discourse, Beliefs, and Intentions: Semantic Defaults and Propositional Attitude Ascription. Oxford: Elsevier Science. Jaszczolt, K. M. (2000). The default-based context-dependence of belief reports. In K. M. Jaszczolt (Ed.), The Pragmatics of Propositional Attitude Reports (pp. 169–185). Oxford: Elsevier Science. Jaszczolt, K. M. (2002). Against ambiguity and underspecification: Evidence from presupposition as anaphora. Journal of Pragmatics, 34, 829–849. Kalisz, R. (1981). The Pragmatics, Semantics and Syntax of the English Sentences with ˙ Complements: A Contrastive Indicative That Complements and Polish Sentences with Ze ´ Uniwersytet Gdanski. ´ Study. Gdansk: Kamp, H., & Reyle, U. (1993). From Discourse to Logic. Dordrecht: Kluwer. Krzeszowski, T. P. (1989). Towards a typology of contrastive studies. In W. Oleksy (Ed.), Contrastive Pragmatics (pp. 55–72). Amsterdam: John Benjamins. Krzeszowski, T. P. (1990). Contrasting Languages: The Scope of Contrastive Linguistics. Berlin: Mouton. Lakoff, G. (1987). Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. Chicago: University of Chicago Press. Levinson, S. C. (1983). Pragmatics. Cambridge: CUP. Levinson, S. C. (1995). Three levels of meaning. In F. R. Palmer (Ed.), Grammar and Meaning: Essays in Honour of Sir John Lyons (pp. 90–115). Cambridge: CUP. Levinson, S. C. (2000). Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge, MA: MIT Press. Nicolle, S., & Clark, B. (1999). Experimental pragmatics and what is said: A response to Gibbs and Moise. Cognition, 69, 337–354. Nida, E. A. (1964). Toward a Science of Translating: With Special Reference to Principles and Procedures Involved in Bible Translating. Leiden: E.J. Brill. Nida, E. A., & Taber, C. R. (1969). The Theory and Practice of Translation. Leiden: E. J. Brill. Oleksy, W. (1984). Towards pragmatic contrastive analysis. In J. Fisiak (Ed.), Contrastive Linguistics: Prospects and Problems (pp. 349–364). Berlin: Mouton. Quine, W. V. O. (1960). Word and Object. Cambridge, MA: MIT Press. Récanati, F. (1989). The pragmatics of what is said. Mind & Language, 4, 295–329. Reprinted in S. Davis (Ed.), Pragmatics: A Reader (pp. 97–120). Oxford: OUP. Récanati, F. (1993). Direct Reference: From Language to Thought. Oxford: Blackwell. Récanati, F. (1997). Can we believe what we do not understand? Mind and Language, 12, 84–100. Sajavaara, K. (1984). Psycholinguistic models, second language acquisition, and contrastive analysis. In J. Fisiak (Ed.), Contrastive Linguistics: Prospects and Problems (pp. 379–408). Berlin: Mouton. Schiffer, S. (1992). Belief ascription. Journal of Philosophy, 89, 499–521.
TSL[v.20020404] Prn:4/11/2002; 10:16
F: PBNS41.tex / p.22 (462)
K. M. Jaszczolt
Sperber, D. (1985). Apparently irrational beliefs. In On Anthropological Knowledge (pp. 35– 63). Cambridge: CUP. Sperber, D. (1996). Explaining Culture: A Naturalistic Approach. Oxford: Blackwell. Sperber, D. (1997). Intuitive and reflective beliefs. Mind and Language, 12, 67–83. Sperber, D., & Wilson, D. (1995). Relevance: Communication and Cognition (2nd edition). Oxford: Blackwell. van Deemter, K. (1998). Ambiguity and idiosyncratic interpretation. Journal of Semantics, 15, 5–36. Wierzbicka, A. (1991). Cross-Cultural Pragmatics: The Semantics of Human Intraction. Berlin: Mouton.
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.1 (463)
Translation equivalents as empirical data for semantic/pragmatic theory Bergljot Behrens and Cathrine Fabricius-Hansen University of Oslo, Norway
.
Introduction
The present paper aims to contribute to the explication of the discourse theoretic concept Elaboration and the semantic description of relevant connectives in German and English. At the outset the relation is taken in a broad sense, i.e. more as a family of relations than as a uniquely identifiable relation. We are interested in establishing the semantic criteria for the members of that family, and see how they relate to each other. The idea we are going to present is that the variation we find in linguistic expressions used to signal Elaboration across languages contributes to define its semantic space in terms of a set of semantic features, some of which are central to the identification/recognition of the relation and others being more peripheral. The variation is established on the basis of translational data, and the features are identified on the basis of the translational properties of the expressions. Methodologically we follow Dyvik’s proposal (Dyvik 1998) that translational properties may be used to define semantic properties of lexical units. This idea requires specification and will be a central issue in the paper. We will present current descriptions and definitions of Elaboration and demonstrate that existing formal definitions are inadequate in that they cover only part of what Elaboration seems to encompass. We will suggest that further study may be based on data elicited from machine readable translational corpora. The translational method of eliciting relevant data will be explicated, and examples elicited will be discussed in terms of their relevance to the problem of explicating the discourse semantic relation.
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.2 (464)
Bergljot Behrens and Cathrine Fabricius-Hansen
. Elaboration as a semantic/functional category: Examples and preliminary explication Elaboration is defined in the literature as a relation holding at clause level (level of expression) or at the semantic level. Thus Michael Halliday defines it as a restatement relation at clause level, whereby “one clause expands another by elaborating on it or some portion of it by restating in other words, specifying in greater detail, commenting, or exemplifying” (Halliday 1994: 220; our italics). Another definition is presented by Asher (1993: 267 ff.), who takes Elaboration to be a discourse structural as well as a semantic relation. Semantically, elaboration holds if the events described in the two related clauses are understood as the same event, i.e. the relation is one of event identity. Discourse-semantically, Elaboration is subordinate, i.e. one clause (the matrix clause) discourse dominates the other. This part of Asher’s definition is far from clear, and we shall therefore not dwell on it in abstractum. In fact, this part of the definition is one that we would like to explicate by way of our study.
. English BY-clauses corresponding to German INDEM-clauses: Partial event identity in Elaboration In pure semantic terms Elaboration is a semantic discourse category obtaining as a semantic relation between propositions or events. In German the prototypical example of Elaboration is expressed in a subordinate finite clause introduced by the subordinator INDEM ‘in that’, as in (1) below:1 (1) a.
Daß Leni selbst eine “Partikularistin” ist, beweist sie täglich, indem sie sämtliche Brötchenkrümel vom Teller aufliest und in den Mund steckt. (OMC/HEB 1) ‘That Leni herself is a “Particulist”, she proves daily, INDEM she gathers all bread crumbs from her plate and puts [them] into her mouth.’ b. That Leni herself is a “Particulist” is something she proves daily by gathering up all the bread crumbs from her plate and putting them into her mouth. (OMC)
In (1a) the matrix verb is an accomplishment: beweisen/prove. The verb denotes the effect of some action; the action/activity causing it is undefined. In event semantics we would represent the accomplishment as a causal event complex (ec), as in Figure 1 below. The representation includes a causing event (e) which is assigned a set of semantic roles, but which appears without any descriptive properties beyond the semantic roles defined. The caused event is specified:
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.3 (465)
Translation equivalents as empirical data
ec:
e´ e CAUS(e´, e) Ag(x, e´) Pat(y, e´) Pat(y, e) beweisen/prove (e, y)
Figure 1. A causally structured event complex, as represented in Discourse Representation Theory (DRT) (Kamp & Roßdeutscher 1994)
The elaborating subclause defines the causing event. For further details as to the procedure by which the event in the subordinate clause (auflesen und in den Mund stecken/gather up and putting in mouth) is updated with the underspecified event represented as part of the matrix event, see Behrens (1998). Elaboration is here understood as a relation between two propositions by which the events in the elaborating clause make up part of the event in the matrix.
. BY-clauses infelicitous translation of INDEM-clauses (2a) is an example of an INDEM-clause in which a BY-clause is an unsuccessful translation into English. (Norwegian VED ‘by’ + infinitive clause is the closest equivalent to an English BY-clause, and that construction would also yield an unsuccesful translation.) (2) a.
