The Yearbook of South Asian Languages and Linguistics: 2006 9783110186628, 9783110186611

Citation preview

The Yearbook of South Asian Languages and Linguistics 2006

EDITORIAL BOARD Chief Editor: Rajendra Singh, De´pt. de Linguistique, Universite´ de Montre´al, C.P. 6128 Succ. Centre-ville, Montre´al, P.Q., Canada H3C 3J7 (rajendra.singh@ umontreal.ca). Senior Associate Editor: Probal Dasgupta, Centre for Applied Linguistics, University of Hyderabad, Hyderabad, India 500046 ([email protected]). Associate Editors: R. K. Agnihotri, Department of Linguistics, The University of Delhi, Delhi, India 110007 ([email protected]); Josef Bayer, Universita¨t Konstanz, Konstanz, Germany ([email protected]); P. E. Hook, Department of Linguistics, University of Michigan, Ann Arbor, Mich., USA ([email protected]). Regional Editors: Tej K. Bhatia, Syracuse University, NY, USA; Baber S. A. Khan, UAE Air Force Academy, Al-Ain, United Arab Emirates; Rajend Mesthrie, University of Cape Town, South Africa; France Mugler, University of the South Pacific, Fiji; John Peterson, Universita¨t Osnabru¨ck, Germany; Tariq Rahman, National Institute of Pakistan Studies, Islamabad, Pakistan; A. P. Saleemi, National Chi Nan University, Taiwan; Udaya Narayana Singh, Central Institute of Indian Languages, Mysore, India; E´. Ti¤ou, Universite´ de Montre´al, Montreal, Canada; Yogendra P. Yadav, Royal Nepal Academy, Kathmandu, Nepal. Editorial Advisors: E. Annamalai, Mysore, India; Bernard Comrie, Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany; Wolfgang U. Dressler, Universita¨t Wien, Vienna, Austria; Hans Heinrich Hock, Univeristy of Illinois at Urbana—Champaign; Aravind Joshi, University of Pennsylvania, Philadelphia, USA; Ashok R. Kelkar, Pune, India; Paul Kiparsky, Stanford University, USA; E. F. Konrad Koerner, University of Ottawa, Canada; Bh. Krishnamurti, Hyderabad, India; Jayant K. Lele, Queen’s University, Kingston, Canada; Marvin Minsky, MIT, Cambridge, MA, USA; Pieter Muysken, Universiteit van Amsterdam, The Netherlands; N. S. Prabhu, Bangalore, India. Editorial Assistants: Montre´al.

Stephen Moran and Shishir Bhattacharja, Universite´ de

South Asia is home to a large number of languages and dialects. While the number of linguists working on South Asia has grown considerably in the recent past, there is as yet no recognized international forum for the exchange of ideas among them. The Yearbook of South Asian Languages and Linguistics is designed to be just that forum. It will consolidate empirical and theoretical research and provide a testing ground for the articulation of new ideas and approaches grounded in a study of South Asian languages but which have universal applicability.

The Yearbook of South Asian Languages and Linguistics 2006 Editor-in-chief Rajendra Singh

Mouton de Gruyter Berlin . New York

Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter GmbH & Co. KG, Berlin.

y Printed on acid-free paper which falls within the guidelines z of the ANSI to ensure permanence and durability.

ISSN 0971-9539 ISBN-13: 978-3-11-018661-1 ISBN-10: 3-11-018661-6 Bibliographic information published by Die Deutsche Bibliothek Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available on the Internet at hhttp://dnb.ddb.dei. > Copyright 2006 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording or any information storage and retrieval system, without permission in writing from the publisher. Cover: Christopher Schneider, Berlin. Southasian letters design by Bharati Mirchandani. Typeset by Asco Typesetters, Hong Kong. Printed in Germany.

Contents

Commentum Editoris A.

Articles Elena Bashir Change in Progress: Negation in Hindi and Urdu Elabbas Benmamoun and Rajesh Kumar The Overt Licensing of NPIs in Hindi Shishir Bhattacharja Reduplication in Bengali Annie Montaut Mirative Meanings as Extensions of Aorist in Hindi/Urdu Terry Varma Some Aspects of the Meaning of the Hindi Particle hii

B.

vii

3 31 49 71 87

Regional Reports, Reviews and Abstracts Regional Reports Elena Bashir Pakistan: Research and Developments in Linguistics and Language Study Omar N. Koul Linguistic Studies on Kashmiri Yogendra P. Yadava Linguistic Activities in Nepal (1999–2004) Reviews Rajeshwari V. Pandaripande, Sociolinguistic Dimensions of Marathi: Multilingualism in Central India by Francesco Gardani

125 145 153

165

vi

Contents Wali Kashi, Marathi by Leena Kelkar-Stephan Bernhard Hurch (ed.), Studies on Reduplication by Lionel Wee Paul St-Pierre and Prafulla C. Kar (eds.), In Translation: Reflections, Refractions, Transformations by Carmen Wickramagamage

C.

173

178

Dialogue Probal Dasgupta Language Policies and Lesser-Known Languages in India Stephen Moran The Baby in the ARSA Bathwater: A Response to Dasgupta

D.

172

193 207

Announcements The Gyandeep Prize Housekeeping

213 213

Notes on Contributors

215

Commentum Editoris

We are delighted to put this, the ninth, volume of The Yearbook of South Asian Languages and Linguistics in your hands. It is the fourth volume to be brought out by Mouton. We are delighted to note that the previous issues brought out by Mouton have been found to be rich and extremely well produced. We are grateful to Dr. Ursula Kleinhenz and Mr. Wolfgang Konwitschny for their continued cooperation. We are also grateful to Stephen Moran, who did a wonderful job of copy-editing, and Shishir Bhattacharja, who provided considerable last-minute help with files for digital transmission. At the suggestion of some of our readers, in this issue we suspend the distinction Invited/Open. Running a yearbook of this sort is, however, a bit like running a multi-national country—issues of regional and subdisciplinary representation are important in both. Should our readers find and should it be decided that abandoning the distinction in question can result in or has in fact resulted in some asymmetries, it may be necessary to go back to that distinction, originally intended to avoid such asymmetries. Be that as it may. We are happy to present this volume. It contains four papers on Hindi (syntax and pragmatics) and a paper on Bengali morphology. We are particularly happy to point out that although the four papers on Hindi deal with the same language, they deal with very di¤erent things, and from very di¤erent perspectives. The regional reports are on Nepal and Kashmir. Although we had hoped that the latter would deal with research in Kashmiri and be something of the sort Annamalai’s report on Tamil was in YB-2004, Prof. Koul felt that there was not enough done on Kashmiri in Kashmiri. As we thought that the report he did send us was useful, we have included it. We would, however, like to take this opportunity to invite our colleagues, particularly in South Asia, to consider preparing reports on research on South Asian languages done and published in South Asian languages. We think it is a valuable resource that needs to be tapped into and made public.

viii

Commentum Editoris

As for the reviews, we are happy to note that both the linguistic and disciplinary diversity represented is quite rich. We are particularly pleased to include an important review of an important book on translation, a very vibrant field in South Asia. We had expected to include two more reviews in this issue, but will have to include them in the next issue due to unexpected delays. It is important for our readers and contributors to know that this is the last stand-alone issue of the Yearbook. Although R.S. will continue to edit it, it will appear as an integral part of the Trends in Linguistics series of Mouton, edited by Walter Bisang and Hans Heinrich Hock, from next year. We also have the somewhat unpleasant task of reminding our authors that even their initial submissions must conform to our style-sheet. Most of the submissions this year, unfortunately, did not do so. In future, the editor will regretfully have to return such submissions unprocessed. R.S., P.D., R.K.A., J.B., and P.E.H.

A

Articles

Change in Progress: Negation in Hindi and Urdu* ELENA BASHIR

The question of the choice of negative particle (na or nahI¯) in the simple perfective arises frequently in pedagogical contexts. Students want to know why, when most textbooks prescribe nahI¯ as the negative particle for the simple perfective, they frequently see na in this function. This paper addresses this question, examining diachronic patterns of use for these two particles and variables correlated with the appearance of na or nahI¯. The Hindi and Urdu situation is compared with that of Bengali, and a reanalysis in progress is described, by which nahI¯ has come/is coming to be perceived as a single semantic unit, NEG, rather than as containing two elements—NEG þ EMPH or NEG þ AUX. 1. Introduction This study investigates factors correlating with the distribution of the negative particles na and nahI¯ 1 in Hindi and Urdu. The impetus for this work is a pedagogical problem: students often ask about the choice between na and nahI¯ in those tenses where both are possible. They soon notice that, despite what is stated in many textbooks, simple perfective and future sentences are sometimes negated with na rather than nahI¯. Most grammars are not very helpful in answering this question, some saying something like ‘‘nahI¯ is more emphatic than na’’, which suggests that na is the unmarked form and that nahI¯ is ‘‘na plus something special.’’ The most specific account I have found (Barker 1967: 269) says: ‘‘The negative for the simple past tense is either na or nahI¯. Both are correct, but the latter carries slightly stronger connotations of emphasis and negation.’’ At the same time, we find generalizations for the distribution of na and nahI¯, to the e¤ect that na is used in certain specified finite contexts, mainly subjunctive and imperative, nahI¯ elsewhere. For example, Bhatia (1973: 5) says: ‘‘nahI¯ is the representative NEG-particle. The other two NEG-particles appear in specific environments. m t appears with the imperative form of a verb, whereas n shows up in imperative, optative, participial, gerundive and conjoined constructions.’’ e

e

4

Elena Bashir

The implication of this is that nahI¯ is the default or unmarked form, and that na is ‘special’ or of more restricted distribution. These two apparently contradictory implications invite explanation. Another question concerns the absence of the auxiliary in negative sentences with the present imperfect, perfect and sometimes progressive tenses, as well as in present-tense copular and existential sentences. While no reason is o¤ered for the absence of the auxiliary in these contexts, the explanation o¤ered for when the auxiliary is present is that it is retained in ‘emphatic’ contexts. This question is directly related to the distribution of na and nahI¯. In this study I explore these questions by analyzing naturally-occurring negative sentences in context. Particular focus is on the use of na and nahI¯ with tense-aspect forms which allow both these negative elements. First, I outline the methodology of the study, then discuss diachronic considerations, some relevant recent research in Hindi and Bangla, and the variables examined in this study. I conclude with a summary of the diachronic developments and shifting analyses of negation in Hindi and Urdu. 2. Methodology The overall frequency and the distribution of na and nahI¯ are investigated in one complete Hindi novel (Premchand’s Goda¯n, ca. 1936).2 To add a diachronic dimension, the occurrence patterns in a contemporary Urdu newspaper corpus (Daily Jang) are examined.3 Only indicative sentences were counted: imperatives, subjunctives, protasis clauses of realis, and both protasis and apodosis clauses of irrealis conditionals were excluded. In counting occurrences of na in relation to specific variables, sentences in which na occurred in the parallel constructions na {[S]/[NP]} . . . na {[S]/[NP]} or nahI¯ {[S]/[NP]} . . . na {[S]/[NP]} ‘neither {S/NP} nor {S/NP}’ constructions were excluded, since na is obligatory in this context. na in expressions like kabhı¯ na kabhı¯ ‘sometime or other’, and na(a) used to form tag questions were also excluded. Similarly, in counting occurrences of nahI¯, occurrences in the pattern V(a‰rmative) OR NOT (e.g. . . . karoge ya¯ nahI¯ ‘will you do it or not’) were excluded. 3. Diachronic considerations 3.1. Origin of nahI¯ Forms like Hindi and Urdu nahI¯, and Bangla ni are generally analyzed as amalgamations of the Old Indo-Aryan (OIA) negative marker na with some form of ‘be’; for example, Kellogg (1938: 281): ‘‘The common negative

Change in Progress: Negation in Hindi and Urdu

5

nahı¯n, Braj na¯hi has arisen from the combination of the negative na with the 3 rd singular a¯hi of the substantive verb.’’ It is also likely, as Turner (1931: 337) suggests, that this formation has been influenced by a combination of na and hı´ (EMPHATIC).4 Masica (1991: 393) concludes: ‘‘Quite possibly the nahı¯@ / na¯hı¯ forms owe something to both na þ hı´ and na þ present AUX, with verb-like features predominating in some languages (Marathi) and particle-like features in others (Hindi).’’5 Here I will call these two simultaneous and interacting origins of nahI¯: nahI¯-E (< na þ EMPH) and nahI¯-T (< na þ ‘be’ [PRESENT TENSE]). 3.2. Changing relative frequency of na and nahI¯ Simple raw counts of occurrences of na and nahI¯ in texts at varying time depths show a marked decline in the ratio of na to nahI¯. In the 16 th century Awadhi represented in Ja¯yasi’s Paduma¯vati (Dhar 1949), the ratio of occurrence of na to nahI¯ was 5.10, whereas in Premchand’s Goda¯n it was .85 and in the Jang corpus, which represents contemporary Pakistani newspaper Urdu, it is .24. Interestingly, in a corpus developed from the BBC Urdu website the overall ratio of na to nahI¯ is .32.6 This result, as well as di¤erences in ratios between approximately contemporaneous literary texts indicates that both genre and authorship are significant variables needing investigation. Counts of several recent short story texts indicate that there appear to be significant di¤erences in individual writers’ styles with regard to use of na and nahI¯. For example, in the story Tı¯srı¯ Kasam by Phanisvaranatha Renu the ratio na : nahI¯ is much lower than that of other contemporaneous writers.7 In fact, in this text, only 19 finite negative sentences are negated with na. Of the 19 occurrences of na in the story, 12 occur in the phrase na ja¯ne ‘who knows’, 4 in tags to elicit an a‰rmative response, 1 with an infinitive. In the only two finite sentences negated with na the finite verbs are sakna¯ and pa¯na¯ ‘to be able to, manage to’. Table 1 presents these results. 3.3. Negation cycles The cyclical evolution of negative markers has been recognized since Jespersen (1917: 4): ‘‘The history of negative expressions in various languages makes us witness the following curious fluctuation: the original negative adverb is first weakened, then found insu‰cient and therefore strengthened, generally through some additional word, and this in its turn may be felt as the negative proper and may then in course of time be subject to the same development as the original word.’’ Givon (1978) argues as follows that the tendency of negation to be associated with emphasis is inherent in the

6

Elena Bashir

Table 1. Comparative frequencies of na and nahI¯ in selected texts and corpora

Corpus Paduma¯vati (Awadhi) Braj Bhasha (Indrajit of Orcha¯)a Us ne kaha¯ tha¯ (C.S. Guleri) Goda¯n (Premchand) Toba Tek Singh (Sa¯dat Hasan Manto) Tisrı¯ Kasam (Phanisvaranatha Renu) Cı¯f kı¯ da¯vat (Bhisham Sahni) Va¯psı¯ (Usha Priyamvada) Hindi (Web dunia.com) – Martynyuk Jang – Martynyuk Jang – Bashir BBC Urdu News – Becker

(Approx.) date of lg. variety

Corpus size (number of tokens)

Freq. rank (1 ¼ Freq. ratio: highest): rank: raw n: raw n: na/ na nahI¯ na nahI¯ nahI¯

1540

—

—

—

158

31

5.10

ca. 1600

—

—

—

59

151

.39

1883–1922

4,285

29

13

24

39

.62

1880–1936 191,766 1912–1955 2,866

11 63

8 13

1,952 9

2,285 35

.85 .26

1921–1977 (1959)

3,003

30

2

19

122

.16

1915–2003

3,225

22

14

22

30

.73

mid 20th century 2002

6,695

31

22

47

47

1.00

@440,000

117b

18

350

3,003

.12

2002 2003 2003–4

@440,000 @187,500 685,574

55c 94

17 25 ?

609 218 1,106

2,499d 850 3,406

.24 .26 .32

?

a. McGregor (1968) b. Martynyuk (p.c.) c. The di¤erence in the rank of na in Martynyuk’s Hindi and Urdu corpora, both belonging to the same time period and both of the same genre, is striking. This deserves further investigation. d. Based on my count including variant forms of nahI¯ in Martynyuk’s corpus. Martynyuk (2003: 139) gives 2,479.

semantics of negation: since to negate something presupposes its possibility or existence, negation occurs in presuppositional contexts—hence the impetus to emphasize non-occurrence. This accounts for the tendency of negative markers to be augmented with emphatic markers, which are gradually bleached of their emphatic force, and then further augmented with another

Change in Progress: Negation in Hindi and Urdu

7

layer of emphasis. Croft (1991) identifies three stages in the development of negative markers in their relation to existential predicates, along with transitional phases, in which synchronic variation is found. His schema follows. A: Negation of existential predicate is performed by the general verbal negator. A@B: The general verbal negator negates the existential predicate; also a special negative existential predicate begins to be used. B: Special negative existential predicate, distinct from the verbal negator. B@C: The special negative existential form begins to be used for ordinary verbal negation. C: Special negative existential predicate, identical to the verbal negator. C@A: The verbal negator/negative existential is reanalysed as only a negator, and a regular positive existential verb comes to be used with it in the negative existential construction and as an auxiliary in negated presenttense sentences. Table 2 illustrates these developments in Hindi. Table 2. Negation cycle in the development of Modern Hindi and Urdu Stage

Structure(s)

Meaning

Form(s)

A B

NEG NEG NEG þ EMPH NEG þ AUX NEG NEG EXISTENTIAL NEG þ AUX

‘not’ ‘not’ ‘not at all’ ‘is not’ ‘not’; ‘no’ ‘is not’ ‘is not’

na na nahı´ (na þ hı´ ) nahI¯ (na þ a¯hi) nahI¯

C C@A

nahI¯ hai

The language varieties mentioned in this paper appear to illustrate various stages in the cycle: Awadhi (Paduma¯vati)-A; Bangla-B(?); Goda¯n-B@C; Jang corpus-C@A It is also possible that di¤erent registers of a language may be at di¤erent stages of the cycle. 4. Relevant recent research 4.1. Bangla negation Ramchand (2002) analyzes the two Bangla negation markers na and ni, which are in complementary distribution depending on the tense/aspect of the verb to be negated: ni only with the perfect tense, and na elsewhere. Examples (1)–(4) (Ramchand 2002: 3) illustrate the Bangla system.

8

Elena Bashir

(1)

ami am-Ta kheye-ch-i. I mango-CLASS eat(PERF)-PRES-1 ‘I have eaten the mango.’

(2)

*ami am-Ta kheye-ch-i na. I mango-CLASS eat(PERF)-PRES-1 NEG ‘I have not eaten the mango.’

(3)

ami am-Ta kha-i ni. I mango-CLASS eat-1 NEG ‘I have not eaten the mango.’

(4)

ami am-Ta khel-am na. I mango-CLASS eat(PAST)-1s NEG ‘I didn’t eat the mango.’

Ramchand argues that since sentences like (3) use a verbal form that is unmarked for tense or aspect, ni must be analyzed as indicating past tense and perfective aspect as well as negation.8 4.1.1. Temporal relative clauses. Ramchand presents examples showing that na is selected with subordinate temporal clauses, in which tense is indicated by the verb of the main clause (5). (5)

a.

b.

[jokhon Mary amTa khel-o na] tokhon When Mary the mango eat (PAST)-3 NEG then John khub rege gælo. John very angry got ‘When Mary didn’t eat the mango, then John got very angry.’ (2002: 3) *[jokhon Mary amTa kha-y ni] tokhon When Mary the mango eat-3 NEG then John khub rege gælo. John very angry got ‘When Mary didn’t eat the mango, then John got very angry.’

4.1.2. Duration adverbials. ni is consistently preferred over na with certain time adverbials (Ramchand 2002: 21). Two key examples are reproduced below as (6) and (7). (6)

a.

tin ghonTa dhore ami kichu boli three hours for I-NOM anything say-1

Change in Progress: Negation in Hindi and Urdu

9

ni NEG/PERF/PAST ‘For three hours I didn’t say a thing.’

(7)

b.

??tin ghonTa dhore ami kichu bollam three hours for I-NOM anything say(PAST)-1 na NEG ‘For three hours I didn’t say a thing.’

a.

kale gaRi ‘start’ hoi ni yesterday car start become-3 NEG/PERF/PAST ‘Yesterday the car didn’t start at all.’

b.

kale gaRi ‘start’ hol-o na yesterday car start become(PAST)-3 NEG ‘Yesterday the car didn’t start.’ (but then the neighbour came and fixed it).

Ramchand explains the di¤erence between (6a)/(6b) and (7a)/(7b) as follows: ‘‘If the event to be negated is claimed not to occur at every moment during the adverbial time span, then ni must be used. On the other hand, if the time span contains only one failure of the event to occur, or constitutes the whole span of time during which one instantiation of the event could have occurred (but didn’t), the na negation is felicitous’’ (2002: 21). In formal terms, the di¤erence between (6a)/(6b) and (7a)/(7b) is argued to result from ni binding a time variable, while na binds a situation variable. Reporting a perception of di¤erence of emphasis for Bengali like that I have found in pedagogical materials for Hindi and Urdu na and nahI¯, Ramchand concludes: ‘‘More generally, the feeling of emphatic-ness that speakers report for the ni form is related to this universal negation of all the times in the context.’’ (2002: 23)9 Given that Hindi and Urdu nahI¯ and Bangla ni share a partially similar historical origin, it will be appropriate to investigate some of the variables specifically discussed by Ramchand in connection with the Hindi and Urdu negatives. 4.2. Mirativity Montaut (2001: 351) argues that the Hindi simple perfective (consisting of the -(y)a¯ participle) is disjunct from the speaker’s present (moment of speech). This property results from the lack of a tense-marking auxiliary, which would anchor the reported event to the speaker’s reference time. Thus the nuance of mirativity, which refers to meanings which ‘‘are grasped

10

Elena Bashir

through a sudden irruption in the consciousness,’’ e.g. surprise or puzzlement, which are not (yet) anchored to the speaker’s mental world and time frame, emerges for the simple perfective form. Montaut gives examples such as (8a) and (8b), contrasting the meaning of surprise in the simple perfective (8a) with the matter-of-fact statement in (8b), which is rooted in the speaker’s (prior) connection to the event. (8)

a.

are ! kitna¯ baRa¯ ho gaya¯! interj. how much tall become-aor (Montaut’s morphemic analysis retained.) ‘Oh, he has grown so tall! / how tall he has grown!’ (Montaut 2001: 352)

b.

vah ka¯f ¯ı baRa¯ ho gaya¯ hai 3s fairly/rather tall become-perfect ‘He has grown quite tall.’

In fact, similar considerations seem to hold for the present progressive as well. Compare the following two examples: (9)

koı¯ uTha¯ nahI¯ raha¯ anyone lift NEG remain(PFV-ms) ‘No one is answering (the telephone).’ (Contemporary conversational Pakistani Urdu)

(10)

mehta¯ ne avisˇva¯s-bhare svar mE kaha¯ —yah a¯p dil se nahI¯ kah rahı¯ hAI mis ma¯latı¯ ! G-170-13 ‘Mehta said with a voice full of distrust, ‘‘You aren’t saying this from your heart, Miss Malati.’’ ’

In (9), the speaker is telephoning, expecting someone to answer. Thus non-answering is an unexpected development, while in (10) the speaker presumes prior knowledge of the addressee’s mental state. 5. Variables related to the distribution of na and nahI¯ Considering the historical origin of nahI¯, and the analyses in Givon (1978), Montaut (2001) and Ramchand (2002), I examine the occurrence of na and nahI¯ with respect to the following variables: (1) tense (since nahI¯ < na þ present tense of ‘be’); (2) occurrence with temporal relative clauses (Ramchand 2002); (3) temporal adverbials (Ramchand 2002); (4) emphasis: contrastive emphasis, presuppositional status (Givon 1978); (5) mirativity (Montaut 2001). In the course of this investigation, modal-like elements are seen to show unique behavior.

Change in Progress: Negation in Hindi and Urdu

11

5.1. Tense The Goda¯n data indicate a very strong preference for na in those tenses which include a past existential or past tense auxiliary (past imperfect and past perfect), and for nahI¯ with those tenses whose a‰rmative forms include a present tense auxiliary (present imperfect and present perfect). This is clear from Table 3. 5.1.1. Future. In the future, na is the default negator, occurring even in some contexts involving strong emphasis (11). (11)

yuvatı¯ ne pati kı¯ or ghu¯r-kar kaha¯ —mAI na ja¯U¯gı¯ , na j ja¯U¯gı¯ , na ja¯U¯gı¯ / puruS ne jaise alTimeTam diya¯ —na ja¯yagı¯? ‘‘na ja¯U¯gı¯’’ / ‘‘na ja¯U¯gı¯?’’ ‘‘na ja¯U¯gı¯’’ / G-138-51¤ ‘The young woman glared at her husband and said, ‘‘I won’t go, I won’t go, I won’t go.’’ The man gave a sort of ultimatum to her, ‘‘You won’t go?’’ (She replied) ‘‘I won’t go.’’ ‘‘You won’t go?’’ ‘‘I won’t go!’’

Most of the infrequent future tense sentences negated with nahI¯ have emphatic V NEG word order, e.g. (12). Some involve questions, with their inherent stress on the presupposed element (13, bold-face instances); and some involve clearly emphatic utterances (13, underlined elements).10 (12)

unko jo kuch kahna¯ hoga¯ mujhe kahEgı¯, tumse to bolEgı¯ bhı¯ nahI¯ / G-135-13 ‘She will say whatever she wants to say to me; she won’t speak to you at all.’

(13)

da¯ta¯dı¯na ne suratı¯ phA¯kte hue kaha¯ ‘‘ka¯m kaise nahI¯ karEge? sa¯l ke bı¯c mE ka¯m nahI¯ choR sakte / jeTh mE choRna¯ ho choR dE, karna¯ ho karE / uske pahle nahI¯ choR sakte’’ / gobar ne jamha¯¯ı lekar kaha¯, ‘‘unhOne tumha¯rı¯ gula¯mı¯ nahI¯ likhı¯ hai / jab tak iccha¯ thı¯, ka¯m kiya¯ / ab nahI¯ iccha¯ hai, nahı¯ karEge / ismE koı¯ zabardastı¯ nahI¯ kar sakta¯’’ / ‘‘to horı¯ ka¯m nahI¯ karEge?’’ ‘‘na¯!’’ G-221-32¤ ‘Tossing tobacco into his mouth, Datadin said, ‘‘What do you mean he won’t work? He can’t quit work in the middle of the year. If he wants to quit in (the month of ) Jeth, he can quit; he can’t quit before that.’’ Yawning, Gobar said, ‘‘He hasn’t signed up to be your slave. So long as he wanted to he worked; now he doesn’t want to (and) he won’t work. No one can force him to.’’ ‘‘So Hori won’t work?’’ ‘‘No!’’ ’

12 Elena Bashir

Table 3. Occurrence of na and nahI¯ by tense (Goda¯n and Jang)

Corpus

Imperfective [-AUX] na

Goda¯n (n) Ratio na/nahI¯ Jang(n) Ratio na/nahI¯

11

nahI¯

490 .02 0 348? 0

Present Imperfect [þAUX] na 0

nahI¯

na

3

135

0 0

? 0

Past Imperfect [þAUX] nahI¯

15 9.00 0 3 0

Perfective [-AUX] na 210

nahI¯

188 1.12 43 327 .13

Present Perfect [þAUX] na 0

nahI¯

na

10

60

0 0

63 0

Past Perfect [þAUX] nahI¯

8 7.50 1 16 .06

Future nahI¯

na 178

33 5.39 1 279 .004

Change in Progress: Negation in Hindi and Urdu

13

5.1.2. nahI¯ with past imperfect and perfect. In non-typical cases where nahI¯ appears with a past tense auxiliary or na is found with the unmarked imperfective there are clearly identifiable conditioning elements. Occurrences of nahI¯ together with a past-tense auxiliary show the contrastive emphatic function of nahI¯ (nahI¯-E) (14–16). (14)

jab tak bacca¯ tha¯, du¯dh pila¯ diya¯ / phir la¯va¯ris kı¯ tarah choR diya¯ / jo sabne kha¯ya¯, vahı¯ mAIne kha¯ya¯ mere lie du¯dh nahI¯ a¯ta¯ tha¯ , makkhan nahI¯ bAdha¯ tha¯ / G-228-24¤ ‘So long as (I) was a child (she) gave (me) milk (to drink). Then she left me like an orphan. Whatever everyone else ate, I ate the same thing. There was no milk or butter for me.’

(15)

‘‘kaha¯ rahe itne din?’’ gobar ne sˇarma¯te hue kaha¯ ‘‘kahI¯ du¯r nahI¯ gaya¯ tha¯ ammA¯, yah lakhnau¯ mE to tha¯’’ / G-210-04¤ ‘ ‘‘Where were you for so many days?’’ Gobar said bashfully, ‘‘I didn’t go anywhere far, Mother, I was here in Lucknow.’’ ’

(16)

kahI¯ naukar the lakhnau¯ mE?’’ gobar ne hekaRı¯ ke sa¯th kaha¯ ‘‘lakhnau¯ gula¯mı¯ karne nahI¯ gaya¯ tha¯.’’ / G-213-45¤ ‘ ‘‘Did you have a job in Lucknow?’’ Gobar said forcefully, ‘‘I didn’t go to Lucknow to be a slave!’’ ’

In (14) there is contrastive focus on milk’s not coming after the speaker grew up in comparison with how he was treated as a child—du¯dh pila¯ diya¯. In (15) du¯r ‘far’ is contrasted with yah lakhnau¯ mE to ‘here in Lucknow’; note also the contrasting particle to. In (16) gula¯mı¯ ‘slavery’ is contrasted with naukar ‘employee’. 5.1.3. na with unmarked imperfective. Most sentences which appear to be possible instances of the present imperfect negated with na are in fact negated simple imperfectives with the function of habitual, backgrounded action in past time: as in the proverb-like collocation marta¯ kya¯ na karta¯ ‘A dying person will do anything’ in (17), or as in (18). (17)

kahne ka¯ man to ca¯hta¯ hai marta¯ kya¯ na karta¯ ; lekinU¯kah ga¯ nahI¯ / G-155-02 ‘I feel like saying it; A dying person will do anything. But I won’t say it.’

(18)

gobar ko vahA¯ baRe sabere ja¯na¯ paRta¯ , aur din-bhar ke ba¯d jab vah diya¯-jale ghar lauTta¯, to uskı¯ deh mE zara¯ bhı¯ ja¯n na rahtı¯ / G-279-23

14

Elena Bashir ‘Gobar had to go there very early in the morning and when he would come home in the evening after a full day’s work there wasn’t a bit of strength left in his body.’

Next I eliminate the tense variable by comparing only sentences in the unmarked perfective. Since the kind of variation between na and nahI¯ that is my object of study here is found only with the verb sakna¯ ‘to be able’ in the Jang corpus, most of the analysis depends on the Goda¯n text. 5.1.4. Unmarked perfective. A fundamental consideration for this analysis is that an unmarked (i.e. [-AUX]) negated perfective can represent either an original simple perfective, or a negated present perfect with AUX omitted. In a negated present perfect, nahI¯ rather than na would appear because of the association of nahI¯ with present tense. Thus many of the sentences having an unmarked perfective with nahI¯, e.g. (19) and (20), are actually negated present perfects rather than simple perfectives. In addition, (19) and (20) are tag questions, which expect a negative answer; this is indicated by the sequence to nahI¯. (19)

mehta¯ ne pu¯cha¯ —a¯p ko zya¯da¯ coT to nahI¯ a¯yı¯ ? G-294-17 ‘Mehta asked, ‘‘You weren’t badly hurt, were you?’’ ’

(20)

kya¯ kahtı¯ hai tu¯? baura¯ to nahI¯ gayı¯ ? G-123-29¤ ‘What are you saying? You haven’t gone crazy, have you?’

Since the expected negative reply is be relevant to the present conversation, the sentence in (19) represents a negated present perfect. (20) is also a negated present perfect; note the explicit present tense in the preceding sentence. Since negated present perfects fall together with negated simple perfectives, in order to determine the number of true simple perfectives negated with nahI¯, one must analyze each sentence in context to determine whether it is a negated present perfect, or a negated simple perfective in which the presence of nahI¯ is due to some emphatic factor. As an illustrative exercise, some emphatic elements were counted for nahI¯-negated sentences with bolna¯ ‘to say, speak’ and kahna¯ ‘to say’. Of the 8 perfectives of bolna¯ negated with nahI¯, 7 have emphatic, inverted V NEG word order, forcing nahI¯, as in (21). Of these, 6 combine this with an emphatic element like kuch ‘any’, QUESTION, or an element indicating contrast like lekin ‘but’. Example (22), with uninverted word order, contains the emphatic elements kisı¯ se ‘to anyone’ and kuch ‘anything’.

Change in Progress: Negation in Hindi and Urdu

15

(21)

usne jhuniya¯ ko jaga¯ya¯ nahI¯ / kuch bola¯ bhı¯ nahI¯ / cupke se khicRı¯ tha¯lı¯ mE nika¯lı¯ G-281-07-09 ‘He didn’t wake Jhuniya, and he didn’t say anything either. Quietly he put some rice in a plate.’

(22)

mAI kisı¯ se kuch nahI¯ bola¯ / sabO ne ana¯ya¯s mujhe ma¯ra¯ / G-28732-33 ‘I didn’t say anything to anyone. They all beat me for no reason.’

Similar considerations apply to the perfectives of kahna¯ ‘to say’. Of the 21 perfectives negated with nahI¯, 8 have emphatic V NEG order, 8 have the (semi) lexicalized combination kuch nahI¯, and 5 are questions. Thus, though at first glance the counts for unmarked perfective would seem to indicate that na and nahI¯ are approximately equally likely, a closer examination of the sentences with nahI¯ indicates that in most cases they are actually negated present perfects (nahI¯-T). In other cases one of the indicators of emphasis—V NEG order, emphatic element, QUESTION—is found, indicating nahI¯-E. Thus for the simple perfective in Goda¯n, it is clear that na is the default choice, and nahI¯ does indeed indicate ‘something extra’. 5.2. Temporal relative clauses Examining jab ‘when (relative), since’ clauses with an unmarked perfective verb in Goda¯n, (cf. Ramchand’s example [5]), I found 8 examples of jab þ na, and 5 examples of jab þ nahI¯, all of which are negated present perfects. One example of each type follows. (23)

jab horı¯ ne na ma¯na¯ aur Tokrı¯ sir par rakhne laga¯ to dhaniya¯ ne donO ha¯thO se pu¯rı¯ sˇakti ke sa¯th Tokrı¯ pakaR lı¯ aur bolı¯ . . . G-132-40 ‘When Hori didn’t agree and started to put the basket on his head, Dhaniya grabbed the basket with both hands with full force and said . . .’

(24)

tumhE jab ¯ısˇvar ne buddhi nahI¯ dı¯ , to kyO mujh-se nahI¯ pu¯ch liya¯ ? G-243-09 ‘Since God hasn’t given you understanding, why didn’t you ask me?’

(24) is a negated present perfect with causal/conditional rather than temporal meaning of jab in the protasis and contrastive stress on a situation contrary to an expected norm. Note also the emphatic placement of kyO ‘why’ and the appearance of a negated compound verb in the apodosis.

16

Elena Bashir

A count of jab tak ‘until’ clauses yielded the following: 28 examples (2 imperfective, 17 subjunctive, 7 future, 2 perfective) of jab tak þ na; and 2 examples (1 present existential, 1 imperfective [a negated present imperfect]) of jab tak þ nahI¯. Since, with the exception of the two imperfective examples, the jab tak . . . na clauses involve conditional meaning, which requires na, there remain two non-conditional jab tak clauses with na and one with nahI¯. One example of each is given (25) and (26). (25)

use dekhte hı¯ uske pı¯che laga¯ ja¯ta¯ aur jab tak miTha¯¯ı na leta¯ , uska¯ pı¯cha¯ na choRta¯ / G-337-41¤ ‘As soon as he would see her he would follow after her and not stop following her until he got some sweets.’

(26)

jab tak sTej par sˇikSit abhinetriyA¯ nahI¯ a¯tI¯, hama¯rı¯ na¯Tya-kala¯ ka¯ uddha¯ra nahI¯ ho sakta¯ / G-056-19 ‘So long as the well-trained actresses don’t come onto the stage our drama can’t be salvaged.’

In (25) the imperfectives in both clauses background habitual action in past time. Thus (25) is similar to the case in (17) and (18) above. Example (26) is a negated present imperfect in the subordinate clause, with contrastive force on the non-arrival of the main actress. 5.3. Temporal adverbials Table 4 shows counts of several temporal adverbials in Goda¯n. Table 4. na and nahI¯ in unmarked perfective sentences with various temporal adverbials in Goda¯n Temporal adverbial

na

nahI¯

kabhı¯ ‘ever’ Abhı¯ tak ‘up to now’ Abhı¯ (excluding abhı¯ tak) ‘right now’ tak (excluding abhı¯ tak) ‘up to’

12 1 2 7

37* 9 19 17

* Of these 37, 5 are in the immediately contiguous (lexicalized) collocation kabhı¯ nahI¯ ‘never’. It should also be noted that kabhı¯ and abhı¯ tak are also emphatic elements.

5.3.1. abhı¯ tak. I found a significant di¤erence in the proportion of occurrences of na and nahI¯ with abhı¯ tak ‘not yet’ < ‘up to now-EMPH’. As opposed to 9 occurrences of abhı¯ tak with nahI¯ there was only 1 with abhı¯ tak and na. Sentences involving abhı¯ tak negated with nahI¯ are negated

Change in Progress: Negation in Hindi and Urdu

17

present perfects. An example with nahI¯ is given as (27), and the single example with na is reproduced as (28). (27)

siliya¯ ke pa¯s a¯ kar bolı¯ —kyO rı¯ siliya¯, mahı¯na¯-bhar raMg la¯ye ho gaya¯, abhı¯ tak paise nahI¯ diye / G 251-19 She came to Siliya and said, ‘So, Siliya, it’s been a full month since you brought the dye (and) you haven’t yet given (me) the money.’

(28)

sˇugar ke hisse abhı¯ tak a¯p ne na xarı¯de / ab bahut thoRe-se hisse bac rahe hAI / ha¯th malte rah ja¯iega¯ / insˇyorens kı¯ pa¯lisı¯ bhı¯ a¯p ne na lı¯ / G-146-05¤ ‘You didn’t yet buy sugar shares. Now there are only a few left. You will regret (not buying any). You didn’t buy an insurance policy either.’

In (28) na appears with abhı¯ tak. Here, with tak, the entire interval is being viewed as in the past, externally, in a point-like manner. This is an instance of Ramchand’s (2002: 21) insight that ‘‘if the time span contains only one failure of the event to occur, or constitutes the whole span of time during which one instantiation of the event could have occurred (but didn’t), the na negation is felicitous’’ (see [7a] and [7b] above). In order to further investigate the temporal factors discussed by Ramchand for Bangla, I examined occurrences of temporal adverbials of duration other than abhı¯ tak, with both na and nahI¯. A selection of such sentences follows. 5.3.2. Other time adverbials—with nahI¯. The time adverbials in examples (29)–(36) refer to continuous intervals of time during which no instance of the event occurred, and thus fit Ramchand’s criterion: ‘‘If the event to be negated is claimed not to occur at every moment during the adverbial time span, then ni must be used.’’ (Ramchand 2002: 21). (29)

sˇobha¯ zamı¯n par leTa¯ hua¯ bola¯ —mAI to maha¯ra¯j a¯Th din se ba¯har nahI¯ nikala¯ / G-114-02 ‘Lying on the ground, Shobha said, ‘‘Sir, I haven’t left the house for eight days.’’ ’

(30)

govindı¯ ne a¯grah kiya¯ —tı¯n din se nahI¯ utara¯ / G-193-10 ‘Govindi was insistent, (and) didn’t relent for three days.’

(31)

ra¯y sa¯hab apne is ila¯qe mE bahut dinO se nahI¯ gaye the / udha¯r mA¯gne cale hAI, tı¯n sa¯l se dhela¯ su¯d nahI¯ diya¯, us par udha¯r diye ja¯o / G-080-14 ‘The Rai Sahib hadn’t gone to this area for a long time. He is going to ask for a loan; he hasn’t paid even half a paisa interest; go on giving him loans on top of it.’

18

Elena Bashir

In (29) we have a negated present perfect. Also, each moment of the eight days is presented as a possible moment of occurrence of leaving the house. Similarly, in (30) each moment of the eight days is a potential moment for relenting. In (31) the presence of the time adverbial of duration bahut dinO se ‘for many days’ appears to override the presence of the past tense auxiliary. A similar adverbial tı¯n sa¯l se ‘for three years’ combined with a negated present perfect elicits nahI¯ in the second clause. Examples (32)–(36) are similar in that the time interval is presented as multiple potential times for an event to occur. (32)

do sa¯l se ek dhela¯ su¯d nahI¯ diya¯, paca¯s to mere su¯d ke hote hAI / G-186-16 ‘For two years (you) haven’t given a half a paisa of interest; my interest (alone) is fifty (rupees).’

(33)

itne dinO mE usne ek paisa¯ bhı¯ ghar nahI¯ bheja¯ / G-203-24 ‘In so many days he didn’t send a single paisa home.’

(34)

jı¯van mE aisa¯ to koı¯ din hı¯ nahI¯ a¯ya¯ ki laga¯n aur maha¯jan ko dekar kabhı¯ kuch baca¯ ho / G-122-07 ‘During his life no day ever came when having paid the land rent and made payment to the moneylender there was anything left over.’

(35)

horı¯ ko ra¯t bhar nI¯d nahI¯ a¯yı¯ / G-030-18 ‘Hori couldn’t sleep the whole night.’

(36)

itne dinO ba¯t nahI¯ pu¯chı¯, to sa¯s ka¯ pA¯v daba¯ne ke lie tel lie dauRtı¯ thı¯ / G-229-35 ‘For so many days she didn’t ask anything and she was running with oil to massage her mother-in-law’s feet.’

Note that the atypical negation of the past perfect gaye the ‘went’ with nahI¯ (31) occurs with the adverbial bahut dinO se ‘for many days’ and also that the only instance of a past perfect of dena¯ ‘to give’ with nahI¯ in Goda¯n occurs with the adverbial, kaı¯ mahı¯ne se ‘for many months’ (37). Thus it appears that a specific time adverbial or an emphatic factor can override the tense variable. In (37) both the duration adverbial with se (> nahI¯-T) and a strong presupposition or expectation that the rent is to be paid (> nahI¯-E) are present. (37)

idhar kaı¯ mahı¯ne se maka¯n ka¯ kira¯ya¯ nahI¯ diya¯ tha¯ / G-334-17 ‘But he hadn’t given the house rent for several months.’

Change in Progress: Negation in Hindi and Urdu

19

Not all adverbials indicating duration, however, necessarily co-occur with nahI¯. The following sentences negated with na also contain duration adverbials. 5.3.3. Other time adverbials—with na. (38)

a¯Th din tak uske ha¯th sı¯dhe na ho sake / G-348-39 ‘For eight days his hands couldn’t straighten out.’

(39)

jab cauthe din sA¯jh tak bhı¯ jhuniya¯ ke du¯dh na utara¯ , to cuhiya¯ ghabra¯yı¯ / G-282-29 ‘When by evening of the fourth day Jhuniya¯’s milk had not come, Cuhiya got worried.’

(40)

abhı¯ ha¯l mE itna¯ ma¯ra¯ tha¯ ki vah kaı¯ din tak kha¯T se na uTh sakı¯ / G-33-26 ‘Now she was so weak that for many days she couldn’t get up from her bed.’

(41)

usne do sa¯l tak ek pa¯¯ı na dı¯ / G-179-13 ‘He didn’t give a single pai for two years.’

(42)

vah us se phir a¯ne ka¯ va¯da¯ kar ke kabhı¯ us ke pa¯s na gaye / G-330-10 ‘Having promised to come again, he never went to see him.’

(43)

laRka¯ pardes kama¯ne laga¯, ek din nevta¯ bhı¯ na khila¯ya¯ . . . G-248-12 ‘The boy started to earn away from home, he didn’t even extend one invitation . . .’

Note that in 5 of the 9 examples (29)–(37) with nahI¯, the time adverbial involves the postposition se ‘from’, in 2 of them the postposition mE ‘in’, and in 2 of them an oblique without explicit postposition. However, of the 6 examples (38)–(43) with na, 4 include the postposition tak ‘up to’, one has kabhı¯ ‘ever’, and one has an unmarked oblique. The correlation of se with nahı¯, and tak with na is consistent with an analysis that with se and mE the interval is viewed from an internal perspective, as consisting of multiple sub-intervals, and nahI¯ tends to occur, whereas when the interval is viewed from an extermal perspective, as a point, as indicated by tak, na occurs. This is also consistent with Ramchand’s analysis (see 7 above). 5.4. Emphatic elements Co-occurrences of na and nahI¯ with various emphatic elements other than time adverbials in unmarked perfective sentences were also counted (see Table 5).

20

Elena Bashir

Table 5. na and nahI¯ in perfective sentences with emphatic elements in Goda¯n Emphatic element11

na

nahI¯

koı¯ / kisı¯ ‘any’ kuch bhı¯ ‘any at all’ zara¯ bhı¯ ‘even a bit’ V tak ‘even’ NP tak ‘even’

20 0 3 0 6

25 3 4 5 1

5.4.1. koı¯ / kisı¯. The distribution of na and nahI¯ in finite simple perfective clauses including the elements koı¯ or kisı¯ is approximately equal. Representative sentences of each type (44) and (45) illustrate some variables associated with the choice between na and nahI¯. (44)

ham ba¯qı¯ cuka¯ne ko paca¯s rupae mA¯gte the, kisı¯ ne na diya¯ / G-118-34 ‘We asked for twenty-five rupees to pay o¤ the debt, but no one gave (it to us).’

(45)

dhaniya¯ ghar mE akelı¯ hı¯ rahtı¯ thı¯; lekin kabhı¯ kisı¯ ne use kisı¯ chaila¯ kı¯ or ta¯kte nahI¯ dekha¯ / G-300-40 ‘Dhaniya stayed alone in the house, but no one ever saw her looking at any handsome youth.’

In (44), with na, the event of asking for money is restricted to an instance of asking for a specific amount at a specific time, at which time no one gave the requested amount. In (45), with nahI¯, note the additional emphatic elements hı¯ and kabhı¯ ‘ever’. 5.4.2. zara¯ bhı¯. Since the meaning of zara¯ bhı¯ is ‘even a bit/at all’, one might expect that if emphasis alone selects nahI¯ over na, there should be a significant preponderance of nahI¯ over na with zara¯ bhı¯. However, this is not the case here; as seen in Table 5, there were 3 occurrences with na and 4 with nahI¯. Comparable unmarked perfective sentences for zara¯ bhı¯ with na and nahI¯ are shown as (46)–(49). (46)

gobar zara¯ bhı¯ garam na hua¯ , kisı¯ praka¯r ka¯ roS uske mUh par na tha¯ / G-360-02 ‘Gobar didn’t get angry at all; there was no kind of anger at all on his face.’ [zara¯ bhı¯ þ adjective]

(47)

mehta¯ zara¯ bhı¯ nirutsa¯h na hue / G-240-15 ‘Mehta didn’t get at all disheartened.’ [zara¯ bhı¯ þ adjective]

Change in Progress: Negation in Hindi and Urdu

21

(48)

mirza¯ sa¯hab ne unkı¯ or hika¯rat se dekh kar kaha¯ —mAI aise rupae par aur a¯p par la¯nat bhejta¯ hU¯ / misTar taMkha¯ ne zara¯ bhı¯ bura¯ nahI¯ ma¯na¯ / ma¯the par bal tak na a¯ne diya¯ / G-101-01 ‘Mirza Sahib looked at him scornfully and said, ‘‘I curse such money and I curse you.’’ Mister Tankha didn’t mind (this) at all; he didn’t turn a hair. [Contrastive focus on reaction unexpected in the context; zara¯ bhı¯ þ V, nahI¯-E.]

(49)

a¯Th din tak uske ha¯th sı¯dhe na ho sake / us din vah zara¯ bhı¯ nahI¯ laja¯ya¯ , zara¯ bhı¯ nahI¯ jhijhaka¯ / G-348-39¤ ‘For eight days his hands couldn’t straighten out. On that day he didn’t feel shy and didn’t hesitate a bit.’ [zara¯ bhı¯ þ V; contrastive stress on us din compared to a¯Th din tak; nahI¯-E.]

Note that in (46) and (47), with na, zara¯ bhı¯ modifies an adjective, while in (48) and (49), with nahI¯, it modifies the verb. 5.4.3. NP þ tak. The behavior of negation in sentences with tak plus NPs referring to present time (50) and those with spatial (51) or other reference (48 above) di¤ers. (50)

laRkı¯ to hama¯rı¯ bira¯darı¯ mE a¯j tak kabhı¯ kUa¯rı¯ nahI¯ rahı¯ / G-129-33 ‘To this day no girl in our community has ever remained unmarried.’

(51)

yah kahne ke sa¯th vah ret mE dauRe aur bandu¯q kina¯re par rakha¯ gaRa¯p se pa¯nı¯ mE ku¯d paRe aur baha¯v kı¯ or tairne lage; magar [a¯dha¯ mı¯l tak pu¯ra¯ zor laga¯ne par bhı¯] ciRiya¯ na pa¯ sake / G-085-19 ‘While saying this he ran into the sand and putting his rifle on the shore, jumped into the water with a splash and began to swim in the direction of the current; but in spite of exerting full e¤ort for a half a mile, he wasn’t able to find the bird.’

In (50) the present-time oriented duration adverbial a¯j tak ‘until today’ points to this sentence being a negated present perfect. Additionally, the emphatic kabhı¯ ‘at any time’ includes all instants of the interval a¯j tak as possible moments for occurrence. Hence negation with nahI¯. In (51), full e¤ort toward a desired outcome and contrastive nonrealization of that outcome are emphasized. Still, na rather than nahI¯ occurs, perhaps because the e¤ort exerted over an extended distance is conceptualized as a point-like entity (note the postposition par ‘on’). Note also

22

Elena Bashir

that this sentence involves sakna¯ ‘to be able’, which we will see preserves na even in the Jang corpus. 5.4.4. V þ tak. This construction consists of the emphatic order [V tak] nahI¯. In two of the five examples, emphasis is reinforced by kabhı¯ ‘never’. Example (52) gives one instance. (52)

da¯da¯ itne acche hAI ki kabhı¯ mujhe DA¯Ta¯ tak nahI¯ / G-357-02 ‘Grandfather is so nice that he has never so much as scolded me.’

Examination of emphatic words and particles and adverbial expressions indicates that several variables influence the choice of na or nahI¯. (a) time adverbial, interval considered internally, multiple potential occurrences focussed on > nahI¯; (b) present-time oriented adverbial > nahI¯; (c) time adverbial, interval viewed externally as point-like whole > na; (d) presence of sakna¯ > na; (e) NP tak > na; (f ) V tak > nahI¯. 5.5. Mirativity Accepting Montaut’s argument regarding the e¤ects of the presence or absence of the present auxiliary, and given that the historical di¤erence between na and nahI¯ is the presence of an incorporated present-tense form of ‘be’ in nahI¯ it seems likely that the di¤erence between na and nahI¯ should be similar to the di¤erence between the simple perfective and the present (or past) perfect. In fact, this does seem to a certain extent to be the case (in contemporary Pakistani Urdu). For example, compare the following two typical utterances. (53)

a. b.

kal ga¯Rı¯ xara¯b ho gayı¯ to mAI na a¯ saka¯. ‘Yesterday (my) car broke down so I couldn’t come. kal mujhe DakTar ke pa¯s ja¯na¯ tha¯; is lie nahI¯ a¯ saka¯. ‘Yesterday I had to go to the doctor; that’s why I couldn’t come.’

In (53-a), the speaker explains that his inability to come was the result of an unforeseen circumstance, while in (53-b) the speaker knew that he had a doctor’s appointment and wouldn’t be able to come. Although Goda¯n yields examples with sakna¯ in which na indicates involuntary action, e.g. (54), given that na is the default negative marker in the perfective, I was not able to associate a unique mirative sense with na. (54)

govindı¯ apnı¯ hAsı¯ na rok sakı¯ / G-197-32 ‘Govindi couldn’t stop laughing.’

Change in Progress: Negation in Hindi and Urdu

23

Since in the contemporary language unmarked perfective forms of hona¯ ‘to become’ often occur in mirative contexts, e.g. kya¯ hua¯?! ‘What happened?’, in looking further for possible association of mirative meaning with the simple perfective in the language of Goda¯n, I counted negated unmarked perfective sentences with hona¯ to see whether there is a correlation of na or nahI¯ with simple perfectives of hona¯. I found 33 instances of na and 21 of nahI¯. I reproduce for comparison one each of the negated perfective sentences with na and nahI¯ (55 and 56). (55)

us ke da¯da¯ kı¯ yah la¯lsa¯ kabhı¯ pu¯rı¯ na huı¯ / G-361-09 ‘This longing of his grandfather’s was never fulfilled.’

(56)

a¯j tak kul mE kabhı¯ aisa¯ nahI¯ hua¯ / G-353-10 ‘Until today nothing like this has ever happened in (our) family.’

(55) (cf. 28 above) is another instance of Ramchand’s generalization (2002: 21) that ‘‘if the time span contains only one failure of the event to occur, or constitutes the whole span of time during which one instantiation of the event could have occurred (but didn’t), na is felicitous.’’ In (56) the presenttime oriented interval adverbial a¯j tak ‘until today’ points to a negated present perfect. Examining the sentences with na and simple perfective of hona¯, we find a mixed situation. Some of them have elements of contrast, while others do not. Some negate events that are located in a causal chain; some seem to have a weak mirative sense. Consistent with the status of na as default negator in the simple perfective, no generalization applies to all of them. Looking at sentences with nahI¯, however, the picture is clear: most of these examples are negated present perfects. 5.6. Modal-like forms 5.6.1. sakna¯ ‘to be able’. In the Jang corpus, the preponderance of nahI¯ is overwhelming, with the single exception of the simple perfective of sakna¯. In Jang, 42 examples of unmarked negated perfectives with sakna¯ occur with na and 70 with nahI¯, whereas in Goda¯n, the ratio is 60 na to 6 nahI¯. In an attempt to discover why na has persisted with the perfective of sakna¯, I looked at negated indicative sentences in the Jang corpus with NEG þ kar- sak- (PFV). There are 9 such sentences negated with na and 8 with nahI¯. Focussing on the unmarked perfective sentences in na and nahI¯, I reproduce one of each type here for comparison (57 and 58).12 (57)

sadar ne kaha¯ ki woh referendum se jo nata¯ij ha¯sil karna¯ ca¯hte the, who badqismatı¯ se ba¯d ke ha¯la¯t ke ba¯is ha¯sil na kar sake.

24

Elena Bashir ‘The president said that unfortunately, because of circumstances developing afterwards, he wasn’t able to get the results he had wanted from the referendum.’

(58)

loya¯ jirga¯ jangju¯ kama¯nDarO se alag authority qa¯im nahI¯ kar saka¯ The Loya Jirga has not been able to establish an authority separate from the warlords.’ [Followed by the context:] International Crisis Group ne kaha¯ hai ki loya¯ jirga¯ jangju¯ koma¯nDarO se alag ek authority qa¯im karne, civil qaya¯dat paida¯ karne aur jamhu¯rı¯ ta¯sur paida¯ karne mE na¯ka¯m ho gaya¯ hai. ‘The International Crisis Group has said that the Loya Jirga has been unsuccessful in establishing an authority separate from the warlords, establishing a civilian leadership, and creating a democratic impression.’

In (57) the inability is due to unavoidable circumstances. Perhaps this, in that the expected results were not achieved, can be considered a weak form of mirativity. In (58), however, we clearly have a negated present perfect. The context following the example sentence makes present time reference explicit. One observation that can be made about this Jang corpus is that the simple perfectives of kar sakna¯ negated with na, state a fact either objectively or with a slight mirative nuance. If mention of context is given, it tends to stress the unavoidability of the inability. The e¤ect of this is that the collocation na kar saka¯, having been bleached of some of its potential volitionality, means something between ‘didn’t do’ (unavoidably, unfortunately, unexpectedly) and ‘wasn’t able to do’. Perhaps because not being able to do something is often unanticipated and due to circumstances beyond the agent’s control, na persists in the perfective with sakna¯. Those perfectives negated with nahI¯, on the other hand, typically state a context in which the inability was contrary to expectations and despite volitional e¤orts. However, since most of the examples of nahI¯ kar sak-PFV are negated present perfects, this may be su‰cient reason for the occurrence of nahI¯. These tendencies toward present perfect tense and volitional, non-mirative contexts, and toward non-volitionality and mirativity to co-occur are consistent with Montaut’s analysis of the present perfect compared to the simple perfective in contemporary Hindi. 5.6.2. ca¯hiye ‘to be needed/wanted’. It is striking that in Goda¯n, ca¯hiye ‘to be needed/wanted’ is always negated with na, while in the Jang corpus it is always negated with nahI¯ (see Table 6). Possibly this is a legacy of the origin of ca¯hiye in a passive formation. The use of na and nahI¯ in passive vis-a`-

Change in Progress: Negation in Hindi and Urdu

25

vis non-passive sentences in Goda¯n was not examined here; this might shed light on this question. Table 6. Negation of ca¯hiye (Goda¯n and Jang) nahI¯

na

Goda¯n Jang

Present

Past

Present

Past

11 0

4 0

0 20

0 1

6. Conclusions The origin of nahı¯ in NEG þ ‘be’ (PRES) has led to several developments: (1) Absence of the present AUX in negated present existential, imperfect and perfect sentences, and sometimes in present progressives (Goda¯n and Jang). (2) Preference for na with verb forms including the past AUX, and the converse association of nahı¯ with present-marked forms (Goda¯n). (3) Preference for nahI¯ with certain duration adverbials (Ramchand 2002 for Bangla; also observed in Goda¯n). (4) Mirative e¤ects associated with na (observed in Jang and contemporary spoken Urdu). I did not find a clear mirative meaning associated with na in Goda¯n, while it seems to be present to some degree in simple perfectives of sakna¯ negated with na in the Jang corpus and in contemporary spoken Urdu. Since in the Goda¯n language na was the default choice for negating the simple perfective; it was nahI¯ which conveyed additional information (e.g. emphasis). It appears that, with the loss of its unmarked/default status, na, where it appears in finite indicative sentences, is specializing to some degree in the contemporary language into the negative marker associated with mirativity or non-volitionality. 6.1. Bangla situation contrasted with Hindi and Urdu The situation with regard to distribution of the Bangla negatives is clear, ni having a clearly defined and restricted occurrence. In contemporary Hindi and Urdu, however, nahI¯ occurs with present and past imperfective, present and past progressive, present and past perfect, simple perfective and future, i.e. all tense/aspect forms except the subjunctive and imperative (although there is a tendency among some speakers now to use nahI¯ even with imperatives, especially with emphatic word order, e.g. karo nahI¯ ‘don’t do it’). nahI¯ is, in fact, now the default negative particle in contemporary Hindi and Urdu.

26

Elena Bashir

In the language of Goda¯n a significant number of the contrasts between na and nahI¯ in the simple perfective do seem to follow the predictions of Ramchand’s analysis of Bangla, but there is a residue which remains to be explained. It seems clear that in the language of Goda¯n the two homophonous nahI¯s—nahI¯-E (< na-hı´, NEG þ EMPH), and nahI¯-T (< NEG þ present tense of ‘be’)—are identifiable. This partially accounts for the complex picture of the use of na and nahI¯ in the language of that period (70–80 years ago). In a text like Goda¯n, in which na is the default, it is possible in many cases to tease apart the distribution e¤ects due to nahI¯-E and those due to nahI¯-T. For instance, with the future, where na is the default negator, and the question of an omitted AUX is not involved, the presence of nahI¯, as in (12) and (13), can be accounted for as an e¤ect of nahI¯-E. Similarly, in the exceptional examples of nahI¯ with the past perfect (15), (16), (31), (37) or past imperfect (14), nahI¯ is nahI¯-E. However, in the case of negated present imperfects and present perfects, both of which normally delete the AUX, we have nahI¯-T, associated with present tense. When the present-tense AUX does appear together with nahI¯ in these tenses, we are seeing a new accretion of a present-tense bearing element, which has the potential to add emphatic meaning. In contemporary Hindi and Urdu, to the extent that variation still exists, the analysis underlying nahI¯-E is replacing/has replaced that of nahI¯-T. One reason for the the present tendencies (1) to include the present auxiliary in negated present imperfects and present perfects, and (2) to include a present-tense form of hona¯ ‘be’ after nahI¯ in the negative existential sense may be the overall diachronic increase in frequency of nahI¯ over na, which has led to reanalysis of nahI¯ as the default negative marker, freed from its linkage to present tense. The widespread bilingualism of Hindi and Urdu speakers in English may be exerting additional pressure in this direction. Thus nahI¯ is now analyzed as a single semantic unit, i.e. NEG, and is ripe for the accretion of an additional AUX element, as in Stage C > A of the negation cycle. Notes * This paper is a revised version of a talk presented at the 23 rd annual meeting of the South Asian Languages Analysis Roundtable (SALA 23), at the University of Texas, Austin, October 10–12, 2003. 1. In this paper, nasalized vowels are represented with upper-case letters. Retroflex consonants are represented with upper-case letters. Abbreviations occurring in this paper are: 1—first person, 2—second person, 3—third person, aor—aorist, AUX—auxiliary, CLASS—classifier, m—masculine, NOM—nominative, NEG—negative, p—plural, PFV—perfective, PERF—perfect, PRES—present, s—singular.

Change in Progress: Negation in Hindi and Urdu

27

2. For this study I used a Roman transcription of Premchand’s Goda¯n prepared at the Institute for the Study of Languages and Cultures of Asia and Africa in Tokyo under the direction of Prof. T. Nara, refined by Prof. K. Machida, and further corrected by P.E. Hook. I am grateful to Peter Hook for providing me the electronic file of the transcription. Examples from this text are cited as follows: G-xxx-yy, where G indicates Goda¯n, xxx the page number, and yy the sentence number. Page numbers refer to the 1965–6 edition published in Ilahabad by Sarasvati Publishers, but probably apply also to other editions by the same publisher. 3. Roman Urdu news stories are from the Daily Jang (May–November, 2002), which represents contemporary Pakistani newspaper Urdu. The idea for using these texts as an easily accessible source of electronically searchable Roman Urdu texts is due to Stanislav Martynyuk, who utilized this source in his article on the contemporary frequency of Urdu and Hindi words (Martynyuk 2003). In addition, I am indebted to him for allowing me to use the corpus as prepared by him for analysis in this paper. For counting various forms in context I used the Simple Concordance Program, Version 4.0.5. 4. Monier-Williams ([1899] 1984: 523): ‘‘na ‘ ind. not, no, nor, neither . . . (as well in simple negation as in wishing, requesting and commanding . . . ; in successive sentences or clauses either simply repeated, . . . or strengthened by another particle, esp. at the second place or further on in the sentence. . . . Among these strengthened uses given is na-hı´ (1984: 532).’’ The entry for na-hı´ gives: ‘‘na´-hı´, (MaitrS. and S´Br.) ind. for not, surely not, by no means, not at all (often strengthened by other particles) . . .’’ The entry for hı´ (1984: 1297) includes the collocations with na ‘for; indeed, assuredly . . .’, ‘. . . na hı´ or na hı¯ ‘ ‘for not; not at all’, RV etc.’ 5. Two negation markers are also found in other New Indo-Aryan (NIA) languages: Hindi, Panjabi, Gujarati, Marathi, Orissa, Bangla, Nepali (Masica 1991: 392). Schwarzschild (1959) is a detailed treatment of these developments. 6. This Unicode-encoded xml corpus was developed by Dara Becker, who has kindly granted me permission for its use in academic research. 7. Romanized searchable texts of the stories Toba Tek Singh, cˇ¯ıf kı¯ da¯vat, Va¯psı¯, and Us ne kaha¯ tha¯ were found at the web site http://www.abhivyakti-hindi.org/ gauravgatha/.htm. I am grateful to Alice Davison for providing me with a searchable Romanized text of Tı¯srı¯ Kasam. 8. Basu (1976) is a detailed discussion of the historical origins of ni (with its [dialectal] variants na¯i and nei). In this article, Basu says that na¯i is derived from na¯sit ‘there was not’ (1976: 80). Thus the origins of Bangla ni and Hindi nahI¯ are di¤erent in that Bangla ni incorporates a past-tense form of ‘be’, whereas Hindi nahI¯ incorporates a present-tense form of ‘be’. Ramchand’s (2002) analysis appears consistent with this historical origin. I thank Probal Dasgupta for pointing me to the work of Basu. 9. The nuance of unwillingness or refusal associated with negation of some verbs in the present imperfect (Sharma 1958: 124) is consistent with this, in that if nahI¯ implies the consideration of every moment in the potential time interval, the sense of ‘don’t ever’ or ‘don’t want to’ develops. Sharma’s examples are (Romanized and normalized, but retaining Sharma’s glosses):

28

Elena Bashir (i)

mAI yeh bojh nahI¯ uTha¯-t-a¯ I this load NEG lift-IMPFV-ms ‘I won’t lift this load.’

(ii)

mAI ba¯za¯r nahI¯ ja¯-t-a¯ I baza¯r NEG go-IMPFV-ms ‘I don’t like to go to the market.

(iii)

yeh ghoRa¯ nahI¯ cal-t-a¯ this horse NEG move-IMPFV-ms ‘This horse won’t move.’

10. See also (17) below, in which denial of a future action is explicitly contrasted with desire to do it. 11. In the entire Goda¯n corpus there was only one occurrence of kahI¯ bhı¯ ‘anywhere’, which occurred in a negated present imperfect sentence with nahI¯. 12. The spelling of transcribed examples from Jang has been normalized. English words and names used as such in the text are given in their English spelling, as was done in the Romanized Jang corpus.

References Barker, Muhammad Abd-al Rahman 1967 A Course in Urdu, Volumes I, II, III. Montreal: Institute of Islamic Studies, McGill University. Basu, Dwijendranath 1976 Negative Auxiliary in Bengali. In Indian Linguistic Researches: A Pentalogue. Calcutta: Basudha. Bhatia, Tej K. 1973 On the Scope of Negation in Hindi. Studies in the Linguistic Sciences, 3: 1–27. Croft, William 1991 The Evolution of Negation. Journal of Linguistics 27: 1–27. Dhar, Lakshmi 1949 Paduma¯vatı¯: A Linguistic Study of the 16 th Century Hindi (Avadhi). London: Luzac & Co., Ltd. Givon, Talmy 1978 Negation in Language: Pragmatics, Function, Ontology. Syntax and Semantics, 9. Pragmatics, Peter Cole (ed.), 69–102. New York: Academic Press. Jespersen, Otto 1917 Negation in English and Other Languages. Copenhagen: Royal Danish Society of Science and Letters. Kellogg, S. H. 1938 A Grammar of the Hindi Language. London: Kegan Paul, Trench Trubner & Co., Ltd. Martynyuk, Stanislav 2003 Statistical Approach to the Debate on Urdu and Hindi. Annual of Urdu Studies, 18 (1): 132–144. Masica, Colin P. 1991 The Indo-Aryan Languages. Cambridge: Cambridge University Press.

Change in Progress: Negation in Hindi and Urdu

29

Monier-Williams, M. 1984 A Sanskrit-English Dictionary. Delhi: Motilal Banarsidass. Original edition, Oxford, Oxford University Press, 1899. Montaut, Annie 2001 On the Aoristic Behaviour of the Hindi/Urdu Simple Past: From Aorist to Evidenciality. In Tohfa-e-Dil: Festschrift Helmut Nespital, D. Loenne (ed.), 345–364. Reinbek: Wezler. Premchand 1965–6 Goda¯n. Ilahabad: Sarasvati Press. Ramchand, Gillian C. 2002 Two Types of Negation in Bengali (draft, May 2002) http://users.ox.ac.uk/~gcram/ research/ (downloaded May 2003). Schwarzschild, L. A. 1959 Some Aspects of the History of Modern Hindi nahıˆn ‘‘no’’, ‘‘not’’. Journal of the Royal Asiatic Society 1959: 44–50. Sharma, Aryendra 1958 A Basic Grammar of Modern Hindi. New Delhi: Govt. of India, Ministry of Education and Scientific Research. Turner, Ralph Lilley 1931 Nepali Dictionary. London: Routledge and Kegan Paul.

The Overt Licensing of NPIs in Hindi* ELABBAS BENMAMOUN and RAJESH KUMAR

This paper provides an analysis of the phenomena of Negative Polarity Item (NPI hereafter) licensing in Hindi and argues for overt licensing. Based on the occurrence of NPIs in di¤erent syntactic positions and reconstruction effects, we argue for the licensing of NPIs in the course of the derivation prior to the mapping to the phonetic component/PF. 0. Introduction On par with structural Case and Binding, one issue that arises in the context of Negative Polarity Licensing concerns the level of representation (Sstructure or LF) or point in the derivation (before or after Spell-out, i.e., overtly or covertly, using minimalist terms) where licensing takes place (Linebarger 1987, Ladusaw 1992, Progovac 1994, Uribe-Echevarria 1994, Haegeman 1995, Benmamoun 1996, among others). In this paper, we provide empirical evidence based on the syntactic distribution of NPIs, and specifically the NPI ek bhii in Hindi, following the characterization given in Mahajan (1990), Lahiri (1998, 2001), Vasishth (1999) and Kumar (2003), to argue for overt licensing of NPIs. We will provide new evidence based on reconstruction e¤ects to argue for overt licensing under c-command by (and/or possible Spec-head relation with) sentential negation. We will also re-analyze evidence that has been put forward to challenge the overt ccommand condition on the overt licensing relation between the negative and the NPI. The empirical facts discussed in this paper have important implications for the current views of the architecture of grammar and the nature of dependency relations involve in licensing relations involving Case, agreement and negation, among others. The paper is organized as follows. In section 1, we discuss sentential negation in Hindi and the representation of negation in clause structure. In section 2, we provide a brief typology of NPIs in Hindi. In section 3, we focus on the distribution of one type of NPI, namely the ek bhii phrase in various positions in the sentence and its licensing options ek bhii. We will

32

Elabbas Benmamoun and Rajesh Kumar

show, on the basis of reconstruction e¤ects, that licensing is carried out overtly prior to spell-out in a c-command or Spec-head relation with sentential negation and the ek bhii phrase. In section 4, we turn our attention to some facts that seem to challenge this analysis and show how they actually end up supporting it. We conclude with a discussion of the implications of the present analysis. 1. Sentential negation in Hindi The sentential negative nahiiN in Hindi occurs adjacent to the verb, specifically to the left of the verb (Bhatia 1973).2 For example, in (1) the negative marker nahiiN ‘not’ occurs to the left of the verb dekh ‘see’. (1)

raajiiv film nahiiN dekh-taa hai. rajiv film NEG watch-HAB PRES ‘Rajiiv does not watch movies.’

Assuming that the functional category, negation, heads its own projection (Pollock 1989, Dwivedi 1991, Zanuttini 1997, Mahajan 2000, Kumar 2003), it seems that, as in French, English and many other languages, NegP is located between the projection of tense and the verb. We will follow Kumar (2003 and 2004) and assume the structure in (2). In (2), V moves to Asp. Then the complex VþAsp adjoins to Neg. Finally, the NegþvþAsp complex moves to T.2 Thus, we get the desired word order of the verbal complex in (1). (2)

The Overt Licensing of NPIs in Hindi

33

The idea is that the verb moves to the Asp first and then to Neg and finally to the T. However, the exact location of the sentential negative nahiiN relative to the functional categories located between tense and the verb is not crucial to our paper.3 This brief background about the syntactic position of sentential negation should su‰ce for the main topic of this paper, which is the syntactic licensing of NPIs. 2. Typology of NPIs in Hindi There are two distinct types of NPIs in Hindi: a group of NPIs that require a clause mate c-commanding negative licensor. We call them type I NPIs and a group of NPIs that are permitted in long distance licensing contexts. We will call them type II NPIs.4 Type I NPIs are not licensed in the context of questions, modality, conditionals or adversative predicates whereas type II NPIs are licensed in such contexts. Type I NPIs are illustrated in (3) and type II NPIs are illustrated in (4). (3)

hargiz ‘at any cost’ baal baaNkaa karnaa ‘to disturb hair’ muNh kholna ‘to open mouth’

(4)

koi bhii ‘anybody’ kisii bhii ‘anybody’ kuch bhii ‘any thing’ ek bhii ‘even one/any’

All type I NPIs in (3) can be licensed only by sentential negation as illustrated in (5–7).5 (5)

a.

b.

(6)

a.

maiN apnaa ghar hargiz nahiiN I self ’s house at any cost NEG becuN-gaa. sell-FUT-MASC-ISG ‘I will not sell my house at any cost.’ *maiN apnaa ghar hargiz becuN-gaa. I self ’s house at any cost sell-FUT-MASC-ISG *‘I will sell my house at any cost.’ raajiiv tumhaaraa baal baaNkaa nahiiN kar rajiv your hair disturb NEG do sake-gaa. can-FUT-MASC-IIISG ‘Rajiv cannot harm you.’

34

Elabbas Benmamoun and Rajesh Kumar b.

(7)

a.

b.

*raajiiv tumhaaraa baal baaNkaa kar rajiv your hair disturb do sake-gaa. can-FUT-MASC-IIISG ‘Rajiv can harm you.’ raajiiv-ne apnaa muNh nahiiN khol-aa. rajiv-ERG self ’s mouth NEG open-PERF ‘Rajiv did not say a word.’ *raajiiv-ne apnaa muNh khol-aa.6 rajiv-ERG self ’s mouth open-PERF ‘Rajiv said a word.’

On the other hand, type II NPIs can be licensed by sentential negation as well as elements other than the sentential negative such as questions, as illustrated in (8–11).7 (8)

(9)

(10)

(11)

a.

us kamre meN koi bhii sTuDeNT nahiiN thaa. that room in any EMPH student neg was ‘There was no/not even one (there wasn’t any) student in that room.’

b.

us kamre meN koi bhii sTuDeNT thaa? that room in any EMPH student was ‘Was there any student in that room?’

a.

maiN-ne kisii bhii sTuDeNT ko nahiiN dekh-aa. I-ERG any EMPH student to neg saw ‘I did not see any student.’

b.

maiN-ne kisii bhii sTuDeNT ko dekh-aa? I-ERG any EMPH student to saw ‘Did I see any student?’

a.

maiN-ne kuch bhii nahiiN khaa-yaa. I-ERG anything EMPH NEG ate ‘I did not eat anything.’

b.

maiN-ne kuch bhii khaa-yaa? I-ERG anything EMPH ate ‘Did I eat anything?’

a.

us kamre meN ek bhii sTuDeNT nahiiN thaa. that room in one EMPH student NEG was ‘There wasn’t any student in that room.’

The Overt Licensing of NPIs in Hindi b.

35

us kamre meN ek bhii sTuDeNT thaa? that room in one EMPH student was ‘Was there was any student in that room?’

In this paper, we will focus on the NPI ek bhii. While this NPI can be licensed in non-negative contexts as discussed in great detail by Lahiri (1998), for us what is important is its interaction with the negative licensor nahiiN. This interaction raises important questions concerning the structural conditions on NPI licensing and the level of representation or point in the derivation where such licensing takes place. 3. The Syntactic Distribution of ek bhii and Licensing Options Ek bhii occurs in a variety of contexts. It can occur with subjects (12a), objects (12b), indirect objects (12c) and genitives (12d). (12)

a.

ek bhii laRkaa klaas meN *(nahiiN) aa-yaa. one EMPH boy class in NEG come.PERF ‘None of the boys came to the class.’

b.

raajiiv ek bhii kitaab *(nahiiN) paR-taa hai. rajiv one EMPH book NEG read.HAB PRES ‘Rajiv does not read any books.’

c.

raajiiv-ne ek bhii laRke-ko kitaab *(nahiiN) rajiv-ERG one EMPH boy-DAT book NEG di-yaa. give.PERF ‘Rajiv did not give a book to any boy.’

d.

raajiiv ek bhii dost-ke ghar *(nahiiN) rajiv one EMPH friend-GEN house NEG ga-yaa. go.PERF ‘Rajiv did not go to any friend’s house.’

Interestingly, the NPI ek bhii does not have to be in the same clause as the negative licensor. In particular, it can be scrambled long distance as illustrated in (13) where the DP containing the NPI ek bhii scrambles to the matrix clause, to a position higher than the clause containing its negative licensor. (13)

[ek bhii laRke-ko]i sariitaa-ne khahaa ki raajiiv-ne t i one EMPH boy-to sarita-ERG said that rajiv-ERG yahaaN nahiiN dekh-aa. here NEG see-PERF ‘Sarita said that Rajiv did not see any boy here.’

36

Elabbas Benmamoun and Rajesh Kumar

The fact that the NPI ek bhii does not have to co-occur with its negative licensor in the same clause raises questions about how it is licensed. We will discuss two main options. 3.1. Option I: Covert Negative Movement One way to license the NPI in contexts where it is not overtly in the ccommand domain of negation is to covertly move the negative to a position where it can end up c-commanding the NPI. This is what Mahajan (1990) suggests. Assuming that in (14) the negative does not overtly c-command the NPI subject he puts forward an analysis whereby the negative nahiiN moves at LF to license the NPI. (14)

a.

thaa sabziii . koi bhii ti nahiiN khaa-taa any one EMPH NEG eat-PRES-HAB was vegetable ‘No one used to eat vegetables.’ (Mahajan 1990)

Specifically, the negative licensor moves at LF and adjoins to the finite IP, and thereby licenses the NPI. (15)

The Overt Licensing of NPIs in Hindi

37

One argument that Mahajan provides for the movement of the negative comes from its interaction with scrambling of objects to positions lower than the target position of the negative movement at LF, namely IP. Thus, in (16) the object sabzii is scrambled to a position lower than IP. The sentence is ungrammatical presumably because the negative cannot move across the scrambled element to adjoin to IP and thus license the NPI. (16)

sabziii thaa. *koi bhii ti nahiiN khaa-taa anyone EMPH NEG eat-PRES-HAB vegetable was ‘No one used to eat vegetables.’ (Mahajan 1990)

(16) is ungrammatical because of the presence of the right scrambled NP sabzii, which is adjoined to AGRP. The idea is that a scrambled element blocks the movement of the negative licensor leaving the NPI unlicensed at LF. More specifically, the presence of the right adjoined scrambled phrase that is lower than the finite I in (16) introduces a barrier that blocks the movement of the negative licensor at LF (Laka 1990).8 However, (16) turns out to be ungrammatical for reasons independent of the requirement on NPI licensing. As illustrated in (17) rightward scrambling is blocked in the context of negation even when the letter does not license NPIs. Similarly, in (18) rightward scrambling yields degraded sentences even in non-negative sentences. (17)

sabzii thaa. ?/*raam ti nahiiN khaa-taa ram NEG eat-PRES-HAB vegetable PST ‘Ram did not used to eat vegetables.’

(18)

?/*raam ti khaa-taa sabzii thaa. ram eat-PRES-HAB vegetable PST ‘Ram used to eat vegetables.’

While an explanation for the facts in (17–18) is beyond the scope of this paper, the main point is that absence of rightward scrambling in the context of sentential negation does not necessarily argue for the abstract movement of the negative. There are further problems with any analysis that relies on the LF movement of negation. First, the movement of the negative licensor is an instance of head move-ment, which is considered to be local, whereas an NPI in Hindi can scramble long distance, as in (13) repeated here. If the movement of negation is necessary for the licensing of NPIs, this analysis predicts that heads may move out of its clause in long distance scrambling contexts such as (13) repeated here. But this movement clearly violates the head movement constraint (Travis 1984, Rizzi 1990).

38 (13)

Elabbas Benmamoun and Rajesh Kumar [ek bhii laRke-ko]i sariitaa-ne khahaa ki raajiiv-ne t i one EMPH boy-to sarita-ERG said that rajiv-ERG yahaaN nahiiN dekh-aa. here NEG see-PERF ‘Sarita said that Rajiv did not see any boy here.’

In (13) the NPI containing ek bhii is scrambled from the lower clause to the higher clause. For it to be licensed the negative nahiiN must move from the lower clause to the higher clause. Since nahiiN is a head, this implies that head movement can be long distance, which is not widely attested. Second, this analysis makes the incorrect prediction that the NPI can be generated in a higher clause and the negative licensor generated in a lower clause. The NPI can be licensed by LF movement of the negative. In other words, we should expect sentences such as (19) to be possible. (19)

*[raajiiv-ne ek bhii laRke-ko kah-aa] ki sariitaa rajiv-ERG one EMPH boy-to say-PERF COMP sarita nahiiN aa-yii thii. NEG come-PERF PST-FEM *‘Rajiv told any boy that Sarita did not come.’

Third, if indeed the negative can undergo LF movement to license a higher NPI, we should expect it to be able to take scope over quantifiers in the clause where the scrambled NPI is located. In Hindi, negation can take scope over a clause mate quantifier subject as illustrated by (20), from Nevins and Anand (2003). (20)

har vidyarthii-ne koi kitaab nahiiN paRhii. every student-ERG some book NEG read-PERF ‘Every student did not read some book.’ (Neg > E > b, *Neg > b > E)9

A negative in a lower clause clearly does not take scope over a quantifier in a higher clause as illustrated in (21), where the higher QP is not within the scope of negation. (21)

har vidyarthii jaantaa hai ki raajiiv-ne kitaab every student know-m-HAB COMP rajiv-ERG book nahiiN paRhii. NEG read-PERF ‘Every student knows that Rajiv did not read the book.’ (*Neg > E)

Crucially, the scope domain of negation does not extend even when the higher clause contains a quantifier subject and a scrambled NPI.

The Overt Licensing of NPIs in Hindi (22)

39

ek bhii kitaab har vidyarthii jaantaa hai ki one EMPH book every student know-m-HAB COMP raajiiv-ne nahiiN paRhii. rajiv-ERG NEG read-PERF ‘Every student knows that Rajiv did not read any book.’ (*Neg > E)

In our opinion, this clearly demonstrates that the reach of negation does not extend higher than its clause, casting doubts on any analysis that would rely on LF movement of negation to higher clauses to license moved NPIs. Finally, it is not conceptually clear what motivates the movement of the negative licensor other than the need for it to be in a position where the ccommand requirement is maintained.

3.2. Option II: Reconstruction Another option is to zero in on the scrambled element rather than the negative. Since the NPI can be scrambled and assuming that licensing of NPIs is an LF condition, one way to bring the NPI into the c-command domain of negation is to reconstruct it to its pre-scrambling position. For example the NPI in (23), can be licensed at LF by reconstruction to the prescrambling position. (23)

[ki [ek bhii kitaab]i [ramesh nahiiN soctaa hai one EMPH book ramesh NEG think be-PRES COMP raajiiv-ne ti paRhii.]] rajiv-ERG read.PERF ‘Ramesh does not think that Rajiv read any book.’

However, there is evidence that the process of reconstruction cannot apply to the licensing of NPIs. One crucial piece of evidence involves the interaction of scrambled NPIs and Binding. It is well known that binding conditions are sensitive to reconstruction (Barss 1986). For example, in Hindi an anaphor can be scrambled long distance. If the licensing of anaphors obtains at LF then reconstruction or the copy of the scrambled element is what saves the representation. (24)

[ek dusre ke saathk ]i raajiiv-ne saritaa se kah-aa ki each other with rajiv-ERG sarita to said that [ramesh aur mohan]k tI laR rahe the. ramesh and mohan fight PROG PST-PL ‘Rajiv told Sarita that Ramesh and Mohan were fighting with each other.’

40

Elabbas Benmamoun and Rajesh Kumar

However, reconstruction is not an obligatory process as is well known. Sometime reconstruction must be blocked in situations where it may lead to violations (Aoun and Benmamoun 1998). Consider the following sentence. (25)

raajiiv-ne kah-aa ki [vo pikcars jo sariitaai ne rajiv-ERG said that that pictures which sarita ERG lii]j uskoi pasand hai tj took she likes PRS ‘Rajiv said that she likes the pictures that Sarita took.’

Reconstruction in (25) would lead to a principle C violation as the pronoun usko c-commands its antecedent sariitaa as shown in (26). (26)

*rajiiv-ne kah-aa ki uskoi pasand hai [vo pikcars rajiv-ERG said that she likes PRS that pictures jo sariitaai -ne lii]. which sarita-ERG took ‘Rajiv said that she likes the pictures that Sarita took.’

Now crucially, suppose we combine such cases where reconstruction is blocked with cases where it must apply in order to license NPIs at LF. Under such a scenario we predict the outcome to be ungrammatical. However, this is not the case, as illustrated in (27). Notice that reconstruction or access of the copy under the copy theory of movement would lead to a violation of principle C as illustrated in (28) because the pronoun usko ends up c-commanding its antecedent sariitaa. (27)

raajiiv-ne kah-aa ki rajiv-ERG said that sariitaai -ne lii]j uskoi sarita-ERG took she ‘Rajiv said that she does

[vo ek bhii pikcar jo that one EMPH pictures which pasand nahiiN hai tj likes neg PRS not like any picture that Sarita took.’

(28)

*raajiiv-ne kah-aa ki rajiv-ERG said that [vo ek bhii pikcars that one even pictures ‘Rajiv said that she does

uskoi pasand nahiiN hai. he likes neg PRS jo sariitaa-nei lii] which sarita-ERG took not like any picture that Sarita took.’

Therefore, it seems reasonable to conclude that the NPI within the scrambled clause cannot be licensed through reconstruction or access of the original copy under the copy theory of movement. In short, the argument for LF/Covert/Post-spell-out licensing of NPIs seems to be problematic. The question then is: how are they licensed?

The Overt Licensing of NPIs in Hindi

41

3.3. Overt Licensing of NPIs We would like to argue for the third option, namely that NPIs are licensed overtly prior to scrambling (i.e., at S-structure), where the clause mate negative licensor c-commands the NPI. This is not a novel idea as others have provided arguments for a similar conclusion on the basis of di¤erent evidence from di¤erent languages (Benmamoun 1996, among others). Let’s go through the cases we started with one by one. The licensing of NPIs in the object position follows straightforwardly. The negative nahiiN c-commands the NPI in the object position. The same analysis carries over to datives and NPIs within complement and genitive positions. (29)

a.

raajiiv ek bhii film nahiiN dekh-taa hai. rajiv one EMPH film NEG watch-HAB PRES ‘Rajiiv does not watch any movies.’

b.

However, the licensing of the NPIs in the subject position raises questions. Assuming that the subject is generated in the Spec of VP one could argue that the c-commanding negative licenses it prior to its movement to the functional layer (such as IP/TP). However, this raises problems for languages such as English where A-moved elements in general (including subjects of passives and of subject-to-subject raising constructions) must be licensed in the final (surface) A-position.10 (30)

a. b.

*Anyone was not arrested. *Anyone is not likely to come.