Schwellen ausfindig zu machen und zu beschreiben, is meine Leidenschaft geworden. Ich beschäftige mich damit auch oft an den Nachmittagen, während des Schuljahrs, indem ich bei Ausgrabungen in der näheren Umgebung mithelfe, wie auf dem keltischen Dürnberg bei Hallein, oder eben, wie vor kurzen, am “Römerweg” in Loig. (OMC/PH 1) ‘. . . . I also often occupy myself with that in the afternoons, during the school year, INDEM I help at the excavations in the immediate vicinity, such as. . . ’ b. Discovering and describing thresholds became a passion with me. During the school year I often devoted an afternoon to it, helping on digs in the immediate vicinity, such as the Celtic Dürnberg near Hallein or, only recently, the “Roman Road” in Loig. (OMC) c. ??Discovering and describing thresholds became a passion with me. During the school year I often devoted an afternoon to it, by helping on digs in the immediate vicinity. . .
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.4 (466)
Bergljot Behrens and Cathrine Fabricius-Hansen
Example (2) suggests that the semantic space of INDEM is wider than that of English BY-clauses (and Norwegian VED + infinitive clauses). In this case the event denoted by the sub-clause is the same as that denoted by the matrix clause.
. Relevant semantic features Closer study (Giesebrecht 1999) suggests that the instrumental or Causing relation displayed in (1a) is the central reading of INDEM, yet examples like (2a) above, satisfy only some of the criteria. Furthermore, the English non-finite ING-clause has an even wider span, as is attested in (3) below, in which a German INDEMclause would yield an unsuccessful translation. (3) a. Philip followed behind, carrying a carved bedhead. (OMC/ST 1) b. Philip folgte ihnen mit einem geschnitzten Kopfteil in der Hand. ‘Philip followed them with a carved bedhead in his hand’ c. ??Philip folgte ihnen, indem er einen geschnitzten Kopfteil in der Hand hatte/trug. ‘Philip followed them, INDEM he had/carried a carved bedhead in his hand’ INDEM-clauses, according to Giesebrecht’s study, require identity of Space and Time, identity of Agent, and in addition some event dependency – not necessarily causality as in (1) – between matrix and subordinate clause. The spatial, temporal and agentive requirements are clearly met in (3a) above. But there is no causality involved, and there is little evidence that there is any event dependency beyond these role identities. This is the type of example which is typically classified as one of “Accompanying Circumstance” in English (Kortmann 1991), and a subordinate INDEM-clause is unsuccessful. For a clearer understanding of the sort of features involved, more data is needed. We would like to determine the features which restrict the use of INDEM relative to English ING-participial subclauses. The semantic features may be depicted as overlapping sets, as in Figure 2, in which the intersection of the three features mentioned above restricts the space of Elaboration. Within this space we find the Causal type of Elaboration discussed above, which is expressed by subordinate INDEM-clauses in German and nonfinite BY-clauses in English. The rest of the space does not distinguish between English ING-participial clauses and German INDEM-clauses: The part of the figure filled with vertical lines is our center of interest. Clearly, INDEM-clauses have a span which is larger than the span of Causal Elaboration. English non-finite ING-clauses have a wider span than German INDEM-clauses. The translational data is valuable input for the identification of the semantic features we are looking for.
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.5 (467)
Translation equivalents as empirical data
Figure 2. Semantic features involved in Elaboration
. Widening the perspective The translational method of study as developed in Dyvik (1998) (see below) shows that Elaboration (in a broad sense) is not confined to INDEM-structures in German. Studies like Pusch (1980) and Giesebrecht (1999) discuss structures with other subordinating conjunctions such as WOBEI ‘whereby’ and DADURCH DASS ‘therethrough that’. DABEI ‘thereby’ is yet another connective relevant for the study of this relation. In the following we shall see that the translational method enables us to elicit examples in German in which two separate sentences without a connective have been used in an Elaborating sense. This may come as a surprise, since German has such a variety of conjunctions to choose from and since German generally appears more synthetic than English (and Norwegian). Before we look at the examples, we will present the method used to elicit these examples.
. The translational method2 Dyvik (1998) has designed a translational procedure by which we may define semantic features of lexical items. The semantic features we are after, he suggests, are those which are needed to distinguish the meaning of words. Such distinctions may be elicited on the basis of contrasts between languages. The more languages that are compared and contrasted, the more features will be needed. In Dyvik’s work the semantic features of a word are actually represented as sets of lexeme pairs. The
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.6 (468)
Bergljot Behrens and Cathrine Fabricius-Hansen
theory behind this will not be discussed here, nor will we name features in terms of crosslinguistic lexical pairs. However, we use his method to elicit semantically related data. A contrastive study based on translational data is obviously very useful for illuminating the concept of Elaboration. Generally speaking, signs in one language which have different translational properties may have their meaning broken down into specific features on the basis of their translational properties. The idea behind this is discussed in Dyvik: genuine ambiguity is in some sense “accidental” – it is a contingent property of a word in a language that it happens to be associated with one or more different meanings. Therefore we do not expect to find the same ambiguity duplicated by signs in a number of different languages. The multiple possibilities arising from ambiguity are an accidental property pertaining to the way a given language happens to be. The multiple possibilities of vagueness or indeterminacy, on the other hand, seem to have more to do with what is being denoted. Vague words denote a family of things that have something in common irrespective of language. To see the way translational properties denote semantic properties, consider the English noun TRUNK. Our intuition tells us that this word is ambiguous, and we expect the different readings to have different expressions in other not too closely related languages. The Oslo Multilingual Corpus yields four occurrences of this item in English source texts, two of which are translated into German as STAMM, one as STUMPF and one as KOFFER. The first translational image of TRUNK with respect to German is illustrated in Figure 3. STAMM TRUNK
STUMPF KOFFER
Figure 3. 1st translational image (t-image) of TRUNK with respect to German
In order for translational correspondences to operate as semantic mirrors, we need to check that they are all within the same semantic space. If we reverse the translation, i.e. check the English translations of each of the German expressions respectively, we will be suspicious of ambiguity if there is no overlap between them beyond the source word (here: TRUNK). Our searches gave the results in Figure 4.
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.7 (469)
Translation equivalents as empirical data
STEM TRIBE TRUNK SUITCASE
STAMM
STUMP
STUMPF KOFFER
Figure 4. The inverse t-image of TRUNK with respect to German
A third round is in order. Each of the English translations of each of the German words for TRUNK (except TRUNK itself, which has already been checked) are checked. As it turns out (see Figure 5) the German translation of SUITCASE does not have any expressions within the German set which overlaps with any of the other sets. This means, in fact, that the SUITCASE- meaning of TRUNK is semantically unrelated to the other meanings of TRUNK (with respect to German). Similarly with STEM and TRIBE – there are no translations of these words which overlap with the translations of SUITCASE. We therefore have translational means to identify the ambiguity of a word.
STIEL STEM TRIBE
STAMM
TRUNK SUITCASE
KOFFER
Figure 5. The restricted 2nd t-image of TRUNK with respect to German3
The translational method demonstrates that TRUNK is ambiguous. This means that the mechanical procedure yields results which are in accordance with our linguistic intuition. The results have different implications once the translations yield overlapping sets. If there is full synonymy, there should be a one to one relationship between the words. The English word SON, for example, does not occur in the translational
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.8 (470)
Bergljot Behrens and Cathrine Fabricius-Hansen
corpus where the German translation is not SOHN, and vice versa. The two are fully synonymous according to the corpus. Finally, partial synonymy, which is central to our domain of study, obtains when there is a large amount of overlap. The procedure by which we identify semantically related items will be demonstrated in the following: Consider the translational properties of the adverb FAST with respect to German. In fictional prose from the OMC, we find translations into RASCH and SCHNELL. Back translation yields the picture in Figure 6.
HASTILY SOON SCHNELL
QUICKLY SWIFTLY FAST SHARPLY RAPIDLY
RASCH
Figure 6. The inverse t-image of FAST with respect to German
The overlap of sets of translations is a strong indication that the meanings of the two German expressions are related. This is also supported by the second timage, in which every lexical item in the translation of SCHNELL and RASCH yield translational sets which include both expressions, viz. the overlapping sets in Figure 7.4
HASTILY SOON
...