42

Elabbas Benmamoun and Rajesh Kumar

On the other hand, since in Hindi there is plausible evidence that the negative nahiiN ends up on the head of IP/TP as we assumed above (Dwivedi 1991; Kumar 2003, 2004), one could argue that the NPI in Spec IP is licensed in a Spec-head relation with the negative in I (following many people who have argued that NPIs can be licensed in a Spec-head configuration: Haegeman 1995; Benmamoun 1996).11,12 Turning to the crucial case, namely scrambled NPIs, the immediate analysis that comes to mind is that they are licensed overtly under c-command by (or Spec with) negation prior to scrambling. (31)

[ek bhii laRke-ko]i sariitaa-ne khah-aa ki raajiiv-ne one even boy-to sarita-ERG said that rajiv-ERG nahiiN dekh-aa thaa NEG saw PST ‘Sarita said that Rajiv did not see any boy.’

Assuming that licensing of NPIs applies overtly prior to scrambling allows us to avoid the problems that arose in the context of the analysis that relies on LF licensing through reconstruction to LF movement of the negative. However, there is one more case we would like to discuss that has been argued to show that negation does not need to c-command or be in Spec-head relation with negation. The situation arises in the context of what we would refer to as restructuring predicates for reasons that will become obvious. 4. NPI Licensing in the Context of Non-finite complements Nevins and Anand (2003) provide one case where it appears that the ek-bhii NPI does not need to be in the c-command domain of negation overtly. The data appear to show that a negative in the lower non-finite clause can license the NPI subject in the higher clause (32). (32)

ek bhii laRkaa [dillii nahiiN jaanaa] cahtaa one EMPH boy [Delhi NEG go-INF] want-IMPF (hai) (be-PRES) ‘Not one boy wants to go to Delhi.’

However, there is strong evidence that actually the negative is in the higher clause within a complex verb that contains the matrix verb and the embedded infinitive that has undergone head movement and incorporation.13 The evidence comes from the fact that while an adverb can intervene between the matrix verb and the embedded infinitive when negation immediately precedes (and hence forms a complex with) the matrix verb (33a), such

The Overt Licensing of NPIs in Hindi

43

intervention by the adverb is blocked if negation precedes the infinitive (33b). (33)

a.

ek bhii laRkaa dillii jaanaa [janvarii se pahle] one even boy Delhi go-INF January from before nahiiN cahtaa (hai) NEG want-IMPF (be-PRES) ‘Not one boy wants to go to Delhi earlier than January.’

b.

*ek bhii laRkaa dillii nahiiN jaanaa [janvarii se one even boy Delhi NEG go-INF January from pahle] cahtaa (hai) before want-IMPF (be-PRES) ‘Not one boy wants to go to Delhi earlier than January.’

This readily follows from the assumption that negation in these constructions is always in the higher clause and that the infinitive may optionally undergo head movement to the higher clause. If movement takes place, the negative will end up preceding the entire complex (as it does with the verbþaux and nounþverb complexes), hence (34a). On the other hand, if the infinitive does not undergo head movement to the matrix clause the negative will only merge with the matrix verb, hence (34b). In short, the facts in (32–34) end up supporting the present analysis, namely that the ek-bhii NPI must be licensed overtly in a c-command relation or Spec-head relation with clause-mate sentential negation. (34)

a.

ek bhii laRkaa dillii [nahiiN [jaanaa cahtaa]] one EMPH boy Delhi NEG [go-INF want-IMPF] (hai) (be-PRES) ‘Not one boy wants to go to Delhi.’

b.

ek bhii laRkaa dillii jaanaa nahiiN cahtaa one EMPH boy Delhi go-INF NEG want-IMPF (hai) (be-PRES) ‘Not one boy wants to go to Delhi.’

5. Conclusions and Implications In this paper, we argued that NPIs in Hindi are licensed overtly. We provided evidence based on reconstruction e¤ects that support our analysis and argue against alternative analyses that in terms of covert/LF licensing of NPIs. However, this analysis obviously raises questions about scrambling

44

Elabbas Benmamoun and Rajesh Kumar

are beyond the scope of this paper. It may turn out that scrambling of NPIs (and probably scrambling in general) involves non-syntactic movement.14 Under such a scenario overt licensing of NPIs follows. The attractive part of this analysis is that we are not forced to require that NPIs be licensed overtly, which in turn amounts to a condition that is specific to overt syntax (basically re-introducing S-structure). All that is required is that at some point in the derivation to LF the NPI be in the c-command domain of negation or in Spec-head agreement relation with it. On the other hand, if scrambling takes place in narrow syntax in the sense of Chomsky (1995 and subsequent work) then one must come up with a stipulation to the e¤ect that NPIs must be licensed overtly. The desirable situation would be one where we could locate a particular feature or property (such as an uninterpretable feature akin to Case) that forces overt licensing/checking. It is not clear to us what that feature is. In short, the theory that NPIs must be licensed overtly has important implications for the architecture of grammar. Similar issues arise in the context of Case and agreement.

Notes * We are grateful to Professor Rama Kant Agnihotri of Delhi University for his suggestions and judgments. 1. Unlike the sentential negative, the constituent negative follows the element it negates. 2. One problem with this derivation is that the movement of verb to Asp is through left adjunction while the movement of VþAsp to Neg is through right adjunction. This is obviously undesirable and an alternative account is called for. We will put this problem aside as the analysis of NPIs below does not crucially hinge on the exact mode of head adjunction. However, see Mahajan (2000) for an alternative to head movement in Hindi relying on phrasal movement. 3. Dwivedi (1991) posits a slightly di¤erent position for the NegP in the clause structure of Hindi. In her system, she suggests that NegP is located below AspP. However, see Kumar (2003, 2004) for arguments that NegP is higher than AspP. 4. This dichotomy echoes the typology of Serbo-Croatian NPIs discussed in Progovac (1994). 5. There are other NPIs such as ek phuuTii kauRii ‘a red cent’ which in a simple sentence appears to be licensed only within the scope of negation. Consider the following example: (i)

raajiiv ke pitaajii-ne usko ek phuuTii kauRii *(nahiiN) rajiv of father-ERG to him one broken penny NEG dii gave ‘Rajiv’s father did not give him a red cent.’

The Overt Licensing of NPIs in Hindi

45

We thank the anonymous reviewer who suggests that NPIs such as ek phuuTii kauRii ‘a red cent’ can be licensed in the absence of a negative licensor. Consider the following sentence as suggested by the reviewer: (ii)

6. 7. 8.

9. 10.

tum samajhte ho, rupaye ghar hii meN to haiN: magar you think PRES money home-EMPH in PRES but dekh lena, tumheN jo ek phuuTii kauRii *(bhii) mile see-IMP you one broken penny-EMPH find-SUBJ ‘You think that it’s OK, the money will remain at home, but wait and see, you won’t get a red cent.’

NPI ek phuuTii kauRii ‘a red cent’ in (ii) is in the context of imperative and subjunctive. It has been suggested (Lahiri 1998; Kumar 2003, among others) that contexts such as questions, conditionals, imperatives, subjunctives and adversative predicates license NPIs in the absence of an overt negative marker. Also, the NPI ek phuuTii kauRii ‘a red cent’ appears to be licenced in these contexts in the presence of the polarity marked bhii ‘also/even’ (Lahiri 2001). In the absence of the polarity marker bhii the sentence in (ii) will be unacceptable as illustrated above. In the light of the above discussion, it appears that in the typology of NPIs there is a third type where an NPIs may not be licensed in the absence of an overt negative marker in a simple sentence and also in some of the contexts that license NPIs otherwise. But such NPIs can possibly be licensed in the context of imperative and subjunctives with added ploarity marker. We, once again, thank the reviewer. This sentence (7b) is grammatical in its literal meaning where it means ‘Rajiv opened his mouth’. See Lahiri (1998) for a detailed analysis of the semantic contexts where Hindi NPIs are allowed. Mahajan (1994) assumes that that adjunction to a maximal projection creates an A-bar node and involves a non L-related position. He argues that scrambling operations in Hindi involve both Argument Shift (movement to an A/L position) and Adjunction (movement to a non A/L position). The scrambling in (16) is an instance of adjunction to AGRP. Since adjunction creates an A-bar node, this particular instance of scrambling is A-bar scrambling. Elements moved to an Abar position can reconstruct (they can be interpreted in their original position) at LF. Thus, we have to make sure that the scrambled element does not reconstruct so that it can continue to block LF movement of the negative. According to Nevins and Anand (2003) the object can take scope over the absolutive subject but not over the ergative subject in (21). It seems that NPIs must be licensed in the highest A-position where they occur on the surface. Of course one could take this to argue against A-movement altogether and instead generate elements in their highest A-position. The facts from Moroccan Arabic discussed in Benmamoun (1996) seem to support the same conclusion. Thus, an NPI can be licensed in the pre-verbal/pre-negative position (i). However, an NPI contained within a subject NP cannot be licensed if it is in the Spec of the head that contains negation (namely TþV) in the SVO order (ii) but is allowed if the subject is postverbal (iii)

46

*
mm ⁄ tta wa⁄ d ma-zˇa uncle even one neg-came

(iii)

ma-zˇa
mm ⁄ tta wa⁄ d. neg-came uncle even one *‘The uncle of anyone did not come.’ e

e

e

e

e

14.

(ii)

e

13.

⁄ tta wa ⁄ d ma-zˇa. even one neg-came.3MS *‘Anyone didn’t come.’ e

12.

(i)

e

11.

Elabbas Benmamoun and Rajesh Kumar

All these facts could follow if the subject is merged directly in the highest Aposition or if NPI licensing is tied to some property that the NP acquires in the highest A-position (Case for example). In fact, this is essentially how Borer (2003) handles the generation of arguments. They are base generated directly in the Spec of the (aspectual) functional projections rather than within a lexical shell. If this is the case then in Hindi the subject is probably generated in a position higher than negation. Incidentally, the Moroccan Arabic facts in (ii) show that licensing cannot be carried out by accessing the copy since the copy is in the same position as the NPI in (iii). This would imply that in English the negative is not in I/T at the stage where the NPI subject must be licensed. Another important issue is whether one can dispense with one of the structural conditions, Spec-head or C-command (Agree). We will put this issue aside as it is not relevant to the main topic of this paper. However, we believe that both may be needed (see Benmamoun 1996 and note 7 above), which has its own implications in light of recent discussions on agreement, such as Chomsky (2000). Another option that we will leave open would be to generate sentential negation in a projection higher than the projection that hosts subjects. This would obviously have implications for head movement and the ordering of functional categories above the VP (see Ouhalla 1990; Zanuttini 1997; and Cinque 1999 for discussion of possible positions of sentential negative heads). For details about infinitives in Hindi/Urdu see Kachru (1980) and Butt (1995) and references cited there. However, the issue of where scrambling applies remains a thorny issue. See Mahajan (1994) and Kidwai (2000) for di¤erent analyses of scrambling constructions in Hindi. See also Erteschik-Shir and Strahov (2004) for a more detailed exploration of a non-syntactic account of scrambling in Danish and Russian.

References Aoun, J. and Elabbas Benmamoun 1998 Minimality Reconstruction and PF Movement, Linguistic Inquiry 29: 569–597. Barss, A. 1986 Chains and Anaphoric Dependence. D Ph. D. diss., MIT: Cambridge, MA. Benmamoun, Elabbas 1996 Licensing of Negative Polarity Items in Moroccan Arabic, Natural Language and Linguistic Theory, 15: 263–287.

The Overt Licensing of NPIs in Hindi

47

Bhatia, Tej K. 1973 On the Scope of Negation in Hindi. In Studies in the Linguistic Sciences, Vol. 3 (2): 1–27. Urbana: University of Illinois. Borer, Hagit 2003 Structuring Sense: An Exo-Skeletal Trilogy. Oxford University Press USA. Butt, Miriam 1995 The Structure of Complex Predicates in Urdu. Stanford: CSLI Publications. Chomsky, Noam 1995 The Minimalist Program. Cambridge, MA: MIT Press. 2000 Minimalist Inquires: The Framework. In Step by Step: Essays on Minimalist Syntax in Honor of Howard Lasnik, Roger Martin, David Michaels and Juan Uriagereka (eds.), pp. 89–155. Cambridge, MA: MIT Press. Cinque, Guglielmo 1999 Adverbs and Functional Heads: A Cross-Linguistic Perspective. Oxford: Oxford University Press. Dwivedi, Veena 1991 Negation as a Functional Projection in Hindi. In Proceedings of the Western Conference on Linguistics, K. Hunt, T. Perry, and V. Samiian (eds.), 1991. Vol: 88–101. Erteschik-Shir, Nomi and Natalia Strahov 2004 Focus structure architecture and P-syntax. Lingua 114: 301–323. Haegeman, Liliane 1995 The Syntax of Negation. Cambridge: Cambridge University Press. Kachru, Yamuna 1980 Aspects of Hindi Grammar. New Delhi: Manohar Publications. Kidwai, Ayesha 2000 XP-adjunction in Universal Grammar: Scrambling and Binding in Hindi-Urdu. New York: Oxford University Press. Kumar, Rajesh 2003 The Syntax of Negation and the Licensing of Negative Polarity Items in Hindi. Ph. D. diss., University of Illinois at Urbana-Champaign. Kumar, Rajesh 2004 Negation in the Clause Structure of Hindi. (Presented at Chicago Linguistic Society 40 at the University of Chicago. To appear in CLS 40 Proceedings.) Ladusaw, William A. 1992 Expressing negation. In Proceedings of Semantics and Linguistic Theory (SALT) 2, C. Baker and D. Dowty (eds.), pp. 237–59. Columbus: Ohio State University. Lahiri, Utpal 1998 Focus and negative polarity in Hindi, Natural Language Semantics 6: 57–123. Lahiri, Utpal 2001 Even-incorporated in NPIs in Hindi definites and correlatives. In Perspectives in Negation and Polarity Items, Jack Hoeksema, Hotze Rullman, Victor SanchezValencia, and Tom van des Wouden (eds.), 237–264. Amesterdam/Phladelphia: Benjamins. Laka, Itziar 1990 Negation in syntax: On the nature of functional categories and projections. Ph. D. diss., MIT. Linebarger, Marcia C. 1987 Negative Polarity and Grammatical Representation, Linguistics and Philosophy 10, 325–387. Mahajan, Anoop K. 1990 LF conditions on negative polarity item licensing, Lingua 80: 333–348.

48 1994

Elabbas Benmamoun and Rajesh Kumar

Toward a unified theory of scrambling. In Studies on Scrambling: Movement and Non-Movement Approaches to Free Word-OrderPhenomenon, N. Corver and Henk van Riemsdijk (eds.), 301–330. Berlin/New York: Mouton de Gruyter. 2000 Eliminating Head Movement. The GLOW Newsletter # 44, 44–45. Nevins, Andrew and Pranav Anand 2003 Some AGREEment Matters. In WCCFL 22 Proceedings, G. Gardina and M. Tsujimura, (eds.), 101–114. Somerville, MA: Cascadilla Press. Ouhalla, Jamal 1990 Sentential Negation, Relativized Minimality and Aspectual Status of Auxiliaries. Linguistic Review 7: 183–231. Pollock, Jean-Yves 1989 Verb Movement, UG and the Structure of IP. Linguistic Inquiry, 20: 365–424. Progovac, Ljilijana 1994 Negative and Positive Polarity. Cambridge: Cambridge University Press. Rizzi, Luigi 1990 Relativized Minimality. Cambridge, MA: MIT Press. Travis, Lisa deMena 1984 Parameters and e¤ects of word order variation. Ph. D diss., MIT. Uribe-Echevarria, M. 1994 Interface licensing conditions on negative polarity items: A theory of polarity and tense interactions. Ph. D. diss., The University of Connecticut. Vasishth, S. 1999 Surface structure constraints on negative polarity and word order in Hindi and English. (Presented at the 11 th European School in Logic.) Zanuttini, Ra¤aella 1997 Negation and Clause Structure: A Comparative Study of Romance languages. New York: Oxford University Press.

Reduplication in Bengali* SHISHIR BHATTACHARJA

The purpose of this paper is to describe the word-level morphological processes of reduplication in Bengali. The description follows a particular model of morphology called Whole Word Morphology (WWM). In order to describe these processes, a clear distinction between ‘pattern’ and ‘process’, suggested in Singh (2003), has been used to expunge mere patterns of reduplication from the description. The paper also argues that WWM can handle the phenomena in question more adequately than some other theories of morphology.

1. Introduction Although South Asian languages are known to be rich in reduplication, not too many detailed studies of this phenomenon in individual South Asian languages exist (for some notable exceptions, see Abbi (1992 and Singh 2003)). As reduplication has recently become an important domain of study for morphological theory1, it is important to undertake such studies so that these theories can be tested against the rich data made available by South Asia. The purpose of this paper is to attempt to do so by examining the relevant facts of Bengali (a.k.a. Bangla). Apart form being briefly mentioned in Chatterji (1926 (1985) and (1945 (1988)) and in Abbi (1992), to the best of my knowledge reduplication in Bengali has not as yet been extensively studied. For the descriptive task at hand, I shall use the framework of W(hole) W(ord) M(orphology), summarized, with appropriate references, in section-2. I find WWM more adequate than other models of morphology, particularly for Bengali (for reasons spelled out in Bhattacharja (forthcoming)). Although I shall not attempt to provide a justification for the model chosen for the description in section-4 it can be easily found in the publications mentioned in section 2. Moreover, the problems presented by Bengali for other models of morphology I discuss in section-5 necessarily constitute arguments for it.

50

Shishir Bhattacharja

2. WWM: A brief description According to Ford and Singh (2003: 18), WWM views morphology ‘‘as the study of formal relationship between words’’.2 They (2003: 19) claim that ‘‘any morphological relationship between a non-unique pair of words of a language can be described by a rule, to be called a Morphological Strategy (MS) or a Word Formation Strategy (WFS)’’. The formulation proposed is given below (reproduced verbatim from Singh (2006: 1413– 1417)): ‘‘/X/a $ /X 0 /b where a. /X/a and /X 0 /b are words and X and X 0 are abbreviations of the forms of classes of words belonging to categories a and b (with which specific words belonging to the right category can be unified or onto which they can be mapped) b. 0 represents (all the) form-related di¤erences between /X/ and /X 0 / that fall outside of automatic phonology c. a and b are categories that may be represented as feature-bundles d. the $ represents a bidirectional implication (if X, then X 0 and if X 0 , then X) e. the interpretation of /X/a is a semantic function of /X 0 /b , and vice versa f. 0 can be null i¤ a 0 b’’ If a lexicon has two pairs of words like child $ childish and ghoul $ ghoulish, the association between these pairs is captured by the morphological strategy /X/ $ /XIS/.3 It can be, as Martohardjono (1986: 22) puts it, ‘‘used in subsequent word formation, for example in lexical innovation [e.g. Benladenish], as well as in the analysis of newly encountered items [e.g. Jihadish].’’ (Parenthetically inserted examples here and throughout the paper are mine). Thus exploitations of morphological strategies help the speakerhearer, as Singh and Ford (2000: 305) claim, ‘‘bridge the gap between actual words she happens to know and the possible words she can be said to know—actually their existence makes the known merely a subset of the knowable.’’ (1) below instantiates a morphological strategy of English for it is licensed by at least two pairs of words based on i) the same formal di¤erence: X/XnIs ([nIs] is a realization of the prime), ii) semantic relatedness: ‘/X/-ness’ and iii) categorial a‰liation: adjective/noun. According to Singh (2006) ‘‘morphological complexity is a matter of the analyzability (0 segmentability) of a word into a variable [kind, bright] and a constant [ness]’’ provided, of course, as Singh and Ford (2000: 308) put it, ‘‘there are strategies that license such analyses.’’

Reduplication in Bengali (1)

51

/X/adj $ /XnIs/n ‘/X/-ness’ kind $ kindness bright $ brightness

Strategies are morphological processes which capture, as Ford and Singh (2003: 19) argue, ‘‘the morphological relatedness amongst the words that happen to be in a lexicon’’. The WWM view would be that in any lexicon, a good number of words are formally and/or categorically di¤erent and semantically related to each other. Strategies emerge on the basis of the generalization of some complex combination of such formal di¤erences as well as semantic relatedness (cf. Singh, 1992) and the schema /X/a $ /X 0 /b based on these generalizations assures that, as Singh (2001: 344) claims, ‘‘all that needs to be said about word-structure in any language (of any type whatsoever)’’.4 WWM is a ‘holistic’ theory of morphology because it claims that words have no (non-phonological) hierarchical structure5. Singh and Starosta (2003: 12) argue that words are ‘‘in essence seamless wholes’’ and therefore, as Ford, Singh and Martohardjono (1997: 3) claim, there can be ‘‘No morphological operations on units other than the word.’’ Hence, no ‘atomistic’ category smaller than the word, for example, ‘a‰x’, ‘root’, ‘stem’, ‘lexeme’, etc. can be regarded, as Ford, Singh and Martohardjono (1997: 3) point out, ‘‘as an object of a morphological enquiry.’’6 3. The distinction between pattern and process of reduplication Sequences of the following sort have been used as examples of reduplication in the Bengali literature (cf. Chatterji 1988: 195–199): /naduSnuduS/adj ‘fat’ /kicirmicir/n ‘sounds of birds’ /baba/n ‘father’

(2)

a. b. c.

(3)

/aSepaSe/p=adv ‘around’

(4)

/luÒpaÒ/n; plu ‘looting, etc.’

(5)

/golgal/adj ‘round and alike’

(6)

/paSepaSe/p=adv ‘always beside’

(7)

/gurumgarum/ ‘big sounds of gunshot’

(8)

/m
Òm
Ò/ ‘sounds of breaking branches’

52 (9)

Shishir Bhattacharja a. b. c.

/biy eSadi/n ‘marriage and similar occasions’ /cakofi/n ‘hot drinks’ /dinrat/adv ‘all the times’

Although reduplication is generally assumed to be a process of wordformation, disagreements prevail regarding whether or not the input of such a process must be a word. For example, according to Abbi (1992), words in (2) can be analyzed back to meaningless sequences like [naduS], [ba] or [kicir] which have been used as input in a process that she calls morphological reduplication7. In her (1992: 12) view, ‘‘in spite of the fact that the part which is repeated is neither a lexical item nor a constituent of a lexical item’’, such a sequence acquires ‘‘this status only after it is being reduplicated.’’ Now, if we accept (2) as complex words obtained from the process of reduplication, we must accept words like /ŒanŒa/ ‘rod’, /lala/ ‘saliva’, /SiSir/ ‘dew-drops’, /l
nÒh
n/ ‘lantern’, /c
nd
n/ ‘sandal wood’ as examples of reduplication. But these words are unanimously accepted as simplex words and no one has ever claimed that they result from the iteration of sequences like [Œa], [la], [Si], [l
n] or [c
n]. As the status of ‘things’ from which the words in (2) are said to be derived is extremely dubious, none of these words can really be claimed to have been obtained through the process of reduplication. The di¤erence between WWM and other theories is that whereas according to WWM both the input and the output of a morphological process must be word, most other theories require only the output to be a word.8 As Singh (2003: 156) puts it, ‘‘Tomato must be ruled out as an example of reduplication because neither toma nor mato are words in English.’’ Therefore, a reduplicated word must be a complex word formed with an input which is also a word. As none of the words in (2) can be analyzed back to another word, according to WWM they cannot be seen as morphologically complex. (10)

/paSe/p=adv ‘beside’ ! /aSepaSe/p=adv ‘around’

(11)

/luÒ/n ‘looting’ ! /luÒpaÒ/n; plu ‘looting, etc.’

Although (3) and (4) can be analyzed back to simple words neither (10) nor (11) represents a process because the formal di¤erence and semantic relatedness between the pair-mates manifest themselves in no other pair. This means that, as is the case with (2), the speaker-hearer has to memorize both (3) and (4) and once forgotten, morphology cannot help her to retrieve them given the input [paSe] or [luÒ]. On the other hand, (5) and (6) do not need to be memorized because the speaker-hearer can retrieve or form these

Reduplication in Bengali

53

words by mapping /gol/ and /paSe/ onto (12) and (13), motivated by the memory of another pair which manifests the same formal di¤erence and semantic relatedness. (12)

AASMAD: /C1 Vþround; þhigh C2 /adj $ /C1 Vþround; þhigh C2 C1 aC2 /adj ‘/C1 Vþround; þhigh C2 / and alike’9 /gol/ ‘round’ $ /golgal/ ‘round and alike’ /b h ul/ ‘false’, ‘erroneous’ $ /b h ulb h al/ ‘false and alike’, ‘erroneous and alike’

(13)

P/AdvP/AdvSVAD: /Xe/p=adv $ /XeXe/p=adv ‘always /Xe/’ /paSe/p ‘beside’ $ /paSepaSe/adv ‘always beside’ /dh ire/adv ‘slowly’ $ /dh iredh ire/adv ‘always slowly’

For Chatterji (1988) both (2a–b) and (7) are examples of reduplicated O(nomatopoeic) W(ord). But in my view, not all OW that show some sort of iteration of phonemic sequence can be described as examples of ‘reduplication’ if this term denotes a morphological process. Some of them like /m
Òm
Ò/ and /gurumgarum/ are indeed complex words obtained from the processes like (14) and (15) whereas others like /naduSnuduS/ (2a) and /kicirmicir/ (2b) show only patterns of reduplication because they have access to no strategy. (14)

OWOWAD: /X/OW $ /XX/OW ‘multiplicity of /X/’ [m
Ò] ‘sound of breaking one single branch’ $ /m
Òm
Ò/ ‘sounds of breaking branches’ [Òh
k] ‘sound of one single knocking on the door’ $ /Òh
kÒh
k/ ‘sounds of knocking on the door’

(15)

OWOWSMAD: /CuX/OW $ /CuXCaX/OW ‘multiplicity of /CuX/’ [gurum] ‘big sound of one single gunshot’ $ /gurumgarum/ ‘big sounds of gunshot’ [Òh uS] ‘light sound of one single gunshot’ $ /Òh uSÒh aS/ ‘light sounds of gunshot’

Words like /biy eSadi/ ((9a) and (16a)) constituted of synonyms: /biy e/ and /Sadi/ ‘marriage’, /cakofi/ ((9b) and (17a)) constituted of hyponyms: /ca/ ‘tea’ and /kofi/ ‘co¤ee’ and /dinrat/ ((9c) and (18a)) formed with antonyms /din/ ‘day’ and /rat/ ‘night’ can be related to either of these synonyms, hyponyms and antonyms. But the pairs in (16), (17) and (18) cannot justify a process because the formal di¤erence they manifest can be found in

54

Shishir Bhattacharja

no other pair. However, many other pairs show the same semantic relatedness that exists between the pair-mates of (16), (17) and (18) but semantic relatedness alone does not su‰ce to justify a process10. Again, in order to form such words, for example, the word meaning ‘books and similar objects’ with the input /b
j / ‘book’, the speaker-hearer must know a number of synonyms of the latter: /pust
k/, /gr
nt h o/, /kitab/. Such synonymy (and also antonymy or hyponymy) motivated constraints on the formulation of a strategy remain unmotivated. (16)

a. b.

/biy e/ ‘marriage’ ! /biy eSadi/ ‘marriages and similar occasions’ /b
j / ‘book’ ! /b
j pust
k/ ‘books and similar objects’

(17)

a. b.

/ca/ ‘tea’ ! /cakofi/ ‘hot drinks’ /kok/ ‘coca cola’ ! /kokspra j Ò/ ‘cold drinks’* (*Sprite is the name of a cold drink)

(18)

a. b.

/din/ ‘day’ ! /dinrat/ ‘all the time’ /S
kal/ ‘morning’ ! /S
kalbikal/ ‘all the time’* (*/bikal/ means ‘afternoon’)

I claim (pace Chatterji 1988 and Abbi 1992) that mere repetition of some phonemic sequence observed in some words cannot reasonably mean that they are examples of reduplication. It is true that reduplication is basically repetition but not vice versa. All the examples mentioned above (2–9) manifest some sort of repetition but only (5–8) are complex words obtained from a ‘processes of reduplication’ whereas the others manifest mere ‘pattern of reduplication’. The distinction is important but unfortunately not made in any work except Singh (2003). Without such a distinction, it is hard to draw a clear line between morphologically complex words and simple words which, for some reason or other, look like complex words but have undergone no morphological process.11 This said, all processes are basically patterns but not all patterns necessarily qualify as processes because not all of them are justified with at least two pairs of words showing the same formal di¤erence, categorial a‰liation and semantic relatedness. Therefore, (2), (3), (4) and (9) which are traditionally described as reduplicated words cannot be claimed, as it has been clearly shown, to have undergone any morphological process. In order to avoid confusion, such examples are eliminated hereinafter from this discussion. I claim that they have nothing to do with morphology and doubt whether any other theoretical approach can satisfactorily account for them.

Reduplication in Bengali

55

4. Morphological strategies and reduplication in Bengali Like any other word, a reduplicated word can be analyzed into a variable and a constant subcomponent by mapping it onto a relevant strategy. In both (1) and (19), the mechanism involved is adjunction-deletion but the di¤erence between the two is that in (1), the speaker-hearer adjoins or deletes a constant subcomponent (nIs) whereas in (19) he adjoins or deletes the variable itself and this clearly shows that there is nothing special with strategies like (19)12. (19)

AAAD: /X/adj; sing $ /XX/adj; plu ‘plural of /X/’ /b
ro/ ‘big’ $ /b
rob
ro/ ‘all big’ /S
bu/ ‘green’ $ /S
buS
bu/ ‘all green’

(20)

aaSMAD: /(C)VX/a $ /(C)VXÒVX/a ‘/(C)VX/, etc.’* /b
j / ‘book’ $ /b
j Ò
j / ‘books and similar things’ /akaS/ ‘sky’ $ /akaSÒakaS/ ‘sky and similar things’ (*a represents here any syntactic category except conjunction and interjection.)

As is the case with other strategies, the variable can be either i) totally unspecified (19) or ii) partly specified (20). The specification of the variable is a question of degree and as we see below, the variable in (20) is less specified as compared to the variable in (21), which is again less specified than the variable in (22).13 (21)

aaSMAD:/XCV(C)s1- s4 /a $ /XCV(C)s1- s4 ÒVs1- s4 /a ‘/XCV(C)s1- s4 / and similar things’ /porik
lpona/n ‘plan’ $ /porik
lponaÒona/n or /p
rik
lp
naÒ
lpona/n ‘planning, etc.’ /raniti/n ‘politics’ $ /ranitiÒiti/n ‘politics, etc.’

(22)

AASMAD: /C1 iC2 /adj $ /C1 iC2 C1 aC2 /adj ‘perfectly /C1 iC2 /’ /fiÒ/ ‘well-dressed’ $ /fiÒfaÒ/ ‘perfectly well dressed’ /Òh ik/ ‘alright’, ‘okay’ $ /Òh ik Òh ak/ ‘perfectly alright or okay’

When the speaker-hearer feels the need to form or retrieve a certain word, the only thing he needs to do is to map an already existing word like /b
ro/ or /b
j / onto relevant strategies (19–20) and the latter will automatically output /b
rob
ro/ and /b
j Ò
j / respectively and nothing more is needed for word formation. As both the input and the output of a strategy are words, a complex word like /b
j Ò
j / can be formed from the simple word /b
j / or the latter can be back-formed from /b
j Ò
j / (because the bidirectional nature of the strategies allows the speaker-hearer to do so).

56

Shishir Bhattacharja

Like any other strategy of Bengali, (19–22) transform a simple word /X/a into a complex one /X 0 /b . Through which particular way or mechanism the formal di¤erence [ 0 ] is obtained is not a relevant question for the theory itself and therefore, such strategies do not need to be put in a particular category or given a di¤erent name. Therefore, as is the case with atomistic distinctions like Derivation/Inflection (cf. Ford, Singh and Martohardjono 1997) and Derivation/Compounding (cf. Singh and Dasgupta 1999) or categories like a‰x or stem, terms like reduplication and categories such as reduplicant or base have no particular status in WWM. According to Ford and Singh (2003: 19) ‘‘The listing of the morphological strategies of a language constitutes a part of the description of that language. It is, therefore, an aspect of linguistic competence, a component of grammar.’’ With the discussion above as background, I shall now proceed, on the basis of the data available to me, to formulate the remaining strategies of Bengali which, together with the strategies mentioned above, should hopefully yield an exhaustive list of strategies involving repetition of the entire variable or some specified part of it. (23)

AASMAD: /CX/adj $ /CXSX/adj ‘/CX/ and alike’ /b
ro/ ‘big’ $ /b
roS
ro/ ‘big and alike’ /moÒa/ ‘fat’ $ /moÒaSoÒa/ ‘fat and alike’

(24)

AAdvSMAD: /CV-high; þback Xa/adj $ /CV-high; þback XaCVþhigh Xi/adv ‘in a /CV-high; þback Xa/ way’ /l
mba/ ‘tall’ $ /l
mbalombi/ ‘lengthwise’ /Soa/ ‘straight’ $ /SoaSui/ ‘straightly’

(25)

NNAD: /X/n; sing $ /XeXe/n; plu; loc ‘in each and every /X/’ /Œal/ ‘branch’ $ /ŒaleŒale/ ‘on each and every branch’ /maS/ ‘month’ $ /maSemaSe/ ‘in each and every month’

(26)

NNSMAD: /XC/n $ /XCaXCi/n ‘reciprocal action involving two or several /XC/ of di¤erent persons’ /kan/ ‘ear’ $ /kanakani/ ‘whispering to one another’, ‘spreading rumours’ /hat/ ‘hand’ $ /hatahati/ ‘hand-to-hand fight’

(27)

NNSMAD: /XVC/n $ /XCaXCi/n ‘fighting that involves exchange of /XVC/’ /kam
r/ ‘bite’ $ /kamrakamri/ ‘biting each other’ /ach ar/ ‘a fall or throw to the ground with force’ $ /ach raach ri/ ‘mutual knocking to the ground as in wrestling’

Reduplication in Bengali

57

(28)

NNSMAD: /C1
C2 oC3 /n $ /C1
C2 C3 aC1 oC2 C3 i/n ‘continuous and reciprocal instances of /C1
C2 oC3 /’ /b
dol/ ‘change’, ‘exchange’ $ /b
dlabodli/ ‘several (reciprocal) instances of exchanging’ /dh
mok/ ‘rebu¤ ’ $ /dh
mkadh omki/ ‘several (reciprocal) instances of rebu‰ng’

(29)

NNSMAD: /CX/n; sing $ /CXCaÒi/n ‘several instances of /CX/’ /kanna/ ‘crying’ $ /kannakaÒi/ ‘several instances of crying’ /h
gra/ ‘dispute’ $ /h
grah aÒi/ ‘several instances of dispute’

(30)

NNSMAD: /CX/n $ /hXCX/n ‘/CX/ and similar things’ /b h ab/ ‘mental state’, ‘mood’ $ /habb h ab/ ‘gestures deportment’ /cal/ ‘manners’ $ /halcal/ ‘manners and conditions’

and

(31)

NNSMAD: /XC/n; sing $ /XCaXC/n; plu ‘multiplicity of /XC/’ /mal/ ‘merchandise’, ‘goods’ $ /malamal/ ‘several goods’ /gal/ ‘insult’ $ /galagal/ ‘insults’

(32)

NNAD: /X/n $ /XXanto/n; plu ‘/X/ and beyond’ /bon/ ‘forest’ $ /bonbonanto/ ‘forest and beyond’ /dur/ ‘far’ $ /durduranto/ ‘far and beyond’

(33)

NAAD: /X/n $ /XX/adj ‘like /X/’ /b
r/ ‘bridegroom’ $ /b
rb
r/ ‘bridegroom-like’ /mey e/ ‘girl’ $ /mey emey e/ ‘girl-like’

(34)

NAAD: /X/n $ /XkeX/adv ‘/X/ after /X/’ /din/ ‘day’ $ /dinkedin/ ‘day by day’ /b
ch
r/ ‘year’ $ /b
ch
rkeb
ch
r/ ‘year after year’

(35)

VNSMAD: /C1 V-high C2 /v; pres imp; 2nd intim=derog $ /C1 V-high C2 aC1 Vþhigh C2 i/n ‘continuous instances of the action asked for in /C1 V-high C2 /’ /ch oÒ/ ‘run!’ $ /ch oÒach uÒi/ ‘running here and there’ /meS/v ‘frequent!’ $ /meSamiSi/n ‘act of frequenting’

(36)

VNSVAD: /XC/v; pres imp; 2nd intim=derog $ /XCaXCi/n ‘several reciprocal instances of the action asked for in /XC/’ /mar/ ‘beat!’ $ /maramari/ ‘several acts of beating each other’ /kar/ ‘seize by force or by tactics!’ $ /karakari/ ‘several reciprocal acts of seizing by force or by tactics’

58

Shishir Bhattacharja

(37)

VNSMAD: /Xa/v; pres imp; 2nd intim=derog $ /XaXi/n ‘several (reciprocal) instances of the action asked for in /Xa/’ /palÒa/ ‘change!’ $ /palÒapalÒi/ ‘exchanging something between each other’ /dh akka/ ‘push!’ $ /dh akkadh akki/ ‘several acts of pushing each other’

(38)

VNSMAD: /CV-high Xa/v; pres imp; 2nd intim=derog $ /CV-high XaCVþhigh Xi/n ‘several (reciprocal) instances of the action asked for in /CV-high Xa/’ /b
dla/ ‘change!’, ‘exchange!’ $ /b
dlabodli/ ‘exchanging something between each other’ /hE cra/ ‘drag or trail along forcefully!’ $ /hE crahe cri/ ‘act of dragging or trailing along each other forcefully’

(39)

VAdvSVAD: /Xe/compl=perf verbal $ /XeXe/adv ‘during or after continuous instances of action referred to in /Xe/’ /lik h te/ ‘to write’ $ /lik h telik h te/ ‘by writing’ /lik h e/ ‘having written’ $ /lik h elik h e/ ‘by writing’

(40)

MWMWAD: /X/MW; sing $ /XX/MW; plu ‘multiplicity of /X/’ /bindu/ ‘drop’ $ /bindubindu/ ‘drops’ /muÒh o/ ‘handful’ $ /muÒh omuÒh o/ ‘handfuls’

(41)

aaSMAD: /(C)VX/a $ /(C)VXfVX/a ‘/(C)VX/ and similar trashes’ /akaS/ ‘sky’ $ /akaSfakaS/ ‘sky and similar trashes’ /paSe/ ‘beside’ $ /paSefaSe/ ‘beside and in similar other disgusting positions’

(42)

aaSMAD: /XC1 aC2 /a $ /XC1 aC2 ÒaC2 /a ‘/XC1 aC2 / and alike’ /rahaS/ ‘goose’ $ /rahaSÒaS/ ‘goose and alike’ /b h alobaS/ ‘love!’ $ /b h alobaSÒaS/ ‘love and do similar things!’

As far as reduplication in Bengali is concerned, the description above allows us to construct a picture of the following sort:

Reduplication in Bengali

59

Distribution of primary strategies involving repetition of (partly specified or totally unspecified) variable14 Morphological operation

Number of strategies

Type of variable

Number of strategies

Noun-Noun

8

Specified

21

Adjective-Adjective

4

Totally unspecified

7

Verb-Noun

4

aa

4

Onomatopoeic Word— Onomatopoeic Word

2

Verb-Adverb

1

Adjective-Adverb

1

Postposition/Adverb— Postposition/Adverb

1

Noun-Adjective

1

Noun-Adverb

1

Measure Word—Measure Word

1

Total

28

28

5. Problematic examples for other morphological theories In this section, I briefly describe how the phenomenon of reduplication is treated in some other morphological theories and cite some examples of processes of reduplication which seem to be di‰cult to handle in these frameworks but satisfactorily handled in WWM. For example, the outputs (here and throughout this section output will mean the rightward output only) of (29) (repeated here as (43)) would be problematic for Marantz (1982: 436), who considers reduplication as ‘‘normal a‰xation processes’’ in which, instead of some normal a‰x, a copy of the phonemic melody of the stem is attached to the stem itself. But as we can see in the outputs of (43), the reduplicational a‰x [ h aÒi] or [kaÒi] does not represent a copy of the phonemic melody of the stem [ h
gra] or [kanna].

60 (43)

Shishir Bhattacharja NNSMAD: /CX/n; sing $ /CXCaÒi/n ‘several instances of /CX/’ /kanna/ ‘crying’ $ /kannakaÒi/ ‘several instances of crying’ /h
gra/ ‘dispute’ $ /h
grah aÒi/ ‘several instances of disputes’

O(ptimality) T(heory) treats reduplication with the identity constraint that requires the reduplicant to be as faithful as possible to the base. According to McCarthy and Prince (1995: 1) ‘‘reduplication is a matter of identity: the reduplicant copies the base’’ But, the identity constraint is violated in all instances of incomplete reduplication and the reduplicant is more of less unfaithful to the base, a common tendency in South Asian languages (cf. Vijaykrishnan 1999 and Singh 2003). However, violation of constraints is not a problem for OT because all OT constraints are in principle violable. In my view, OT will have di‰culties in accounting for the outputs of (20) (repeated here as (44) with di¤erent pairs) and (21) (repeated here as (45)) in which di¤erent pieces of prosody: [t
na], [Ò
lp
na] or [Òorik
lpona] are a‰xed to the base resulting in several optimal outputs, a fact that OT can hardly allow. (44)

aaSMAD: /(C)VX/a $ /(C)VXÒVX/a ‘/(C)VX/, etc.’ /porik
lpona/n ‘plan’ $ /porik
lponaÒorik
lpona/n ‘planning, etc.’ /raniti/n ‘politics’ $ /ranitiÒaniti/n ‘politics, etc.’

(45)

aaSMAD: /XCV(C)s1- s4 /a $ /XCV(C)s1- s4 ÒV(C)s1- s4 /a ‘/XCV(C)s1- s4 / and alike’ /porik
lpona/n ‘plan’ $ /porik
lponaÒona/n or /p
rik
lp
naÒ
lpona/n ‘planning, etc.’ /raniti/n ‘politics’ $ /ranitiÒiti/n ‘politics, etc.’

According to Kiparsky’s (2002: 395–397) L(exical) P(rosodic) M(orphology) ‘‘the size and melody of the base is predictable from the morphology of reduplication’’ and ‘‘the reduplicant gets its form from the base’’ which ‘‘determines both how much of the reduplicated word’s melody (the root, stem, or the whole word) is accessible to the reduplication process, and what the form of the copied melody is.’’ It would be di‰cult to account for the outputs of (35) (repeated here as (46)) and (47) within the framework of LPM because several sequences compete for the base-hood in this words. For example, both /ch oÒ/ and /ch uÒi/ can be considered as the base for /ch oÒach uÒi/ and /meSamiSi/ can be formed with either /meS/ or /miSi/. This eventually shows that categories like base and reduplicant are not easy to identify. The outputs of (43) also represent problematic cases for LPM because little of the base /h
gra/ is apparently accessible to the morphology of reduplication, which means that the base /kanna/ or /h
gra/

Reduplication in Bengali

61

can determine neither the prosody nor the melody of the reduplicant /kaÒi/ or /h aÒi/. (46)

VNSMAD: /C1 V-high C2 /v; pres imp; 2nd intim=derog $ /C1 V-high C2 aC1 Vþhigh C2 i/n ‘continuous instances of the action asked for in /C1 V-high C2 /’ /ch oÒ/ ‘run!’ $ /ch oÒach uÒi/ ‘running here and there’ /meS/v ‘frequent!’ $ /meSamiSi/n ‘act of frequenting’

(47)

VNSMAD: /CVþhigh Ci/v; simp pres; 1st $ /CV-high CaCVþhigh Ci/n ‘several reciprocal instances of the action referred to in /CVþhigh Ci/’14 /ch uÒi/ ‘I run’ $ /ch oÒach uÒi/ ‘several acts of running to and fro’ /miSi/ ‘I frequent’ $ /meSamiSi/ ‘act of frequenting’

In M(orphological) D(oubling) T(heory), a morpho-semantic approach by Inkelas (2005), a reduplicated word is comprised of two daughters, i) reduplicant and ii) base, both generated by the morphology from a mother input recoverable morphologically and semantically from the daughters. ‘‘The key assumption of MDT is that daughters are semantically identical’’ whereas ‘‘phonological identity is not presupposed or required because each of the daughters may be subject to special phonology’’ (Inkelas 2005:67). MDT assumes that the same morphological structure is assigned to partial and to total reduplication, which di¤er solely in whether or not one of the daughters is phonologically truncated. (48)

NNSVAD: /X/n $ /XXanto/n ‘/X/ and beyond’ /bon/ ‘forest’ $ /bonbonanto/ ‘forest and beyond’ /dur/ ‘far’ $ /durduranto/ ‘far and beyond’

The outputs of (32) (repeated here as (48)) are problematic for MDT because in neither of them are the two sisters semantically identical (/bonanto/ means ‘end of the forest’ and /duranto/ means ‘remote or far-o¤ place’). If my observations are right, the two sisters in (43) [kanna] and [kaÒi] have not been assigned the same morphological structure and their formal contrast can be explained neither by truncation nor by any phonological rule of Bengali if phonology includes only automatic alternations. The outputs of (44) and (45) can also be problematic for MDT because it does not determine where or how much of the reduplicant has to be truncated or whether the same reduplicant can be truncated in di¤erent ways. The base /porik
lpona/ can have three di¤erent reduplicants: /Ò
rik
lpona/ (44) or /Ò
lpona/ or /Òona/ (45) depending on whether the reduplicant is truncated and/or how it is truncated.

62

Shishir Bhattacharja

Travis (2001) argues that reduplication is always triggered in syntax and, like other types of a‰xation, results from feature checking head movement: the head of the NP (e.g. /h
gra/ ‘dispute’ in the reduplicated word /h
grah aÒi/ ‘disputes, etc.’) moves to adjoin to the head representing a quantity feature which is realized as the (reduplicational) a‰x (/h aÒi/). Reduplication is necessarily realized through this quantity feature checking because, according to her, in many cases reduplication represents a quantity of some sort. Outputs of (28) (repeated here as (49)) which involve several a‰xes ([a] and [i]) as well as segmental modification (deletion of /o/) may seem to be problematic for Travis (2001). However, this problem can be overcome if the su‰x is represented by the relevant prosodic template (aC1 oC2 C3 i) and the segmental modification is assumed to be part of the process of a‰xation (i.e. as a ‘morpho-phonological’ change as some approaches usually describe such non-automatic changes). (49)

NNSMAD: /C1
C2 oC3 /n $ /C1
C2 C3 aC1 oC2 C3 i/n ‘continuous and reciprocal instances of /C1
C2 oC3 /’ /b
dol/ ‘change’, ‘exchange’ $ /b
dlabodli/ ‘several (reciprocal) instances of exchanging’ /dh
mok/ ‘rebu¤ ’ $ /dh
mkadh omki/ ‘several (reciprocal) instances of rebu‰ng’

As is the case with weak lexicalist approaches in general, Travis deals exclusively with inflexional morphology because she is only concerned about the checking of the number feature of NP. But as the outputs of (23), (24) and (26) (repeated here as (50–52)) show, words belonging to syntactic categories other than noun can undergo reduplication. In (50) for example, if

Reduplication in Bengali

63

the reduplicated adjective needs to check any feature, it will be its quality feature. On the other hand, the adverb in (51) and the derived singular nouns in (52) do not need to check any number feature at all. (50)

AASMAD: /CX/adj $ /CXSX/adj ‘/CX/ and alike’ /b
ro/ ‘big’ $ /b
roS
ro/ ‘big and alike’ /moÒa/ ‘fat’ $ /moÒaSoÒa/ ‘fat and alike’

(51)

AAdvSMAD: /CV-high Xa/adj $ /CV-high XaCVþhigh /CV-high Xa/ way’ /l
mba/ ‘tall’ $ /l
mbalombi/ ‘lengthwise’ /Soa/ ‘straight’ $ /SoaSui/ ‘straightly’

(52)

NNSMAD: /XC/n $ /XCaXCi/n ‘reciprocal action involving two or several /XC/ of di¤erent persons’ /kan/ ‘ear’ $ /kanakani/ ‘whispering to one another, spreading rumours’ /hat/ ‘hand’ $ /hatahati/ ‘hand-to-hand fight’

Xi/adv

‘in

6. Conclusion In this work, I have tried to sketch out a particular type of word-formation in Bengali traditionally known as ‘reduplication’ and I have shown how it can be satisfactorily handled in the theoretical framework of WWM. Following Singh (2003), I make a distinction between the processes (¼ strategies) of reduplication and the patterns that exhibit reduplication and claim that the latter fall outside the domain of morphology. Bengali would seem to confirm the WWM claim that there is only one morphology (cf. Ford, Singh and Martohardjono 1997: 3) and there is nothing in reduplication which makes it radically di¤erent from other bits of morphology, except the fact that strategies activated for forming or retrieving the socalled reduplicated words consist of repeating the (partly specified or totally unspecified) variable. Notes * This paper is a substantially revised version of a chapter of my recently submitted doctoral thesis. I am grateful to my supervisors for their comments on that chapter and to Stephen Moran for the modifications he proposed. The usual disclaimers apply. 1. To cite some examples, Optimality theory (McCarthy and Prince 1995), Neo optimality theories (Yip 1999; Golston and Thurgood 1999; Struijke 2000), Morphological doubling theory (Inkelas 2003), Lexical prosodic morphology (Kiparsky 2002).

64

Shishir Bhattacharja

2. WWM was, as noted in Singh (2006), first outlined in Ford and Singh (1991) followed by a fuller, monographic sketch in Ford, Singh and Martohardjono (1997). Various aspects of the model have been elaborated in Singh and Ford (2000), Ford and Singh (2003), Singh and Neuvel (2003) and Singh and Starosta (2003) among others. The model as such has been tested in the light of external evidence in Martohardjono (1986) and Ford, Singh and Marotohardjono (1997) and against the empirical facts of several languages: Hindi in Singh and Agnihotri (1997), Bengali in Singh and Dasgupta (1999) and Bhattacharja (forthcoming), various South Asian languages including Khasi, Kashmiri and Sanskrit in Singh and Ford (2000), German in Becker (2000), Armenian in Baronian (2002), Latin in Bender (2003) and West Greenlandic in Neuvel (2003). 3. The apparently ambiguous term word has a clear-cut definition in WWM: an expression can be considered a word if and only if it possesses at least the following three properties among others which remain to be defined: i) a phonological structure, ii) a (syntactic and/or morphological) category, and iii) a semantic use (i.e. a meaning). Ford, Singh and Martohardjono (1997: 5) remind us that ‘‘whether these properties are su‰cient or not to identify the word in any context is a question that we shall leave open. They are necessary, but insofar as we wish to draw a distinction between syntax (formal relationships between linguistic units other than the word) and morphology, they are probably not su‰cient.’’ 4. A WFS is called a strategy rather than a rule because, i) they are invoked, as Singh and Ford (2000: 305) point out ‘‘only in moments of crisis i.e. when the speaker needs to analyze or fashion a word she needs for the purpose at hand, often to meet a syntactically enforced requirement’’ and ii) they are not as automatic as linguistic rules are generally claimed to be, for example, if the speaker-hearer already knows a certain word, (s)he does not need to activate any strategy to retrieve it. 5. The holistic view of morphology goes back to Bhartrihari, an eminent critic of Paninian morphology, who lived in India around 8th century AD. In my view, a true holistic theory would reject the idea that words can be divided into smaller sub-parts. Holistic models of morphology are intrinsically word-based but the contrary is not true which means that not all word-based models of morphology are necessarily holistic. The inputs and the outputs of the word-formation rules proposed by a word-based model are words but the model of morphology remains atomistic (see footnote 6) if it does not categorically reject units smaller than the word (root, stem, a‰x, etc.). In some word-based approaches, such units are considered as part of word-formation rules and in some others, they are listed in the lexicon. 6. The atomistic view of morphology goes back to Panini, who lived in the northwest region of Indian subcontinent (now in Pakistan) around 6th century B.C. (cf. Katre 1989). A good number of the existing models of morphology are largely influenced by or are in the line of the Paninian school as all of them encourage morphology to be a matter of divide and rule (cf. Ford, Singh and Martohardjono, 1997) which means that one must divide words into smaller subparts (root, stem, a‰x, etc.) in order to find out the rules of their concatenation. Quintessentially, the atomistic or Paninian approaches sees morphology, as Singh (2001: 344) puts it,

Reduplication in Bengali

7.

8. 9.

10.

65

‘‘as a combinatorics of units smaller than the word, involving word-internal syntax in some versions, and allows morphological operations on those units.’’ Therefore, by ‘atomistic’ or ‘Paninian’ I point to a certain type of morphological theory which shares a common belief that i) words have internal hierarchical structure and that ii) words result from the combination of di¤erent types of ‘word-parts’ labeled as stem, radical, root, a‰x, etc. Abbi (1992: 12) defines reduplication as ‘‘repetition of all or a part of a lexical item carrying a semantic modification’’. She classifies reduplication into two di¤erent types: i) Morphological and ii) Lexical. Abbi (1992: 12) argues that ‘‘morphological reduplication refers to the minimally meaningful and segmentally indivisible morphemes which are constituted of iterated syllables’’ and lexical reduplication ‘‘refers to the repetition of any sequence of phonological units comprising a word.’’ For her (1992: 13) ‘‘onomatopoeic constructions, imitatives, certain instances of sound symbolisms, mimic words are all examples of morphological reduplication.’’ Needless to say that the concern here is word-level reduplication. Reduplication of units larger than the word is clearly non-morphological. A tag like AASMAD means that the strategy changes an adjective into another adjective (AA), involves segmental modification (SM) (Vþround; þhigh is replaced by /a/) and the mechanism of adjunction-deletion (AD). In some tags, in stead of (SM) appears (SV) which means specified variable. (SM) is necessarily (SV) but not vice versa. For example, if a speaker-hearer tries to form the word meaning ‘without fear’ with /b h
y / ‘fear’, (a) will output either */niSb h
y / or /nirb h
y /. The pairs in (b) show that the formal di¤erence alone does not su‰ce either to justify a process. (a)

(b)

2.

/k
l
Nko/n ‘bad reputation’ ! /niSk
l
Nko/adj ‘without any bad reputation’ /
h
Nkar/n ‘pride’ ! /nir
hNkar/adj ‘without pride’, ‘humble’

1. 2.

/ma/n ‘mother’ ! /upoma/n ‘metaphor’ /har/n ‘necklace’ ! /upohar/n ‘gift’

1.

11. Certain homophonous sequences like /b
rb
r/ is interesting in the sense that they can be described either as example of process or of pattern. One of the versions of /b
rb
r/ ‘bridegroom like’ can be mapped onto (33) (repeated here as (c)) and be thereby analyzed into /b
r/ ‘bridegroom’. If only this version is considered, /b
rb
r/ is a complex word that has undergone the process of reduplication. But the other version of the word /b
rb
r/ ‘barbarian’ cannot be analyzed back into any word and if this version is considered, /b
rb
r/ would be a simple word showing only the pattern of reduplication. (c)

NAAD: /X/n $ /XX/adj ‘like /X/’ /b
r/ ‘bridegroom’ $ /b
rb
r/ ‘bridegroom-like’ /mey e/ ‘girl’ $ /mey emey e/ ‘girl-like’

12. Each strategy involves some ‘change in the category of the word’ which I label as morphological operations (such as adverbialization, pluralization, gender change,

66

Shishir Bhattacharja etc.). These operations are realized through formal means which I call morphological mechanisms or simply mechanisms. The di¤erent mechanisms are exemplified below:

Identity

Adjunction-deletion

Substitution

/X/n; sing $ /X/v

/X/n; sing $ /Xz/n; plu

/X«nt/adj $ /X«ns/n

rhyme $ rhyme

dog $ dogs

absent $ absence

fight $ fight

rose $ roses

important $ importance

Segmental modification B adjunction

Suprasegmental Modification

/Xk/adj $ /Xsiti/n

/ss/n; sing $ /s 0 s/v

electric $ electricity

import $ im’port

opak $ opasiti

protest $ pro’test

13. In approaches other than WWM, the outputs of strategies like (19) are described as completely reduplicated words and those of (20–22) as partially reduplicated ones. Unlike completely reduplicated words which generally have a distributive interpretation, partially reduplicated words are usually endowed with either an ‘Et Cetera’ and/or an ‘X and alike’ interpretation (cf. Singh 2003) in Bengali. 14. If a strategy i) involves the mechanism of substitution and ii) all of its outputs can be obtained from some other strategy, I consider it as a secondary strategy. No such strategy is included in the list of 28 strategies in section 4. For example, (47) (repeated here as (d)) is not a primary strategy because all of its outputs can be obtained from (46) (repeated here as (e)). On the other hand, (28) and (38) (repeated here as (f–g)) are both primary strategies because one of the outputs of (g) /hE crahe cri/ cannot be obtained from (f ) for */hE c
r/ is not a word of Bengali. However, this should be kept in mind that WWM grants no theoretical status to the di¤erence between primary and secondary strategies. As long as there is a lexicon, the morphological module automatically makes various obvious links among them and therefore, all strategies are part of the morphological module. But, as I do not have su‰cient space, I have only listed the primary strategies and have left aside the secondary ones. (d)

VNSMAD: /CVþhigh Ci/v; simp pres; 1st $ /CV-high CaCVþhigh Ci/n ‘several reciprocal instances of the action referred to in /CVþhigh Ci/’ /ch uÒi/ ‘I run’ $ /ch oÒach uÒi/ ‘several acts of running to and fro’ /miSi/ ‘I frequent’ $ /meSamiSi/ ‘act of frequenting’

(e)

VNSMAD: /C1 V-high C2 /v; pres imp; 2nd intim=derog $ /C1 V-high C2 aC1 Vþhigh C2 i/n ‘continuous instances of the action asked for in /C1 V-high C2 /’

Reduplication in Bengali

67

/ch oÒ/ ‘run!’ $ /ch oÒach uÒi/ ‘running here and there’ /meS/ ‘frequent!’ $ /meSamiSi/ ‘act of frequenting’ (f )

NNSMAD: /C1
C2 oC3 /n $ /C1
C2 C3 aC1 oC2 C3 i/n ‘continuous and reciprocal instances of /C1
C2 oC3 /’ /b
dol/ ‘change’, ‘exchange’ $ /b
dlabodli/ ‘several (reciprocal) instances of exchanging’ /dh
mok/ ‘rebu¤ ’ $ /dh
mkadh omki/ ‘several (reciprocal) instances of rebu‰ng’

(g)

VNSMAD: /CV-high Xa/v; pres imp; 2nd intim=derog $ /CV-high XaCVþhigh Xi/n ‘several (reciprocal) instances of the action asked for in /CV-high Xa/’ /b
dla/ ‘change!’, ‘exchange!’ $ /b
dlabodli/ ‘exchanging something between each other’ /hE cra/ ‘drag or trail along forcefully!’ $ /hE crahe cri/ ‘act of dragging or trailing along each other forcefully’

Abbreviations: 1st: First person 2nd: Second person A/adj: Adjective Adv/adv: Adverb compl: Complement derog: Derogative imp: Imperative interj: Interjection intim: Intimate, loc: Locative

MW: Measure Word n/N: Noun p: Postposition perf: Perfective plu: Plural pres: Present sing: Singular simp: Simple V/v: Verb.

References Abbi, Anvita 1992 Reduplication in South Asian Languages: An Aerial, Typological and Historical Study, New Delhi: Allied Publishers ltd. Baronian, Luc 2002 No morphemes in my pocket, lexemes up my sleeves or stems under my hat: Western Armenian verbal morphology, Chicago Linguistic Society 37. Becker, Thomas 2000 The non-hierarchical structure of compounds: a reply to Singh and Dasgupta. The Yearbook of South Asian Language and Linguistics 2000: 283–292. Bender, Byron W. 2003 A perfect strategy for Latin. In Explorations in Seamless Morphology, Rajendra Singh and Stanley Starosta (eds.), 301–327.

68

Shishir Bhattacharja

Bhattacharja, Shishir Forth. Word formation in Bengali: A whole word morphological description and its theoretical implications, Ph.D. diss., Department of Linguistics and Translation, Universite´ de Montre´al (submitted May 2006). Chatterji, Suniti Kumar 1985 The Origin and Development of the Bengali Language. Calcutta: Rupa and Co. Reprint. Original edition, Calcutta University, 1926. 1988 Bhashaprakash Bangla Beakoran [Grammar of Bengali]. Calcutta: Rupa and Co. Reprint. Original edition, Calcutta University, 1945. Ford, Alan, Rajendra Singh, and Gita Martohardjono 1997 Pace Panini, Towards a Word-based Theory of Morphology. New York: Peter Lang. Ford, Alan, and Rajendra Singh 1991 Prope´deutique morphologique. Folia Linguistica 25: 549–575. 2003 Prolegomena to a theory of non-Paninian morphology. In Explorations in Seamless Morphology, Rajendra Singh and Stanley Starosta (eds.), 18–42. Golston, Chris, and Elzbieta Thurgood 1999 Reduplication as echo: evidence from Bontok and Chumash. Rutgers Optimality Archive: 456-0801. Inkelas, Sharon 2003 Morphological doubling theory: evidence for morphological doubling in reduplication. In Studies in Reduplication, Bernard Hurch (ed.), 65–88. Berlin and New York: Mouton de Gruyter. Kiparsky, Paul 2002 Paradigms and Opacity. Stanford: CSLI Publications. Katre, Sumitra M. 1989 Astadhyayi of Panini (translated). New Delhi: Motilal Banarsidass. Marantz, Alec 1982 Re Reduplication. Linguistic Inquiry 13: 435–482. Martohardjono, Gita 1986 Morphology and external evidence: a study of the Ford and Singh model of the lexicon. M.A. diss., De´partement de Linguistique et Traduction, Universite´ de Montre´al. McCarthy, John, and Alan S. Prince 1995 Faithfulness and reduplicative identity. Rutgers Optimality Archive 216-0997. Neuvel, Sylvain 2003 Metamorphology: A word-based account of polysynthesis and other multivalent morphological relations. Ph.D. diss., Department of Linguistics, University of Chicago. Singh, Rajendra 1992 Two morphological notes: a sharp and a flat. Revue Que´be´coise de Linguistique 21: 182–191. 2001 Constraints, preferences, and context-sensitivity in morphology. In Constraints and Preferences, Katarzyna Dziubalska-Kolaczyk (ed.), 339–358. Berlin: Mouton. 2003 Reduplication in modern Hindi and the theory of reduplication. The Yearbook of South Asian Language and Linguistics 2003: 156–170. 2006 Whole word morphology, Elsever Encyclopedia of Linguistics, 2nd edition: 1413– 1417. Singh, Rajendra, and Rama K. Agnihotri 1997 Hindi Morphology, A Word-based Description. Delhi: Motilal Banarsidass Publishers. Singh, Rajendra, and Probal Dasgupta 1999 On So-called Compounds. The Yearbook of South Asian Languages and Linguistics 1999: 265–292.

Reduplication in Bengali

69

Singh, Rajendra, and Alan Ford 2000 In praise of Sakatayana: some remarks on whole word morphology. The Yearbook of South Asian Languages and Linguistics 2000: 303–310. Singh, Rajendra, and Stanley Starosta (eds.) 2003 Explorations in Seamless Morphology, New Delhi: Sage Publications. Singh, Rajendra, and Sylvain Neuvel 2003 When the whole is smaller than the sum of the parts: the case of morphology. Chicago Linguistic Society 38. 2: 299–306. Struijke, Caro 2000 Why constraints conflict can disappear in reduplication. Rutgers Optimality Archive: 373-0100. Travis, Lisa D. 2001 The syntax of reduplication, NELS 31. Vijaykrishnan, Kumaralingam G. 1999 Conflict resolution in Optimality theory: The case of antifaithfulness in reduplicative morphology. Paper presented at GLOW 1999. Hyderabad, India. January 20–23, 1999. Yip, Moira 1999 Reduplication as alliteration and rhyme. Rutgers Optimality Archive: 377 0200.

Mirative meanings as extensions of aorist in Hindi/Urdu ANNIE MONTAUT

Mirative meanings (surprise, sudden awareness, high degree, polemic) have recently been described as distinct from evidentiality. Languages with evidential markers, such as Nepali, Kalasha or Khowar, are already known to have grammaticalized the expression of such meanings. Hindi/Urdu, which has no specific marker, displays nontheless a wide set of such meanings systematically attached to its aorist (the simple form used for narrative past). This paper attempts to test the claim that mirativity is a category distinct from evidentiality, a claim supported by languages that attach mirative extensions to verbal forms other than evidentials. I first define the standard meanings of the –standard form, argue in favour of the aoristic behaviour of the whole set of meanings, then try to relate the aoristic e¤ect to the special (mirative) meanings, and finally suggest an interpretation derived from the enunciative (utterance) theory of Culioli.

Typology started about twenty years ago (Chafe and Nichols 1986, then Guentcheva 1996 and Aikhenvald 2004) to describe evidentiality as a linguistic category.1 More recently, it started to describe mirativity as a distinct linguistic category (Delancey 1997; Aikhenvald 2004), although evidentials are now well known to display ‘‘mirative extensions’’ (to behave as markers of mirative meanings too). The empirical facts which were described as a consequence of the new interest in evidentials are certainly responsible for the change in the representation of the category (focusing less on the distance from the source of information) and eventually for the distinction between two categories. At the same time descriptive grammars now get a proper frame for describing related forms or meanings and are more and more aware of the existence of the category. The reason why this category (or these categories) has long been ignored in languages where it existed is simply because it was unknown in most familiar European languages. Nepali grammar, for instance, may now reconsider the ‘‘inferential’’ as a marker for evidentiality and/or mirativity, and there are many

72

Annie Montaut

other examples of such welcome results of interactions between typology and description. However, Hindi/Urdu have not yet been described as displaying an evidential system, which is understandable since there is no morphologically specific marker for it.2 As noted by Aikhenvald (2004: 210), ‘‘a major argument in favour of mirative meanings as independent from evidentials and information source comes from languages where mirative extensions are characteristic of categories other than evidentials.’’ I will try here to support this claim, pointing at the same time to the di‰culties in identifying the ‘‘path’’ and origin of extension due to the labelling of these ‘‘other categories’’. If the first attempts to identify mirative meaning as a linguistic strategy in Hindi/Urdu are recent (Montaut 2003, 2004), the reason is that no specific form in the verbal paradigm could previously lead to building a non tenseaspect-mood category to account for it, whatever its label. As opposed to Hindi/Urdu, Nepali has a second (underived) perfect which long ago drew attention under the label ‘‘inferential perfect’’. Apart from this second perfect, the past system in Nepali is very close to that of Hindi/Urdu,3 two closely related Indo-Aryan languages, but is described with very di¤erent labels, so that comparison is blurred for those who compare the existing grammar or descriptions, specially for typologists having access to second hand sources only. I will then first give a general presentation of the aorist in the global economy of the Hindi verbal system, before focusing on its specific meanings in view of understanding what is really at stake in the operation and relating it to the ‘aoristic e¤ect’. 1. The Hindi/Urdu aorist in the verbal system Hindi/Urdu, as well as Nepali, are head final languages, with auxiliaries after the radical (R) in compound forms. Whether written in one word (Nepali) or separately, most of the verbal forms are compound, and use the verb be for marking tense (H/U ho, N cho for present, th- for past). The main verb is in the participial form (R-a/o in the accomplished, R-taˆ/dai in the unaccomplished).4 Let us first briefly justify the label ‘‘aorist’’, and give at the same time its principal meanings in standard Hindi. The Hindi/Urdu simple form for expressing anteriority was traditionally labelled bhuˆt (past), and now in most of the English literature ‘‘perfective’’ (as opposed to the past imperfective), but it had also been labelled indefinite (Kellogg [1875] 1972), because the form contains no mark for tense and can be used in non past meanings including future. Besides, there is in the verbal system an opposition close to the Slavic perfective/imperfective: the simple verb (in all TAM) contrasts with the compound verb (þ semi-

Mirative meanings as extensions of aorist in Hindi/Urdu

73

Table 1. Indicative tense-aspect system in Indo-Aryan HINDI/URDU Unaccomplished Aspect Tense

NEPALI Accomplished Aspect Tense

Unaccomplished Aspect Tense

R-aˆ aorist

Accomplished Aspect Tense

Inferential

R-yo aorist

R-taˆ hai present (R rahaˆ hai progressive present) R-taˆ thaˆ imperfect (R rahaˆ thaˆ progressive)

R-aˆ hai Perfect

R-cha, R-dai cha present

R-eko cha perfect

R-aˆ tha pluperfect

R-rahe(ko)-cha progressive present R-dai thyo imperfect R-rahe(ko) thyo progressive imperfect

R-e(ko) thyo pluperfect

R-e cha inferential perfect

auxiliarized small set of movement verbs) in a way similar to the Russian preverbs, and the compound verb has been analyzed as a perfectivizer by quite a few scholars, from Porizka (perfective: 1968, 1982) to Hook (relative completion, 1974) and Nespital (perfective, 1997) although it is also used to indicate predicate orientation (transitive verbs with ‘give’ acquire an externalised meaning, with ‘take’ a self benefactive meaning or internalized orientation) and in depicting the manner (þ violence, impulsivity, etc.):5 (1)

a.

ve log kab aˆenge? / kab aˆe? (simple verb) these people when come-FUT / when come-AOR ‘When will these people come? When did they come?’

b.

jo kal aˆnevaˆle the ve aˆkhir aˆ REL yesterday come-su¤ were they finally come gae. go-AOR (compound verb) ‘Those who were to come yesterday finally arrived.’

c.

maˆr Daˆlo use ! strike throw-imper him ‘Kill him!’ (cf. maˆro use ‘beat him, strike at him’)

True, the simple form (similar to the past participle6 which is also used in the formation of the whole perfect system) is widely used in narratives to represent an anterior event or a sequence of anterior events:

74 (2)

Annie Montaut a.

tab vah bolaˆ then 3s speak-AOR-3ms ‘then he said’

b.

ghar se niklaˆ, postaˆfis gayaˆ, kuch samaˆn house of get-out-AOR post o‰ce go-AOR some things kharıˆdaˆ, phir socaˆ . . . . buy-AOR then think-AOR ‘He (or I) left the house, went to the post o‰ce, bought some things, then thought . . .’

However, it also has non-past non-narrative uses, the most common in the hypothetical system: in (3a) the temporal reference of the form cal gayaˆ in the protaxis is clearly future as suggested by the apodosis, even when the latter is omitted as in (3b). (3)

a.

maˆn ko pataˆ cal gayaˆ to kyaˆ mother dat knowledge walk go-AOR then what hogaˆ? be-FUT ‘If Mother comes to know, what will happen?’

b.

aur barsaˆt aˆ gaıˆ to? phisal paRoge? aur and shower come go-AOR then slip fall-FUT ThaNDıˆ lagıˆ to? and cold touch-AOR then? ‘And in case it rains? You will slip and fall? And in case it gets cold?

Less commonly it may express non-temporal static statements such as the conventional expression indicating the death of somebody, X nahıˆn rahe (‘stay’: aorist) ‘X is no longer here.’ Similarly, in (4) the occurrence of Thahrnaˆ ‘stay’ in the aorist does not mean that the state has come to an end or is accomplished, but alludes to a general truth like the definition of the subject as a poet, by nature or essence, akin to the so-called gnostic truths. (4)

A: B: A: B:

shair kahenge? poetry say-FUT-pl kyon nahıˆn kaenge, shaˆyar jo Thahre why not say-FUT-pl poet REL stay-AOR-pl ‘Will he (honorific) tell a poem?’ ‘Why not, since he is a poet? (he who is a poet)’

Given such facts, this last example being attested in many languages with an aorist (e.g. Greek), the label of aorist seems to better fit the concrete

Mirative meanings as extensions of aorist in Hindi/Urdu

75

behaviour of this tense, specially keeping in mind the meaning of this label (with no definite limits), which suits Kellogg’s (1972) intuition of ‘‘indefiniteness’’. Simply, in a narrative context, as is the case for other languages, the form is used to relate past events with no relation (non-incidence) on the present hence not in relation with the speaker. The selection of aorist for ‘‘historical’’ or objective narratives has been described by Benveniste (1996) for the French simple past, which he calls ‘‘aoriste du re´cit’’ (narrative aorist). Since in the standard narrative, and particularly in the representation of historical events, which are not supposed to be related to the speaker, and which are presented as objective statements, the use of the aorist allows events to speak for themselves.7 This contrasts with the perfect (and its present relevance) which suits subjective discourse and personal assumption by the speaker. The aorist is the form par excellence in disjunction from the speaker and time of utterance, so that the process is not represented as validated by the speaker but ‘‘auto-validated’’ (de Vogu¨e´ 1995), particularly fitting for an objective representation of past. According to this radical di¤erence, we can propose the following topological representation of both tenses (adapted from Descle`s, 1991): (5)

topological representation: of the aorist ---[////]T1 -----To of the perfect ----[--- ](///// T1 ----- To

Whereas the Hindi perfect can have experiential connotations (vah amerıˆkaˆ gayaˆ hai hai ‘He has been to America: He knows what it is like’) and resultant connotations (vah baˆzaˆr gayaˆ ‘He has gone to the market: He is not home, you cannot see him’8), the aorist cannot. For the same reasons the perfect, which has current relevance, is selected for summarizing the results of past processes, as in (6), though the various processes (‘do’ in the perfect, ‘provoke’, ‘cause’, ‘work’, ‘drive’ in the aorist) refer to the same temporal strata. The first two predicates take stock of what happened, the three following give the detailed sequence of events, in a non-chained enumeration leading to the initial balance-sheet. (6)

bahut kar liyaˆ hai aˆpne; gaˆnv men kyaˆ kuch kam much do take PFT you-ERG village in what anything kiyaˆ hai ? bhanDaˆriyon ko raˆvton se bhiRvaˆ less do PFT Bhandari ACC Ravat ABL split diyaˆ; khandhuriyon aur jyoshiyon men pushtainıˆ-dushmanıˆ give-AOR Khanduri and Jyoshi in enduring-hostility karvaˆ di, aur kuch na banaˆ, to kha-ba do-CAUS give-AOR and nothing be-made-AOR then kha-ba kaˆ hıˆ cakkar calaˆ diyaˆ . . . of just turn drive give-AOR

76

Annie Montaut ‘You have done (PFT) a lot; what haven’t you done (PFT) in the village? You provoked (AOR) a clash between the Bhandaris and the Ravats; you caused (AOR) enduring hostility between the Khanduris and the Jyoshis, and since/when nothing worked (AOR), you conducted (AOR) personal quarrels.’

Most of the above meanings of perfects and aorists (1–6) were noted long ago—whatever the labels used for each form—but we still lack a systematic explanation of the system. What is the reason why the simple form kiyaˆ represents anterior events, eventuality, general truth and other non-past meanings? To say it is an aorist answers the first part of the question (1–6), since it is expected from a standard aorist to map precisely this configuration of meanings. But there are less frequently mentioned meanings which may help better characterize the function of this tense. 2. Special meanings of the aorist: mirative extensions The following meanings are mainly found in oral interaction, most of them with a quasi-exclamative intonational pattern. We may broadly assign them to four categories: surprise, argumentative and polemic, saliency, and subjective intensity, although there is overlap. 2.1. Surprise Example (7) occurs in the context of a couple and their fifteen-year-old son visiting their old friend after a long time. The friend hardly recognizes the boy whom he had known as a child: (7)

Are! kitnaˆ baRaˆ ho gayaˆ ! (?* ho gayaˆ hai) Hey! how-much tall be go[become]-AOR (?* be go-PFT) ‘My! How tall he has become!’

The speaker confronted with an unexpected fact or situation (here the size of the boy) uses the aorist and not the perfect or present, which would only mean a neutral statement (less natural in the context and intonational pattern here).9 Similarly, when some expected event happens at an unexpected moment (earlier or later), as in (8) for delayed guests almost no longer expected: (8)

aˆkhir aˆ gae (*aˆ gae hain) mahemaˆn! finally come go-AOR (*come go pft) guests ‘Here they are finally, our guests’ (French Les voila`)

Mirative meanings as extensions of aorist in Hindi/Urdu

77

The above examples are found with the inferential perfect in Nepali, the form (echa) used to represent inference but also surprise. They are classical mirative extensions of evidential markers in languages with evidential markers, and can be explained by a contrast between what is expected P 0 and what occurs P, a contrastive validation of P rather than P 0 , possible and distinct from P (Donabedian 1996; Donabedian and Bonnot 1995). Languages which express such extensions through a special perfect (Armenian, Nepali) typically use the space adjacent to the time of utterance in the perfect (cf. 5). This space allows the speaker to comment on, approve or disapprove of, the event retold, and this is why so many languages derive evidential markers from perfects. The aorist, and not the perfect, conveys such contrasts in Hindi/Urdu. Similarly, the Hindi aorist is used in many occurrences where languages with an evidential system use evidential markers. 2.2. Argumentative and polemic use In (9), there is a first (neutral) statement in the perfect aˆ gayaˆ hai, and a requalification of the statement in the aorist aˆ gayaˆ. This requalification adds an emphasis which is a personal judgement in contrast with the bare statement, here aiming to counter speaker A who accepts the situation, and to express stronger disapproval. (9)

A: ghor kaliyug aˆ gayaˆ hai : chokraˆ kahtaˆ hai, dire iron-age come go[become] PFT young says, low Dom bıˆt sabhıˆ to baraˆbaˆr haiN. castes all but equal are. B: He raˆm! kyaˆ zamaˆnaˆ aˆ gayaˆ! Hey Ram! what time come go [become]-AOR A: ‘The worst of dark ages has come: Youngsters say, ‘‘Low or high cast, all are equal.’’ ’ B: ‘Hey Ram! What times we are in!’

Similar contrastive requalifications often result in polemical meanings such as (10), an almost lexicalized insulting expression conveying extremely aggressive connotations baRaˆ aˆyaˆ /baRıˆ aˆyıˆ ‘big/great come-AOR’, occurring as a requalification of tuˆ ‘you’ speaker B, an expression which is always in the aorist although always with present relevance. Note that the aorist in the first sentence of (10) is of the intensive type, allowing subjective emphasis (perceived as stronger than a perfect sharam nahıˆn rahıˆ hai). (10)

A: Behiyaˆ, terıˆ aˆnkhoN meN sharam nahıˆN rahıˆ ? Shameless, your eyes in shame neg remain-AOR

78

Annie Montaut B: A:

A: B: A:

MaˆN, apnıˆ hizzat apne haˆth hotıˆ hai. Mother, self honour self hand be pres Arıˆ, chal! baRıˆ aˆyıˆ mujhe dhamkiyaˆn denevaˆlıˆ. Hey, go! big come-AOR I-DAT threats give-er (*baRıˆ aˆyıˆ hai, *baRıˆ aˆ rahıˆ hai)10 (*great come PFT *great come PROG PRS ‘Shameless, and you aren’t even ashamed!’ ‘Mother, our honour is in our hands.’ ‘Get out of here, who are you to threat me (what a high horse you are riding).’

Examples (9) and (10) can be explained by the same operation of contrast as (7–8), since a requalification necessarily involves two distinct notions, the initial statement or thought and its rephrasing. 2.3. Saliency The contrast in (11) is not so obvious, when a speaker hearing a knock at the door opens it and finds that a totally unexpected person has come. This is a very common expression found in Nepali (11b) as well as in Hindi (11a). (11)

a.

are! kaise aˆe? Hey! how come-AOR ‘How come you are here?’ / ‘What’s up?’ / ‘What a surprise!’

b.

timıˆ po rahechau (Nepali) you but be-inf.pft ‘Here you are, how comes?’ (Fr. C’est toi que voila`: Michailovsky) (Clark 1963: 247)

(11) certainly expresses surprise, but it does not represent a fact as contrasting with some previous expectation, unless we admit as the contrasting frame an initial undefined situation. The statement in the aorist then would contrast with the mere indeterminacy of the previous state. Significantly, such statements simply representing a salient event occur in a kind of blank situation and do not so much deliver information as comment on its salient (exciting, interesting) character, as in (12). The first exclamation ‘Sun is here!’ in the aorist occurs as an expression of sudden awareness, not yet integrated into the store of knowledge. When it is integrated in the speaker’s conceptual framework (second statement), it is expressed in the perfect. In the context, it is integrated as a fact allowing for roof reparation after the rains, that is to say it is reframed into a causal logic, entailing other processes: the aorist is ruled out.

Mirative meanings as extensions of aorist in Hindi/Urdu (12)

79

‘‘dhuˆp nikal gaıˆ! dhuˆp nikal gaıˆ!’’ kıˆ aˆvaˆz se ‘sun get-out go-AOR sun get-out go-AOR’ of voice by maiN ekdam uTh baiThaˆ thaˆ. maˆlik ko batlaˆ I at-once get-up sit PPFT landlord DAT say duˆN dhuˆp nikal gaıˆ hai, kal tak kaˆrıˆgar give-SUBJ sun leave go pft, tomorrow till worker bhijvaˆ deNge? send-CAUS give-FUT ‘ ‘‘Here is the sun, here is the sun (aor)!’’ Hearing this I got up at once. Should I tell the landlord that the sun has come out (pft), (that) he may have the workers sent by tomorrow?’’ ’

Suddenness of awareness is the feature central to the Turkish evidential mis according to Aksu-Koc¸ & Slobin (1986). DeLancey (1985) too cites it as the reason for selecting the Tibetan inferential marker red, fit for new information which has not yet been integrated into the speaker’s conceptual marker (as opposed to yin/yod, fit for well integrated information). As has been pointed by Bonnot (2004) with respect to Russian statements with non-final accent in similar contexts (sputnik letit11 ‘Look, a sputnik flies’), the statement here is given as a preconstructed relation. With the terms preconstruct and preconstruction, borrowed from Culioli’s (1999) enunciative theory, I am not referring to presupposition, but to the fact that usually a well formed statement has to be constructed as such, through an operation of localization (or situation) of the predication in relation to a localizer (the situation involving in a more or less direct way a relation to the speaking subject S)12. In statements such as (12), the predicative relation hsun
risei is given as a block already constructed (outside S)—no pause is possible— and not as an information being constructed by S. It is a comment on a fact, saying that it is remarkable, rather than the construction of the representation of the event. This relation taken as a whole is in disjunction (Culioli: coupure, rupture) with the indeterminacy of the initial situation. Significantly, French translations would tend towards nominal structures rather than predications (which predicate something about something): here the purpose of the speaker is to not predicate something about something but to point to a fact as remark date. Examples (13a) and (13b) in Hindi/Urdu too express simple saliency. This is not real information, but a hint towards the interest/saliency of a fact given as a whole, suddenly coming to consciousness, as if out of the blue, for both speaker and hearer (unprepared mind). (13)

a.

are dekho, karghosh niklaˆ ! hey look-imper, rabbit leave-AOR ‘Look at that! (There is) a rabbit coming out!’ (Fr un lapin qui de´boule!)

80

Annie Montaut b.

aˆhaˆ! kasto raˆmro pokhrıˆ rahecha (Nepali) ah! what beautiful lake be-inf.pft ‘What a beautiful lake.’ (Clark 1963: 244)

One can discuss whether or not there really is a contrast between the previous indeterminacy of the situation (which could have been characterized by di¤erent possible relations P 0 ) and the new fact. What is clear is the meaning, akin to open focalization (an operation involving extraction, hence the intuition of contrast although non restrictive) and therefore to exclamation. The following set of values is close to Aikhenvald di¤ered realization: ‘‘post factum interpretation of a fact that they may have observed in some way.’’ (2004: 209). This typically occurs with cognitive verbs like understand, remember or forget, when the process occurs as a sudden flash not yet processed by the mind, like an Eureka exclamation. (14)

a.

i.

ii.

b.

i.

ii.

c.

i.

ii.

d.

samajh gayaˆ! understand go-AOR ‘Now I understand, get it.’ (Fr C ¸ a y est, j’ai pige´.) tumhaˆrıˆ baˆt samajh rahaˆ huˆN your speech understand progr prst ‘I understand your argument.’ jay maˆtaˆ kıˆ! chhatrıˆ bhuˆl gayaˆ! glory mother of umbrella forget go-AOR ‘By god! I forgot my umbrella!’ yaˆd nahıˆN maiN sab kuch bhuˆl gayaˆ huˆN memory neg I everything forget go pft ‘Now I have forgotten all and everything.’ yaˆd aˆyaˆ! vahıˆ laRkaˆ hai, jo. . . . memory come-AOR this-same boy is who . . . ‘Now I remember! This is the very boy who . . .’ bacpanaˆ yaˆd aˆyaˆ hai /aˆ rahaˆ hai childhood memory come pft come prog pst ‘I remember my childhood.’

tyo kaˆgat ta birsen bhaneko ta, khaˆltıˆ maˆ this paper top forget-AOR having said top, pocket in pa haˆlechu (Nepali) but put-inf.pft-1s ‘I thought I had forgotten this paper, but I have put it in my pocket.’ (sudden discovery)13

While Nepali (14d) is clearly a di¤ered realization, (14a.i) does not seem to involve any contrast nor even post factum realization (in questions, this

Mirative meanings as extensions of aorist in Hindi/Urdu

81

form may convey overtones of threat, sudden intimation) because the realization is immediately contemporaneous to the fact. And while (14b.i) allows for a di¤ered realization of a previous fact (gap possible), this does not seem crucially relevant compared with the sudden illumination in disjunction with the blank state of mind of S. S is in a way confronted with a fact (pre-constructed relation) which simply occurs to him and that he cannot process as a conscious process or elaborate (whereas in the present or perfect the speaker constructs the relation and states it as a processed information). Hence the e¤ect of unwillingness, non-participation of S, undeliberateness. Similarly, psychological/feeling predicates often occur in the aorist to express a high degree of inner emotion in almost lexicalized expressions (15) instead of a progressive present or perfect which gives a flat matter-of-fact statement. (15)

mazaˆ aˆ gayaˆ fun come go-AOR ‘great fun, I do enjoy it so much!’ baRıˆ kushıˆ huıˆ big happyness be-AOR ‘so pleased’ kamaˆl ho gayaˆ miracle be go-AOR ‘that is a miracle (fantastic!)’ mar gaıˆ die go-AOR ‘that is the end, I am finished’

Here the speaker is overwhelmed by the feeling, as if discovering it at the same time as he expresses it, hence expressing it by means of a (somewhat externally) preconstructed relation, more akin to a ready-made result than to a dynamic construction (hence the translations by nouns or nominal predications). The widely used kyaˆ huaˆ when A is shocked by the appearance of B (what be-AOR; ‘What is the problem?’), as opposed to the perfect kyaˆ huaˆ hai ‘What has happened?’, a more neutral question asking for information, also points to the emotional implication of the speaker rather than asking for information.