...
QUICKLY SWIFTLY
SCHNELL
FAST
RASCH
...
SHARPLY RAPIDLY
...
...
Figure 7. The restricted 2nd t-image of FAST with respect to German
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.9 (471)
Translation equivalents as empirical data
The translational method thus helps us extract examples which fall within the same semantic space across languages. The procedure is therefore very valuable for the extraction of comparable data in the study of Elaboration.
. The translational method applied to the study of Elaboration The OMC corpus is designed in such a way that comparison may be made across more than two languages. It consists of between 20 and 30 texts in English, German and Norwegian with their corresponding translations into two of the languages. The corpus is aligned in such a way that we come up with the two translations of any sentence in which our search word occurs. Since there are source texts in all three languages, an occurrence in a translation may be checked against its occurrence in a source language text. The corpus allows us to make comparisons between English, German and Norwegian, and pre-studies indicate that such a comparison may be very valuable for further study into the semantics of Elaboration. The translational method, however, has been designed for language pairs. In the present study we will therefore limit ourselves to the extraction of bilingual data. Let us consider INDEM, the prototypical expression for Elaboration in German. To decide whether the expressions used as translations of INDEM, (Figure 8), are within the same semantic space and not the result of a true ambiguity of INDEM, we need to make a back-translation. The picture we then come up with is illustrated in Figure 9. This figure represents the inverse t-image of INDEM. The overlap patterns in the inverse t-image already indicate a division into different senses. The expressions or structures which do not intersect: do they represent independent meanings, or do they fall within the same semantic space? To determine this question we may check the translational properties of each expression found in the inverse t-image,
BY + -ING INDEM
, -ING WHEN AS
Figure 8. 1st t-image of INDEM with respect to English
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.10 (472)
Bergljot Behrens and Cathrine Fabricius-Hansen
DAMIT DASS DURCH ZU VON MIT
VP
WOBEI DABEI WIE WÄHREND
BY + -ING S INDEM
BEI DA WENN
, -ING WHEN
NACHDEM AS
WANN SOBALD IN AUF
Figure 9. The inverse t-image of INDEM with respect to English
and include only those which appear with one of the expressions in the first timage. Such a procedure yields a grouping of the first t-image of INDEM into its senses with respect to English. I.e. the senses of INDEM are represented in terms of signs. The overlap we get indicates that the senses are related, although not clearly how. We note that the 2nd t-image – finding the English t-images of the German options in Figure 9 – may not be completed due to the fact that some of the translations are not lexically identifiable, notably VP-structures and S-structures as demonstrated in the examples in Section 5 below. We find it interesting, however, that there are examples in which German translation chooses a paratactic structure where English uses a dependent clause. Table 1 shows that out of 600 examples of ING-participial adverbial adjuncts in English, about half have been translated by a paratactic verb phrase (VP) in German and about 7% appear as full sentences: Table 1. German translations of English ING-participial adverbial adjuncts German
English ING-participial adverbial adjuncts: 600 occurrences ,S .S , VP VP conj. 14 29 107 196 Sum: 42 occurrences: 7.1% Sum: 303 occurrences: 50%
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.11 (473)
Translation equivalents as empirical data
. Elaboration in two separate sentences In focussing on paratactic expressions of Elaboration, we are aware that our data includes examples in which the translational choice may be motivated by nonsemantic forces. However, the examples make up an important set of data in view of the forces that motivate linguistic form and expression. In (4) below we note that translation by a subordinate INDEM-clause is an acceptable option for the first part of the elaboration. However, once the second part of the elaboration is added, the choice of INDEM does not seem quite felicitous: (4) a.
Like all successful villains, he kept a low profile around the manor, driving an unobtrusive car, his sole indulgence being the elegance of his apartment. (OMC/ID 73) b. Auch verhielt er sich wie alle erfolgreichen Ganoven in seinem Revier unauffällig. Er fuhr einen bescheidenen Wagen, und die einzige Schwäche, die er sich gestattete, war die elegante Wohnung. (OMC) ‘. . . . He drove an unobtrusive car, and the only indulgence, that he allowed himself, was the elegant apartment’ c. Auch verhielt er sich wie alle erfolgreichen Ganoven in seinem Revier unauffällig, indem er einen bescheidenen Wagen fuhr. Die einzige Schwäche, die er sich gestattete, war die elegante Wohnung. ‘. . . , INDEM he drove an unobtrusive car. The only indulgence. . . ’ ?? c Auch verhielt er sich wie alle erfolgreichen Ganoven in seinem Revier unauffällig, indem er einen bescheidenen Wagen fuhr und die einzige Schwäche, die er sich gestattete, die elegante Wohnung war. ‘. . . , INDEM he drove an unobtrusive car and the only indulgence, that the allowed himself, was the elegant apartment’
The second conjunct in (4c ) does not have an agent subject and does not denote an action. Since INDEM has scope over the conjunction – both conjuncts are verb-final subordinate clauses – it may be the agent identity requirement (see Section 2.3) that is responsible for the infelicity of the translation. We therefore conclude that information structure may inflict on translational choice within the space of Elaboration. We do not aim to give a full account of the use of two separate sentences in German for the expression of Elaboration in this short paper. However, the examples we have come up with by way of our method clearly point to the problem of specifying the semantic space of this relation, and of determining the constraints on the expressions we use to denote it. The following three examples are listed to illustrate the relevance of paratactic structures in this study.
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.12 (474)
Bergljot Behrens and Cathrine Fabricius-Hansen
(5) a.
When thwarted he would rage, shutting his eyes tight, opening his mouth and holding his breath before letting out a bellow of such terrifying power that Neil half expected the whole of Lydsett to come running to see which of them was tormenting the child. (OMC/ID 241) b. Wenn ihm das misslang, begann er freilich zu toben. Er kniff die Augen zusammen, öffnete den Mund, hielt einen Augenblick die Luft und stiess dann einen so gellenden Schrei aus, daß Neil jedesmal befürchtete, ganz Lydsett würde dahergelaufen kommen, um nachzusehen, wer denn das Kind mißhandle. (OMC) ‘. . . . He shut his eyes tight, opened his mouth, held his breath for a moment and then let out a bellow. . . ’
(6) a.
He cribbed from other guidebooks, seizing small kernels of value and discarding the rest. (OMC/AT 1) b. Er schrieb aus anderen Reiseführern ab, pickte aber nur wertvolle Körnchen heraus und verschmähte die Spreu. (OMC) ‘He copied from other guidebooks but picked only the valuable kernels out and discarded the junk.’
(7) a.
Now it was the old man who wanted to stick with Vic and Roger and it was the kid (by this time forty years old) who wanted to jettison them, arguing with some logic that it would be madness to hand their account over to a two-bit ad agency six hundred miles north of the New York pulsebeat. (OMC/ID 617) b. Jetzt war es der Alte, der an ihnen festhalten wollte, und der Junge (inzwischen vierzig Jahre alt) wollte sie zum Teufel Jagen. Er argumentierte nicht ohne Logik, dass es heller Wahnsinn sei, den Etat einer Zweimann-Agentur sechshundert Meilen nördlich vom Zentrum New Yorks überlassen. (OMC) ‘Now it was the old (man) who wanted to hold on to them, and the boy (. . . ) wanted to send them to the devil. He argued not without logic that (. . . )’
(8) a.
It was clear, he said, that the winds which brought the cold down from Russia were arriving in Provence with greater velocity than before, taking less time to reach their destination and therefore having less time to warm up en route. (OMC/ID 402) b. Es sei klar, so behauptete er, daß die Winde, die die Kälte von Russland herübertrugen, mit höherer Geschwindigkeit als früher in der Provence einträfen, also weniger Zeit brauchten, um ihr Ziel zu erreichen, und folglich weniger Zeit hätten, sich en route zu erwärmen. (OMC)
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.13 (475)
Translation equivalents as empirical data
‘It was clear . . . that the winds . . . arrived in Provence with greater velocity than before, ALSO needed less time to reach their goal and consequently had less time to warm up en route’ The examples above have the same subordinate structure in English, and their interpretation is one of Elaboration, both in terms of Halliday and in terms of Asher. INDEM-subjunction seems an infelicitous alternative in all of these cases. The examples therefore offer important data for further study. It seems to us that the examples invite different types of restrictions. Manner specification in the German matrix clause is one, as has already been noted by Pusch (1980). (6), on the other hand, invites us to consider other semantic aspects of the verbs chosen in the translations. Finally, (7) and (8) are of a different kind, and require a study into the worlds in which the propositions in the two sentences/clauses may be evaluated.