3. Conclusions To sum up, most of the mirative meanings which are associated with the inferential perfect of Nepali are associated with the aorist in Hindi/Urdu, except inference (Michailovsky 1996: 113) and high degree (intensive) in subjective predicates. In both languages, the putative (hypothetic inferential in Nepali) is expressed by a di¤erent device (future of copula, hogaˆ/holaˆ), hearsay by still a di¤erent device. Michailovsky concludes that the mirative

82

Annie Montaut

category is crucial to the evidential system—as do Slobin and Aksu (1982) for Turkish—further proposing that the extension to hearsay in some languages like Turkish is the result of a further grammaticalization of evidentials and in no way their core meaning. Hearsay, ignored by languages with less grammaticalization of evidentiality, such as Nepali, where the narrated fact remains asserted as true and is never questioned, is the far end of the grammaticalizing chain of the inferential perfect (Michailovsky 1996: 119). The Hindi/Urdu data provide further arguments for mirativity as a distinct category in its own right. Now the question is: Why should an aorist get mirative extensions? The Nepali perfect (past participle < Sanskrit passive past participle þ copula) developed inferential and various mirative meanings (surprise, sudden awareness, saliency, di¤ered realixation), while a derived perfect (þ nominalizing ko su‰xed to the participle) conveyed the standard meaning of perfect, both distinct from the aorist. The ko distinction, which occurs with progressives and pluperfect, is meaningful only in the perfect, with present auxiliary: the older (shorter) form, pertaining more to the action than the more stative new form, came to acquire the meaning of present awareness of a past event (Peterson 2000: 23.24). The relation of perfects and evidentials or miratives has been explored enough not to require further comment. But the Hindi aorist does. Since no other form was available in the language, mirativity could only be attached to a non-specific marker and therefore is more easily ignored in descriptions. But why to the aorist and not the perfect? There may be a couple of reasons. One is related to the nature of aorist itself, the other to its morphology and origin. The aorist per se is a non-tensed form, whose specific feature is the radical disjunction from speaker and time of occurrence (which confers to it the ability to be used in non-assertive statements such as hypothesis, with no independent localization, and in narratives, to fit the historical objectiveness of events speaking ‘‘for themselves’’). Being radically disjuncted from S, it is also fit for expressing relations that occur as pre-constructs to the speaker, as if S himself takes no part in the construction and does not use his own spatiotemporal correlates to construct a predication. Here it suggests a path similar to the one described by Aikhenvald as non-participation of the speaker > distancing e¤ect > surprise. All the examples in section 2 are perceived as ‘‘intensive’’, intonationally close to exclamations, thus highly subjective but without intellectual implication of S. Confronted with a fact that he has not processed, which is given as a whole preconstructed relation, and which he is unable to integrate (construct as a predication), the speaker utters a simple comment (hence the emotional tones) rather than delivers information. The ‘‘suddenness of his awareness’’ or ‘‘immediacy’’ (Michailovsky 1996) paradoxically compels S to resort to not-really assertions,

Mirative meanings as extensions of aorist in Hindi/Urdu

83

nominal like sentences with no assumptive S, S being only a medium delivering a statement constructed outside of him, hence the tendency to lexicalize as idiomatic phraseology (14, 15). The form too, clearly nominal in Hindi where the aorist has no personal endings (as opposed to the Nepali aorist) and without tense-aspect mark (like Nepali), may be considered as best suited to expressing auto-validated facts, results presented as a static whole and not as dynamic constructs (hence the nominal translations in French, often using no predicate). Aikhenvald (2004) notes that nominals added to the verb and copula display mirative extensions in Kham. Etymologically the Sanskrit passive past participle first grammaticized into a resulting past in middle Indian and then into an anterior (in narratives). At that time the periphrastic perfect developed, but both kept their ancient syntactic meaning of static localizing predications, still materialized by the ergative construction (‘he did that’: literally by him that done in Hindi). This is a long story, very similar to the birth of modern Romance ‘‘possessive’’ perfects (Benveniste 1996), which were also initially localizing stative predications and not representations of ‘‘actions’’ (mihi id factum to-me this done, like Skr maya tat kritam of-me done and modern Indo-Aryan main-ne yah kiya by-me this done), too complicated to fully explore here (for further development, Montaut 1996). Pointing to the relevance of diachrony in descriptions, may be insightful both for labelling and comparing forms and for studying paths of grammaticalization in mirativity as well as other categories: Hindi shows the propersity of originally nominal forms to grammaticize into mirative extensions. But we have to know what is really a ‘‘narrative’’, a perfect, a perfective, an aorist in a given language (ie: we have to take into account all the meanings of the form and not only those required by pre-established models) before concluding to any cognitive paths.

Abbreviations ACC—accusative AOR—aorist CAUS—causative DAT—dative ERG—ergative FUT—future PFT—perfect PRS—present PROG—progressive PPFT—pluperfect

84

Annie Montaut

Notes 1. Or non-testimonial form, since it is widely used to related events not directly observed by the speaker (reported events). Aikhenvald (2004) however rightly points to the fact that markers for evidentiality relate in many languages to the mode of perception of the event (directly observed, heard, felt, even smelled) rather than to the distance from the direct source of information. 2. Apart from the use of the future of the verb hogaˆ ‘be’ as an auxiliary for marking presumption or probability (hence the terminology of presumptive or putative, sometimes used to define it). Note that the use of hogaˆ is limited to cognitive inference (not requiring a material trace or observed event to infer from it the past event). 3. The present verbal paradigm as described in Clark (1963), Michailovsky (1996: 110–111), and Peterson (2000) in fact strongly echoes the Hindi system in the nineteenth century: a synthetic form still expressing non past (present-subjunctivefuture) with a concurrent periphrastic present (unaccomplished participle þ copula) and still expressing both general and progressive meanings, with a concurrent periphrastic progressive (radical þ rah-‘stay’ þ copula). The su‰xing of ko to accomplished participles is worth discussing: do we consider it as a genitive su‰x (Clark, followed by Michailovsly and Peterson) or do we relate the form to the ‘enlarging su‰x’ –kaˆ widely used in Indo-Aryan after a nominal base (homonymous in Hindi with the genitive marker too). For –o/–aˆ alternation, we find the –o ending equivalent to masculine singular Hindi –aˆ in all Pahari dialects. 4. The choice of the term ‘‘accomplished’’ instead of the more usual ‘‘perfective’’ in the local tradition, will become clearer in the next paragraph. The opposition accomplished/unaccomplished has been used for a long time, particularly to oppose both paradigms in Roman languages (present, imperfect vs present perfect, pluperfect, simple past) in many respects very similar to the Hindi system. It is used for instance by Cohen (1989) to describe other language aspectual oppositions, as well as Creissels (1995). 5. Apart from his numerous papers, one can refer to Nespital’s landmark Dictionary of Hindi Verbs (with Hardev Bahri, 1997). 6. The only di¤erence with the participle in the plural feminine (nasalized in the predicative use, un-nasalized in the participial use). Whereas in Nepali it adds personal endings to the participal form garyo. 7. Les e´ve´nements parlent d’eux-meˆmes. 8. Examples from Narang (1984), who calls the latter ‘‘inferential’’. 9. The use of the same form in the aorist would of course have a di¤erent meaning in a narrative: ‘He grew very tall.’ 10. baRaˆ aˆyaˆ, paNDit kaˆ bacchaˆ ‘you are riding such a high horse, son of a pandit (¼ faked pandit ¼ don’t pretend to know)’. 11. Whereas the neutral statement would bear a final accent (first syllable of the verb). Non-final accent in Russian occurs in the same contexts as the Armenian evidential perfect, except for hearsay and inference (Donabedian and Bonnot 1995). 12. For a more detailed explanation of the operation of localization (qualification, quantification), see Culioli 1999 (a few of the chapters are in English).

Mirative meanings as extensions of aorist in Hindi/Urdu

85

13. Cf. dhaˆNTyachan ‘(Now I realize) they lied.’ (di¤ered discovery) vs. bare statement (aor: DhaˆNTe, pft: DhaˆNTekaˆ chan).

References Aikhenvald, Alexandra Y. 2004 Evidentiality, Oxford: Oxford University Press. Aksu-Koc¸, Ahyan and Slobin, Dan 1986 A Psychological account of the development and use of evidentials in Turkish. In Wallace Chafe and Johanna Nichols (eds.): 159–167. Benveniste, Emile 1966 Proble`mes de linguistique ge´ne´rale. Paris: Gallimard. Bonnot, Christine 2004 Repre´sentation pre´construite et focalisation : pour une analyse unitaire des e´nonce´s a` accent non final en russe moderne, Slovo 30–31: 211–270. Chafe, Wallace and Johanna Nichols (eds.) 1986 Evidentiality: The Linguistic Coding of Epistemology. Norwood, New Jersey: Ablex. Clark, T. W. 1963 Introduction to Nepali. Cambridge: W. He¤ers & Sons. Cohen, David 1989 L’Aspect verbal. Paris: Presses universitaires de France (PUF). Creissels, Denis 1995 Ele´ments de syntaxe ge´ne´rale. Paris: Presses universitaires de France. Culioli, Antoine 1999 Pour une linguistique de l’e´nonciation. 3 vols. Paris: Ophrys. DeLancey, Scott 1997 Mirativity, the grammatical marking of unexpected information, Linguistic Typology 1.1: 33–52. 1985 Lhasa Tibetan Evidentials and the Semantics of Causation, Proceedings of the Eleventh Annual Meeting of the Berkeley Linguistics Society (BLS) 11: 65–72. de Vogu¨e´ Sarah 1995 L’E¤et aoristique. In Langues et langages. Proble`mes et raisonnement en linguistique, J. Bouscaren, J.J. Frankel, and S. Robert (eds.), 247–259. Paris: Presses universitaires de France (PUF). Descle`s, Jean-Pierre 1991 La pre´dication ope´re´e par les langues (ou a` propos de l’inte´raction entre langue et perception). Langages 103: 83–96. Donabedian, Anaı¨d 1996 Pour une interpre´tation des diverses valeurs du me´diatif en arme´nien occidental. In Zlatka Guentche´va (ed.), 87–108. Donabedian, Anaı¨d and Christine Bonnot 1995 Lorsque la morpho-syntaxe rencontre la prosodie: accent non final en russe et me´diatif en arme´nien, Faits de Langue 13: 182–190. Guentcheva, Zlatka (ed.) 1996 L’Enonciation me´diatise´e. Louvain: Peters. Hook, Peter E. 1974 The Compound Verb in Hindi. Chicago: University of Michigan Press. Kellogg, Rev. S. H. 1972 Reprint. Grammar of the Hindi Language. Delhi: Orient Reprints. Original 1875.

86

Annie Montaut

Michailovsky, Boyd 1996 L’Infe´rentiel du ne´palais. In Zlatka Guentcheva (ed.), 109–123. Montaut Annie 2004 Hindi Grammar. Mu¨nchen: Lincom Europa. 2003 On the evidential meanings of the Hindi aorist. In Tohfa-e-dil, Festschrift Helmut Nespital, Dirk W. Lo¨nne (ed.), Vol. 1, 345–364. Reinbek: Verlag fu¨r Orientalistische Fachpublikationen. 1996 La Gene`se des syste`mes aspecto-modaux en indo-aryen moderne, Journal Asiatique 284 (2): 325–360. Narang, Vaishna 1984 Communication of Situation/Time in Hindi. Delhi: Garimashri Prakashan. Nespital, Helmut (with Hardev Bahari) 1997 Hindi Kirya Kosh. Allahabad. Lokhbarati. Peterson, John 2000 Evidentials, Inferentials and Mirativity in Nepali. Linguistics of the Tibeto-Burman Area 23-2: 13–37. Porizka, Vincenz 1968 On some Aspectual Functions of Simple Verbs and Verbal Expressions in Hindi. In Readings in Hindi-Urdu Linguistics. K. S. Singh (ed.): 158–164. Delhi: National Publishing House. 1982 On the Aspectual Contrast in Hindi. In Topics in Hindi Linguistics, Omkar N. Koul (ed.), Vol. 2: 122–134. Chandigarh-Delhi: Bahri Publications. Slobin, Dan I. and Ahyan A. Aksu 1982 Tense, Aspect and Modality in the Use of Turkish Evidential. In Tense-Aspect: Between Semantics and Pragmatics, Paul J. Hoper (ed.), 185–200. Benjamins: Amsterdam/Philadelphia.

Some aspects of the meaning of the Hindi particle hii TERRY VARMA

This paper is an investigation into a range of meanings of the Hindi particle hii. It is shown that many of these meanings can be related to those of exclusive focus particles in other languages and that these di¤erent meanings of hii can be understood within the parameters of focus and exclusion. Meaning is shown to vary according to what is being focused on, what alternatives are excluded, whether the context is scalar or not and what sort of scales are involved. hii is also shown to play the role of a discourse marker of preclusion and to be involved in the marking of epistemic and deontic modality. 1. Introduction Hindi hii is a versatile particle which plays a variety of di¤erent roles in the language. However, little work has been done on the meaning of this particle and none has yet been devoted to its study. Dictionaries and grammar handbooks present a confusing picture with a plethora of terms being used and numerous diverse meanings being suggested.1 In this paper I discuss a number of meanings of hii in some depth and I will show that many functions and uses of hii can be shown to be interconnected in a systematic way, based on a few parameters of meaning. These are focus, alternative sets, exclusion and scale. These features of meaning of hii are exploited in di¤erent ways in a variety of important areas of the language such as information structuring and the expression of aspect. This set of meaning parameters also places hii with the exclusive focus particles, a linguistic category that has already been established and compared across languages (see in particular Ko¨nig 1991). However, it will be seen that the range of focusing functions of hii only partially overlaps with the descriptions of exclusive particle meanings in other languages. It is also argued that not all functions and meanings of hii can be satisfactorily accounted for within the exclusive focus particle parameters, hii also has modal and discourse marking functions. In the next section of the paper I will discuss hii in the context of information structure. In the third I make a proposal to distinguish the cases

88

Terry Varma

where hii can be considered equivalent to the exclusive focus particle only in English and where it cannot. Following this I present the function of hii as a marker of central importance or an extreme, that is, as an intensifier. In the fifth part I bring to attention cases of hii as a marker of an expression of identity. In the sixth I will argue that hii plays a role in various types of verum focus in Hindi. Next I investigate the role of hii and time and space scales, arguing that hii can be a phase and aspect marker. Following this I propose that hii in addition functions as a discourse marker of preclusion. In the last section of the paper I discuss the role of hii in marking epistemic and deontic modality.2 2. hii and information structure In regard to the role of hii in information structuring, two types of focus need to be distinguished: the focus which demarcates the utterance into focus and presupposition, and a second type which evokes alternatives to the expression which is in focus. The domain of the first meaning is information structuring. The second is the meaning which is involved when focus is associated with a focus particle. It is defined by Ko¨nig (1991) in the following manner: ‘‘a focus establishes a relation between the value of a focus expression and a set of alternatives’’ (p. 32). Focus particles indicate the relation of the expression to the alternatives, which in the case of hii, is one of exclusion. I will first clarify some points regarding information structure relevant to my analysis before discussing the contribution made by the focus particle hii. The distinction between focus and presupposition seems to be an essential feature of information structure, marked in all languages, though it may be expressed in di¤erent ways. The most common is for the focus of new information to be highlighted in some way through prosodic prominence. For example in: (1)

He caught the FIVE-FIFTEEN to Madras.

Five-fifteen is prosodically marked as the information focus of the utterance, and the rest of the sentence is marked as presupposed. The focus marks the new informational contribution and the presupposition is presented as given or shared information (note that focus is shown by the use of capital letters in English, for the Hindi examples I will use boldface). Some examples from Hindi will set the scene. In Hindi too, prosodic prominence is used to distinguish focus from presupposition. (2)

a.

aaj mAA ne lalitaa kaa acaar khaayaa. today mother erg Lalita poss pickle eat-perf ‘Today MOTHER ate Lalita’s pickle.’

Some aspects of the meaning of the Hindi particle hii b.

aaj mAA ne lalitaa kaa acaar khaayaa. ‘Today mother ate LALITA’S pickle.’

c.

aaj mAA ne lalitaa kaa acaar khaayaa. ‘Today mother ate Lalita’s PICKLE.’

89

Change of stress placement results in a change in information structure. For example, in (2a) someone having eaten Lalita’s pickle is part of the presupposition and mother is the focus of new information. In (2b), however, the speaker takes it to be common ground that mother ate someone’s pickle and the new information is that it was Lalita’s. I will refer to the focus of focus and presupposition as the information focus of the sentence (IF) in recognition of its role in information structuring or packaging, and to distinguish it from the other type of focus associated with a focus particle. This type has no general label but in the context of information packaging ‘contrast focus’ is a commonly used term and I will use it in this section for convenience. The prosodic marking of the IF is not without its ambiguities. In (2c) for example, highlighting the final part of the direct object may lead to three di¤erent focus construals. IF can be checked for by reconstructing the whquestion that the sentence in which it appears would answer, the IF supplies the missing information asked for by the question. (2a), for instance, would answer the question: Who ate Lalita’s pickle? The sentence (2c), on the other hand, could be the answer to three di¤erent questions: 1) What did mother eat? Answer: Lalita’s pickle (IF); 2) What did mother do? Answer: (she) ate Lalita’s pickle (IF); 3) What happened? Answer: Mother ate Lalita’s pickle. In the latter case the whole sentence constitutes the IF and there is no presupposition indicated, except possibly, that something happened. There are two things to notice at this point. The first is that the default construal of information structure is what has been called the predicate focus (PF) type (Lambrecht 1994), where new information is predicated of a given argument. This is one of the readings of (2c) where it is predicated of mother that she ate Lalita’s pickle. Special marking may be needed to prevent inappropriate PF construal. One example of this sort of noncanonical marking is the English it-cleft construction. (3)

It was Lalita’s PICKLE that mother ate.

Here a PF construal is out of the question. The second thing to observe is that the standard word order in Hindi is SOV (with a number of caveats). But word order is quite flexible and word order may also be used as an indicator of the focus-presupposition partitioning of the sentence. In particular, constituents may be focused by

90

Terry Varma

displacing them to the pre-verb focus position (Masica 1993). (4) for example has the reading where ‘mother’ is the IF: (4)

lalitaa kaa acaar mAA ne khaayaa. ‘MOTHER ate Lalita’s pickle.’

A further layer of information structure that will be relevant in this paper concerns the level of familiarity or the discourse status of the elements in the sentence for the speaker and hearer. Various items may be marked in some way to indicate this status, and some structures of the language may be sensitive to this status, various cleft constructions in English are again an example (Prince 1986; Ward and Birner 2004). An item in a sentence may be discourse-old or discourse-new. Discourse-old are items whose referents the speaker assumes to be known to the hearer either because they have been previously evoked in the discourse or because they can be inferred from what has been evoked. A further category is that of hearer-old items. These the speaker assumes the hearer knows or can infer independently of the current state of the discourse. Looking back at (2a) or (4), we can see that newness in the focus-presupposition sense need not overlap with newness in the familiarity sense. As we observed, here mother is the new information but mother may be very much discourse-old. In fact mother may have a high level of saliency, which is to say it may be active in the hearer’s consciousness or short-term memory at the moment of utterance, having just been mentioned, say. To return now to the two types of focus, information focus and contrast focus, it has been noted in the literature that these may often be intertwined. In the Hindi example (2b), repeated here for convenience as (5), (5)

aaj mAA ne lalitaa kaa acaar khaayaa. ‘Today mother ate LALITA’S pickle.’

we saw that with the prosodic highlighting on Lalita it is clear that the sentence is not to be given a PF construal. The preferred construal is now what Lambrecht (1994) terms argument focus, or AF construal. In AF sentences we find a given open proposition with a variable (x) and the IF instantiates the variable, i.e. it provides the missing information represented by the variable. The open proposition for (5) is: mother ate x’s pickle. Lalita is the instantiation of the variable, the IF. But in addition to marking the IF status of Lalita, the prosodic stress here may imply contrast focus. It may imply alternatives to Lalita—other people’s pickle that mother could potentially have eaten, given the discourse context. Furthermore, this focusing may also suggest that the alternatives to Lalita are excluded.3 In the first of the hii functions that I will

Some aspects of the meaning of the Hindi particle hii

91

describe, both types of focus are also intertwined but the contrast meaning is actively established with hii, a meaning which is only implied in the case of prosodic accent. It also gives a further, clearer indication that we are dealing with a non-canonical, i.e. a non PF information construal of the sentence. Take for example: (6)

sTeshan par sonuu ne hii paapaa ko pahacaanaa. station on sonu erg hii dad acc recognize-perf-m sing ‘At the station it was SONU who recognized Dad.’

Here hii is associated with an element which is the IF of the sentence. hii focusing evokes potential alternatives to the IF Sonu, the other people who went along to the station to fetch father in this case. It also clarifies that it is being presupposed that somebody recognized Dad and the new information is that this was Sonu. This means that we are dealing here with an AF and not a PF information structuring. Even in a written text, with no indication of the placement of intonational prominence, it is clear from the hii that Sonu represents the new information. This is because hii is placed adjoining the end of the information focus constituent. I do not wish to go into detail regarding the position of hii.4 But a few comments may be in order. hii normally stands after the constituent which is in focus. This in turn means that hii, like other focus particles (Ko¨nig 1991), shows considerable positional variability in the sentence. It never stands alone, nor does it ever appear sentence initially. As we shall see later, in association with an element in the verbal complex it usually appears within the verbal complex, not at the end of it, with exceptions such as verbs in the future tense. This means that, since Hindi is an SOV language, hii does not often appear in sentence final position. To sum up so far, hii follows a focused element, which in these cases is the IF element in the sentence, and evokes alternatives to the expression in focus. It also indicates that the alternatives are excluded, it was Sonu in (6), not the others, who recognized Dad. I will call this type of focus which involves these three elements contrast IF. For contrast IF to be used felicitously I also claim that there are requirements on the discourse/hearer status of various elements in the sentence. To appreciate this, obviously more of the context is needed than the sentence itself. For this reason, in this section I have taken care to use attested spoken and written examples where the context is known. The requirements referred to involve the expression in focus, the set of variables evoked, and the open proposition, which need to be discourse-old or hearer-old. The presence of felicity requirements involving the status of these elements is reminiscent of the requirements which have been studied in information packaging in relation to non-canonical information constructions in

92

Terry Varma

English such as preposing, and various cleft constructions (Lambrecht 2001; Ward and Birner 2004). In fact hii can often be suitably translated by a cleft and I will often do so in the examples. However, it is also clear that a cleft is not always a good choice and it will take further, closer comparison with contrast IF to discern why and when it is suitable.5 Let us proceed then to see some of these points in action with examples. The most straightforward cases are where the open proposition, the focused expression and the alternative set are salient because they have just been mentioned. Here is a typical example of hii. The sentence is spoken by the domestic help who wants to know how many cups of tea she should make. There is confusion, at issue is whether one or two cups should be made, there are just two people at home. Finally she asks: (7)

do hii kap banaaUU ? two hii cups make-subj-1psing ‘Should I make TWO cups?’ / ‘Is it TWO cups that I should make?’

In this sentence, the expression in focus is discourse-old and active, so is the alternative, one cup, and so is the open proposition: x kap banaaUU. In the next example, a written one, the writer is reflecting on the subject of the murder of old people, why people do it and how the involvement of servants is often suspected. So in the context, the open proposition, the expression in focus and the alternative set have just been mentioned. (8)

zyaadaatar buuRHO ko jaayadaad ke laalac mE hii most often old person-pl-infl acc property gen greed in hii maar diyaa jaataa hai. kill lv—perf p-ms go-impf-ms be-aux-pr-3ps ‘Mostly it’s because of greed for PROPERTY that old people get killed.’ (NKD)

In the case of a newspaper headline, where the question of immediately previous mention does not arise, the writer can rely on the reader’s knowledge of past articles or current knowledge, as in the following recent example: (9)

briTen ne bhii kahaa, hamale ke lie Laaden hii Britain erg too say-perf attack-infl for Laden hii jimmedaar. responsible ‘Britain too has said that LADEN is responsible for the attack.’ (J)

Some aspects of the meaning of the Hindi particle hii

93

Here the open proposition, the expression in focus and the alternatives to the expression in focus are well known, especially in the context of the news media. It should be noted again that the hii leaves no room for a PF reading, that something is being predicated of Laden. In other cases it can be that a smaller or bigger inferential leap needs to be made to accommodate the open proposition as discourse-old. To illustrate this consider (10). In this written example a girl hears the expression candaa maamaa. candaa maamaa is the common way to refer to the moon when talking to small children and it means Uncle Moon. At that moment the girl starts wondering: (10)

kyaa gaarATii hai ki candaa maamaa hii what guarantee be-pres that moon uncle hii hAI, maamii nahII hAI, daadii nahII be-pres-3pl hon aunty neg be-pres-3pl-hon grandma neg hAI maamaa hii hAI ? be-pres-3pl-hon uncle hii be-pres-3pl-hon ‘What is the guarantee that it’s UNCLE moon, not aunty or grandma moon?’ (C)

Here the focused expression has just been mentioned, the alternative set immediately evoked is family members, which is spelled out in the next part, and the open proposition—it (the moon) is x Moon—is inferable from the discourse context. The child wonders why alternatives have been excluded. In (11) also some inference will have to be made: (11)

par meraa phaayadaa sessions hii mE hai, na? but my benefit sessions hii in be-pres-3s no ‘But it’s the SESSIONS that are of use to me, right?’

The context for this is that there is going to be a conference and the speaker has been o¤ered a pass. There is a discussion of all the di¤erent o¤erings at the conference and it is revealed that the pass does not cover entry to the sessions. This was the remark made at that moment. Again the open proposition, x is of use to me, has not yet been mentioned but it is directly inferable from the discourse context that one goes to a conference for what use it is. The open proposition is discourse-old, the instantiation of the variable sessions is active and so are the alternatives, the other events at the conference. The familiarity requirements in the following written example rely on general or cultural knowledge: (12)

meri mAA ko mere pitaajii se my-f mother acc my-mpl-hon father-hon with

94

Terry Varma milaanevaale tumhaare pitaa introduce-inf-infl-nomin-mpl-hon your-mpl-hon father hii the. hii be-past-mpl-hon ‘It was YOUR FATHER who introduced my mother to my father.’ (KVB)

The speaker and the hearer have been talking about the marriage of the speaker’s parents. The familiarity of the open proposition relies on it being assumable that someone introduced the couple. The expression in focus is discourse-active though not just mentioned and alternatives can be construed as the usual likely people to do the introducing. The alternatives are excluded. In some cases the open proposition, the expression in focus and the alternative set may have been mentioned some time earlier, several pages in the following. The sentence is thus linked back to a previous discourse topic, in this case, the issue of who would light the funeral pyre: (13)

aag surendar bijalii vaale ne hii dii. fire-f Surendar electricity nomin erg hii give-perf-fsing ‘It was SURENDAR THE ELECTRICIAN who lit the funeral pyre.’

Sometimes it takes more inferential processing to relate the open proposition to the current discourse status than going further back in the discourse: (14)

unakaa hii upanyaas (. . .) vahAA mere her-hon-msing hii novel-m there my-msing-infl rikshe ke siiT ke niice paRaa rickshaw-m-infl poss-msing-infl seat beneath lie-perf p thaa. be-past-3psing-m ‘It was HER novel that was lying under the seat of my rickshaw.’ (KBN)

In this sentence the open proposition is: x’s novel is there under the seat of my rickshaw. The reader has not been informed of there being any novel under the seat. What we learned earlier is that this rickshaw puller is crazy about reading and reads before work, during work and after work. This is then considered by the speaker to be su‰cient to take the open proposition as discourse-old, as indicated by the hii. Again the information in the sentence is linked back to earlier discourse and also it is made clear that it is not to be given a PF reading.

Some aspects of the meaning of the Hindi particle hii

95

Finally I would like to look at two examples where it may be even harder to accommodate an open proposition as discourse-old. In (15) the writer has been speaking of the poisonous nature of lead: (15)

siise kii viSHaktataa ko dekhate hue hii haal mE lead-m-infl poss toxicity acc considering hii recently peTrol mE siisaa milaanaa band kar diyaa petrol in lead mix-inf stopped do lv-perf p-msing gayaa hai. go-perf p-msing be-aux-pres 3sing ‘It is BECAUSE OF LEAD’S TOXICITY that adding lead to petrol has been banned.’ (S)

It is arguable here that the reader may not be aware of the banning, that is to say the speaker may know that the hearer may not be aware of the open proposition. A second case where the open proposition may not only be not salient but also in no way knowable by the hearer is (16): (16)

vah shaayad apane muTaape kii kshatipuurti ke she probably own-msing-infl obesity gen compensation for lie hii baal chhoTe kaTavaatii thii, bagair hii hair-mpl short-mpl get cut-impf be-past-fsing without bAAh kii kamiizE pahantii thii aur (. . .) sleeve gen kameezes wear-impf-fsing be-past-fsing and (. . .) ‘It was probably TO COMPENSATE FOR HER FATNESS that she cut her hair short, wore short-sleeved kameezes, and (. . .)’ (MRSK)

At this point in the story the hearer has not been told about the lady protagonist’s hair or clothes and there seems to be no reason to imagine that they would be as described here. Is this then a matter of exceptions or of a di¤erent type of hii? In the case of English it-clefts, there is some controversy in the literature over sentences which pose similar problems, the most commonly cited example being: (17)

It was just about 50 years ago that Henry Ford gave us the weekend.

As Prince (1986) points out, (17) might well be uttered in a context where the speaker cannot assume that the hearer is aware of Henry Ford giving us the weekend. Prince posits the existence of a separate category of it-cleft, the informative-presupposition cleft, or IP cleft. Prince’s interpretation is that an IP-cleft marks what is considered to be the presupposition as a well known fact that the hearer just happens not to be aware of, in this way she is creating the impression that this is a generally accepted, non-controversial

96

Terry Varma

fact. On the other hand, on syntactic grounds, Lambrecht (2001) posits that there is no separate IP type of cleft, instead all clefts have in common that they specify the variable in a discourse or hearer-old open proposition. This fact or rule may then be exploited for rhetorical e¤ect as the hearer accommodates her interpretation to it. The feeling that IP are special is because they are more di‰cult to accommodate. They represent one end of a range of more or less easy to accommodate presupposed open propositions in itclefts. In either case it seems that a given presuppositional structure is exploited for pragmatic purposes in a way that has become a more or less conventionalized rhetorical device. Another possibility is that the same requirements do not apply to certain types of focus expression. In (15) and (16) for example, the focused expressions are both reason adjuncts, it is possible that such adjuncts are an exception. In English, for example, Ward and Birner (2004) suggest that locative preposing is an exception to the rule that preposing requires a salient open proposition. Whatever the case may be it is sure that the use of hii provides information as to the correct focus-presupposition construal of the sentence. Hooper (1901) notes the following example: (18)

us ne ham se giRagiRaake bintii kii aur he-infl erg we from entreat-conj p request do-perf-f and ham ne usakii na sunii, isii kaaraN ab we erg his-f neg listen-perf-f this-inflþhii reason now ham is sAkaT mE pare hAI. we this-infl trouble in lie-perf p-mpl be-pres-3ppl ‘He besought us and we did not listen to him, for THIS reason we are now in this trouble.’

His explanation is in terms of the rather vague notion of ‘‘emphasis’’. He says that if it had been is kaaraN ‘for this reason’ rather than isii kaaraN ‘for this-hii reason’, the emphasis would have been laid on the trouble instead of the trouble being assumed as evident and the statement being that the cause of it was their previous misconduct. I see this in terms of the focus and presupposition partitioning of the sentence. Since the contrast IF is constrained in its use by certain familiarity requirements it is not surprising that it is not generally used in an answer to a wh-question, although it is clear that in the case of a wh-question there is a salient open proposition waiting for the instantiation of the variable represented by the Q-word. The use of contrast IF further requires the discourseold status of the focus expression and alternative set. If these conditions are met then the use of hii in the answer to a Q-question becomes appropriate as in this attested example:

Some aspects of the meaning of the Hindi particle hii (19)

A:

A: B: B:

97

kaunsaa caahie? choTaa 20 watt kaa which-msing is wanted small-msing 20 watt gen hai aur baRaa 40 watt. be-pres-3sing and large-msing 40 watt ‘Which do you want? The small one is 20 watts and the large is 40 watts.’ baRaa hii THiik hogaa. large-msing hii OK be-fut-3msing ‘The LARGE one will be fine.’

To summarize this section, we can say that hii may play a role in the information packaging of a sentence. It interacts with the information focus in a sentence, sharpening the focus-presupposition distinction and indicating a non-PF construal. It points to the presence of discourse-old alternatives to the expression in focus and these alternatives are excluded. 3. hii and only In this section I will discuss the meaning of hii in the context of proposals in the literature concerning English only and similar particles in other languages. I will argue that we have here a related but di¤erent meaning of hii which overlaps with the first, as well as overlapping with the meaning of only. I will discuss two di¤erent interpretations of only and argue that the type of meaning defined for example by Ko¨nig (1991) is not adequate for an understanding of the di¤erent usages of hii. The overlap in meanings that I have referred to becomes apparent in an interesting and pervasive feature of Indian English: the overuse or misuse of the word only.6 Here is an example attested recently. A picks up a sweater and asks: (20)

A: B:

Is this yours, Neelu? No, aunty, it’s GAUTAM’S only.

Not only has only been wrongly placed after rather than before the expression in focus but also the use of only is inappropriate here. On the other hand, only is often a perfectly acceptable translation of hii, as in (21)

ve kavitaa hii likhate hAI. they poetry hii write-imp-mpl be-aux-pres-pl ‘They only write POETRY.’

Only has been described as an exclusive or restrictive focus particle which indicates exclusion of the alternatives in the evoked alternative set. In the

98

Terry Varma

words of Ko¨nig (1991), ‘‘a sentence with only presupposes the relevant sentence without the particle and entails that none of the alternatives under consideration satisfies the open sentence obtained by substituting a variable for the focus expression.’’ An example he gives is: (22)

Only JOHN came.

This meaning is illustrated in Hindi in (21). The similarities between this hii/ only meaning, which for convenience I will call the exclusive focus particle meaning and the meaning of contrast IF hii are obvious. Both are focus particles and both indicate exclusion of the evoked alternatives. What then is di¤erent between the two meanings of hii? A possibility could be that exclusive focus particles like only and its equivalents in other languages may have scalar interpretations. The definition given above is of what Ko¨nig terms the non-scalar meaning of only. He also distinguishes a scalar use in order to do justice to cases where the focus value and the alternative values can be ordered according to some sort of scale.7 In general in the case of only, the values that are excluded are higher on the scale than the focus value. Here is an example from Ko¨nig: (23)

I only bought THREE apples.

Since higher values are being excluded, the expression in focus is often at the same time evaluated as low or minimal. hii too may be scalar. It may evoke a set of variables ordered on a scale, excluding higher values to that of the expression in focus. In (21), for example, hii could result in a scalar reading where poetry and other forms of writing under consideration are ordered high to low according to some scale, say a scale of literary value, and where poetry is ranked low. It is important to realize that the scale on which the focus expression and the alternatives are ranked in relation to each other is highly context dependent and what may naturally appear high scale in one context can be considered low where the alternatives are even higher. In (24) for example a gold necklace can be considered low on a scale of jewelry when the alternative in mind was a diamond necklace: (24)

usne isko ek sone kaa haar she-infl-erg she-infl-dat a gold-infl gen-msing necklace hii diyaa. hii give-perf-msing ‘She only gave her a gold NECKLACE.’

It should also be noted that a naturally low scale focus expression with hii is not necessarily translatable by only as example (7) shows.

Some aspects of the meaning of the Hindi particle hii

99

The scalar exclusive particle usage of only coincides with that of hii. The scalar meaning, then, does not distinguish between only and hii, though it allows hii to always be translated appropriately as only when the reading is scalar and the focus expression is ranked as low. The translation problem does not arise in these cases of overlap. Moreover, the presence of scalar potential as part of the meaning of the exclusive focus particle hii, but not of contrast IF hii, distinguishes between these two uses. Let us go back then to the non-scalar meaning of the exclusive focus particle and the definition given in the discussion of (22). This definition is not adequate to distinguish between exclusive focus particles and contrast IF particles. Nor, then, does it help distinguish whether only is an appropriate translation for hii or not. If we look back at example (20), where the focus expression is Gautam and the set of alternatives consists of either Neelu or the rest of the potential owners of the sweater around at the time, hii clearly excludes all the alternative members of the set, all the alternative potential owners of the sweater. But only is not acceptable here. Or we can consider example (7) again, reproduced here as (25): (25)

do hii kap banaaUU ? two hii cups make-subj-1psing ‘Should I make TWO cups?’ / ‘Is it TWO cups that I should make?’

In the given context where the choice was between one or two cups of tea being made, all alternatives, that is ‘‘one cup’’, are excluded in the answer, yet an only here would be inappropriate. A possible alternative view of only sometimes found in the literature is that only should be considered to be some sort of contra-additive focus particle. Umbach (2004), for instance, points out that at first sight, the following examples may appear to be equivalent, both convey the information: Ronald went shopping and nobody else did. (26)

Things have changed at the Miller family. a. Tonight, RONALD went shopping. b. Tonight, only RONALD went shopping.

But on reflection, (26a) implies that in the past some other person than Ronald went shopping whereas (26b) implies that in the past Ronald went shopping along with some other person(s). Umbach argues that two types of exclusion need to be distinguished which have di¤erent presuppositions: a) presupposition: x went shopping; assertion: x ¼ Ronald. b) presupposition: Ronald went shopping; assertion: None of the alternatives under consideration went shopping too.

100

Terry Varma

In a) alternatives to Ronald are excluded. In b) others in addition to Ronald are excluded. If we take the exclusive focus particle hii to have this contra-additive interpretation we can explain why hii in (20) and (25) are not examples of the exclusive focus particle usage and why hii cannot be translated as only in these cases. In example (20), since sweaters normally only have one owner, the question of owners in addition to Gautam does not arise, so hii cannot be excluding them. In (25) it is not additional cups of tea that are being excluded by hii, so only is not an appropriate translation. Translation of hii as only in the non-scalar meaning is only appropriate in those contexts where alternatives in addition to the focus expression are a plausible possibility. Because of the overlap in the di¤erent uses of hii as an exclusive focus particle and a contrast IF particle, it is predictable that there will be cases of ambiguity where the meaning remains indeterminate between the two. In fact, there is no need to reject the possibility that both meanings may be intended at the same time. This is the case in (27): (27)

mAI umiid karataa hUU ki asalii I hope do-impf-msing be-aux-pres-1pmsing that real-fsing sinduur kaa hii upayog kar rahii hAI. vermilion gen hii use do stay-perf p-f be-pres-3ppl a. ‘I hope it’s REAL vermilion that they are using.’ b. ‘I hope they’re only using REAL vermilion.’ c. ‘I hope it’s only REAL vermilion that they’re using.’ (S)

The context is a discussion of why women put vermilion on their hair parting. Earlier a popular theory was quoted that vermilion is good for controlling a woman’s electrical energy as it contains large amounts of important minerals such as lead which penetrate the soft area on the female head. After the sentence in (27) he points out that vermilion is not made from lead at all and in fact lead is dangerous for a person’s health. In this context hii may be evoking and excluding the discourse-old lead vermilion (27a), or any vermilion in addition to the real one (27b). In fact both may be combined to give (27c). Scalar cases of only in propositions which express su‰cient conditions are seen as problematic in the literature. Take (28) for example. (28)

Only a MIRACLE can save us.

In such cases it appears that the excluded alternatives are those which are lower on the scale than the focus expression and not higher. A contraadditive reading works but the focus expression in this case strongly suggests a scale and a high extreme on that scale. There seems to be no problem in a

Some aspects of the meaning of the Hindi particle hii

101

su‰cient condition context if the focus expression is not obviously high scale. For example: (29)

sac kyaa hai, yahto koii ciRiyaa hii bataa truth what be-pres-3psing this to some bird hii tell sakataa hai. can-impf-msing be-aux-pres-3psing ‘What the truth is, only a BIRD can tell.’

A suggestion for further investigation is that perhaps in (28) only is actually contra-additive but its argumentative role is to indicate a proposition which is lower in argumentative strength than some stated or predicted alternative. That only a miracle can save us is a pretty minimal statement, not much can save us, certainly less than we would hope for. In this section we have seen that hii is used as an exclusive focus particle. In this case its meaning overlaps with but is not the same as that of the contrast IF particle. The information it conveys does not answer the question as to what the instantiation of a variable x in an open proposition is but whether there are any further instantiations of the variable other than x. The exclusive focus particle hii may also receive a scalar reading with the interpretation of the focus expression as low. 4. Exclusion and scalar extremes However, there are also frequent examples where scalar hii indicates a different ordering of the focused element and its alternative set. In these cases hii seems to mark its focus expression as high scale, not low scale and it is often felt by native speakers that even, a scalar focus particle indicating a high ranking on some scale, is the appropriate translation of hii in such cases. To clarify this use of hii, I will relate it to another type of focus particle discussed in the literature. In English and many other languages focusing particles are used to mark the particular importance of the focused expression in relation to the alternatives in the discourse context, emphatic reflexives are a case in point. For example: (30)

I would like to talk to the MANAGER himself.

In Ko¨nig and Siemund (1999) emphatic reflexives are grouped under a category which they call ‘‘intensifiers’’, choosing this term as it is simply the ‘‘least misleading among the many that have been proposed in the literature.’’ I will also use this term in relation to this third use of hii. Creswell (2002b) identifies two di¤erent properties of intensifiers (which she calls

102

Terry Varma

‘‘emphatic reflexives’’) that are relevant here. In some cases the intensifier marks the focus as a central entity among a set of related entities. This means that being able to interpret the sentence then depends on a shared understanding that in the discourse context certain entities are more central and others more peripheral, as in example (30). Sometimes, however, she has established that it is only the entity’s rank as an extreme on a scale that gives it prominence. An example from Creswell (2002b) is: (31)

That boy in my project group is such a loser. He’s dragging us all down. MY PINKIE TOE itself has more creative energy than he’ll ever have.

hii too has intensifier uses.8 That is, it marks discourse central or extreme scale entities in a discourse context. The following examples will serve by way of illustration. In some cases the ordering of the entities depends on general knowledge such as the central importance of life in (32) or of the air in (34), whose warning appeared on a poster campaigning against air pollution. In other cases it is a specific discourse context as in (36) where the conversation was a discussion as to various reasons why the kebabs had not been a success. In (35) NT, the Navbharat Times, is the morning newspaper which has been stopped for a few days because of a strike. (32)

jab ek baar dhurii gaRabaRaa jaatii when one time axis-f be disturbed lv-impf-fsing hai to zindagii hii laRakhaRaa jaatii be-aux-pres-3sing then life-f hii wobble lv-impf-fsing hai. be-aux-pres-3sing ‘When the axis is disturbed, LIFE itself starts wobbling.’ (AB)

(33)

yah kaisii jiit hai, jab this what kind of-fsing victory-f be-pres-3psing when jiitanevaalaa hii nahII rahaa? win-inf-infl-nomin-msing hii neg remain-perf-msing ‘What kind of victory is this when the VICTOR himself is dead?’

(34)

havaa hii maut ban gaii to! air-f hii death become lv-perf-fsing then ‘What if the air itself becomes death!’

(35)

NT kii NT poss-f gaayab ho gone be

kamii aisii lack-f such-f gayaa lv-perf-msing

khalii jaise khaane mE namak hii empty as if food in salt-m hii ho. be-aux-sub-pres

Some aspects of the meaning of the Hindi particle hii

103

‘The absence of the NT feels as if the SALT itself is missing from the food.’ (NT) (36)

miiT hii THiik nahII thaa. meat hii alright neg be-impf-past-msing ‘The MEAT itself wasn’t right.’

(37)

ham to jaat-paant se uupar uTH cuke we to lineage above rise be finished-perf-mpl the na paramparaa se, jaati se, be-aux-impf-past-mpl no tradition from, caste from, desh se hii uTh cuke country from hii rise be finished-perf-mpl the. be-impf-aux-past-mpl ‘We had risen above matters of lineage hadn’t we? Above tradition, above caste, above the COUNTRY itself.’ (M)

Hindi also has a bhii and a tak which may often be used in the same extreme scalar environments as hii. I have discussed this aspect of bhii and tak in Varma (2003).9 Schwenter and Vasishth (2000) attempt to establish a distinction between two types of even, an inclusive and a non-inclusive, drawing on Spanish and Hindi to support the distinction, though their findings for Hindi have been questioned. For Hindi they claim that bhii is inclusive and tak is not. This possibility is also suggested by their origins, bhii as an additive non-scalar particle and tak as indicating only an end point. What hii has in common with bhii and tak, besides their all being scalar focus particles is that they may all mark an item as extreme on a scale. I propose that what distinguishes hii here from bhii and tak is that it is exclusive. It excludes lower values on the scale or less discourse central values from consideration. Even in contexts where it might be argued that it could be read as inclusive, such as in (37), it marks a separate, independent point. By focusing on a centrally important or extreme case the message is that other members of the set can be discarded, this one ‘‘says it all’’. Another separate use of hii adds some support to this interpretation, as can be seen in the following two examples: (38)

gaNit hii ko liijie. maths hii acc take-imp-hon ‘Take MATHS for example.’

(39)

ab yah kAATe, cammac, chhurii kii hii baat now these forks, spoons, knives gen hii matter

104

Terry Varma lo. take-imp-fam ‘Take the example of these KNIVES, FORKS and SPOONS.’

These examples show the use of hii in marking items which are selected to represent a set of alternatives, discarding other items from the need for consideration. To sum up, in this section I have identified an intensifier use of hii which marks its focus as centrally important or extreme in relation to the evoked alternative set, with other values being excluded. This extreme usage relates hii to meanings of even. This usage shares meaning features with the exclusive focus particle usage but unlike the latter, it is always scalar and the scale evoked is di¤erently structured. 5. Identity by exclusion Before turning attention to the interpretation of hii in relation to the verbal complex, I would like to mention a further use of hii where the elements of focus and exclusion are exploited to yield a di¤erent meaning. I will refer to this function of hii as the emphatic assertion of identity to relate it to the discussion of focus particles with comparable meaning in Ko¨nig (1991). This is a common use of hii and it can be understood alongside the functions I have discussed so far. However, it most often involves pronouns for reasons which will become apparent, and this study has not gone into issues of hii with pronoun focus and my interpretation of the identity meaning may need to be reconsidered at a later point. There are three cases I wish to mention, one where the sentence will often contain the focus additive particle bhii ‘also’ as well as well as hii. For example in (40) after describing a fan as slow the author says of another fan: (40)

vah bhii dhiime hii cal rahaa that one also slowly hii move stay-aux-prog-msing thaa. be-aux-past-msing ‘That one was going at the same slow SPEED.’

In the second there is no bhii, but here too the focused expression in the scope of hii either has the same reference as that in another proposition mentioned or is inferable in the discourse context: (41)

par baad mE unake saare guN hii unake doSH but later her-mpl all-mpl qualities-m hii her-mpl faults

Some aspects of the meaning of the Hindi particle hii

105

ban gae. become lv-perf-mpl ‘But later all her very QUALITIES became her faults.’ (KVB) (42)

ek din yahii sharaab aapako pii jaaegii. one day thisþhii alcohol-f you-acc drink lv-fut-f ‘One day this very same DRINK will destroy you.’ (Government poster)

The third shows a standard use of vahii. (43)

iishvar vahii rahataa hai, praarthanaaE god thatþhii stay-impf-msing be-aux-3psing prayer-fpl badal jaatii hAI. change lv-impf-fpl be-aux-3ppl ‘God stays the SAME, the prayers change.’

In vahii we find the demonstrative vah ‘that’ þ hii of exclusion. When vah is combined with hii it results in a fused form. In the case of (43) it shows the conventionalized meaning of ‘the same’. Like vah, it can be used as a determiner as in vahii shakal ‘the same face’ or as a pronoun as in shakal vahii hai, ‘the face is the same.’ What the examples given above have in common is that the expression in focus is anaphoric in its reference. The hii adds an element of exclusion of the alternatives in the evoked set of alternatives. In this way the identity indicated by the anaphora is strengthened by the exclusion of any alternatives there may be. There is no requirement here on the discourse status of the open proposition or alternative set. The particle hii in these examples allows for the emphatic assertion of identity, identity by exclusion. 6. Verum focus, a‰rming the truth Usually focus is on the lexical content of a particular constituent, but sometimes it involves, rather, the truth value of the proposition. This property has given rise to the term verum focus, (Creswell 2004). This type of focus can involve the polarity of the verb. It may occur, for example, when the opposite polarity has been mentioned or is inferable from the discourse context. This is the case in (44) where the speaker has been discussing with others whether to take a particular flat on rent or not. She decides: (44)

mAI socatii hUU flat le hii I think-impf-f be-aux-1psing flat take hii lUU. lv-subj-pres-1psing ‘I think I WILL take the flat.’

106

Terry Varma

In this sentence the new information is the positive polarity of the sentence, that she will take the flat rather than not. The alternatives consist only of the opposite, that she will not, and this alternative is excluded. In English the focus is placed on an auxiliary verb rather than the lexical verb, which serves to indicate that it is not the lexical content which is in focus. In Hindi, as is normal when the focus is within the verbal complex, it is the lexical verb which receives the intonational accent and the focus particle hii is placed immediately after the lexical verb, before all auxiliaries of tense and aspect and before light verbs. Nevertheless the new information focus in cases like (44) is the positive a‰rmation of the proposition and the alternative in the set evoked by hii is the negation of the proposition and this alternative is excluded. In (44) the alternative has been specifically evoked earlier in the discourse context. At other times it may only have been implied or presupposed as in: (45)

A: A: B: B:

saaraa khuun pii jaatii hAI. all-msing blood drink lv-impf-f be-aux-pres-3ppl ‘They drink up all your blood.’ khuun hai hii kahAA, beTii? blood be-pres-3sing hii where, daughter ‘I HAVE no blood, child.’ (KBN)

Here speaker B is countering the presupposition in A’s statement that she has some blood to be drunk. She excludes the positive and asserts the negative. In (46) the speaker has been asked if she has paid for her niece’s medical treatment. She says she has not and then, to clear up any inference in this regard, she says the following, adding the clarification that she doesn’t want a possible wrong implication: (46)

jo kar sakate hAI kar hii rahe what do can-impf-mpl be-aux-ppl do hii stay-aux-mpl hAI . . . yahnahII ki ham nahII kar rahe. be-aux-3ppl this neg that we neg do stay-aux-mpl ‘We ARE doing what we can . . . it’s not that we’re not.’

What is in common is the requirement that the proposition with the reverse polarity be discourse- or hearer-old, which includes inferable from the discourse context. Otherwise verum focus hii is inappropriate. It is possible that in some instances hii focusing is related to other elements of the inflected verb such as aspect. For example: (47)

A: A:

raagii khatam ho rahii hai. millet-f finished be stay-aux-pres-f be-aux-pres-3psing ‘The millet flour is finishing.’

Some aspects of the meaning of the Hindi particle hii B: B:

107

khatam hii ho gaii. finish hii be lv-perf-f ‘It HAS finished.’

It should be noted that khatam is actually the lexical element of this conjunct verb and hii is placed after it. There has been little research on this type of aspectual focus but it has generally been categorized together with verum focus.10 To summarize, hii may be used within the verb complex to focus polarity or aspect, excluding the alternatives of polarity or aspect which are discourse-old. 7. Scale, phase, aspect The last use of hii that I wish to relate to the elements of meaning and function discussed so far in this paper brings together examples noted here and there in grammar books of Hindi with glosses that seem to indicate aspectual e¤ects.11 These examples are instances where hii appears within the verbal complex. I will propose that these examples can also be explained in terms of focus, exclusion and scale. The following are typical examples from the grammar books: (48)

mAI dekh sun hii rahii thii. I watch listen hii stay-aux-prog-f be-aux-past-f ‘I was just watching and listening.’

(49)

daftar pahuncaa hii thaa. o‰ce arrive-perf-m hii be-aux-past ‘He had just arrived at the o‰ce.’

Languages often reflect the fact that time and space relations may be interpreted on a scale, from earlier to later or nearer to farther and so on. hii may be used with such scales to focus on a particular point and exclude others. As it takes time to cover space the two are often linked, as in the following example which is said in the context where someone who has left the house is needed now in the house: (50)

saamane kii galii mE hii hogaa. front gen-f lane-f in hii be-fut-3psing-m ‘He must be in the front LANE hii.’

In this case hii evokes alternatives to the focus expression ordered on a scale of distance/time representing the progress of the person who left, away

108

Terry Varma

from the house. The focus is in the lane and excluded are alternatives further along the scale. As in the case of scalar hii/only, there is an implied evaluation of lower or here nearer or earlier on the scale than some stated or inferred alternative(s), or of minimality of the distance. In English one might use phasal ‘still’ or ‘just’ as a translation.12 An interesting example to further illustrate features of the meaning of hii in temporal and spatial contexts is the following, for which I have o¤ered two possible translations (there may be others): (51)

vah he a. b.

ghar se hii khaanaa khaa aayaa. home from hii meal eat come-perf-msing ‘He had his meal BEFORE starting o¤ and coming here.’ ‘He had his meal JUST before starting o¤ and coming here.’13

We can order the di¤erent phases of the state of a¤airs that is described in or can be inferred from this sentence along a time/place scale: at home ate meal

/on the way /

/ /

here arrived

There is a transition point involved between being at home and starting on the way. hii clearly limits the event of eating the meal to the at home phase and excludes later points. But interestingly, as in (51b), there can be a further meaning implied in which the eating event is shifted to very close to the transition point, without crossing it: just before leaving the house. In this reading a minimal distance to the transition point is indicated but crossing the transition point and moving on to other phases is excluded. The following example of a temporal post-position may be interpreted in a similar vein. The context is watching television and someone is asking about program X. The answer given is: (52)

isake baad hii hai. this-infl-gen after hii be-pres-3psing ‘It’s just after this.’

Here the event of the program is located ‘after this one’ ke baad, the hii shifts the point to very close after and excludes later points. I propose that cases like (48) and (49) can be analyzed in a similar scalar way. Events may have internal structure with di¤erent phases. They may be viewed as at di¤erent phases, for example: imminent, initial and later stages, completion and the consequent state of a¤airs. I will examine the use of hii in verbal complexes which are marked to show di¤erent aspects or phases. The first uses a construction to indicate imminence:

Some aspects of the meaning of the Hindi particle hii (53)

109

ham use jagaane vale the. we he-infl-acc wake-inf-infl imm.asp-mpl be-past. ‘We were going to wake him.’

Compare this with the addition of hii (54)

ham use jagaane hii vale the. ‘We were just going to wake him.’

The e¤ect of hii is to focus on the imminence, a moment closer to the starting point of the verb is indicated (a minimal distance away from it) and further phases are excluded. Next consider the case of verbal complexes that include a progressive aspect marker indicating that the event is ongoing at the reference time. Interestingly, hii may have three di¤erent meanings with progressives. The progressive itself, it should be noted, may show a shift in meaning from ongoing to imminent. It is also important to note another related tendency (Ko¨nig 1991). The tendency for adverbs referring to the moment of utterance to shift their reference to a time immediately preceding or following that point in time. These examples of meaning shift from ongoing to imminent in the use of the progressive aspect are from Shapiro (1989): (55)

laRake gaanaa gaa rahe hAI. boys song sing stay-aux-prog-mpl be-aux-3ppl ‘The boys are singing songs.’ (ongoing)

(56)

kal mAI dillii jaa rahaa tomorrow I Delhi go stay-aux-prog-msing hUU. be-aux-pres-1psing ‘I am going to Delhi tomorrow.’ (imminent)

Let us look now at the e¤ect of the addition of a hii to a verb marked progressive. (57)

a.

mAI apanii diidii ko de rahii thii. I own-f sister to give stay-aux-prog-f be-aux-past-f ‘I was giving it to my sister,’ or ‘I was going to give it to my sister.’

b.

mAI apanii diidii ko de hii rahii thii. ‘I was just going to give it to my sister.’

In (57b) the hii focuses the imminent phase to close to the onset (minimal distance) and excludes further phases. In (58b) the speaker has been stopped by the police for speeding and tries to convince them that it is not possible that they were speeding.

110 (58)

Terry Varma a.

ham nikal rahe hAI. we leave stay-aux-prog-mpl be-pres-3psing ‘We are leaving (the house).’

b.

ham nikal hii rahe hAI. ‘We are just leaving the house.’

Here the hii focuses on the early phase of the event of leaving and going on the road, close to the initial point, and excludes later phases along the time scale. A third example with the progressive is: (59)

a.

meeting cal rahii hai. meeting-f go stay-aux-prog-f be-aux-pres-3psing ‘The meeting is going on.’

b.

meeting cal hii rahii hai. ‘The meeting is still going on.’

In (59) with the hii present, the next phase, the ending of the meeting is excluded and the focus is on the ‘‘ongoingness’’. There was an expectation in this instance that the meeting would have ended, no longer be ongoing. A similar example is (60) where someone was extolling the merits of his mobile server. The card lasted well beyond the point where he saw that the time he had paid for was over. He says that at that point: (60)

a.

cal rahii thii. go stay-aux-prog-f be-aux-past-f ‘It was working.’

b.

cal hii rahii thii. ‘It was going on and on working.’

Mention should be made here of the inflected invariable use of what is referred to as the present participle of the verb together with hii. This present participle is the imperfect participle in Hindi. From the verb lauTanaa ‘to return’, for example, the imperfect participle is lauTataa ‘returning’ whose ending
taa can be changed to indicate feminine gender and plurality. The adverbial form of ‘returning’ is formed by inflecting the ending to invariable
e, lauTate. An example can be seen in the following sentence: (61)

ghar lauTate (hue) usane apanii patnii ko phon home returning (stative) he-infl-erg own-f wife to phone kiyaa. make-perf-msing ‘As he was returning home he phoned his wife.’14

Some aspects of the meaning of the Hindi particle hii

111

An interesting use of this adverbial form of the imperfective particle is with hii. hii changes the meaning expressed by the imperfective aspect: (62)

ghar lauTate hii usane apanii patnii ko phon kiyaa. ‘As soon as he returned home he phoned his wife.’

Now rather than seeing the returning as ongoing, incomplete, it is seen as recently completed. The point in time has been shifted forward to just after but at a minimal distance from the transition point, the completion of the event, later points are excluded. This is a common usage of hii and the common way to indicate almost, but not quite simultaneous phases of two events, the end of one and the beginning of the next. It is interesting to compare this with the use of the conjunctive particle, which does indicate completion, with the
te hii construction in the following pair of sentences: (63)

a.

vah girate hii, mar gayaa. he fall-impf p hii die lv-perf-msing ‘He died as soon as he hit the ground.’

b.

vah girakar mar gayaa. ‘He fell and died.’

Consider next the case of hii where verbs have auxiliaries of continuous aspect. Here the continuity is focused on and the next phase, the ending of the event, is excluded. (64)

a.

hilataa rahaa. shake-impf-msing stay-aux-cont-perf-msing ‘It continued to shake.’

b.

hilataa hii rahaa. ‘It went on and on shaking, without stopping.’

As for hii with verbs marked for the perfective, it may have more than one meaning. It may move the focus to a point close after the completion (minimal distance) and exclude later moments of the consequent state. The following is an example: (65)

a.

tab mAI ne yah auto liyaa then I erg this autorickshaw-m buy-perf-msing thaa. be-aux-past-msing ‘Then I had bought this autorickshaw.’

b.

tab mAI ne yah auto liyaa hii thaa. ‘Then I had just bought this autorickshaw.’15

However, with a perfective marker, hii may focus rather on the completion of the event, as when there might be some reason why it might not have

112

Terry Varma

been completed. This is reminiscent of verum focus. Often this calls for a translation with an auxiliary stress in English as with verum focus and often ‘finally’, or ‘anyway’ seem to be appropriate additions. An example is the following said by a neighbor addressing a relative of the house she has been talking to for a few minutes: (66)

bhaabhii ne bhojpurii siikh hii lii. sister-in-law erg Bhojpuri-f learn hii lv-perf-f ‘Bhabhi DID finally learn Bhojpuri.’

Or as in this example which is a newspaper headline: (67)

mulaayam haar hii gae. Mulayam lose hii lv-aux-perf-m-plhon ‘Mulayam DID lose in the end.’ (J)

In such cases hii focuses on the completion of the event and excludes the negative possibility. This use would seem to overlap with verum focus and how it might di¤er, if it does, is not clear.16 8. Summary At this point I will summarize some of the di¤erent functions of hii in the form of a chart which brings together its di¤erent focusing and exclusion meanings. It does not include various requirements on some uses such as the discourse status of various elements.

Table 1. Di¤erent focusing and exclusion meanings of hii Type of focus

Alternatives excluded

Approximate equivalent in English

Contrast information focus Exclusive focus Scalar exclusive focus Scalar, central, extreme focus Assertion of identity focus Verum focus

Discourse-old

Clefts

Any in addition Any higher on the scale Any less central, extreme

Only Only Emphatic reflexive, the very, even The same Auxiliary stress

Aspect focus

Any alternative The opposite polarity Other tense or aspects Later phase Negative phase

Varies with the aspect involved

Some aspects of the meaning of the Hindi particle hii

113

If we compare hii with the exclusive focus particles and their uses identified in the literature, we see that the basic elements of meaning of focus, exclusion and scale are the same. But the range of functions of hii overlaps with but is not the same. Generally speaking hii seems to have a wider range of functions than other exclusive focus particles in other languages, where functions such as verum focus marking or the emphatic assertion of identity are carried out by other means or other particles. In the final section of this paper I would further like to extend my discussion of hii to capture two other dimensions of its use. 9. A discourse marking function: precluding Shapiro (1999) has already established that the ‘‘emphatic particle’’ to is often not a marker of emphasis but rather serves as a discourse marker (DM) in the sense that it relates the meaning of the proposition it appears in to one which is expressed in, or is inferable from the discourse context.17 I propose that this is also true of hii, though its DM functions are probably not as varied as those of to. The DM uses of hii may be linked in meaning to the exclusive focus properties, but they form a category and subcategories based rather on their discourse relating function. In this section I will discuss one of the DM functions of hii which I will call one of preclusion where hii marks the proposition as precluding some other state of affairs which is discourse-old or discourse-inferable. As I mentioned, sometimes this discourse marking function overlaps with some of the functions of hii that I have discussed already. Verum focus in which a presupposition of the hearer is countered is a case in point. Take (45), for instance, repeated here as (68): (68)

A: A: B: B:

saaraa khuun pii jaatii hAI. all-msing blood drink lv-impf-f be-aux-pres-3ppl ‘They drink up all your blood.’ khuun hai hii kahAA, beTii? blood be-pres-3sing hii where, daughter ‘I HAVE no blood, child.’ (KBN)

Having no blood precludes the possibility of the witches drinking it. This is a case of preclusion by pointing out the error in a basic assumption of the hearer’s previous point. The marking of the focus as extreme or central as in (36), repeated here as (69) may also be a way to indicate preclusion. (69)

miiT hii THiik nahII thaa. meat hii alright neg be-impf-past-msing ‘The MEAT itself wasn’t right.’

114

Terry Varma

In (69) we saw that the meat is marked by hii as being of central importance in the situation. But also if the central element, the meat, was no good this precludes any chance of the kebabs being good. Of all the points to explain the poor kebabs this may be the strongest point as it precludes them being good anyway. Similarly in (33), repeated here as (70), if the victor is dead, it precludes there being a true victory: (70)

yah kaisii jiit hai, jab this what kind of-fsing victory-f be-pres-3psing when jiitanevaalaa hii nahII rahaa? win-inf-infl-nomin-msing hii neg remain-perf-msing What kind of victory is this when the VICTOR himself is dead?

It will be recalled that when hii is a marker of an extreme it may sometimes have the same meaning as even. However most often hii cannot replace even, i.e. bhii or tak in Hindi. In the following example, for instance, hii would be inappropriate: (71)

ghar mE picchale ek maah se kaid madhu ko house in last one month since imprisoned Madhu to bhaaii- bhaabhii khaanaa tak nahII dete brother sister-in-law food even neg give-impf-mpl the. be-aux-past-3ppl ‘Her brother and sister-in-law didn’t even give any FOOD to Madhu, who has been a prisoner in her home for a month now.’

The context is a description of how badly treated the handicapped Madhu is. But interestingly, in the same proposition hii can be used if the relationship of the proposition to the context is one of preclusion. For example: (72)

vah moTii kaise ho jaatii, usako khaanaa hii nahII she fat-f how be lv-subj-f she-infl-to food hii neg dete the. give-impf-mpl be-aux-past-3ppl ‘How could she have become fat, they didn’t even give her any FOOD.’

In other instances of preclusion marking, hii does not seem to be any of the focus types. The following example is from an ad in the newspaper put there by another newspaper where the writer is trying to defend his newspaper’s generally frivolous content and catchy headlines. He defends the catchy headlines in the following sentence:

Some aspects of the meaning of the Hindi particle hii (73)

115

agar headline mE mazaa nahII aayaa, to pej hii if headline in pleasure-m neg come-perf then page hii palaT liyaa. turn lv- perf-msing ‘If they are not attracted by the headline they turn the PAGE hii.’

This hii becomes interpretable when seen as marking a proposition as bearing the relation of preclusion to the context. Here the message is that if the reader turns the page because of a boring headline, it will preclude him from reading the paper anyway, which defeats the object of any newspaper. In questions hii often indicates that this is a fundamental, basic question that makes what has gone before in the context seem less important, and precludes the need for any other question. (74)

paapaa ko vivaah kii baat suujhii hii kaise? dad to marriage gen matter-f occur-perf-f hii how ‘How did Dad GET the idea of marriage in the first place?’

(75)

aap vahAA gaye hii kyO? you there go-perf hii why ‘Why did you go there anyway?’

Because these are also examples of rhetorical questions—a very common form of actual negation in Hindi—they also convey other preclusive implications. These two questions imply first that Dad should never have got this idea and that you should not have gone there. Then a further implication is that none of the issues we are grappling with would have arisen, they would have been precluded. Given an out of context proposition with hii a native speaker may well try to frame it as having some sort of preclusive implication. For example (76)

merii gaaRii to kharaab hii ho gaii. my-f car-f to bad hii be lv-perf-f ‘My car’s broken down hii.’ could be interpreted to imply ‘‘we can’t do anything now.’’

Finally, a number of common expressions with hii use the preclusive implication to give an e¤ect of obviousness. In each of the following examples the proposition marked with hii actually precludes the need to say what is about to be said: (77)

kahaa hii gayaa ki . . . say-perf-p hii go-perf-msing that . . . ‘It has been said hii that . . .’

116

Terry Varma

(78)

dekh hii rahe ho . . . see hii stay-perf pl be-aux-2ppl . . . ‘You can see hii . . .’

(79)

jaanate hii hAI ki . . . know-impf hii be-aux-3plhon that . . . ‘You know hii that . . .’

It can be seen that the translation of hii is very varied, but all are di¤erent ways of implying a discourse relation of preclusion. In (77)–(79) the best way to translate hii is probably already, which can also be used to indicate that the need to do something else is precluded by this proposition. In the question examples I suggested in the first place as a translation, indicating a very basic question. Anyway is also often a good translation for this hii because it too can mark a statement that makes others appear less important and less needed. 10.

Modal functions

Finally I would like to discuss uses of hii where it interacts with modal aspects of the sentence. I believe that it would be worthwhile to consider the modal functions of Hindi focus particles as a separate group. I will consider two clear modal functions of hii in this section, an epistemic and a deontic function. Modal markers are involved in expressing something about the speaker’s attitude to what she is saying; in the case of epistemic markers they express judgments about the degree of likelihood or factuality of an event. Accompanying a verb, hii often indicates a high degree of certainty; the speaker is definite about what she is saying. This is illustrated in the following three examples: (80)

laaluu ne kahaa ham banaaEge to Lalu erg say-perf-msing we build-fut-3pmpl then banaaEge hii. build-fut-3pmpl hii ‘If Lalu said that he will build it he will definitely build it.’

(81)

vah hogaa hii. that be-fut-3pmsing hii ‘That will definitely be present.’

(82)

patrakaar vahAA aaEge hii. journalists there come-fut-3pmpl hii ‘Journalists are sure to come there.’

Some aspects of the meaning of the Hindi particle hii

117

The expression of definiteness supplied by hii is particularly common in the future tense. This may reflect the importance to the hearer of whether the event will actually take place, how certain the speaker is that it will. Deontic modal expressions concern obligation and necessity. Adding hii to verb forms which have this meaning asserts a high degree of necessity. This can be seen if we compare the following: (83)

(84)

(85)

a.

use ek na ek din samaapt honaa it-infl-ind one day or the other end be-inf thaa. be-past-impf-msing ‘One day or the other it had to end.’

b.

use ek na ek din samaapt honaa hii thaa. ‘One day or the other it was bound to end.’

a.

subsidy kam karanii paRegii. subsidy less make-inf-f have to-f ‘The subsidy will have to be reduced.’

b.

subisdy kam karanii hii paRegii. ‘The subsidy will just have to be reduced.’

a.

vah bataa rahii thii kyaa karanaa she explain stay-aux-prog-f be-past-f what do-inf hai. be-pres-3psing ‘She was explaining what had to be done.’

b.

vah bataa rahii thii kyaa karanaa hii hai. ‘She was explaining what absolutely had to be done.’

It may be pointed out that in both these modal uses of hii we can see links to the meanings of focus and exclusion. It is possible to argue that what is being excluded is the opposite polarity of the verb, as in the case of verum focus, though it seems to me that in the modal cases there is no requirement for the opposite polarity to be discourse-old. Still, the di¤erence between these uses and those of the first part of the paper may not be enough to claim a di¤erent category of functions. Ko¨nig (1991) considers there to be a separate Modal Particle category in German and other languages. On the other hand there is no consensus on what modal means in the case of such particles. Most interpretations include the idea that modality belongs on the interactional level of language involving degrees of certainty but also the perspective of the speaker. An investigation of other modal uses of hii along with that of other Hindi particles should help clarify these issues.

118 11.

Terry Varma Conclusion

hii is polysemous, in the sense that it has a number of di¤erent meanings which can be seen as interconnected in a motivated way, based on the core parameters of focus, exclusion and scale. Meaning varies according to what is focused on, what alternatives are excluded, whether the context is scalar or not and what sort of scales are involved. These meanings overlap and may result in ambiguity. hii may be compared to focus particles of exclusion in other languages, which can also display several di¤erent meanings, but the range in di¤erent languages only partly overlaps. hii plays a role in several important systems of the language, interacting with information packaging, aspect and modality and acting as a discourse marker. Further research is needed into other meanings and functions of hii to see how far they relate to the meanings discussed here. It is hoped that the analyses here will serve as a beginning towards a comparison of the di¤erent members of the class of particles in Hindi and as a basis for a typological comparison among the languages of the subcontinent.

Notes 1. Some of the meanings given in the literature are: ‘precisely’, ‘particularly’, ‘one thing rather than another’, ‘only’, ‘merely’, ‘just’, ‘definitely’, ‘even’ etc. Some attempts at generalizations have been made with most agreeing that hii conveys ‘emphasis’ on the word it follows, others see ‘contrast’ as a linking element of meaning. hii has also been variously categorized as a ‘particle’, an ‘emphatic particle’, an ‘enclitic’, a ‘discourse marker’ or ‘discourse clitic’, an ‘adverb’, and an ‘intensive adverbial particle’. Similarly, there is no agreement on which items should be included in the same category as hii, or why. Only Verma (1971) attempts a detailed theoretical discussion. He claims that hii, bhii and to act as ‘scope indicators’ as part of ‘limiter constituents’ which he calls ‘limiter particles’. hii, then, is a scope marker for such limiter particles as sirf and bas. hii is optional. hii’s limiter particle may also be optionally deleted. However it is always assumed in the presence of hii. I find these suggestions interesting, but in any case they seem applicable only to a small section of the range of meanings of hii that I discuss in this paper and I see no reason for the moment to posit any dependence of hii on a stated or understood limiter particle. In a sense the terminology I use is closest to Sharma (1999), who states that ‘‘hii marks exclusive focus, similar in some ways to only in English, . . . hii, bhii and to perform a range of pragmatically and semantically restrictive functions’’. However, Sharma’s study is focused on the syntax of clitics and not on meaning and function, as well as being limited to the nominal domain. My own study attempts rather to spell out and unify some of the meanings and functions that she makes brief reference to in the above quotation.

Some aspects of the meaning of the Hindi particle hii

119

2. I have not considered all the functions of hii in this paper, others await further investigation. In particular I have not examined the interaction of negation with hii in any detail. 3. On the other hand it is claimed that in some languages the distinction between information focus and contrast focus must be maintained. See Molna`r (2001). 4. For further details of the placement of hii within rather than at the end of the NP in focus and its varying levels of acceptability, see Sharma (1999). 5. Creswell (2002) points out that even when the felicity conditions are present in a sentence, it does not guarantee that a cleft will be used, further considerations are at work. 6. It is also interesting to note the preference of English learners of Hindi to seize on only as the meaning of hii and to overuse this understanding (Hooper 1901). This suggests that other meanings of hii may be more subtle for the foreign learner. 7. This is a common feature of focus and focus particles. See Varma (2003) on bhii. 8. In addition Creswell posits that there needs to be available in the discourse, or inferable from it, a predication of the focus expression and a related predication of a less prominent entity. For now I will leave this consideration aside as I am not trying to show a complete overlap between hii and intensifiers such as the emphatic reflexive in this role. 9. In the above examples, hii evokes an alternative set in which the expression in focus is central or extreme in some sense. It can also be seen that often even would serve as a translation for hii. An interesting cross-linguistic observation made by Ko¨nig and Siemund (1999) is that the emphatic reflexive intensifier is often exactly the same word as for even, the scalar additive particle, German selbst and French meˆme for example. 10. For some analysis of verum focus and aspect in relation to Russian see McCoy (2001). 11. To throw further light on the interaction of hii with aspect a better understanding of aspect and the way information structure regarding tense and aspect sequences of verbs is very much needed. For an attempt to outline some of the complexities involved see Butt (2003). 12. For a study of French phasal particles and their modalizing and discoursemarking uses, see Mosegaard (2002). 13. I am grateful to my colleague O.P. Jayaswal for pointing this out to me. O.P. Jayaswal was for years a teacher at the Musoorie Hindi Language School and was involved in Bible translation. 14. Some speakers require the presence of hue. Another interesting di¤erence of opinion is whether the
te form alone, without hii can also be used to mean ‘as soon as’. 15. One of the reviewers of this paper pointed out the following interesting example which can be compared with (65) and (49): (i)

daftar pahunc hii gayaa thaa. o‰ce arrive hii lv.3msg be-aux-past ‘He had almost arrived at the o‰ce.’

120

Terry Varma

A past perfect form with a light verb here receives an interpretation as indicating a point just preceding the completion of the event (native speakers seem to disagree about the interpretation of this example as well as that of (51b), not everyone accepts the ‘close to before the transition point’ reading). 16. It seems to be di‰cult to find parallel instances with plain perfect form verbs with no other perfective marking. It is not clear why this should be. Plain perfect forms followed by hii do freely occur in some types of discourse marking functions of hii. 17. I am using the term DM to refer to a subclass of pragmatic markers which have a connective function, signaling a relationship between the discourse they appear in and the current situation of discourse; cf. Pons (2005) for a discussion of issues of terminology. Fraser (2005) proposes narrowing the definition of DM to connectives which signal only one of four relationships: elaboration, contrast, inference or temporality. hii would, then, belong in the category of ‘inference’.

Sources (AB) (M) (MRSK) (KVB) (KBN) (NKD) (C) (J) (NT) (S)

— Bhandari, Mannu. 1997. aapakaa bATii. New Delhi: Radhakrishna Prakashan. — Gitanjalishri. 1993. maaii. New Delhi: Rajkamal Prakashan. — Mohan Rakesh. 1999. mohan rakesh kii sApuurN kahaaniyAA. Delhi: Rajpal and Sons. — Saraogi, Alka. 2000. kali-kathaa: vaayaa baaipaas. Panchkula (Haryana): Adhar Prakashan. — 2004. koi baat nahII. New Delhi: Rajkamal Prakashan. — Vaid, Krishna Baldev. 2000. ek naukaraanii kii Daayarii. Delhi: Rajpal and Sons. — cakamak: baal vigyaan patrikaa. Bhopal: Eklavya. — janasattaa. Indian Express Newspapers (Mumbai), published from Delhi and Kolkata. — navabhaarat Taaimz. Bennett and Coleman & Co, published from New Delhi. — srot. Bhopal: Eklavya.

References Butt, Miriam 2003 Tense and aspect in Urdu. Journe´e d’e´tudes sur le temps et l’aspect. Paris. ling.uni-konstanz.de/pages/home/butt/pargram04-hnd.pdf 19/7/05 Creswell, Cassandre 2002a Syntactic form and discourse function in NLG. Proceedings of the Second International Natural Language Generation Conference (Student Session). July 2002. 2002b The use of emphatic reflexives with NPs in English. In Information Sharing, Kees van Deemter and Roger Kibble (eds.), Stanford, CA: CSLI Publications. (first presented at Workshop on the Generation of Nominal Expressions, European Summer School on Logic Language and Information, August 1999). 2004 The discourse function of verum focus in wh-questions. ftp://ling.upenn.edu/papers/ students/Creswell/verum.ps. 12/6/04

Some aspects of the meaning of the Hindi particle hii

121

Hooper, Rev. W. 1901 Helps to the Attainment of Hindustani Idiom. London: Christian Literature Society for India. Ko¨nig, Ekkehard 1991 The Meaning of Focus Particles. London: Routledge. Ko¨nig, Ekkehard and Peter Siemund 1999 Intensifiers as targets and sources of semantic change. In Meaning change— Meaning variation, Regine Eckardt and Klaus von Heusinger (eds.), Vol. I, 97– 109. (Workshop held at Konstanz, Feb. 1999). Lambrecht, Knud 1994 Information Structure and Sentence Form: Topic, Focus and the Mental Representations of Discourse Referents. Cambridge: Cambridge University Press. 2001 A framework for the analysis of cleft constructions. Linguistics 39: 463–516. Masica, Colin 1993 The Indo-Aryan Languages. Cambridge: Cambridge University Press. McCoy, Svetlana 2001 Connecting information and discourse structure levels through ‘‘Kontrast:’’ Eviˇ E, and VED 0 . In Information Strucdence from Colloquial Russian Particles -TO, Z ture, Discourse Structure, and Discourse Semantics, Ivana Kruij¤-Korbayova´ and Mark Steedman (eds.), 85–98. Workshop Proceedings. ESSLLI 2001 (13 th European Summer School in Logic, Language, and Information). Molna´r, Vale´ria 2001 Contrast from a contrastive perspective. In Information Structure, Discourse Structure, and Discourse Semantics, Ivana Kruij¤-Korbayova´ and Mark Steedman (eds.), 99–114. Workshop Proceedings. ESSLLI 2001 (13 th European Summer School in Logic, Language, and Information). Mosegaard Hansen, Maj-Britt 2002 From aspectuality to discourse marking: the case of the French de´ja` and encore. Belgian Journal of Linguistics 16: 23–51. Pons Borderı´a, Salvador 2005 A functional approach to the study of discourse markers. In Approaches to discourse particles, Kerstin Fischer (ed.). Amsterdam, Elsevier Press. Prince, Ellen F. 1986 On the syntactic marking of presupposed open propositions. In Papers from the Parasession on Pragmatics and Grammatical Theory, 22 nd Regional Meeting, A. Farley, P. Farley and K.-E. McCullough (eds.), 208–222. Chicago Linguistic Society. Schwenter, Scott A. and Shravan Vasishth 2000 Absolute and relative particles in Spanish and Hindi. To appear in Proceedings of Berkeley Linguistics Society Conference. Shapiro, Michael 1989 A Primer of Modern Standard Hindi. Delhi: Motilal Banarsidass. 1999 Hindi to as a discourse marker. In Vidyopaasanaa: Studies in honour of Harivallabh C. Bhayani, P.J. Mistry and Bharati Modi (eds.). Mumbai: Image Publications Pvt. Ltd. Sharma, Devyani 1999 Nominal clitics and constructive morphology in Hindi. In Proceedings of the LFG99 Conference, Miram Butt and Tracy Holloway King (eds.). Stanford, CA: CSLI Publications. Online. http://www-csli.stanford.edu/publications/ March 2005. Umbach, Carla 2004 On the notion of contrast in information structure and discourse structure. Journal of Semantics 21: 155–175.

122

Terry Varma

Varma, Terry 2003 Scalarity and let alone constructions in Hindi. In The Yearbook of South Asian Languages and Linguistics 2003, Rajendra Singh (ed.), 71–95. Berlin/New York, Mouton de Gruyter. Verma, Manindra K. 1971 The Structure of the Noun Phrase in English and Hindi. Delhi: Motilal Banarsidass. Ward, Gregory L. 1990 The discourse functions of VP preposing. Language 66: 742–763. Ward, Gregory L. and Betty Birner 2004 Information structure and non-canonical syntax. In The Handbook of Pragmatics, Laurence R. Horn and Gregory Ward (eds.) 153–174. Malden, Mass.: Blackwell.

B

Regional Reports, Reviews and Abstracts

Pakistan: Research and developments in linguistics and language study ELENA BASHIR

Since the previous Yearbook report on Indo-European languages in Pakistan (Rahman 1998), there have been many new developments, along several lines: (1) the study of linguistics as a discipline in Pakistan; (2) research on languages of Pakistan; (3) teaching of the languages of Pakistan. This report will discuss only work going on or published exclusively within Pakistan. Thus, following Yearbook guidelines, collaborative research by Pakistani scholars and those based outside Pakistan, work by Pakistanis or Pakistanists residing outside Pakistan and work by Pakistanis published outside Pakistan are not included. With a few exceptions, the bibliography includes works published since 1998; the reader is referred to Rahman (1998) for coverage and discussion of the situation up to that point.

1. The discipline of linguistics in Pakistan 1.1. New developments toward linguistics as an autonomous discipline Rahman (1998) begins his report on languages and linguistics in Pakistan by saying: ‘‘Pakistan does not have a university department or institute of higher education and research in linguistics.’’ The rest of that report details the consequences of that fact. In brief, according to Rahman, as of 1998, linguistics work in Pakistan was dominated by the philological and comparativist traditions or by ideological agendas, and linguists who lived and worked in Pakistan were on the periphery of the field of linguistics. Subsequent to that, Rahman (2005: 403) found ‘‘. . . more courses in linguistics and applied linguistics in several departments of English and other languages than before,’’ but still (2005: 416) concludes that ‘‘most linguists working in Pakistan actually work on the peripheries of the field of linguistics.’’ The present report is more optimistic than Rahman (1998) or (2005), for several reasons. In addition to initiatives for research on and study of Pakistani languages, there are several projects within Pakistan specifically related to linguistics as a discipline. The foremost of these is the work being done at

126

Elena Bashir

the Center for Research in Urdu Language Processing (CRULP) at the National University of Computer and Emerging Sciences (NUCES) campus in Lahore, under the leadership of Sarmad Hussain. The work of CRULP, http://www.crulp.org/, includes projects in three main areas: speech processing, language processing and computational linguistics, and script processing. Their speech processing lab is working on developing an Urdu speech interface for computers, which, crucially for the field of linguistics, involves basic research in phonetics, phonology, speech systems, stochastic pattern matching and other more practically oriented topics. The language processing and computational linguistics section is working on producing language applications for Urdu, including a machine translation system, Urdu grammar and spell checkers and Urdu lexicon development. Work is currently in progress on online Urdu (currently in beta version) and Sindhi (under development) dictionaries, which in addition to word meanings will include essential grammatical information, and examples of usage. Basic research related to these projects includes work on the morphology, syntax, semantics and pragmatics of Urdu and other Pakistani languages. In connection with these projects, a workshop was conducted in the summer of 2003 by Miriam Butt, a computational linguist and syntactician with a research concentration on Urdu. Linguistics topics taught included introductions to grammar development, Lexical Functional Grammar (LFG), and morphology (with an emphasis on Urdu morphology), morphological analysis via finite-state machines, grammatical functions, the Urdu case system, argument manipulating morphology and complex predicates. The architecture of the speech translation program makes use of functional structures (F-Structure) for Urdu and English, an application of the LFG concepts taught in the 2003 summer workshop. The work of the Script Processing Lab has resulted in the development and dissemination of a series of by now widely used Nafees Naskh and Nastaliq fonts for Urdu, a series of keyboard layouts for typing in Urdu, and a speech translation program, all of which are freely available to the public via the CRULP web site. The script processing unit is also working on developing optical character recognition (OCR) systems for Urdu. Underlying these applications is basic research in font development formalisms, image processing, application of rule-based systems and stochastic pattern matching. In addition to the output of these major projects, there have so far been published 70 papers by CRULP students, most of them on phonetic and phonological analysis of several Pakistani languages: Urdu, Panjabi, Pashto, and Siraiki. These are accessible online at the CRULP website: http://www.crulp.org/English%20Site/main.html, and have also been published by the Muqtadarah Qaumi Zaban [National Language Authority],

Pakistan: Research and developments in linguistics

127

Islamabad, in three issues of their journal Axbar-e Urdu (April–May 2002; June 2003; and December 2004). These papers include: Saleem, Kabir, and Khurram (2002); Kabir and Saleem (2002); S. I. Sheikh (2002); Shahid (2002); Aziz (2002); Mazhar (2002); S. Bashir (2002); Khalid (2002); Afzal (2002); Rafiq (2002); Rafique (2002); Riaz (2002); M. J. Sheikh (2002); Rizvi (2002); Baig (2002); Shah (2002); Wyne (2002); Akram (2002); Khan (2002); Kabir (2002); Karamat (2002); Nazar (2002); Ghazali (2002); Nayyar (2002); Saleem (2002); Nawaz (2002); Zia (2002); Sarfraz (2002); Wali (2002); Hussain (2002); Anwar (2002); Samad (2002); Wali et al. (2002a, 2002b); Riaz, Rafique, and Shahid (2002); Naseem (2003); Ijaz and Anwar (2003); Munir and Shahbaz (2003); Khan and Waqar (2003); Nayyer and Madni (2003); Rehman and Awan (2003); Nisar and Baqir (2003); Munir and Mohsin (2003); Ijaz and Moin (2003); Ahmad and Nasir (2003); Bokhari and Pervez (2003); Abbasi and Ahmad (2003); Fatima and Adan (2003); Bashir and Zia (2003); Ijaz (2003); Latif (2003); Azad and Ahmed (2003); Karamat, Nawaz, and Sarfraz (2003); Wali and Hussain (2003); Ahmed et al. (2003), Kabir et al. (2003), bin Shahbaz et al. (2003); Hussain et al. (2004); Sarwar, Ahmed, and Tarar (2004); Khurshid, Usman, and Butt (2004); Bokhari and Bokhari (2004); Dawood, Shahid, and Ahmed (2004); Khan, Iqbal, and Asad (2004); Raza, Pervaiz, and Iqbal (2004); Raza, Zahid, and Raza (2004); Usman, Ali, and Masood (2004); Ejaz (2004); Khan, Anjum, and Usman (2004); Saleem, Nazar, and Muzammil (2004); and Ahmad (2004). An additional aspect of this student work is that, beginning with the 2002–3 Annual Student Report, most of the student papers originally written in English were translated into and simultaneously published in Urdu, both on the CRULP web site and in Axbar-e Urdu. This was done with the aims of demonstrating that technical work in linguistics can be written in Urdu and further developing the technical vocabulary in the field. Computational linguistics work has also begun at the University of Peshawar in the Computer Science Department, under the guidance of M. A. Khan, who says (p.c. 5 February 2006) that he and his students are working on computational linguistics projects but feel the lack of background training in theoretical linguistics. At the time of this communication there are six M. Sc. students in Computer Science who are working on the following computational linguistics related projects in Pashto and Urdu: Pashto and Urdu syntax, discourse analysis of Urdu, corpus-based morphology of Pashto, Urdu, and Hindko, and computational phonology of Pashto. In addition, some B. C. S. and M. Sc. groups are working on topics such as ellipsis resolution and example-based machine translation. Some of this work is described in Abid, Begum and Kulsoom (1995), and Khan and tuz Zuhra (2005a, 2005b).