. Summing up The variation in expressions used in translation is a sign that there is no clear oneto-one correspondence between connectives in the three languages. Furthermore, there is great variation in the choice between a paratactic and a hypotactic mode of expression. However, we think that the cross-linguistic variation contributes to sharpening our intuitions regarding the meaning of the complex clause structures. We have come closer to understanding the semantics of INDEM as a typical expression of Elaboration and by that token in determining the space of Elaboration as a semantic rather than a discourse structuring relation. We believe that further study of the paratactically structured examples will give insight into any interrelation between the semantic category and the discourse structuring category. We have demonstrated the translational method of eliciting semantically related expressions. The method is best suited to the study of truly lexical items, and has certain limitations with respect to discourse connectives since such relations may be expressed structurally rather than lexically. Nevertheless we think that it has given us relevant empirical data which can be subjected to more traditional discourse semantic investigation.
Notes . Examples marked OMC/. . . are taken from the electronic Oslo Multilingual Corpus, which consists of authentic text samples in English, German and Norwegian, and their authorized – sentence-aligned – translations (see http://www.hf.uio.no/german/sprik/ english/index.shtml and Section 4 of this paper).
TSL[v.20020404] Prn:4/09/2002; 15:15
F: PBNS42.tex / p.14 (476)
Bergljot Behrens and Cathrine Fabricius-Hansen . We would like to thank Torgrim Solstad for invaluable help in designing the figures in this and the following section and for discussing the application of the Dyvik method to our data. . The German translation Stiel of English stem is not part of the translational properties of trunk. This translation is therefore irrelevant for the discussion. . Again, the expressions found in the translations are left out as long as they do not fall within the set of translations in the first t-image.
References Asher, N. (1993). Reference to Abstract Objects in Discourse. Dordrecht: Kluwer. Behrens, B. (1998). Contrastive Discourse: An interlingual approach to the interpretationand translation of free ING-participial adjuncts. Unpublished Ph.D. thesis, University of Oslo. Behrens, B. (1999). A dynamic semantic approach to translation assessment. In M. Doherty (Ed.), Sprachspezifische Aspekte der Informationsverteilung (pp. 90–112). Berlin: Akademie Verlag. Giesebrecht, K. L. (1999). Dadurch, daß das eine geschiet, geschiet zugleich das andere. MA thesis, University of Oslo. Dyvik, H. (1998). A translational basis for semantics. In S. Johanssson & S. Oksefjell (Eds.), Corpora and Crosslinguistic Research: Theory, Method, and Case Studies (pp. 51–86). Amsterdam: Rodopi. Fabricius-Hansen, C. (1999). Information packaging and translation: Aspects of translational sentence splitting (German – English/Norwegian). In M. Doherty (Ed.), Sprachspezifische Aspekte der Informationsverteilung (pp. 175–214). Berlin: Akademie Verlag. Fabricius-Hansen, C. (2000). Formen der Konnexion. In HSK Text und Gesprächslinguistik/ Linguistics of Text and Conversation (pp. 331–343). Berlin etc.: Walter de Gruyter. Halliday, M. (1994). An Introduction to Functional Grammar (2nd edition). London: Arnold. Johansson, S. (1997). Using the English-Norwegian Parallel Corpus for contrastive analysis and translation studies. In Practical Applications in Language Corpora. Lodz. Kamp, H., & Roßdeutscher, A. (1994). Remarks on Lexical Structure, DRS-construction and Lexically Driven Inferences. Theoretical Linguistics, 20, 97–164. Kamp, H., & Reyle, U. (1993). From Discourse to Logic. Dordrecht: Kluwer. Kortmann, B. (1991). Free Adjuncts and Absolutes in English. London: Routledge. Pusch, L. F. (1980). Kontrastive Untersuchungen zum italienischen ‘gerundio’. Tübingen: Niemeyer. Zifonun, G., Hoffmann, L., & Strecker, B. (1997). Grammatik der deutschen Sprache, Vol 3. Berlin: de Gruyter.
Language index
A Africans 214–217, 219, 222, 223, 432 Alamblak 294, 301, 306, 308, 310, 311, 313, 316, 317, 321, 322, 325, 329 Ancient Greek 291, 301, 329 B Black South African English 213–215 C Chidigo 3 Chinese 170, 176, 197, 199–201, 203–205, 208–210 Czech 277–279, 282, 285–288 D Danish 286 Digo 3, 5–7, 9–11, 14–16, 18, 19, 21 E English 4, 11, 20, 23, 37, 40, 41, 43, 46, 51, 52, 54, 55, 57, 59–63, 66–68, 70, 71, 73, 76–79, 82–87, 89, 92–104, 109–111, 113, 116, 118–122, 143, 144, 149–152, 154–156, 159, 161–164, 166, 168, 170–173, 175–178, 187, 198, 199, 206, 210, 213–217, 221–224, 228, 237, 239, 251–257, 259, 260, 263, 267, 270, 277, 278, 286, 291, 294, 297–299, 302, 305, 306, 308, 310–313, 315–318, 320, 322, 326,
329, 331, 333, 337–341, 345, 349, 351, 354–356, 359–361, 363–365, 373, 375, 376, 396, 398, 407, 408, 412, 413, 435, 442, 463–469, 471, 472, 475, 476 Estonian 373, 375–385, 387–389, 392–398, 400 Ewe 51, 55, 67–73
F Finnish 76–79, 84, 96, 99, 102, 109–113, 116–121, 390, 400 French 38, 55, 76–79, 82–84, 92, 93, 96, 99, 102, 103, 131–134, 138, 223, 237–242, 246, 247, 249, 277 Fuliiru 3, 5, 10–12, 14–19, 21
G German 21, 76–79, 84, 91, 96, 99, 102, 103, 111, 116, 117, 121, 125, 127, 130, 132–135, 137, 138, 251, 252, 254–260, 277, 286, 349–351, 361–364, 407, 408, 412, 419, 463, 464, 466–473, 475, 476 Giryama 7
H Hungarian 72, 73
51, 55, 59, 61, 63, 66, 67,
I Ifaluk 314, 318, 322, 329 Irish 39, 391
Language index
Italian 23, 24, 30, 34–36, 41–45, 47, 408, 412, 413, 419
J Japanese 197, 198, 200–210, 249, 263–272, 314, 328, 374, 396, 434 Java 303, 304, 312, 313, 319, 321
K Kifuliiru 3 Koine Hawaiian 329 Korean 210 Kwale 329
L Latin 43, 342 Luaniua 329 Lungga 329 Lunqa 302, 310, 319, 329
M Malay 141–150, 155, 156 Maskelyne 298–300, 302, 305, 307–310, 317, 325, 329 Middle English 40, 46, 345, 351, 363–365 Moroccan Arabic 143, 144
N Ndebele 214 Northern Sotho 214 Norwegian 51, 52, 54, 55, 63, 64, 66–68, 71, 73, 111, 116, 117, 119, 121, 466, 471, 475
O Old Italian 36, 43 Old Spanish 345, 351, 353–357, 360, 367 Ontong 303, 304, 312, 313, 319, 321, 329
P Pintupi 319, 329 Pokomo 7 Polish 115, 279, 286, 331, 333, 337–341, 442
R Romanian 237–242, 244, 246–249 Russian 95, 277–288, 318, 329, 442
S Samoan 318, 329 Scottish 39, 154 Sepedi 214, 217–224, 226–228, 230 Sesotho 214 Setswana 214, 217 Sibe 303, 313, 314, 316, 328 Siswati 214 Southern Sotho 214 Spanish 18, 159, 161–166, 168, 169, 172, 173, 175–180, 239, 267, 277, 345, 351, 353–357, 359, 360, 367 Swahili 4, 6, 7, 16, 21 Swedish 75–80, 82–84, 86, 87, 89, 91–93, 96–104, 109, 111, 112, 116–122
T Tahitian 318, 329 Tshivenda 214 Turkish 405, 410–413, 416, 418, 419
U Uare 301, 308, 310, 311, 325 Uliveo 298–300, 302, 305, 307, 309, 317, 329 Utku 299, 307, 309, 315, 329
X Xhosa 214, 216, 218 Xitsonga 214
Language index
Y Yankunytjatjara
301, 308, 329
Z Zulu 215–217
Name index
A Asher, N.