128

Elena Bashir

Importantly, the impetus for these two programs, which are already under way, is coming from the fields of computer science and information technology, and is a part of the Government of Pakistan’s e¤ort to improve basic education and research in science and technology, particularly information technology development and localization in the languages of Pakistan. This direction represents a break with the historical association of linguistics with the humanities (the philological tradition, connection with English language teaching) or the social sciences. Recent communications from Tariq Rahman indicate that a proposal for a Department of Linguistics at Quaid-i-Azam University is also making progress. If this proposal is approved and the Department established, a prerequisite for its sustainability will be assembling su‰cient faculty to teach a variety of basic linguistics courses. In addition, there is a plan to eventually start a Department of Linguistics at the University of Peshawar (Badshah Munir Bukhari, p.c.), the initial stages of which will be a Research Centre for Persian and other languages of NWFP. At present, this is to be a research rather than a teaching center. The Frontier Language Institute, a privately run NGO operating in Peshawar, o¤ers workshops and training in techniques of field research and maintains a web site with articles on languages of NWFP and the Northern Areas: http://www.fli-online.org/. At the University of Balochistan too a plan has been prepared to establish a Department of Linguistics, and it is hoped that it will soon be approved by the Academic Council of the University. According to Razzak Sabir, the originator of the project, it will probably take a year or two to plan for courses and syllabi and to recruit suitable teaching personnel (Sabir p.c. 21 February 2006). 1.2. Linguistics as an ancillary subject In addition to e¤orts toward dedicated departments of linguistics, individual courses in linguistics are increasingly being taught as part of the curricula of other departments, often English. Rahman (2005) contains a useful annex listing the linguistics courses taught and degrees awarded at Azad Jammu and Kashmir University, Bahauddin Zakariya University (Multan), International Islamic University, National University of Modern Languages, Karachi University, Kinnaird College University, Peshawar University and the University of the Punjab. The University of the Punjab, for example, o¤ers a one-year Diploma in Linguistics with papers in Fundamentals, Phonetics and Phonology, Morphology and Syntax, Semantics and Lexicography, Stylistics and Pragmatics, Practical in General Linguistics, and Ear Training and Performance, http://www.pu.edu.pk/course/description.asp? programid=605. Bahauddin Zakariya University in Multan o¤ers a M. Sc.

Pakistan: Research and developments in linguistics

129

in applied linguistics and an M. Phil in linguistics in the English Department. Additionally, the University of Sindh (Jamshoro) includes two courses in linguistics in its two-semester Master of Arts (Hons) in Sindhi, http://arts .usindh.edu.pk/sindhi/Ma-sindhi.htm. 1.3. Desiderata We have seen that there are now at least four Pakistani universities who have either started or plan to start dedicated departments of linguistics. It seems that the conditions necessary for the sustainable development of linguistics in Pakistan may now exist or be on the verge of existing, and linguistics may be, as the development economists say, (almost) ready for ‘‘takeo¤ ’’. Motivation is coming from two quite di¤erent but complementary directions. One is the need for information technology development and the other the need for cultural identity a‰rmation. Impetus from both of these societal forces seems to be converging in the discipline of linguistics. A crucial enabling condition will be the development of a critical mass of qualified people to teach courses in the basic sub-disciplines of linguistics: phonetics, phonology, syntax, semantics, historical linguistics, typological linguistics and computational linguistics. In addition to those persons already mentioned, linguists in Pakistan known to this writer are Nasim Akhtar Raja, with a Ph. D. from the University of Essex (Raja 2000), Khalid Khan Khattak, a linguistics Ph. D. from School of Oriental and African Studies (SOAS, Khattak 1988), and Abdul Razzak Sabir, Ph. D., University of Balochistan (Sabir 1994). The first cohorts of students to emerge from the computational linguistics programs at CRULP or Peshawar will greatly add to the pool of personnel. A comprehensive, long-term approach to this e¤ort will require attention not only to programs to educate linguists within the country, but also to provision for suitable professional employment for them. In this writer’s opinion, coordination and cooperation among the institutions beginning the teaching of linguistics could maximize the utilization of scarce human and infrastructural resources, for example by developing unique areas of specialization. A second necessity will be availability of good textbooks and information resources for students. The publication of a good introductory linguistics textbook in Pakistan is an important priority, since the prices of foreign publications are prohibitive. However, the increasing number of resources available by the internet makes this problem somewhat less severe than it has been in the past. It is important and historically interesting that the technology sector, specifically computer science, is leading the way. However for work in linguistics based upon or with potential applications for Pakistani languages to progress, an informational infrastructure of good descriptive work on

130

Elena Bashir

the languages of Pakistan is necessary. Good empirical data are the foundation for theory building in the natural and social sciences. In the case of linguistics, this means that good descriptive studies are necessary before typological and other theoretical work can be informed and advanced by data from Pakistani languages. Conversely, knowledge of the elements of descriptive theory is necessary in order to produce good descriptions. Dryer (forthcoming) distinguishes between descriptive theories, which describe how languages are, and explanatory theories, which explain why languages are the way they are, and argues that ‘‘basic linguistic theory’’ (Dixon 1997), provides a sound foundation on which to base work in constructing grammars. Basic linguistic theory is an approach grounded in structuralism and integrating newer insights from semantics and typological studies. It expresses its insights in ordinary language rather than formalisms. Descriptive theory, Dixon and Dryer argue, is not to be looked down on as inferior to explanatory theory, but recognized as a di¤erent enterprise. Says Dixon (1997: 134): ‘‘The myth is that the work done by the ‘theoreticians’ is more di‰cult, more important, more intellectual, altogether on a higher plane than the basic work undertaken by the descriptivists. . . . This is wrong. . . .’’ The relevant point here is to argue that one need not be pessimistic about the possibilities for doing good linguistic work in Pakistan, and that good descriptive work should not be considered as merely on the periphery of linguistics. Given that Pakistan is home to many little-studied and in some cases endangered languages, descriptive work on these languages is a basic imperative. In the words of Dixon (1997: 144–5), ‘‘the most important task in linguistics today . . . is . . . to describe languages, while this can still be done. Self-admiration in the looking glass of formalist theory can wait; that will always be possible. Linguistic description must be done now.’’ Another aspect of building an information infrastructure is the development of corpora, i.e. databases of language data/texts, in the languages of Pakistan. Some corpora development work has been done for Urdu outside Pakistan, e.g. the EMILLE corpora project, which has developed corpora of Urdu and Panjabi, among other South Asian languages. A corpus of about one million Pashto words was prepared in the 1980s by Wilma Heston in collaboration with the Pashto Academy. This corpus is a transliteration of mostly Pakistani Pashto texts into ASCII. It includes adult literacy texts, newspaper articles, selections from literary journals, and classical poetry and folk texts. The corpus can be made available for research purposes. Corpus development in the languages of concern is an activity which could be readily integrated into the new linguistics programs which will hopefully emerge, or into the work of existing institutions like the Sindhi Language Authority, Pashto Academy, Balochi Academy, Brahui Academy

Pakistan: Research and developments in linguistics

131

and others. Like descriptive work, corpus linguistics o¤ers rich empirical data on the ways languages actually function. The kind of localization work already in progress for Urdu needs to be extended to the other Pakistani languages. This includes font development, and, importantly, e¤orts for the inclusion in the Unicode standard of all characters necessary for writing Pakistani languages. Font development is an area in which much progress is being made, e.g. the Urdu fonts developed by CRULP, and work on Sindhi fonts and Unicode implementation by Abdul-Majid Bhurgri (Bhurgri n.d.). Although Rahman (2005: 420) points out that e¤orts in Pakistan toward script standardization, vocabulary development and research on languages and literatures have historically been related to identity and ethnic politics, this need not necessarily be the case. These e¤orts can also be regarded as necessary steps toward the construction of a national information infrastructure. Such an infrastructure will contribute not only to the development of the discipline of linguistics, but to educational and economic development as well. In this connection, the contributions of the Muqtadarah Qaumi Zaban, with the leadership of Fateh Muhammad Malik, in encouraging preparation of bilingual and multilingual dictionaries, and publishing research related to all the languages of Pakistan are valuable aspects of the corpus planning aspect of language development, which will allow these languages to function in a wider range of functional domains and maintain and advance the status of the languages. 2. Research on the languages of Pakistan Since Urdu is to date the best studied of the languages of Pakistan, it is not surprising that theoretical work treating Pakistani languages has begun with Urdu. Recent Ph. D. dissertations on Urdu include the following. Lodhi (2004) is a computationally oriented work aimed at developing an Urdu character pattern classification, representation and recognition system which can classify patterns even if they are deformed by transformations or in the presence of noise. Hussain (1997) is an instrumental investigation of how the phonetic properties of Urdu vocalic and consonantal segments change with stress and the theoretical implications of these phonetic changes. Most of the other research on Urdu concerns literature or secondlanguage acquisition. The situation with regard to linguistic work on Panjabi is quite di¤erent. Despite the fact that Panjabi has the largest number of first-language speakers of any single language in Pakistan, there is very little linguistic research on it in Pakistan. Malik (1977), written in Panjabi, has chapters on a general introduction to linguistics, Panjabi phonetics, Panjabi morphology,

132

Elena Bashir

and literary Panjabi. Perhaps the only doctoral dissertation on Panjabi qua Panjabi by a Pakistani is Nasim Akhtar Raja’s work on aspectual complex predicates in Panjabi (Raja 2000). Raja has also published two articles on a similar topic (1992, 2003). In addition to its publications mentioned elsewhere in this article, and to numerous articles in various issues of Axbar-e Urdu on the development of Urdu software, the Muqtadarah Qaumi Zaban has published several titles on the language and literature of various Pakistani languages, e.g. Sindhi (2000?), S. M. Y. Bukhari (2004), Javed (1997) on Panjabi literature and language and Parvez (2001) on Siraiki. The Sindhi Language Authority http://www.sindhila.com.pk/list.htm encourages and publishes work on and in Sindhi. Recent publications listed on their web site include primarily literature or literary studies, but also several dialect descriptions (Brohi 1992; Sindhi 1993; Junejo and Prem 1994; N. B. Baloch 2003), dictionaries (Junejo and Bughio 2004; Brohi 2005; A. S. Baloch 2003) and other language-related publications (Buksh 1992; Brohi 1994; Bolchand n.d.; Chand 2004). Other recent work on Sindhi includes Allana (1998, 2000); Bughio (2001); and Cole (2005). T. Farrell (Summer Institute of Linguistics: SIL) has written on Karachi Balochi (Farrell 1995, 2003). Though not specifically on language or linguistics, the work of Sabir Badalkhan, currently teaching at the University of Naples in Italy, is an important corpus of literary and historical studies on Pakistani Balochi (Badalkhan 2000, 2003). Recent studies on Brahui are Sabir (2003), a comparison of Brahui and Balochi morphology, and Bashir (forthcoming), a discussion of innovations in the Brahui verbal system. B. M. Bukhari’s Ph. D. thesis (Peshawar University) on lexical relationships between Khowar and Urdu has been published by the Muqtadarah Qaumi Zaban (Bukhari 2003). Inam Ullah, a native speaker of Torwali, continues his lexicographical work on Torwali (Inam Ullah 2004), part of which will be accessible on the web site of the Digital Dictionaries project at the University of Chicago (Inam Ullah, in progress). Shina (Gilgit dialect) is being researched by Carla Radlo¤ (SIL). Radlo¤ and Shakil (1998) and Radlo¤ (1999, 2003) are important contributions, especially to analysis of its phonology. Akbar (1992) is a comparative lexical study of Shina and Urdu. Several authors have published primers on Shina. Much new work on Kohistani Shina by Razwal Kohistani and Ruth Schmidt has appeared in recent years, some published in Pakistan, some in European journals (Kohistani 1999; Kohistani and Schmidt 1999, in press; Schmidt 2000, 2001, 2003a, 2003b, 2004a, 2004b; Schmidt and Kohistani 1998, 2001). According to Razwal Kohistani (p.c. 21 February 2006), work on a grammar of Kohistani Shina by Schmidt and Kohistani is well advanced, with publication expected within a year or two. Kohistani is

Pakistan: Research and developments in linguistics

133

also preparing to publish (in Urdu) on a writing system for Shina and other Dardic languages. A major new development in Burushaski studies is the forthcoming publication by the Burushaski Research Academy of the first volume of a projected multi-volume Burushaski-Urdu dictionary (Burushaski Research Academy, forthcoming 2006). According to Dr. Shahnaz Hunzai, the Chairman of the Publication Committee, this dictionary has been compiled by the Burushaski Research Academy in collaboration with the University of Karachi, Urdu Dictionary Board, and the Pakistan Academy of Letters Sindh. In addition, the Burushaski Research Academy has recently published four other books books about the language, two mainly in Urdu (Hunzai 2003, Hunzai 2004), and two employing both Burushaski and Urdu (Hunzai, n.d.-a, n.d.-b). Three new publications on Balti have come to my attention. Baltistani (2003) is a Balti-Urdu dictionary; Abadi et al. (2004) is a primer using Perso-Arabic script; and Lobsang (1995) is a grammatical sketch of Balti. Barki (1999) is an appeal for the preservation of Ormuri, an endangered Eastern Iranian language. 3. Teaching of Pakistani languages Several Pakistani universities o¤er advanced degrees in a few Pakistani languages. For example, the University of Balochistan o¤ers an M. A. in Balochi, Brahui and Pashto. At Sindh University M.A., and at the University of Karachi, M. A. and M. Phil. programs in Sindhi are o¤ered. Bahauddin Zakariya University in Multan o¤ers an M. Phil. and Ph. D. in Urdu. In addition to a large Urdu department, the University of the Punjab now has departments of Pashto, Sindhi, and Regional Languages. It o¤ers the M. A. and Ph. D. in Panjabi. An important recent development is the creation of the Department of Pakistani Languages at Allama Iqbal Open University (AIOU), http:// www.aiou.edu.pk/DeptDetail.asp?DeptID=47. This department o¤ers a M. Phil. program in Pakistani languages, o¤ering national-level courses on languages such as Balochi, Brahui, Pushto, Sindhi, Punjabi, Siraiki, Hindko, Kashmiri, Gojri, Pahari, Balti, Shina, Burushaski, Torwali, Gawri, Khowar and Wakhi. A multi-volume textbook including chapters on these languages is being prepared, with chapters on each language by experts in the respective languages, is in preparation. By summer 2005, the first two volumes, with units on Balochi, Sindhi, Pashto, Panjabi, and Kashmiri were complete (Department of Pakistani Languages, Allama Iqbal Open University 2003). The medium of instruction of this program is Urdu but the students can write their theses in Urdu, English, or their own native language.

134

Elena Bashir

In addition to the programs discussed above several collaborative initiatives for the teaching of Pakistani languages are going on, most of them involving applications of internet technology. One such project on lessertaught languages of Pakistan, based on work carried out in Pakistan, is disseminated through the Digital dictionaries project at the University of Chicago (USA) http://dsal.uchicago.edu/dictionaries/. A Pashto online dictionary by Heston (Heston in progress), a Torwali online dictionary by Inam Ullah (Inam Ullah in progress), and a Khowar online dictionary by Rahmat Karim Baig and Maula Nigah edited by Elena Bashir (Bashir, Baig and Nigah in progress) are in development. Much discussion of language identity and language planning issues by both Pakistan-based and diasporic groups and individuals, and online dissemination of literature and fonts also takes place utilizing the internet. For a discerning reader, these internet sites can be sources of much useful material, as many o¤er access to cultural materials such as canonical literary texts, music, and computer fonts. Such access is an example of the ways in which technology can help to equalize access to information for those physically distant from traditional libraries. Three examples of such sites are: Sindhi, http://www.freewebs.com/majidbhurgri/; Panjabi, http://www .apnaorg.com; Balochi, http://users.tpg.com.au/users/goshti/.

References Abadi, Muhammad Yusuf, et al. 2004 Balti qaida [Balti Primer]. Islamabad: Muqtadarah Qaumi Zaban. Abbasi, Abdul Rafay, and Naveed Ahmad 2003 Urdu viseme identification. Axbar-e Urdu 19(6): 217–221. Islamabad: Muqtadarah Qaumi Zaban. Abid, Mohammad; Rashid Begum and Bibi Kulsoom 1995 Urdu anaphora resolution. Scientific Khyber 8(2): 99–120. Afzal, Muhammad Usman 2002 Speech synthesis for Urdu vowels using HLSYN. Axbar-e Urdu 18(4–5): 82–94. Islamabad: Muqtadarah Qaumi Zaban. Ahmad, Tafseer 2004 angrezi urdu naqal harfi [English to Urdu transliterator]. http://www.crulp.org/ English%20Site/main.html Ahmad, Arslan, and Shakil Nasir 2003 Phonotactic distribution of consonantal sounds in Urdu. Axbar-e Urdu 19(6): 228– 232. Islamabad: Muqtadarah Qaumi Zaban. Ahmed, Fuaad, et al. 2003 Urdu computational lexicon. Axbar-e Urdu 19(6): 163–167. Islamabad: Muqtadarah Qaumi Zaban. Akbar, Akbar Hussain 1992 Urdu aur Shina ke mushtarik alfaz [Words common to Urdu and Shina]. Islamabad: Muqtadara Qaumi Zaban.

Pakistan: Research and developments in linguistics

135

Akram, Bilal 2002 Analysis of Urdu syllabification using maximal onset principle and sonority sequence principle. Axbar-e Urdu 18(4–5): 173–178. Islamabad: Muqtadarah Qaumi Zaban. Allana, Ghulam Ali 1998 Papers on Sindhi language & linguistics: compiled and written by G. A. Allana, ed. Allah Rakhio Butt. Jamshoro: Institute of Sindhology, University of Sindh. 2002 The origin and growth of Sindhi language. Jamshoro: Institute of Sindhology, University of Sindh. Anwar, Muhammad Jamil 2002 Nicknaming patterns in Urdu language. Axbar-e Urdu 18(4–5): 273–282. Islamabad: Muqtadarah Qaumi Zaban. Azad, Muhammad Umer, and Salman Ahmed 2003 Change in short vowel in the context of /h/. Axbar-e Urdu 19(6): 189–197. Islamabad: Muqtadarah Qaumi Zaban. Aziz, Umar 2002 Nasal aspirates in Urdu. Axbar-e Urdu 18(4–5): 51–57. Islamabad: Muqtadarah Qaumi Zaban. Baart, Joan L. G. 1997 The sounds and tones of Kalam Kohistani. Islamabad: National Institute of Pakistan Studies, Quaid-i-Azam University, and Summer Institute of Linguistics. 1999 A sketch of Kalam Kohistani grammar. Islamabad: National Institute of Pakistan Studies, Quaid-i-Azam University, and Summer Institute of Linguistics. 2003 Tonal features on languages of northern Pakistan. In Pakistani languages and society: Problems and prospects, Joan L. G. Baart, and Ghulam Hyder Sindhi (eds.), 132–144. Islamabad: National Institute of Pakistan Studies, Quaid-i-Azam University and Summer Institute of Linguistics. 2004 Contrastive tone in Kalam Kohistani. Linguistic Discovery 2(2): 1–21. Baart, Joan L. G., and Esther L. Baart-Bremer 2001 Bibliography of languages of northern Pakistan. Islamabad: National Institute of Pakistan Studies, Quaid-i-Azam University and Summer Institute of Linguistics. Babrakzai, Farooq 2000 Topics in Pashto syntax. Ph. D. diss., University of Hawaii. Badalkhan, Sabir 2000 Ropes break at the weakest point: Some examples of Balochi proverbs with background stories. In Proverbium: Yearbook of international proverb scholarship, Wolfgang Mieder (ed.), Vol. 17: 43–69. Burlington: University of Vermont. 2003 Language contact in Balochistan and its impact on Balochi personal names. In Jahani and Korn (eds.), 279–312. Baig, Mirza Fahd Arshed 2002 The duration of vowel representing kasre izafat in the compound words of Urdu language. Axbar-e Urdu 18(4–5): 137–149. Islamabad: Muqtadarah Qaumi Zaban. Baloch, Abdul Sattar 2003 Urdu-Sindhi Lughat [Urdu-Sindhi dictionary]. Hyderabad, Sindh: Sindhi Language Authority. Baloch, Nabi Baksh Khan 2003 Jatki boli [The Jatki dialect]. Hyderabad, Sindh: Sindhi Language Authority. Baltistani, Muhammad Ali Shah 2003 Balti Urdu lughat [Balti-Urdu dictionary]. Islamabad: Muqtadarah Qaumi Zaban. Barki, Rozi Khan 1999 Maax a xway zabaan tah goor GaaDah zar zhiin? [in Ormuri, Will we let our language go to the grave?]. Islamabad.

136

Elena Bashir

Bashir, Elena 1999 The Urdu and Hindi ergative postposition ne: Its changing role in the grammar. Yearbook of South Asian Languages and Linguistics 1999, Rajendra Singh (ed.), 11–36. New Delhi: Sage Publications. Forth- Contact-induced change in Khowar. In New perspectives on Pakistan, Saeed Shafqat coming (ed.), 205–237. Karachi: Oxford University Press. Bashir, Elena, Rahmat Karim Baig, and Maula Nigah to A digital dictionary of Khowar with audio: 1000 words. appear Bashir, Maryam, and Fariha Zia 2003 Existence of aspirated l, r and R in Urdu language. Axbar-e Urdu 19(6): 207–212. Islamabad: Muqtadarah Qaumi Zaban. Bashir, Sheraz 2002 Variation between palatal voiced fricative and palatal approximant in Urdu spoken language. Axbar-e Urdu 18(4–5): 70–74. Islamabad: Muqtadarah Qaumi Zaban. Bhurgri, Abdul-Majid n.d. Enabling Pakistani languages through Unicode. accessed February 19, 2006 from: http://download.microsoft.com/download/1/4/2/142aef9f-1a74-4a24-b1f4782d48d41a6d/ PakLang.pdf#search=‘Abdul%20Majid%20Bhurgri’ bin Shahbaz, M. et al. 2003 Corpus based data acquisition system for lexical development using HMM. Axbar-e Urdu 19(6): 147–154. Islamabad: Muqtadarah Qaumi Zaban. Bokhari, I¤at Haider, and Ri¤at Haider Bokhari 2004 Existence of aspirated l and m in Urdu language. Axbar-e Urdu 20(12): 176–180. Islamabad: Muqtadarah Qaumi Zaban. Bokhari, Rafia, and Sadaf Pervez 2003 Syllabification and resyllabification in Urdu. Axbar-e Urdu 19(6): 222–227. Islamabad: Muqtadarah Qaumi Zaban. Bolchand, Dulamal n.d. A manual of Sindhi. Hyderabad, Sindh: Sindhi Language Authority. Brohi, Dad Muhammad 1992 Sibi ji boli [The dialect of Sibi]. Hyderabad, Sindh: Sindhi Language Authority. 1994 Sindhi Brahiwi boli jo taqabili jaizo [A comparison of Sindhi and Brahui]. Hyderabad, Sindh: Sindhi Language Authority. 2005 Sindhi Brahvi lughat [Sindhi-Brahui dictionary]. Hyderabad, Sindh: Sindhi Language Authority. Bughio, Qasim 2001 A comparative sociolinguistic study of rural and urban Sindhi. Munich: Lincom Europa. Bukhari, Badshah Munir 2003 Urdu aur Khowar ke lisani rawabit [Linguistic relationships between Urdu and Khowar]. Islamabad: Muqtadarah Qaumi Zaban. Bukhari, Syed Muhammad Yusuf 2004 Kashmiri aur Urdu zaban ka taqabli mutaliya [A comparative study of Kashmiri and Urdu]. Islamabad: Muqtararah Qaumi Zaban. Buksh, Nabi 1992 Sindhi scripts and orthography [Sindhi soratakhati ain khatati ]. Hyderabad, Sindh: Sindhi Language Authority. Burushaski Research Academy To Burushaski lughat (Burushaski-Urdu Dictionary). Karachi: University of Karachi. appear

Pakistan: Research and developments in linguistics

137

Butt, Miriam 1995 The Structure of Complex Predicates in Urdu. Stanford, CA: Center for the Study of Language and Information Publications. Chand, Muhammad Umar 2004 Sindhi boli: Lisaniyati jagrafi adhariyoon ain lafzi tarteeb. [Sindhi: Linguistic, geographical, and lexical order]. Hyderabad, Sindh: Sindhi Language Authority. Dawood, Haroon, Rehan Shahid, and Touqeer Ahmed 2004 Intonation patterns in Punjabi. Axbar-e Urdu 20(12): 170–175. Islamabad: Muqtadarah Qaumi Zaban. Department of Pakistani Languages, Allama Iqbal Open University 2003 Pakistani adab, Vols. 1 and 2. Islamabad: Allama Iqbal Open University. Dixon, Robert Malcolm Ward 1997 The Rise and Fall of Languages. Cambridge: Cambridge University Press. Dryer, Matthew to Descriptive theories, explanatory theories, and basic linguistic theory. In Catching appear language: Issues in appear Grammar Writing, Felix Ameka, Alan Dench and Nicholas Evans (eds.). Berlin: Mouton de Gruyter. Ejaz, Madiha 2004 Phonological behavior of multiple aspirated consonants in Urdu. Axbar-e Urdu 20(12): 140–149. Islamabad: Muqtadarah Qaumi Zaban. Farrell, Tim 1995 Fading ergativity? A study of ergativity in Balochi. In Subject, Voice, and Ergativity, David C. Bennett, Theodora Bynon, and B. George Hewitt (eds.), 218–243. London: School of Oriental and African Studies, University of London. 2003 Linguistic influences on the Balochi spoken in Karachi. In Jahani and Korn (eds.), 169–209. Fatima, Nakhat, and Rabea Adan 2002 Vowel structure of Urdu. Axbar-e Urdu 19(6): 213–216. Islamabad: Muqtadarah Qaumi Zaban. Ghazali, Muhammad Ahmad 2002 Urdu syllable templates. Axbar-e Urdu 18(4–5): 203–209. Islamabad: Muqtadarah Qaumi Zaban. Hallberg, Daniel G., and Callinda E. Hallberg 1999 Indus Kohistani: A Preliminary Phonological and Morphological Analysis. Islamabad: National Institute of Pakistan Studies, Quaid-i Azam University and Summer Institute of Linguistics. Heston, Wilma 1992 A Dictionary of Basic Pashto: Frequency List I, Project Description and Samples, and Frequency List II. ERIC Document Reproduction Service No. ED 347 841. to A Pashto-English digital dictionary with audio: 1000 words. appear Hunzai, Nasir al-Din Nasir n.d.-a. De´karan [Burushaski research papers]. Karachi: Burushaski Research Academy. n.d.-b. Asqu´rin´e Bası´. Karachi: Burushaski Research Academy. 2003 Shimol Boq [A grammar and pronunciation book of Burushaski language]. Karachi: Burushaski Research Academy. 2004 Burushaski Jawahir Paray [Some glimpses of Burushaski language]. Karachi: Burushaski Research Academy. Hussain, Sara 2002 Prosody in Urdu poetry: A phonological approach. Axbar-e Urdu 18(4–5): 263–272. Islamabad: Muqtadarah Qaumi Zaban.

138

Elena Bashir

Hussain, Sarmad 1997 Phonetic correlates of lexical stress in Urdu. Ph. D. diss., Northwestern University (USA). Hussain, Sarmad, Ahmed Sheraz Butt, Mohammad Asad, and Salahuddin Chaudhry 2004 Rule-based expert system for Urdu nastaleeq justification. Axbar-e Urdu 20(12): 194–198. Islamabad: Muqtadarah Qaumi Zaban. Ijaz, Adnan, and Muhammad Awais Anwar 2003 Acoustic di¤erence between dental and alveolar stops in Urdu. Axbar-e Urdu 19(6): 281–284. Islamabad: Muqtadarah Qaumi Zaban. Ijaz, Fatima, and Muhammad Raihan Moin 2003 Gender and number generation of Urdu nouns. Axbar-e Urdu 19(6): 233–237. Islamabad: Muqtadarah Qaumi Zaban. Ijaz, Madiha 2003 Phonemic inventory of Pashto. Axbar-e Urdu 19(6): 203–206. Islamabad: Muqtadarah Qaumi Zaban. Inam Ullah 2004 Lexical database of the Torwali dictionary. Paper presented at the Asia Lexicography Conference, Chiangmai, Thailand, 24–26 May, 2004. Accessible at http://www .fli-online.org/documents/torwali/Inam_paper_2004.pdf to A digital Torwali-English dictionary with audio. appear Jahani, Carina 1989 Standardization and Orthography in the Balochi Language. Acta Universitatis Upsaliensis, Studia Iranica Upsalensia I. Stockholm: Almquist & Wiksell, International. Jahani, Carina, and Agnes Korn (eds.) 2003 The Baloch and Their Neighbours: Ethnic and Linguistic Contact in Balochistan in Historical and Modern Times. Wiesbaden: Reichert. Javed, Inam ul Haq 1997 Panjabi zaban-o-adab ki muxtasar tarix [A brief history of Panjabi language and literature]. Islamabad: Muqtadarah Qaumi Zaban. Junejo, Abdul Jabbar, and Hidayat Prem 1994 Thar ji Boli [The dialect of Thar]. Hyderabad, Sindh: Sindhi Language Authority. Junejo, Tanveer, and Muhammad Qasim Bughio 2004 Sociology ji lughat [Dictionary of sociology]. Hyderabad, Sindh: Sindhi Language Authority. Kabir, Hammad, Shanza Nayyer, Jahangir Zaman, and Sarmad Hussain 2003 Two pass parsing implementation for an Urdu grammar checker. Axbar-e Urdu 19(6): 155–162. Islamabad: Muqtadarah Qaumi Zaban. Kabir, Hasan 2002 Syllabification rules in Punjabi. Axbar-e Urdu 18(4–5): 183–188. Islamabad: Muqtadarah Qaumi Zaban. Kabir, Hassan, and Abdul Mannan Saleem 2002 Speech assessment methods in phonetic alphabet. Axbar-e Urdu 18(4–5): 35–38. Islamabad: Muqtadarah Qaumi Zaban. Karamat, Nayyara 2002 Phonemic inventory of Punjabi. Axbar-e Urdu 18(4–5): 189–197. Islamabad: Muqtadarah Qaumi Zaban. Karamat, Nayyara, Sadaf Nawaz, and Huda Sarfraz 2003 Unigram analysis of Urdu names for a speech recognition application. Axbar-e Urdu 19(6): 184–188. Islamabad: Muqtadarah Qaumi Zaban.

Pakistan: Research and developments in linguistics

139

Khalid, Nauman 2002 Existence of [v] and [w] in Urdu language. Axbar-e Urdu 18(4–5): 75–81. Islamabad: Muqtadarah Qaumi Zaban. Khan, Amad, Muhammad Arif Iqbal, and Nawab Zada Asad 2005 Existence of voiceless velar nasal in Urdu. Axbar-e Urdu 20(12): 166–169. Islamabad: Muqtadarah Qaumi Zaban. Khan, Amna, and Sahar Waqar 2003 Identification of diphthongs in Urdu and their acoustic properties. Axbar-e Urdu 19(6): 270–275. Islamabad: Muqtadarah Qaumi Zaban. Khan, Ayesha, Bushra Anjum, and Muhammad Usman 2004 Research study: Punjabi (Majhi) orthography & pronunciation contrast. Axbar-e Urdu 20(12): 138–141. Islamabad: Muqtadarah Qaumi Zaban. Khan, Jehangir Zaman 2002 Syllabification rules in Pashto. Axbar-e Urdu 18(4–5): 179–182. Islamabad: Muqtadarah Qaumi Zaban. Khan, Mohammad Abid, and Fatima-tuz-Zuhra 2005a Towards the computational treatment of the Pashto verb. Scientific Khyber 18(1): 123–141. 2005b Pashto verbs in the past temse: A computational view. Scientific Khyber 18(2): 243– 274. Khattak, Khalid Khan 1988 A case grammar study of the Pashto verb. Ph. D. thesis, Department of Phonetics and Linguistics, School of Oriental and African Studies, Faculty of Arts, University of London. Khurshid, Kiran, Salman Ahmad Usman, and Nida Javaid Butt 2004 Possibility of existence and identification of diphthongs and triphthongs in Urdu language. Axbar-e Urdu 20(12): 181–185. Islamabad: Muqtadarah Qaumi Zaban. Kohistani, Razwal 1999 Basic Shina-Urdu Dictionary: A Dictionary of the Palas Dialect. Rawalpindi: Shina Research Forum-Karakorum. Kohistani, Razwal, and Ruth Laila Schmidt 1996 Shina Qaida (Shina primer). Palas (Indus Kohistan): Himalayan Jungle Project. 1999 Shina Proverbs (Shina proverbs of Kohistan). Rawalpindi: Shina Research Forum Karakorum. in press Shina in contemporary Pakistan. In Saxena and Borin (eds.). Latif, Amna 2003 Phonemic inventory of Siraiki language and acoustic analysis of voiced implosives. Axbar-e Urdu 19(6): 198–202. Islamabad: Muqtadarah Qaumi Zaban. Lobsang, Ghulam Hassan 1995 Short Sketch of Balti Grammar: A Tibetan Dialect Spoken in Northern Pakistan. Univ. Bern, Inst. fu¨r Sprachwissenschaft, Arbeitspapiere, 34. Bern: Universita¨t. Bern, Institut fu¨r Sprachwissenschaft. Lodhi, Saeed M. 2003 Robust Urdu character recognition using Fourier descriptors. Ph. D. diss., University of Denver (USA). Malik, Shahbaz 1977 Panjaabii lisaaniyaat [Panjabi linguistics]. Lahore: Maktabah Meri Laibreri. Mazhar, Suleman 2002 Aspirated approximants in Urdu. Axbar-e Urdu 18(4–5): 58–69. Islamabad: Muqtadarah Qaumi Zaban.

140

Elena Bashir

Munir, Irfan Umer, and Muhammad bin Shahbaz 2003 Categorial perception. Axbar-e Urdu 19(6): 276–280. Islamabad: Muqtadarah Qaumi Zaban. Munir, Kashif, and Tayyaba Mohsin 2003 Existence of ingma in Urdu language and its phonological features. Axbar-e Urdu 19(6): 238–243. Islamabad: Muqtadarah Qaumi Zaban. Naseem, Tahira 2003 Acoustic analysis of Punjabi phonemes /L/ and /N/. Axbar-e Urdu 19(6): 282–285. Islamabad: Muqtadarah Qaumi Zaban. Nawaz, Sadaf 2002 Deletion rules in Urdu language. Axbar-e Urdu 18(4–5): 227–236. Islamabad: Muqtadarah Qaumi Zaban. Nayyar, Shanza 2002 Syllable stress in Urdu. Axbar-e Urdu 18(4–5): 210–214. Islamabad: Muqtadarah Qaumi Zaban. Nayyer, Rabia, and Fatima Gull Madni 2003 Analysis of intonation patterns in Urdu. Axbar-e Urdu 19(6): 259–269. Islamabad: Muqtadarah Qaumi Zaban. Nazar, Muhammad Noman 2002 Syllable templates in Urdu language. Axbar-e Urdu 18(4–5): 198–202. Islamabad: Muqtadarah Qaumi Zaban. Nisar, Adnan, and Zeeshan Baqir 2003 Phonological behaviour of aspirated consonants in Urdu. Axbar-e Urdu 19(6): 244– 247. Islamabad: Muqtadarah Qaumi Zaban. Parvez, Sajjad Haider 2001 Siraiki zaban o adab ki muxtasar tarix [Brief history of Siraiki language and literature]. Islamabad: Muqtadarah Qaumi Zaban. Radlo¤, Carla F. 1999 Aspects of the Sound System of Gilgiti Shina. Islamabad: National Institute of Pakistan Studies, Quaid-i-Azam University and Summer Institute of Linguistics. 2003 The solution to a linguistic mystery: The absence of diphthongs in the Shina language of Gilgit. In Pakistani Languages and Society: Problems and Prospects, Joan L. G. Baart and Ghulam Hyder Sindhi (eds.), 145–153. Islamabad: National Institute of Pakistan Studies, Quaid-i-Azam University and Summer Institute of Linguistics. Radlo¤, Carla F., and Shakil Ahmad Shakil 1998 Folktales in the Shina of Gilgit. Islamabad: National Institute of Pakistan Studies, Quaid-i Azam University and Summer Institute of Linguistics. Rafiq, Mustafa 2002 Quality analysis of non-nasal and non-aspirated Urdu stops, synthesized using HLSYN (a high level synthesizer). Axbar-e Urdu 18(4–5): 95–105. Islamabad: Muqtadarah Qaumi Zaban. Rafique, Muhammad Mustafa 2003 Synthesis of /ikis/ (21) and /bais/ (22). Axbar-e Urdu 18(4–5): 106–113. Islamabad: Muqtadarah Qaumi Zaban. Rahman, Tariq 1998 (Regional report on) Pakistan: Indo-European. In Singh (ed.) 1998: 184– 196. 2005 Linguistics in Pakistan: A survey of the contemporary situation. In Social Sciences in Pakistan: A profile, Inayatullah, Rubina Saigol and Pervez Tahir (eds.), 403–426. Islambad: Council of Social Sciences.

Pakistan: Research and developments in linguistics

141

Raja, Nasim Akhtar 1992 Punjabi compounds: A structural and semantic study. Unpublished M. Litt. dissertation, University of Strathclyde. 2000 Aspectual complex predicates in Punjabi. Ph. D. dissertation. University of Essex, Colchester (UK). 2003 Aspectual complex predicates in Punjabi. In Singh (ed.) 2003: 99–129. Raza, Agha Ali, Mansoor Pervaiz, and Waqas Iqbal 2003 Existence of velar nasal stops in Urdu. Axbar-e Urdu 20(12): 161–165?. Islamabad: Muqtadarah Qaumi Zaban. (note: There is a binding error with this article; consult CRULP web site version.) Raza, Sarfraz, Agha Furrukh Zahid, and Usman Raza 2004 Phonemic inventory of Sindhi and acoustic analysis of voiced implosive. Axbar-e Urdu 20(12): 155–160. Islamabad: Muqtadarah Qaumi Zaban. (note: There is a binding error with this article: consult CRULP web site version.) Rehman, Ahmed Abdul, and Salman Bashir Awan 2003 Speaker dependent features in transition from a stop to a vowel. Axbar-e Urdu 19(6): 248–258. Islamabad: Muqtadarah Qaumi Zaban. Riaz, Muhammad Khurram 2002 Synthesis of oral and nasal vowels of Urdu. Axbar-e Urdu 18(4–5): 114–121. Islamabad: Muqtadarah Qaumi Zaban. Riaz, Muhammad Khurram, Muhammad Mustafa Rafique, and Syed Raza Shahid 2002 Vowel insertion grammar. Axbar-e Urdu 18(4–5): 310–312. Islamabad: Muqtadarah Qaumi Zaban. Rizvi, Mustafa Mubashir 2002 Speaker dependent features in approximants of Urdu. Axbar-e Urdu 18(4–5): 130– 136. Islamabad: Muqtadarah Qaumi Zaban. Sabir, Abdul Razzak 1994 Baloci aur Brahvi zabano@ ke rawabit [Relations between the Balochi and Brahui languages]. Unpublished Ph. D. thesis, University of Balochistan, Quetta. 2003 Some structural and lexical similarities between Balochi and Brahui. In Carina Jahani and Agnes Korn (eds.), 211–220. Wiesbaden: Reichert. Sabir, Ghaus Baksh 1997 Balochi zaban o adab ki muxtasar tarix [A brief history of Balochi language and literature]. Islamabad: Muqtadarah-e Qaumi Zaban. Saleem, Abdul Mannan 2002 Assimilation rules in Urdu. Axbar-e Urdu 18(4–5): 215–226. Islamabad: Muqtadarah Qaumi Zaban. Saleem, Abdul Mannan, Hasan Kabir, and Muhammad Khurram 2002 Urdu consonantal and vocalic sounds. Axbar-e Urdu 18(4–5): 30–34. Islamabad: Muqtadarah Qaumi Zaban. Saleem, M. Adil, M. Mudassar Nazar, and Muein Muzammil 2004 Deletion rules in Punjabi language. Axbar-e Urdu 20(12): 132–137. Islamabad: Muqtadarah Qaumi Zaban. Saleemi, Anjum P. 2003 Case and agreement in Hindi-Urdu. In Singh (ed.) 2003: 131–153. Samad, Khurram Shahzad 2002 Phonological analysis of nicknaming in Urdu. Axbar-e Urdu 18(4–5): 283–290. Islamabad: Muqtadarah Qaumi Zaban. Sarfraz, Huda 2002 Formation of generalization words (‘‘Mohmil’’) in Urdu. Axbar-e Urdu 18(4–5): 249–255. Islamabad: Muqtadarah Qaumi Zaban.

142

Elena Bashir

Sarwar, Aiza, Sana Ahmed, and Aymen Aziz Tarar 2004 Diphthongs in Urdu Language and Analysis of their Acoustic Properties. Axbar-e Urdu 20(12): 186–197. Islamabad: Muqtadarah Qaumi Zaban. Saxena, Anju, and Lars Borin (eds.) to Lesser-Known Languages of South Asia: Status and Policies, Case Studies and Appliappear cations of Information Technology. Berlin: Mouton de Gruyter. Schmidt, Ruth Laila 2000 Typology of Shina pronouns. Berliner Indologische Studien 13/14: 201–13. 2001 Compound tenses in the Shina of Indus Kohistan. In Tohfa-e Dil: Festschrift Helmut Nespital, Dirk W. Lo¨nne (ed.). Reinbek: Wezler. 2003b Converbs in a Kohistani Shina narrative. Acta Orientalia 64: 137–152. 2004a A grammatical comparison of Shina dialects. In Himalayan Languages Past and Present, Anju Saxena (ed.), 33–55. Berlin: Mouton de Gruyter. 2004b Compound verbs in the Shina of Kohistan. Acta Orientalia 65: 19–31. Schmidt, Ruth Laila and Razwal Kohistani 1998 Paalus /kostyo’@/ Shina revisited. Acta Orientalia 59: 106–49. 2001 Nominal inflections in the Shina of Indus Kohistan. Acta Orientalia 62: 107–43. Shah, Syed Mustafa 2002 Urdu nasal consonants and their phonological behaviour. Axbar-e Urdu 18(4–5): 150–156. Islamabad: Muqtadarah Qaumi Zaban. Shahid, Syed Raza 2002 A study on glottal stops in Urdu. Axbar-e Urdu 18(4–5): 45–50. Islamabad: Muqtadarah Qaumi Zaban. Shakil, Shakil Ahmad n.d. Gilgiti Shina meN paRh likhna: ek taaruf [Reading and writing in Gilgiti Shina: an introduction]. Gilgit (?): Informal Committee for the Promotion of Shina Language and Culture. Sheikh, Muhammad Jamal 2002 Speaker dependent features in stops and a¤ricates of native Urdu speakers. Axbar-e Urdu 18(4–5): 122–129. Islamabad: Muqtadarah Qaumi Zaban. Sheikh, Saad Inam 2002 Study of aspirated continuants in Urdu. Axbar-e Urdu 18(4–5): 39–44. Islamabad: Muqtadarah Qaumi Zaban. Sindhi, Hyder 2000? Sindhi zaban o adab ki tarix [History of Sindhi language and literature]. Islamabad: Muqtadarah Qaumi Zaban. Sindhi, Memon Abdul Majeed 1993 Shikarpur ji Boli [The dialect of Shikarpur]. Hyderabad, Sindh: Sindhi Language Authority. Singh, Rajendra (ed.) 1998 The Yearbook of South Asian Languages and Linguistics, 1998. New Delhi: Sage. 2003 The Yearbook of South Asian Languages and Linguistics, 2003. Berlin: Mouton de Gruyter. Taj, Abdul Haq 1989 Shina Qaida. [Shina primer]. Gilgit: by author. Usman, Muhammad, Syed Farooq Ali, and Aamir Masood 2004 Syllabification of English words when spoken in Urdu. Axbar-e Urdu 20(12): 150– 154. Islamabad: Muqtadarah Qaumi Zaban. Wali, Aamir 2002 The rules governing the writing-pronunciation contrast in Urdu. Axbare Urdu 18(4–5): 256–262. Islamabad: Muqtadarah Qaumi Zaban.