418, 419, 464, 475
B Bardovi-Harlig, K. 375, 376 Bateson, G. 374 Black, M. 114, 142 Blakemore, D. 58 Boxer, D. 251, 375 C Cameron, L. 114, 116, 147 Chaucer, G. 345, 351, 361, 362, 364, 365 Chick, J. 214–216 Clyne, M. 214 Coulmas, F. 239 Croft, W. 142 D Dagut, M. de Kadt, E. Dyvik, H. Duranti, A.
117 213, 215–217 463, 467, 468, 476 374
E Eco, U. 405, 406, 413 Ervin-Tripp, S. 213 F Farkas, D. 405, 415, 416 Fillmore, C. 81, 84 Fodor, J. 414, 419 Fritz, G. 346, 347, 367
G Goatly, A. 114, 138, 143, 161, 165, 170, 174 Goffman, E. 331, 332, 399 Grice, P. 424, 443–445, 452, 457, 458 Gumperz, J. 214, 399 H Halliday, M. 161, 464, 475 Higginbotham, J. 415 Hintikka, J. 415 Hodge, F. 213 Hofstede, G. 265, 266 Hönig, H. G. 349, 350 Hymes, D. 238 J Jaszczolt, K. 23, 418, 425, 441, 442, 444–446, 456, 459 Jayez, J. 31, 32, 45, 46 Johnson, M. 86, 95, 112, 113, 115, 125, 126, 135, 143, 160, 162 Jupp, T. 214 K Kachru, B. B. 215 Kamio, A. 207 Kamwangamalu, N. 214 Kimura, H. 205, 206, 208 Kussmaul, P. 349, 363, 364 L Lakoff, G. 75, 112, 113, 115, 125–128, 135–138, 162, 333, 447
Name index
Li, C. N. 199, 202, 410 Lindner, S. 88, 90, 103 Locke, J. 125, 138 M Martin, S. 198, 204, 210 Miller, G. 113, 114, 121, 263 Moriyama, T. 201, 205, 206, 208 Muzale, H. 15 Myers, G. 376, 397 N Newmark, P. 120, 159, 160 Nida, E. 305, 442, 454, 455 Nurse, D. 7, 15 Q Quirk, R.
Shea, D. 374 Snell-Hornby, M. 160 Sperber, D. 58, 59, 63, 443, 447–451, 453, 454, 459 Steen, G. 114
T Thompson, S. 47, 199, 202 Toulmin, S. 136 Tovena, L. 31, 45, 46 Traugott, E. 41, 42, 45, 47, 51, 52 Turner, K. 271, 272, 418
U Ungerer, F. 144 Uyeno, T. 208
46, 97
R Richards, I. A. 113, 114 Roberts, C. 214 S Sag, I. 414, 419 Schiffrin, D. 197, 208, 209 Schmid, H. 144, 429, 438 Schwenter, S. 18 Scollon, R. and S. 219, 220, 265 Searle, J. 183, 193, 237, 264, 367, 424, 438
V Vanderveken, D. 183, 188, 193, 194 Visconti, J. 23, 26, 29, 30, 36, 45–47
W Watanabe, S. 374, 376, 394, 399 Wierzbicka, A. 213, 222, 223, 237, 238, 240, 247–249, 291, 292, 309, 311, 312, 314, 317, 323, 326, 328, 349, 367, 443 Wilson, D. 51, 55, 56, 58, 59, 63, 115, 154, 443, 448, 454
Subject index
B ba 197, 199–210, 232, 233 Bank of English 144, 145, 150, 151, 153, 154, 156 breakdown of communication 213 British National Corpus (BNC) 82
conditional 23–26, 33, 36, 40, 41, 43, 44, 46, 51, 55–63, 65–67, 71, 73, 118, 277–283, 285, 286, 288, 384, 399, 416, 417 conducive question particle 71 confirmation 61, 63, 70, 72, 127, 197, 198, 202, 203, 205, 206, 208, 209, 359 consequence marker 67, 70 Contrastive Conversation Analysis 374 contrastive perspective 112, 150, 348 contrastive pragmatics 345–348, 442 conversational strategies 396, 400 style 375 cultural difference model 216 cultural script 237, 248, 249
C Cantar de mio Cid 345, 351, 352, 354, 356, 360, 365, 366 cognitive network model 333, 341 cognitive semantics 75, 141, 142 cognitive status 52–54 cognitivism 126 collaborative message construction 385 commissive 194 communicative style 375 comparative semantics 293 complaints 251, 252, 254–257, 259, 260 conceptual encoding 3 conceptual key 141, 144–146, 149, 150, 152–154
D da 25, 28–30, 33–35, 51, 63–68, 71–73, 120, 244, 245, 279 default semantics 426, 437, 444, 445, 455 deficit theory 215 de re-de dicto 405 deontic attitude 63, 64 deshoo 201, 202, 204–206, 208, 210 direction of fit 183, 193 direction of regulation 186, 189 directives 183, 213, 277, 279, 280, 282, 285, 288 see requests discourse completion tasks (DCTs) 253 Discourse Representation Theory (DRT) 444, 445, 465
A access ritual 332 agreement 197, 199, 201–203, 205, 207–209, 219, 385, 387, 388, 397, 399, 405, 411, 419 disagreements 388, 389, 391, 395 akkor 51, 59–63, 66, 67, 72, 73 argumentative 125, 132, 135–137
Subject index
dissociative attitude 63, 66, 73 distal aspects 3–5, 9–11, 16, 18–20 distance 5, 6, 8, 88, 109, 121, 163, 168, 200, 213, 219, 239, 253, 258, 260, 264, 266–268, 270, 271, 273, 284, 351, 452 distant address 279, 284, 286, 288 see politeness E elaboration 453, 463–468, 471, 473, 475 emotions 291, 292, 296–298, 304, 310, 316, 322–324, 326, 327, 396 empathy 204, 317 encyclopaedic knowledge 144, 156 endpoint focus 84 English-Norwegian Parallel Corpus (ENPC) 110 English Swedish Parallel Corpus (ESPC) 76, 82 entailment 137 epistemic attitude 63, 72 Euro-discourse 127, 135 EuroWordNet 75 everyday rituals 331, 333, 335–337, 341 explanation and modelling 167 explicature 56, 58, 59, 64, 71, 453, 456 F face 76, 115, 120, 175, 185, 213, 221, 245, 248, 251, 252, 263, 332, 334, 348, 384, 388, 391, 393, 394 see politeness Face-Threatening Acts (FTAs) 251 family imagery 125, 127 farewells 332, 333, 335, 337, 339–341 see greetings figurative phrases 141, 142, 144, 145, 152
File Change Semantics 445 filling lexical gaps 161, 164 focus 3, 42, 52–54, 56, 65, 66, 75, 79, 84, 85, 88, 90, 91, 95, 98, 103, 109, 166, 170–172, 178, 215, 218, 278, 300, 303, 307, 345, 357, 361, 363, 375, 380, 424, 425 foreground-background structure (FG-BG) 423, 427, 438 framing 58, 213, 224, 225, 373, 374, 376, 378, 392, 393, 395–399 G genericity 406–409, 419 gestalt analysis 340, 341 grammaticalization 19, 23, 41, 42, 47, 51–54, 59, 63, 66, 67, 72, 73, 96, 103 greetings 228, 332, 333, 335–341, 368 see farewells Grice’s maxims of conversation 445 grounding 134, 136, 423–437 group discussions 373, 398 H high boundary tone 64 higher-level explicature 71 higher-order entity 52–54 historical dialogue analysis 346, 347 humour and games 176 I iconic 67, 155 illocutionary acts 183, 189, 192–195 illocutionary marker 278, 281, 286 imperative 8, 9, 15, 65, 175, 210, 213, 270, 277, 280, 281, 285, 286, 288, 359, 384, 412, 419 implicature 57, 447, 453 imposition 213, 219, 221, 264, 266–269, 271–273 indebtedness 238–243, 246 indirect constructions 277
Subject index
inference particle 55, 59, 63–68, 72, 73 information ownership 197, 205, 208 information territory 208 interaction rituals 331 interactional strategies 373, 375, 398 interrogative structure 350, 353, 357, 358, 360 intonational phrasing 65 intralingual translation 354, 361 “Invariance Principle” 126, 135 involvement 173, 198, 204, 396, 399 K komma 75–80, 82–96, 98–104 L lexical cohesion 165, 168, 170, 173, 177 low boundary tone 64 M Malay Language Agency 145 material implication 64, 72 mental events 89, 90 metaphor 78, 86, 87, 91, 93, 95, 109–121, 125–128, 131, 132, 134–138, 142–144, 147–149, 154, 159–165, 167–171, 173, 175, 176, 178, 179, 438, 443 activation of inactive metaphors 174 conceptual 113, 116, 119, 142, 143 corpus 127 dead 109, 160, 164, 177 diversification of 165 Event Structure Metaphor 86 extension 168, 170, 178 extended 115, 163, 168 family 126, 127, 131 functions of 162 ideological function of 162 in translation 116
inactive metaphorical expressions 165, 167, 178 innovative 109, 113, 120, 121 lexicalized 109, 113 literary 110 love 127 ‘metaphorhood’ 113 ‘metaphoricity’ 113, 114 mixed 170 word-play on a 173, 176 metarepresentation 449 metonym 117, 143, 144, 146, 153, 154 reference 163 modal auxiliary 56, 277, 285 multilingual corpus 109, 468, 475 multi-party talk 373, 376 N narrative 6, 8, 79, 81, 202, 359, 373, 392, 437 Natural Semantic Metalanguage (NSM) 237, 291, 292 ne 31, 85, 197–210, 239, 244, 278, 279, 288 negation 70, 247, 277, 278, 280–287, 364, 409, 419, 444, 457 New Oxford Dictionary of English 120 non-distant address 278, 282, 284, 288 non-truth-conditional inference 56, 58, 73 P personification 163 physical event 89 polarity focus 65, 66 politeness 193, 216–223, 230, 237, 241, 244, 248, 252, 264, 271, 272, 278, 280, 281, 284–288, 331, 333, 346, 347, 375, 379, 381, 383, 384, 389, 391, 392, 394, 395, 398 see distance
Subject index
see face deference 219 lexical politeness marker 280, 281, 284, 286 marker 278, 280, 281, 284–287 positive 217, 375, 379, 391, 392, 395 strategies 219, 252, 381, 383, 391, 392, 394 polysemy 56, 75, 104, 177, 324 ‘popular terms’ 166, 167, 178 popularisation texts 160, 167 power 35, 114–116, 131, 136, 168, 213, 219, 220, 253, 264, 266–273, 326, 353, 359, 474 see distance and status pragmatics 109, 180, 215, 251, 253, 345–350, 352, 354, 357, 361, 362, 365, 366, 373, 375, 398, 418, 423–426, 431, 436–438, 441, 442, 444, 445, 456, 457, 459 historical 345–348, 350, 352, 354, 357, 361, 362, 366 interpersonal 375 pragmalinguistic 215, 223 pragmatic competence 215 pragmatic equivalence 335, 340, 341, 442–447, 449, 454, 455, 459 pragmatic failure 213, 214, 217 pragmatic pattern 352, 358–361 pragmatic remodelling 352, 354, 366 see semantics-pragmatics interface presupposition 136, 137 Prime Minister’s Question Time 350 procedural encoding 3 information 59, 60 pronominal address 278 prototype analysis 334, 337 public discourse 125–127, 135, 137
Q qualora 23–37, 40–45, 47 quando 27, 35, 36, 45 quasi-synonyms 165, 166, 177 questions 27, 28, 60, 64, 71, 114, 208, 247, 252, 255, 277–285, 287, 288, 291, 295, 335, 349, 350, 352, 353, 356, 357, 359–361, 367, 368, 443 R ‘rapport’ 198, 204, 208 register 136, 178, 380, 396 Relevance Theory 51, 115, 455 remedial rituals 332 requests 61, 72, 207, 209, 213–225, 230, 243, 251, 256, 263–265, 267, 268, 270–272, 277–288, 334, 353, 443 conventionalized indirect requests 270, 277, 286, 287 see directives indirect requests 222, 223, 243, 263, 264, 267, 268, 270–272, 277–282, 284–288 rhetorical question 350, 360, 368 ritual 152, 331–334 routine formulae 332 S scenarios 133, 137, 166, 254, 326 segmental attrition 66 selectional restrictions 172 semantic components 6, 178, 291, 292, 294, 296, 297, 323 semantics-pragmatics interface 423, 424, 436 sentence-final particles 197, 198, 201, 210 shared information 198 simile 24, 28, 112–114, 118, 146 socio-pragmatics 215, 223 softening 201, 203, 209 specificity 405–407, 409–419
Subject index
speech act shifts 345, 361 speech act theory 183, 184, 187, 192, 237, 349 speech act verbs 194, 237, 238 speech event 11, 12, 374–376, 392, 393, 398 status 7, 23, 35, 42, 46, 47, 52–54, 130, 136, 149, 161, 174, 193, 202, 207, 209, 213, 214, 219–221, 244, 265, 269, 270, 280, 304, 312, 316, 319, 332, 335, 361, 363, 365, 377, 433, 435, 444, 448 see distance stock exchange markets 159, 169 supportive rituals 332 supposition 46, 199, 204–206 synonyms 92, 165, 166
thanking formulas 239, 241 then 51–73 translatability 117 translation studies 159, 179, 345, 348–351, 366 truth-conditional adverbial 51, 66, 73 turn change 198, 208 turn-management 198
T target language 109, 159, 215, 253, 451 tense-aspect-modality (TAM) 3 tertium comparationis 23, 44, 348, 441, 443, 458 text-structuring function 166 textual approach 113, 160 thank 238 gratitude 238, 239, 241, 242, 245, 247, 333
W warrant 67, 136, 137 whenever 23–25, 27, 28, 33–41, 43, 44, 46, 47, 243 Wife of Bath’s Prologue and Tale 345, 351, 361, 366 WordNet 75
U universal quantifier 30, 31, 414–416
V value judgement 169, 177
Y yo 201–203, 205, 210, 233, 357
Contents of Volume 1
Acknowledgements
ix
Editorial preface Katarzyna M. Jaszczolt and Ken Turner
xi
Negation Distributional restrictions on negative determiners L. M. Tovena
3
Towards a comprehensive view of Negative Concord João Peres
29
Temporality On temporal constructions involving counting from anchor points – Semantic and pragmatic issues Telmo Móia
45
On the semantics and pragmatics of situational anaphoric temporal locators in Portuguese and in English Ana Teresa Alves
61
Remarks on the semantics of eventualities with measure phrases in English and Romanian Ilinca Cr˘ainiceanu
75
The present perfect in English and in Catalan Hortènsia Curell A contrastive reading of temporal-aspectual morphemes in Swahili: The case of ‘-li’ and ‘-me’ Frederick Kang’ethe Iraki
101
117
Contents of Volume 1
Modality Semantic and pragmatic constraints on mood selection Rui Marques Dilemmas and excogitations: Further considerations on modality, clitics and discourse A. Capone
129
147
Evidentiality Inferred evidence: Language-specific properties and universal constraints Sergei Tatevosov
177
Extension of meaning: Verbs of perception in English and Lithuanian Aurelia Usonien˙e
193
Perspectives on eventualities Information structure, argument structure, and typological variation Márta Maleczki
223
The network of demotion: Towards a unified account of passive constructions Andrea Sansò
245
Valence change and the function of intransitive verbs in English and Japanese Mayumi Masuko
261
The transitive/intransitive construction of events in Japanese and English discourse Patricia Mayes
277
Topics in grammar and conceptualization Towards a universal DRT model for the interpretation of directional PPs within a reference frame Didier Maillat
295
Contents of Volume 1
The interaction of syntax and pragmatics: The case of Japanese ‘gapless’ relatives Akiko Kurosawa
307
Constraint interaction at the semantics/pragmatics interface: The case of clitic doubling Javier Gutiérrez-Rexach
335
Cross-language commutation tests and their application to an error-prone contrastive problem – Ger. einige, Fr. quelques, Sp. algunos Eva Lavric
355
Language index
371
Name index
373
Subject index
375
Contents of Volume 2
381
In the PRAGMATICS AND BEYOND NEW SERIES the following titles have been published thus far or are scheduled for publication: 1. WALTER, Bettyruth: The Jury Summation as Speech Genre: An Ethnographic Study of What it Means to Those who Use it. Amsterdam/Philadelphia, 1988. 2. BARTON, Ellen: Nonsentential Constituents: A Theory of Grammatical Structure and Pragmatic Interpretation. Amsterdam/Philadelphia, 1990. 3. OLEKSY, Wieslaw (ed.): Contrastive Pragmatics. Amsterdam/Philadelphia, 1989. 4. RAFFLER-ENGEL, Walburga von (ed.): Doctor-Patient Interaction. Amsterdam/Philadelphia, 1989. 5. THELIN, Nils B. (ed.): Verbal Aspect in Discourse. Amsterdam/Philadelphia, 1990. 6. VERSCHUEREN, Jef (ed.): Selected Papers from the 1987 International Pragmatics Conference. Vol. I: Pragmatics at Issue. Vol. II: Levels of Linguistic Adaptation. Vol. III: The Pragmatics of Intercultural and International Communication (ed. with Jan Blommaert). Amsterdam/Philadelphia, 1991. 7. LINDENFELD, Jacqueline: Speech and Sociability at French Urban Market Places. Amsterdam/Philadelphia, 1990. 8. YOUNG, Lynne: Language as Behaviour, Language as Code: A Study of Academic English. Amsterdam/Philadelphia, 1990. 9. LUKE, Kang-Kwong: Utterance Particles in Cantonese Conversation. Amsterdam/Philadelphia, 1990. 10. MURRAY, Denise E.: Conversation for Action. The computer terminal as medium of communication. Amsterdam/Philadelphia, 1991. 11. LUONG, Hy V.: Discursive Practices and Linguistic Meanings. The Vietnamese system of person reference. Amsterdam/Philadelphia, 1990. 12. ABRAHAM, Werner (ed.): Discourse Particles. Descriptive and theoretical investigations on the logical, syntactic and pragmatic properties of discourse particles in German. Amsterdam/Philadelphia, 1991. 13. NUYTS, Jan, A. Machtelt BOLKESTEIN and Co VET (eds): Layers and Levels of Representation in Language Theory: A functional view. Amsterdam/Philadelphia, 1990. 14. SCHWARTZ, Ursula: Young Children’s Dyadic Pretend Play. Amsterdam/Philadelphia, 1991. 15. KOMTER, Martha: Conflict and Cooperation in Job Interviews. Amsterdam/Philadelphia, 1991. 16. MANN, William C. and Sandra A. THOMPSON (eds): Discourse Description: Diverse Linguistic Analyses of a Fund-Raising Text. Amsterdam/Philadelphia, 1992. 17. PIÉRAUT-LE BONNIEC, Gilberte and Marlene DOLITSKY (eds): Language Bases ... Discourse Bases. Amsterdam/Philadelphia, 1991. 18. JOHNSTONE, Barbara: Repetition in Arabic Discourse. Paradigms, syntagms and the ecology of language. Amsterdam/Philadelphia, 1991. 19. BAKER, Carolyn D. and Allan LUKE (eds): Towards a Critical Sociology of Reading Pedagogy. Papers of the XII World Congress on Reading. Amsterdam/Philadelphia, 1991. 20. NUYTS, Jan: Aspects of a Cognitive-Pragmatic Theory of Language. On cognition, functionalism, and grammar. Amsterdam/Philadelphia, 1992. 21. SEARLE, John R. et al.: (On) Searle on Conversation. Compiled and introduced by Herman Parret and Jef Verschueren. Amsterdam/Philadelphia, 1992.
22. AUER, Peter and Aldo Di LUZIO (eds): The Contextualization of Language. Amsterdam/Philadelphia, 1992. 23. FORTESCUE, Michael, Peter HARDER and Lars KRISTOFFERSEN (eds): Layered Structure and Reference in a Functional Perspective. Papers from the Functional Grammar Conference, Copenhagen, 1990. Amsterdam/Philadelphia, 1992. 24. MAYNARD, Senko K.: Discourse Modality: Subjectivity, Emotion and Voice in the Japanese Language. Amsterdam/Philadelphia, 1993. 25. COUPER-KUHLEN, Elizabeth: English Speech Rhythm. Form and function in everyday verbal interaction. Amsterdam/Philadelphia, 1993. 26. STYGALL, Gail: Trial Language. A study in differential discourse processing. Amsterdam/ Philadelphia, 1994. 27. SUTER, Hans Jürg: The Wedding Report: A Prototypical Approach to the Study of Traditional Text Types. Amsterdam/Philadelphia, 1993. 28. VAN DE WALLE, Lieve: Pragmatics and Classical Sanskrit. Amsterdam/Philadelphia, 1993. 29. BARSKY, Robert F.: Constructing a Productive Other: Discourse theory and the convention refugee hearing. Amsterdam/Philadelphia, 1994. 30. WORTHAM, Stanton E.F.: Acting Out Participant Examples in the Classroom. Amsterdam/Philadelphia, 1994. 31. WILDGEN, Wolfgang: Process, Image and Meaning. A realistic model of the meanings of sentences and narrative texts. Amsterdam/Philadelphia, 1994. 32. SHIBATANI, Masayoshi and Sandra A. THOMPSON (eds): Essays in Semantics and Pragmatics. Amsterdam/Philadelphia, 1995. 33. GOOSSENS, Louis, Paul PAUWELS, Brygida RUDZKA-OSTYN, Anne-Marie SIMONVANDENBERGEN and Johan VANPARYS: By Word of Mouth. Metaphor, metonymy and linguistic action in a cognitive perspective. Amsterdam/Philadelphia, 1995. 34. BARBE, Katharina: Irony in Context. Amsterdam/Philadelphia, 1995. 35. JUCKER, Andreas H. (ed.): Historical Pragmatics. Pragmatic developments in the history of English. Amsterdam/Philadelphia, 1995. 36. CHILTON, Paul, Mikhail V. ILYIN and Jacob MEY: Political Discourse in Transition in Eastern and Western Europe (1989-1991). Amsterdam/Philadelphia, 1998. 37. CARSTON, Robyn and Seiji UCHIDA (eds): Relevance Theory. Applications and implications. Amsterdam/Philadelphia, 1998. 38. FRETHEIM, Thorstein and Jeanette K. GUNDEL (eds): Reference and Referent Accessibility. Amsterdam/Philadelphia, 1996. 39. HERRING, Susan (ed.): Computer-Mediated Communication. Linguistic, social, and cross-cultural perspectives. Amsterdam/Philadelphia, 1996. 40. DIAMOND, Julie: Status and Power in Verbal Interaction. A study of discourse in a closeknit social network. Amsterdam/Philadelphia, 1996. 41. VENTOLA, Eija and Anna MAURANEN, (eds): Academic Writing. Intercultural and textual issues. Amsterdam/Philadelphia, 1996. 42. WODAK, Ruth and Helga KOTTHOFF (eds): Communicating Gender in Context. Amsterdam/Philadelphia, 1997. 43. JANSSEN, Theo A.J.M. and Wim van der WURFF (eds): Reported Speech. Forms and functions of the verb. Amsterdam/Philadelphia, 1996.
44. BARGIELA-CHIAPPINI, Francesca and Sandra J. HARRIS: Managing Language. The discourse of corporate meetings. Amsterdam/Philadelphia, 1997. 45. PALTRIDGE, Brian: Genre, Frames and Writing in Research Settings. Amsterdam/Philadelphia, 1997. 46. GEORGAKOPOULOU, Alexandra: Narrative Performances. A study of Modern Greek storytelling. Amsterdam/Philadelphia, 1997. 47. CHESTERMAN, Andrew: Contrastive Functional Analysis. Amsterdam/Philadelphia, 1998. 48. KAMIO, Akio: Territory of Information. Amsterdam/Philadelphia, 1997. 49. KURZON, Dennis: Discourse of Silence. Amsterdam/Philadelphia, 1998. 50. GRENOBLE, Lenore: Deixis and Information Packaging in Russian Discourse. Amsterdam/Philadelphia, 1998. 51. BOULIMA, Jamila: Negotiated Interaction in Target Language Classroom Discourse. Amsterdam/Philadelphia, 1999. 52. GILLIS, Steven and Annick DE HOUWER (eds): The Acquisition of Dutch. Amsterdam/ Philadelphia, 1998. 53. MOSEGAARD HANSEN, Maj-Britt: The Function of Discourse Particles. A study with special reference to spoken standard French. Amsterdam/Philadelphia, 1998. 54. HYLAND, Ken: Hedging in Scientific Research Articles. Amsterdam/Philadelphia, 1998. 55. ALLWOOD, Jens and Peter Gärdenfors (eds): Cognitive Semantics. Meaning and cognition. Amsterdam/Philadelphia, 1999. 56. TANAKA, Hiroko: Language, Culture and Social Interaction. Turn-taking in Japanese and Anglo-American English. Amsterdam/Philadelphia, 1999. 57 JUCKER, Andreas H. and Yael ZIV (eds): Discourse Markers. Descriptions and theory. Amsterdam/Philadelphia, 1998. 58. ROUCHOTA, Villy and Andreas H. JUCKER (eds): Current Issues in Relevance Theory. Amsterdam/Philadelphia, 1998. 59. KAMIO, Akio and Ken-ichi TAKAMI (eds): Function and Structure. In honor of Susumu Kuno. 1999. 60. JACOBS, Geert: Preformulating the News. An analysis of the metapragmatics of press releases. 1999. 61. MILLS, Margaret H. (ed.): Slavic Gender Linguistics. 1999. 62. TZANNE, Angeliki: Talking at Cross-Purposes. The dynamics of miscommunication. 2000. 63. BUBLITZ, Wolfram, Uta LENK and Eija VENTOLA (eds.): Coherence in Spoken and Written Discourse. How to create it and how to describe it.Selected papers from the International Workshop on Coherence, Augsburg, 24-27 April 1997. 1999. 64. SVENNEVIG, Jan: Getting Acquainted in Conversation. A study of initial interactions. 1999. 65. COOREN, François: The Organizing Dimension of Communication. 2000. 66. JUCKER, Andreas H., Gerd FRITZ and Franz LEBSANFT (eds.): Historical Dialogue Analysis. 1999. 67. TAAVITSAINEN, Irma, Gunnel MELCHERS and Päivi PAHTA (eds.): Dimensions of Writing in Nonstandard English. 1999. 68. ARNOVICK, Leslie: Diachronic Pragmatics. Seven case studies in English illocutionary development. 1999.
69. NOH, Eun-Ju: The Semantics and Pragmatics of Metarepresentation in English. A relevance-theoretic account. 2000. 70. SORJONEN, Marja-Leena: Responding in Conversation. A study of response particles in Finnish. 2001. 71. GÓMEZ-GONZÁLEZ, María Ángeles: The Theme-Topic Interface. Evidence from English. 2001. 72. MARMARIDOU, Sophia S.A.: Pragmatic Meaning and Cognition. 2000. 73. HESTER, Stephen and David FRANCIS (eds.): Local Educational Order. Ethnomethodological studies of knowledge in action. 2000. 74. TROSBORG, Anna (ed.): Analysing Professional Genres. 2000. 75. PILKINGTON, Adrian: Poetic Effects. A relevance theory perspective. 2000. 76. MATSUI, Tomoko: Bridging and Relevance. 2000. 77. VANDERVEKEN, Daniel and Susumu KUBO (eds.): Essays in Speech Act Theory. 2002. 78. SELL, Roger D. : Literature as Communication. The foundations of mediating criticism. 2000. 79. ANDERSEN, Gisle and Thorstein FRETHEIM (eds.): Pragmatic Markers and Propositional Attitude. 2000. 80. UNGERER, Friedrich (ed.): English Media Texts – Past and Present. Language and textual structure. 2000. 81. DI LUZIO, Aldo, Susanne GÜNTHNER and Franca ORLETTI (eds.): Culture in Communication. Analyses of intercultural situations. 2001. 82. KHALIL, Esam N.: Grounding in English and Arabic News Discourse. 2000. 83. MÁRQUEZ REITER, Rosina: Linguistic Politeness in Britain and Uruguay. A contrastive study of requests and apologies. 2000. 84. ANDERSEN, Gisle: Pragmatic Markers and Sociolinguistic Variation. A relevance-theoretic approach to the language of adolescents. 2001. 85. COLLINS, Daniel E.: Reanimated Voices. Speech reporting in a historical-pragmatic perspective. 2001. 86. IFANTIDOU, Elly: Evidentials and Relevance. 2001. 87. MUSHIN, Ilana: Evidentiality and Epistemological Stance. Narrative retelling. 2001. 88. BAYRAKTAROG LU, ArFn and Maria SIFIANOU (eds.): Linguistic Politeness Across Boundaries. The case of Greek and Turkish. 2001. 89. ITAKURA, Hiroko: Conversational Dominance and Gender. A study of Japanese speakers in first and second language contexts. 2001. 90. KENESEI, István and Robert M. HARNISH (eds.): Perspectives on Semantics, Pragmatics, and Discourse. A Festschrift for Ferenc Kiefer. 2001. 91. GROSS, Joan: Speaking in Other Voices. An ethnography of Walloon puppet theaters. 2001. 92. GARDNER, Rod: When Listeners Talk. Response tokens and listener stance. 2001. 93. BARON, Bettina and Helga KOTTHOFF (eds.): Gender in Interaction. Perspectives on femininity and masculinity in ethnography and discourse. 2002 94. McILVENNY, Paul (ed.): Talking Gender and Sexuality. 2002. 95. FITZMAURICE, Susan M.: The Familiar Letter in Early Modern English. A pragmatic approach. 2002. 96. HAVERKATE, Henk: The Syntax, Semantics and Pragmatics of Spanish Mood. 2002.
97. MAYNARD, Senko K.: Linguistic Emotivity. Centrality of place, the topic-comment dynamic, and an ideology of pathos in Japanese discourse. 2002. 98. DUSZAK, Anna (ed.): Us and Others. Social identities across languages, discourses and cultures. 2002. 99. JASZCZOLT, K.M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume 1. 2003. 100. JASZCZOLT, K.M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume 2. 2003. 101. LUKE, Kang Kwong and Theodossia-Soula PAVLIDOU (eds.): Telephone Calls. Unity and diversity in conversational structure across languages and cultures. 2002. 102. LEAFGREN, John: Degrees of Explicitness. Information structure and the packaging of Bulgarian subjects and objects. 2002. 103. FETZER, Anita and Christiane MEIERKORD (eds.): Rethinking Sequentiality. Linguistics meets conversational interaction. 2002. 104. BEECHING, Kate: Gender, Politeness and Pragmatic Particles in French. 2002. 105. BLACKWELL, Sarah E.: Implicatures in Discourse. The case of Spanish NP anaphora. 2003. 106. BUSSE, Ulrich: Linguistic Variation in the Shakespeare Corpus. Morpho-syntactic variability of second person pronouns. 2002. 107. TAAVITSAINEN, Irma and Andreas H. JUCKER (eds.): Diachronic Perspectives on Address Term Systems. n.y.p. 108. BARRON, Anne: Acquisition in Interlanguage Pragmatics. How to do things with words in a study abroad context. n.y.p. 109. MAYES, Patricia: Language, Social Structure, and Culture. A genre analysis of cooking classes in Japan and America. n.y.p. 110. ANDROUTSOPOULOS, Jannis K. and Alexandra GEORGAKOPOULOU (eds.): Discourse Constructions of Youth Identities. n.y.p. 111. ENSINK, Titus and Christoph SAUER (eds.): Framing and Perspectivising in Discourse. n.y.p. 112. LENZ, Friedrich (ed.): Deictic Conceptualisation of Space, Time and Person. n.y.p. 113. PANTHER, Klaus-Uwe and Linda L. THORNBURG (eds.): Metonymy and Pragmatic Inferencing. n.y.p.