Pakistan: Research and developments in linguistics

143

Wali, Aamir, and Sara Hussain 2003 Rhythm dictated resyllabification across word boundaries for Urdu poetry. Axbar-e Urdu 19(6): 168–183. Islamabad: Muqtadarah Qaumi Zaban. Wali, Aamir, Atif Gulzar, Ayesha Zia, Muhammad Ahmad Ghazali, Muhammad Irfan Rafiq, Muhammad Saqib Niaz, Sara Hussain, and Sheraz Bashir 2002a Contextual shape analysis of nastaliq. Axbar-e Urdu 18(4–5): 291–304. Islamabad: Muqtadarah Qaumi Zaban. 2002b Features for Noori Nastaliq. Axbar-e Urdu 18(4–5): 305–309. Islamabad: Muqtadarah Qaumi Zaban. Willson, Stephen R. 1998 Basic Burushaski Vocabulary. Islamabad: National Institute of Pakistan Studies, Quaid-i-Azam University and Summer Institute of Linguistics. Wyne, Nabeel 2002 Languages and their families. Axbar-e Urdu 18(4–5): 157–172. Islamabad: Muqtadarah Qaumi Zaban. Zia, Ayesha 2002 Assimilation and dissimilation rules in Urdu. Axbar-e Urdu 18(4–5): 237–248. Islamabad: Muqtadarah Qaumi Zaban.

Linguistic Studies of Kashmiri OMKAR N. KOUL

1. Introduction Linguistic studies of Kashmiri, comprising of grammars, grammatical studies, lexicography, phonology etc., began in the middle of the 19 th century. The grammatical literature includes a variety of materials written in the form of brief notes, articles, monographs, dissertations, and independent grammatical sketches and grammars. The lexicographical works include di¤erent types of vocabularies, glossaries and dictionaries. The linguistic studies available can be classified in the areas of genealogical classification and dialect surveys, grammars and grammatical studies, phonetics and phonology, lexicography, sociolinguistics and instructional materials. Here an attempt will be made to present a brief survey of certain significant materials prepared in and on this language. 2. Classification The genealogical classification of Kashmiri began with Grierson (1906), placing it in the Dardic group of Aryan languages. Morgenstierne (1961) classifies it among Indo-Aryan languages and is followed by all others in this regard. The classification is reviewed in Kachru (1969), and Koul and Schmidt (1984). There has been no other serious study of the subject since 1990. In more recent work, Afaq Aziz (1994) presents a comparative study of various languages of the Dardic group, with special reference to Kashmiri, Shina, and Kohistani. Koul (1994, 2000) repeats previous stands on the subject. Masica (1991) refers to linguistic characteristics of Kashmiri as compared to other Indo-Aryan languages. 3. Phonetics and Phonology Kashmiri has peculiar phonetic and phonological characteristics, such as the high central and mid vowels, dental a¤ricates, palatalisation, vowel

146

Omkar N. Koul

harmony rules, etc., which it does not share with other Indo-Aryan languages. The description of Kashmiri phonetics and phonology, or of issues related to some of its special characteristics is available in Grierson (1904, 1911, 1919), Bailey (1937), Firth (1939), Morgenstierne (1941), Sidheswar Verma (1964), Kelkar and Trisal (1964), Sar (1970, 1977), Handoo (1973), Zakharyin (1974), Koul (1977, 1985, 1987), Bhat (1987), Wali and Koul (1997) and Koul and Wali (2006). These works present briefly the principal phonological characteristics of Kashmiri. 4. Grammar There has been very significant research in the area of Kashmiri grammar. Kashmiri is Verb 2 language. This is the feature which it shares with German, Dutch and Icelandic. Grammatical works on Kashmiri began as early as mid-19 th century, with Edgworth (1841) and Leech (1944) followed by a complete grammatical description of the language in Ishvar Koul’s monumental work Kashmirshabdamritam, written in Sanskrit in 1879, edited by George A Grierson, and published by Royal Asiatic Society of Bengal in 1889. Grierson calls it ‘‘an excellent grammar of Kashmiri’’ and based quite a few of his works on it. Grierson published his Standard Manual of Kashmiri Language in 1911, and also provided a sketch of Kashmiri grammar in his Linguistic Survey of India (1919: Vol. 8, part 2). He also published papers by Burkhard (1887–1889) in his Essays on Kashmiri Language (1899). Some other grammatical descriptions by European scholars continued till the middle of the 20 th century. Though the tradition of presenting grammatical sketches and descriptions continued till midway through the 20 th century, serious works on the subject commenced from the early sixties following the models of grammars prepared in other Indian languages. Trisal’s doctoral dissertation (1964) is the first descriptive grammar of Kashmiri written in Hindi. Kachru provides the first detailed grammatical description of Kashmiri in his A Reference Grammar of Kashmir (1969). His other work, An Introduction to Spoken Kashmiri (1973), prepared for teaching and learning Kashmiri as a second/foreign language provides notes on Kashmiri grammar and culture. He has also dealt with certain grammatical aspects of the Kashmiri language in his other papers. Kachru’s work stimulated a great interest in the study of various aspects of Kashmiri grammar, including both morphology and syntax. Koul (1977) dealt with various syntactic aspects following new theoretical developments. A few doctoral dissertations have dealt with morphology and syntax in detail. Bhat (1980) provides a detailed description of phonology and morphology; Sar (1981) describes verbal morphology; Andrabi

Linguistic Studies of Kashmiri

147

(1984) discusses syntactic aspects of Reference and Co-reference in Kashmiri; Vijay Kaul (2006) deals with compound verbs in Kashmiri. Peter Hook and Koul jointly worked on various syntactic aspects like word order, pronominal su‰xes, ergativity, transitivity, causatives, modal verbs, etc. Koul and Hook (1984) present certain important grammatical aspects of Kashmiri as contributed by various scholars. The period after 1990 is very significant for the study of various grammatical aspects and for the preparation of grammars dealing in detail with morphology, syntax and semantics. Scholars in India and abroad and also in collaboration have prepared some significant research works, available in the form of dissertations, papers and books. Most of the dissertations deal with di¤erent syntactic aspects of Kashmiri. Asha Tickoo (1990) deals with word order in Kashmiri; Rakesh Bhatt (1994, published 2000) deals with word order and case in Kashmiri; Achla Raina (1993) deals with certain syntactic aspects of Kashmiri using an S-Selectional approach to grammar; Estella Del Bon (2001) deals with clitics in Kashmiri. Wali and Koul (1997), in their Kashmiri: A Descriptive-Cognitive Grammar, provide a detailed description of Kashmiri grammar covering morphology and syntax. This book, widely referred to, has stimulated a number of linguists to take up further research in Kashmiri. Topics in Kashmiri Linguistics, edited by Koul and Wali (2002) is a collection of research papers devoted to syntax contributed by Peter Hook, Ashok Koul, Omkar N. Koul, Achla M. Raina, Estella del Bon and Kashi Wali. Koul (2005) in his Studies in Kashmiri Linguistics provides a description of various linguistic and sociolinguistic aspects of Kashmiri. Kashmiri: A Study in Comparative Indo-Aryan by Hook and Koul (forthcoming) has 20 chapters devoted to various aspects of grammar. The Modern Kashmiri Grammar of Koul and Wali (2006) is pedagogically oriented for teaching/learning Kashmiri as a second language. Hook and Koul (2006) discuss valency sets in Kashmiri. There are very few grammars and grammatical studies written in Kashmiri. Naji Munawar and Shafi Shauq (1976), and Nishant Ansari (1976) provide a very brief description of traditional grammatical terms in Kashmiri. Their main contribution has been in introducing Kashmiri terms for traditional grammatical terms used in Urdu. Adil and Talashi (2002) present the first description of the grammatical aspects of Kashmiri, and Afaq Aziz’s (2005) grammar is the first pedagogically oriented grammar in Kashmiri. 5. Sociolinguistic research Very limited sociolinguistic work has been conducted in Kashmiri thus far. To begin with, Grierson (1911) and Kachru (1969) have listed certain

148

Omkar N. Koul

linguistic characteristics of the speech of Hindus and Muslims. Whereas Grierson uses Hindu and Muslim Kashmiri to distinguish these two varieties, Kachru prefers to use Sanskritised and Persianised Kashmiri for these varieties, respectively. The so-called varieties are not exclusively Hindu and Muslim, but are important from the point of view of registers and diglossia. M. K. Koul (1986) has studied sociolinguistic variables of Kashmiri spoken in the Anatnag district of the state and that of Srinagar. Dhar (1985) has pointed out sociolinguistic variations of Kashmiri spoken in Sopore. Kantroo (1985) has studied variations of Kashmiri by certain minority communities and occupational groups. The first ever sociolinguistic survey, conducted by Koul and Schmidt (1983), studies the language use and language preferences of native speakers of Kashmiri. It reveals the use of Kashmiri in social domains and preferences for its use in education and administration at lower levels. Mahfooza Jan (1993) has studied dialects spoken by certain professional groups. Koul (1994, 1995) in his two papers has analysed personal names, surnames and nicknames in Kashmiri. Koul (1998) has studied language maintenance and language loss of Kashmiri migrant children in Jammu and Delhi. The study reveals loss of Kashmiri in formal domains and its maintenance in certain restricted social domains. There is comparatively more loss of the language in Delhi than in Jammu. His survey on Language Preferences in Education in India (2001) shows preference for the use of Kashmiri as a subject and as medium of education at the elementary level in the state of Jammu and Kashmir. Kashmiri is primarily used in restricted social domains and about 70% of parents prefer to talk to their children in Kashmiri at home. There is a wide scope for serious sociolinguistic research in Kashmiri and for its planning in education, administration and mass media. 6. Lexicography Lexicographical works in Kashmiri fall under di¤erent categories: vocabularies, glossaries and dictionaries. It is believed that Sonti Pandit (1859) prepared a Kashmiri-Persian dictionary in 1859, which is not available now. Ishar Kaul (d. 1883) made a first serious attempt to prepare a KashmiriSanskrit dictionary but could not complete it before his death. Grierson (1916–1932) compiled A Dictionary of Kashmiri Language partly from materials left by Ishar Kaul. This is the first comprehensive Kashmiri-English dictionary available. The Jammu and Kashmir Academy of Art, Culture and Languages has prepared a monolingual dictionary entitled Kashir Dictionary (1972–79) in eight volumes, and a bilingual Urdu-Kashmiri Farhang (1967–80) in nine volumes. Rattan Lal Shant et al. have prepared a Hindi-

Linguistic Studies of Kashmiri

149

Kashmiri dictionary published by Central Hindi Directorate (CHD) in 1980. Several vocabularies have been prepared as a part of grammars and instructional materials. Handoo and Handoo (1975) have prepared a HindiKashmiri common vocabulary. A Kashmiri-English Glossary prepared by Koul et al. in 1976 was published as Kashmiri-English Dictionary for Second Language Learners in its Revised Version in 2000 by the Central Institute of Indian Languages (CIIL). Koul and Talashi have prepared a PunjabiKashmiri Dictionary (1999). Jawahar Tickoo (2006) has recently published a Kashmiri-English Dictionary. Knowles prepared A Dictionary of Kashmiri Proverbs and Sayings as early as in 1885. Koul (1992, 2006) has prepared A Dictionary of Kashmiri Proverbs, which provides Kashmiri proverbs with their literal translations, and idiomatic equivalents or explanations in English. 7. Instructional Materials Kachru (1973) prepared a course in spoken Kashmiri for the learning of Kashmiri as a second/foreign language. Koul (1987) has prepared Spoken Kashmiri: A Language Course as a self-instructional course. Teaching of Kashmiri as a second language to in-service teachers commenced at the Northern Regional Language Centre of the CIIL in 1971. CIIL has published quite a few instructional materials, which include a Kashmiri Phonetic Reader (1973), An Intensive Course in Kashmiri (1985), Kashir Kitab: Level I, Kashir Kitab: Level II, Kashmiri Pictorial Glossary, Intermediate Course Reader in Kashmiri (1995), Tests of Language Proficiency: Kashmiri by Omkar N. Koul, S. N. Raina, R. K. Bhat and M. K. Koul (2000), and A Handbook of Audio-Cassette Course in Kashmiri (with three cassettes, 2002). R. K. Bhat has also edited a Kashmiri Primer and a Kashmiri Reader using Devanagari script published by Sampreti (2003). The above survey brings out clearly that though linguistic research in Kashmiri began about hundred fifty years ago in di¤erent fields, there has been significant interest in the areas of grammars and grammatical studies, preparation of dictionaries and other pedagogical materials in Kashmiri after 1990. There are still a few important areas in which no adequate work has been done so far. This includes work in the area of computational linguistics and application of information technology. References Aziz, Afaq 2005 vitastaa kaashur graamar [Vitasta Kashmiri Grammar]. Srinagar: University of Kashmir.

150

Omkar N. Koul

Bhat, Rookkrishen 1987 A Descriptive Study of Kashmiri. Delhi: Amar Prakashan. Bhatt, Rakesh M 1999 Verb Movement and the Syntax of Kashmiri. Dordrecht: Kluwar Academic Press. Grierson, G. A. 1911 A Standard Manual of the Kashmiri Language 2 Vols. Oxford: Reprinted Rohtak: Light and Life Publishers, 1973. 1919 The Linguistic Survey of India Vol. VIII, Part II. Calcutta: Royal Asiatic Society. Reprint. Delhi: Motilal Banarasidas, 1968. 1932 A Dictionary of the Kashmiri Language. New Delhi: B.R. Publishing Corporation. Reprint 1985. Hook, Peter Edwin, and Omkar N Koul 2006 Valency Sets in Kashmiri. In Voice and Grammatical Relations. In Honor of Masayoshi Shibatani, Tasaku Tsunda and Taro Kageyama (eds.), 43–84. Amsterdam: Benjamins. forth- Kashmiri: A Study in Comparative Indo-Aryan Languages. Tokyo: Institute of Asian coming and Aftrican Languages and Cultures. Kachru, Braj B. 1969a A Reference Grammar of Kashmiri. Urbana: University of Illinois. 1969b Kashmiri and other Dardic languages. In Current Trends in Linguistics Vol. V, Thomas A. Sebeok (ed.), 284–306. The Hague: Mouton. 1973 An Introduction to Spoken Kashmiri. Urbana: University of Illinois. Kak, Adil, and Ratan Talashi 2002 kaashir zabaan: akh graamrii vyetshnay [The Kashmiri Language: A Grammatical Analysis]. Srinagar: Nihaar Publications. Kaul, Vijay Kumar 2006 Compound Verbs in Krshmiei. Delhi: IILS. Koul, Maharaj K. 1986 A Sociolinguistic Study of Kashmiri. Patiala: Indian Institute of Language Studies. Koul, Omkar N. 1977 Linguistic Studies in Kashmiri. New Delhi: Bahri Publications. 1985 An Intensive Course in Kashmiri. Mysore: Central Institute of Indian Languages. 1987 Spoken Kashmiri: A Language Course. Patiala: Indian Institute of Language Studies. 1995 An Intermediate course in Kashmiri. Mysore: Central Institute of Indian Languages. 2000 Kashmiri Language, Linguistics and Culture: An Annotated Bibliography. Mysore: Central Institute of Indian Languages. 2001 Language Preference in Education in India. In Daswani, C. J. (ed.) Language Education in multilingual India, New Delhi: UNESCO. 2003 Kashmiri. In The Indo-Aryan Languages, George Cardona and Dhanesh Jain (eds.), 895–952. London: Routledge. 2005 Studies in Kashmiri Linguistics. Delhi: Indian Institute of Language Studies. Koul, Omar N., and Peter Edwin Hook (eds.) 1984 Aspects of Kashmiri Linguistics. New Delhi: Bahri Publications. Koul, Omar N., and Ruth Laila Schmidt 1983 Kashmiri: A Sociolinguistic Survey. Patiala: Indian Institute of Language Studies. 1984 Dardistan revisited: An examination of relationship between Kashmiri and Shina. In Aspects of Kashmiri Linguistics, Omkar N. Koul and Peter Edwin Hook (eds.), 1–26. New Delhi: Bahri Publications. Koul, Omar N., and Kashi Wali (eds.) 2002 Topics in Kashmiri Linguistics. New Delhi: Creative.

Linguistic Studies of Kashmiri

151

Koul, Omar N., and Kashi Wali 2006 Modern Kashmiri Grammar. Springfield: Dunwoody Press. Morgenstierne, George 1961 Dardic and Kafir Languages. In The Encyclopedia of Islam, Vol. 2, Fasc. 25. Leiden: E.J. Brill. Strand, R.F. 1973 Notes on the Nuristani and Dardic Languages. Journal of the American Oriental Society, 93.3: 297–305. Wali, Kashi, and Omkar N. Koul 1997 Kashmiri: A Cognitive-Descriptive Grammar. London and New York: Routledge.

Linguistic activites in Nepal (1999–2004) YOGENDRA P. YADAVA

1. Introduction The period (1999–2004) witnessed a number of activities in Nepalese linguistics. These activities mainly include the ongoing linguistic projects on documenting endangered languages such as Puma, Chhintang, Hayu, Chhantyal and Bhujel, the compilation of basic trilingual (mother tongue– Nepali–English) dictionaries, corpus linguistics, language enumeration in the national census, and transitional bilingual education. This report is an attempt to summarize the principal of these activities (for details, see the works cited). 2. Linguistic studies With the opening of the Department of Linguistics at Tribhuvan University in 1996 there have been concerted e¤orts on pursuing linguistics studies in Nepal. These e¤orts have been further accentuated in the last five years or so with the strengthening of the Department, which has been now functioning as the hub of linguistic studies and expertise. Since 1999 it has undertaken a number of projects in collaboration with national and foreign agencies. The principal projects include the following. 2.1. Puma Chhintang Documentation Project The documentation project of two endangered languages of Nepal– Chhintang and Puma—sponsored by the Volkswagen Foundation, Germany, under the DOBES program and carried out by the Institute of Linguistics, University of Leipzig in conjunction with the Central Department of Linguistics, Tribhuvan University, constitutes one of the two initial activities within a long-term project, entitled ‘‘Linguistic Survey of Nepal’’ (LINSUN) project, the other being the ‘‘Encyclopedia of Nepal’s Languages’’ project. The LINSUN project, with its national and international

154

Yogendra P. Yadava

committees, was conceived by the Central Department of Linguistics to document Nepal’s languages, with priority given to endangered ones. This project aims to provide linguistic and ethnographic documentation of the two endangered languages—Puma and Chhintang, which belong to the Kiranti group of Tibeto-Burman languages. The Chhintang and Puma documentation project is a positive step in reinforcing the constitutional provision for preservation of endangered languages in Nepal at a time when there have been pressures felt within Nepal as well as concerns expressed by the international community about the loss of linguistic diversity and its appeal for the documentation of this diversity before it is completely lost. This project has been considered as a chain in German support for linguistic research in Nepal. 2.2. Basic Dictionary Project The basic dictionaries of five languages—Baram, Lohorung, Mewahang, Sherpa and Gurung—supported by a grant from the National Foundation for Development of Indigenous Nationalities (NFDIN) for documenting indigenous languages, have been prepared at the Department of Linguistics, Tribhuvan University. These are trilingual (vernacular–Nepali–English) dictionaries. For this purpose, a basic lexical database (including simple words, compounds, idioms, and other phrasal units) was collected from native speakers using a finderlist and texts in these languages. Each of these basic trilingual dictionaries consists of three parts: I. Vernacular-Nepali-English Dictionary; II. Nepali-Vernacular Index; and III. English-Vernacular Index. Within each entry in part I, the headword appears in Devanagari script (except for Sherpa, which has its own Sherpa script), followed by its pronunciation (using IPA symbols), parts of speech, Nepali and English glosses and example sentences (if necessary). These dictionaries support e¤orts to document the five languages, their use in literacy and mother tongue education and the broad needs of the speech community. Such trilingual dictionaries may also serve a mixed audience with di¤erent languages and purposes. Finally, they may help the preservation of a culture encoded in a language. This is an ongoing project for compiling basic dictionaries for Nepal’s indigenous languages. 2.3. Basic Documentation Project With financial support from NFDIN, the Centre for Nepalese and Asian Studies, in conjunction with the Department of Linguistics, completed documentation of three endangered languages: Hayu, Chhantyal

Linguistic activites in Nepal (1999–2004)

155

and Bhujel. This documentation includes information on sociolinguistic setting, grammar, basic vocabulary and some interlinearized texts of each language. In addition, the Department of Linguistics, again in collaboration with NFDIN, carried out the documentation of Kusunda language, which is significant for linguists not only because it is extremely endangered but also because it represents a ‘‘language isolate’’. The population of Kusunda people has been reduced drastically to just 164 (2001 census). With this marginalized strength, the Kusundas are left with no option but to marry non-Kusundas, especially Magars or even Khas (Nepali)-speaking Khetriyas. Their assimilation will eventually lead them to lose their unique ethnic identity, including language. If there exist no Kusundas per se their mother tongue, namely, Kusunda language, will naturally cease to survive and be lost to posterity The Kusunda language thus appears to be an extreme case. Also called Begai, this is a language isolate, without any genetic relation with the other languages spoken in Nepal. Reported earlier to be dead, it has been recently discovered to have a marginal number of speakers. Of the total 164 Kusundas, 86 are reported to speak their mother tongue Kusunda. Left with a handful of fluent but elderly speakers, this language is dying out or moribund. In addition to having few speakers, there is little or no transmission of the Kusunda language to younger generations. This is due mainly to the use of Nepali language in intra- and interhousehold discourse and also to Kusunda children’s negative attitude to their ancestral language. Other factors that threaten the safety of the Kusunda language include the lack of a literate tradition, disuse in domains such as education and mass media, the migration of its speakers from jungle to urban areas and their poor economic status, and the lack of o‰cial recognition. Documentation will no doubt strengthen the language. It is, however, only a part of the story. To preserve and promote the Kusunda language, there is a need for concerted e¤ort by the Kusunda community, government and non-government organizations as well as linguists. Besides documentation, we must encourage awareness among its speakers to preserve their language, introduce a writing system in the language, and extend its domains of use such as education and mass media. These measures for revitalizing the language must be carried out e¤ectively to avert its loss. Otherwise, this language isolate, which is quite distinct from the languages of the four language families (Indo-European, Sino-Tibetan, Austro-Asiatic and Dravidian) spoken in Nepal, and whose genetic a‰liation is yet to be identified, may not survive this threat of extinction. And such a language loss would be a great cultural and intellectual disaster for all of humanity.

156

Yogendra P. Yadava

2.4. NeLRaLEC (Nepali Language Resources and Localization for Education and Communication) Project This project was undertaken under an EU grant to further integrate Nepal and neighbouring regions into the Information Society. The project will: – ‘‘establish a core of human language technologies for Nepali at a level taken for granted in developed countries for the national languages: – develop Nepali corpora, for both text and speech, making these available on the web; – deliver a new Nepali dictionary for public free use, developing standards for spelling and orthography; – develop human language processing software for Nepali. – establish a base of routine o‰ce and home software working in Nepali: – deliver industrial strength Nepali software for desk top publishing; – deliver web access and communications software for Nepali; – deliver speech generation software for Nepali. – ensure that these technologies are adopted and used by critical segments of Nepal’s organizations: – give training in the use of these tools to schools and ministries; – evaluate and demonstrate within education the Nepali language and software technologies; – evaluate the use of the speech generation by non-literate people. – ensure that this Nepal Information Society is sustained and sustainable: – give training in the continued development of human language technologies (HLT) resources; – enable future research in Nepal for HLT for Nepali; – establish university courses on HLT for Nepali; – develop resources (a localization kit) for localizing existing software to use Nepali; – make all language resources and tools available as open source.’’ (Grant Application Form, 2004: 2) This is a collaborative project to be carried out by the Open University (UK), Madan Puraskar Pustakalaya (Nepal), Lancaster University (UK), the University of Go¨teborg (Sweden), the European Language Resources Association (ELRA, France) and Tribhuvan University (Nepal). 2.5. Census The Tribhuvan University Department of Linguistics and its faculty have also been providing consultancy service to government and nongovernment agencies in the field of linguistic issues. For the first time the

Linguistic activites in Nepal (1999–2004)

157

Central Bureau of Statistics (CBS) sought its expertise in sorting out the issue of identifying mother tongues spoken in Nepal. According to the 2001 census, there are 92 languages (compared to 44 (Central Bureau of Statistics 1952/54); 36 (1961); 17 (1971); 18 (1981); and 32 (1991)) spoken in Nepal. However, there still remain quite a few languages (used by as many as 168,340 strong speakers, i.e. 0.74 per cent of the total population) which are lumped together into an ‘‘unknown’’ slot for lack of adequate information. This significant increase is mainly due to the fact that a large number of languages used as mother tongues have been returned for the first time because of the growing awareness of several ethnic minorities about their distinct cultural and linguistic identity. The languages (except Kusunda) enumerated in the 2001 census belong to four language families: Indo-European, Sino-Tibetan, Austro-Asiatic, and Dravidian. Kusunda, a dying language, is a language isolate and its genetic relation is yet to be identified. Most of Nepal’s languages, esp. indigenous ones, have no literate traditions. Recently there has been growing awareness among various ethnic groups to introduce literacy and basic education as well as literature in their native languages. This requires as a prerequisite the development of writing systems. To develop writing systems suited to the sound systems existing in di¤erent non-literate languages, the Department of Linguistics and its faculty members such as Professor Madhav Pokharel and Professor Chuda Mani Bandhu have been providing expertise in this enterprise. 2.6. Nepal’s Indigenous Languages: a Situation Analysis Nepal is a multi-ethnic nation comprising 59 indigenous nationalities. These nationalities have distinct identities characterized by their cultural as well as linguistic diversity. There are approximately 78 distinct indigenous languages of four genetic stocks (with the exception of Kusunda, which seems to be a language isolate): Sino-Tibetan, Austro-Asiatic, Indo-Aryan and Dravidian associated with these various indigenous nationalities. However, their precise number is certainly higher, though this needs to be ascertained through further investigation. The National Foundation for Development of Indigenous Nationalities (NFDIN) has been established under an act of the parliament for ‘‘the development of indigenous nationalities in Nepal.’’ In keeping with the spirit of the 1990 Constitution of the Kingdom of Nepal, one of the principal objectives of the Foundation is to preserve and promote the languages of indigenous nationalities. In consonance with this objective, it is important to investigate and analyze their situation as a feedback to formulate and implement any plan of action for the preservation and promotion of the

158

Yogendra P. Yadava

indigenous languages. It is with this perspective that the study by Yadava and Bhadra (2004) has been undrtaken. This study aims to first analyze the situation of the languages used by various indigenous nationalities as mother tongues in Nepal with a view to recommending measures for their preservation and promotion. To accomplish this task, it tries to: (i) (ii) (iii) (iv) (v) (vi)

identify the languages used as mother tongues by indigenous nationalities in Nepal; show their genetic a‰liation; present their distribution; ascertain their literate traditions; discuss the relation between languages and their speakers’ ethnicity; and identify endangered languages.

To accomplish these aspects of the study, information has been gleaned from various published/unpublished works, esp. by CBS, NFDIN and other relevant sources. Questionnaires and interviews have been used for eliciting information on language endangerment. We hope this study of Nepal’s languages can be beneficial for a number of reasons. First, there have been varying figures about these languages. The proposed study seeks to clarify this number though it is an approximation gathered from available sources. Secondly, it determines the literate traditions of the indigenous languages so as to develop appropriate strategies for using them in developing their writing systems, preparing textbooks and using them in primary education and literacy programs, undertaking their documentation, and promoting their use in conformity with the broad needs of the speech communities. And, it can serve as a baseline study for evaluating further policies and action plans for the preservation and promotion of indigenous languages. Above all, this study can help the preservation of a culture encoded in a language. (Yadava and Bhadra 2004) 2.7. Nepali Language in Information Technology The Nepali Language in Information Technology (NLIT) Steering Committee has been formed under the aegis of the High Level Commission for Information Technology (HLCIT) with a view to localizing information technology in Nepali language through the development of character sets, keyboards, glyphs, fonts, sorting orders, glossaries, standardization, etc. The localization of IT in Nepali has been undertaken with linguists’ academic support both in Microsoft and Linux at Unlimited New Media and Madan Puraskar Pustakalay, respectively.

Linguistic activites in Nepal (1999–2004)

159

2.8. Transitional Bilingual Education Under the Education for All (EFA) program (2004–2009) the government has introduced transitional bilingual education in order to make education easily accessible to linguistic minorities through the introduction of mother tongues as medium of instruction at grades 1–3, and then the gradual shift to Nepali, the national language (Curriculum Development Centre). This program will involve the use of linguistic expertise in preparing mother tongue reading materials and training expertise. 2.9. UNESCO Language Survey Report, Nepal (2000) Keeping in view its rich linguistic and cultural heritage, Toba, Toba, and Novel (2000) carried out the language survey of Nepal using the UNESCO questionnaire. This survey report contains information on the existing situation of sixty languages spoken or at least understood in Nepal. It is expected to help experts in the fields of linguistics, education and administration in formulating appropriate linguistic, educational and administrative policies. 2.10.

Themes in Himalayan Languages and Linguistics

Themes in Himalayan Languages and Linguistics (Kansakar and Turin, eds. 2003) has been an important publication in this period. This volume is a compilation of 24 articles presented at the 5 th Himalayan Symposium held in Kathmandu on 13–15 September, 1999. These articles deal with languages such as Baram, Bote, Chinese, Dhimal, Hmar, Khasi, Kyirong Tibetan, Lhomi, Manipuri, Meiteiron, Nepali, Newar, Rabha, Sherpa, Sunuwar, Tamang, Tangbe, Thangmi, spoken and written Tibetan, and so on. The volume also represents several linguistic subfields—grammatical descriptions of individual languages, sociolinguistic and anthropological studies, and general linguistic theory. 3. Grammatical studies During this period a number of articles by Nepalese linguists have appeared in books and journals. These articles encompass the various fields of linguistics. Bandhu (1999) focuses on the activities for the introduction of terminological neologisms in Nepali and suggests three contexts—international, regional and national—to take into consideration.

160

Yogendra P. Yadava

S. S. Dahal (1999) shows with ample examples how kinship terms in Thami, a Tibeto-Burman language, converge with those in Indo-Aryan Nepali. In his article, B. M. Dahal (1999) attempts to analyze the structure of Nepali baby talk which, he concludes, conforms to the common characteristic features of baby talk in the di¤erent languages studied so far. In an article Kansakar (1999) looks at language endangerment in Nepal and argues for examining the multilingual and multiethnic situation in the country in the context of current demographic processes. Pokharel (1999) illustrates with ample data how almost all the TibetoBurman languages spoken in the catchment area of Kosi river are characterized by a su‰x in the verb to express reciprocity. Lohani (1999) discusses permissive constructions in Nepali within the Lexical Functional Grammar (LFG) framework. Bhadra (2004) examines the turn-taking mechanism in Nepali conversations. Apart from illustrating di¤erent types of turn construction units, this article cites examples from Nepali conversations to show how turns are allocated and how turn transition is accomplished. Also illustrated with examples from Nepali are di¤erent types of silences (pauses and gaps) and the minimized occurrences of overlaps. The article also discusses when overlap occurs, how it is triggered, how it relates to transition-relevance places (TRPs) and what techniques speakers employ to minimize the negative e¤ects of overlap in Nepali conversations. Finally, the article examines the problems regarding what are or are not ‘‘interruptions’’ and describes the di¤erent types of overlap with examples from naturally occurring Nepali conversations. Kansakar (2004) explores the syntactic and semantic properties of nonnominative subjects in Newar. In a joint article Fortier and Rastogi (2004) provide phonological comparisons to further show the relatedness between Khaamchi and Rawti as well as shed light on their historical relationship and a‰liations with both the Proto Tibeto-Burman and other areal languages. Recently, there have been completed doctoral dissertations on two specific Tibeto-Burman languages: Tamang (Poudel 2003) and Athpare (Neupane 2002). Both of them present descriptive grammars of the respective languages. It has been shown how in Maithili, as in several other languages, there exists no one-to-one relation between the grammatical relations of nominals and their case (Yadava 2004). A nominative nominal, for example, can be the subject of a clause, but it can also function as the object of a verb in another construction. On the other hand, though the subject is typically coded with nominative case it can also be realized with other case markings.

Linguistic activites in Nepal (1999–2004)

161

In other words, Maithili clauses also involve non-nominative subjects, apart from the nominative one. This article is an attempt to analyze the non-nominative subjects in Maithili in terms of their morphosemantic and syntactic properties such as anaphoric coreference, pronominal noncoreference, gapping in coordinate structures, verb agreement, and case demotion. Most South Asian languages (including Indo-Aryan, Tibeto-Burman, Dravidian and Austric languages) typically employ non-finite clauses instead of finite clauses to realize clause linkage (cf. Masica 1976). In Yadava (2004) one such construction, the sequential converbal construction, has been analyzed and compared in the three N(ew)I(ndo)-A(ryan) languages: Nepali, Hindi and Maithili from typological perspectives. Other studies include Tamang verbs (Chalise 1999, 2003), complex predicates in Bote (Prasain 2000), Kumal verbs (Gautam 2000, 2004), causativization and anticausativization in Magar (B. N. Regmi 2000), Nepali clause structure (D. R. Regmi 2000), compound verbs in Bote (Poudel 2000), clause combining in Tamang (Chalise 2000), tense and aspect in Danuwar (Bhandari 2003, 2004), Magar-Nepali creole (Ghimire 2004), child language acquisition in Nepali (Pathak 2004), the case system in Dhankute Tamang (Poudel 2004), exclusivity and inclusivity in Chamling Rai (Rai 2004), relativization in Nepali (D. J. Regmi 2004), Thakali noun phrase (Sharma 2004), the state of the art of Maithili linguistics (Yadava 2000), Persian influence on Nepali vocabulary (Raj 2002), case forms in Chhathare verbs (Tumbahang 2004), light verbs in Bote (Prasain 2003), nonnominative subjects in Nepali (Lohani 2003), deontic modality in Thakali (Sharma 2003), personal pronouns in Koyu (Regmi 2003), and influence of Hindi on Nepali (Pathak 2003).

4. Other linguistic activities 4.1. The SCALLA 2004 Working Conference The SCALLA (Sharing Capability in Localization and Human Language Technologies) 2004 Working Conference, held from January 5–7, 2004 with support from the European Union, was organized in Kathmandu to bring the computational linguistics and software localization communities in Europe and South Asia closer together and learn from each other through participation in conferences and workshops. 32 papers were presented by European and South Asian (including Nepalese) computational experts and linguists. The conference was attended by 42 participants from Europe and South Asia. The conference was successful in accomplishing its objective of creating an awareness about crossing the digital divide through

162

Yogendra P. Yadava

sharing software technologies to meet human linguistic needs in South Asia (SCALLA 2004). 4.2. The LSN Silver Jubilee Conference The Linguistic Society of Nepal (LSN) organized its silver jubilee conference attended by more than 100 Nepalese and foreign linguists. 45 papers were presented on diverse linguistic topics including localization of Nepali language and the pattern of language endangerment in the country. The conference resolved to bring out the special volume of its silver jubilee proceedings. The LSN has been holding annual conferences for a quarter of a century without interruption, as well as pursuing occasional seminars and workshops and bringing out its journal Nepalese Linguistics. The Society, the oldest academic organization with maximum involvement of Tribhuvan University faculty members and students, has been providing a vibrant forum for national and foreign linguists to make presentations on various vital issues related to Nepal’s languages in particular and linguistics in general. References Bhandari, Bhabendra 2003 Tense system in Danuwar. Nepalese Linguistics 20: 114–119. 2004 Aspects in Danuwar. Nepalese Linguistics 21: 28–41. Bandhu, Chura Mani 1999 Terminological neologisms in Nepali. Nepalese Linguistics 16: 1–8. Bhadra, Sarada 2004 Turn-taking in Nepali. Nepalese Linguistics 21: 1–21. Central Bureau of Statistics 1952/54 Population census. Kathmandu: National Planning Commission. 1961 Population census. Kathmandu: National Planning Commission. 1971 Population census. Kathmandu: National Planning Commission. 1981 Population census. Kathmandu: National Planning Commission. 1991 Population census. Kathmandu: National Planning Commission. 2001 Population census. Kathmandu: National Planning Commission. Chalise, Krishna Prasad 1999 Classification of the Tamang verb stems. Nepalese Linguistics 16: 74–86. 2000 Clause combining in Tamang. Nepalese Linguistics 17: 77–90. 2003 Causative construction in Tamang. Nepalese Linguistics 20: 110–113. Curriculum Development Centre 2004 Education for All (EFA) programme (2004–2009). Kathmandu: Ministry of Education, HMG/Nepal. Dahal, Ballabh Mani 1999 Nepali baby talk. Nepalese Linguistics 16: 13–18. Dahal, Subhadra Subba 1999 Thami kinship terms. Nepalese Linguistics 16: 9–12.

Linguistic activites in Nepal (1999–2004)

163

Fortier, Jana, and Kavita Rastogi 2004 Sister languages? Comparative phonology of two Himalayan languages. Nepalese Linguistics 21: 42–51. Gautam, Bhim Lal 2000 Verbal a‰xation in Kumal. Nepalese Linguistics 17: 28–35. Ghimire, Laxman 2004 Magar-Nepali creole? Nepalese Linguistics 21: 58–64. Kansakar, Tej Ratan 1999 Multilingualism and the language situation in Nepal. Nepalese Linguistics 16: 19–32. 2004 Non-nominative subjects and the notion of subjecthood in Newar. Nepalese Linguistics 21: 65–86. Kansakar, Tej Ratan, and Mark Turin, eds. 2003 Themes in Himalayan Languages and Linguistics. Heidelberg and Kathmandu: South Asia Institute and Tribhuvan University. Lohani, Ram Raj 1999 Permissive constructions in Nepali: Evidence from complex predicates. Nepalese Linguistics 16: 62–73. 2003 Nonnominative subjects in Nepali. Nepalese Linguistics 20: 54–64. Masica, Colin P. 1976 Defining a Linguistic Area: South Asia. Chicago: Chicago University Press. Neupane, Tank 2002 athpareko yaakaran [A grammar of Athpare], Doctoral diss., Tribhuvan University. Pathak, Lekhnath S. 2003 Influence of Hindi on Nepali. Nepalese Linguistics 20: 131–137. 2004 Child language acquisition: A case study. Nepalese Linguistics 21: 87–93. Pokharel, Madhav P. 1999 Reciprocity in Kiranti. Nepalese Linguistics 16: 33–40. Poudel, Kamal 2000 Compound verb in Bote. Nepalese Linguistics 17: 56–61. Poudel, Kedar Prasad 2003 A descriptive study of Tamang (as spoken in Dhankuta district). Ph. D. diss., Tribhuvan University. 2004 Case system in Dhankute Tamang. Nepalese Linguistics 21: 94–116. Prasain, Balaram 2000 Complex predicates in Bote: An overview. Nepalese Linguistics 17: 6–27. 2003 Some light verbs in Bote. Nepalese Linguistics 20: 37–47. Rai, Bagdevi 2004 Exclusivity and inclusivity in Chamling Rai. Nepalese Linguistics 21: 117–124. Regmi, Bhim Narayan 2000 Morphological causativization and anticausativization in Magar. Nepalese Linguistics 17: 36–50. 2003 Personal pronouns in Kou. Nepalese Linguistics 20: 83–87. Regmi, Dan Raj 2000 An analysis of Nepali basic clauses within TG approach. Nepalese Linguistics 17: 56–61. 2004 The Morphosyntax of relativization in Nepali: A typological perspective. Nepalese Linguistics 21: 125–134. Regmi, Dan Raj, and Yogendra P. Yadava 2002 The ergative morphosyntax in Nepali. Nepalese Linguistics, 19.

164

Yogendra P. Yadava

SCALLA 2004 SCALLA (Sharing Capability in Localization and Human Language Technologies). ms. Sharma, Narayan P. 2003 Deontic modality in Thakali. Nepalese Linguistics 20: 71–77. 2004 Head parameter system in NP of Thakali. Nepalese Linguistics 21: 135–141. Toba, Suyoshi, Ingrid Toba, and Novel K. Rai 2000 UNESCO Language Survey Report. Kathmandu: UNESCO/Nepal. Tumbahang, Govinda B. 2004 Place and order of case markers in the verb forms of Chhathare Limbu. Nepalese Linguistics 21: 154–167. Yadava, Yogendra P. 2000 A fresh look at grammatical relations in Indo-Aryan. Lingua (with B. Bickel). 2004a Non-nominative subjects in Maithili. In Non-nominative Subjects, Vol. 2, Bhaskararao and Subbarao (eds.): 253–264. Amsterdam/Philadephia: Benjamins. 2004b Language, Population Monograph Volume 1. Kathmandu: Central Bureau of Statistics/United Nations Population Fund (UNFPA). Yadava, Yogendra P., and Sarda Bhadra 2004 Languages of Indigenous Nationalities. Kathmandu: National Foundation for Development of Indigenous Nationalities NFDIN.

Reviews

Rajeshwari V. Pandharipande. 2003. Sociolinguistic Dimensions of Marathi: Multilingualism in Central India (Lincom Studies in Sociolinguistics 03). Mu¨nchen: Lincom Europa. 171 p., ISBN 3 89586 713 6, EUR 66,00, USD 79,20. —Reviewed by Francesco Gardani, Universita¨t Wien The book is divided into three parts which are organised around wider thematic centres. Overall, the book is well organised. The introduction briefly sums up the subject of each chapter and each chapter is adorned with a resume´ or conclusive discussion. The book opens with a foreword which illustrates the aims pursued by the study. An overall presentation of the goals of the work follows in the introductory Chapter One. Here Pandharipande puts forward the general theoretical framework referring to most important works on sociolinguistics and postulates the axiom that ‘‘language structure is not monolithic but is invariably full of variation’’ (7). The focus of the research is pointed out in (10): the investigated area is the Marathi region in Central India (state of Maharashtra). (By the way, the author claims that the total population of Maharashtra amounts to more than 50,000,000 and would thus constitute one tenth of India’s population (11). The consultation of an encyclopaedia tells us that India has one billion inhabitants!). The conceptual focus is the ‘‘various dimensions of the form and function of Marathi in the multilingual setting of Maharashtra in general and of Central India in particular’’, i.e. the impact of language pluralism (interaction with Persian, Sanskrit, English) on language change resulting in language variation and linguistic convergence. Further research questions (12) are the way language pluralism may a¤ect the structure of languages, the way languages in contact readjust their structures in the process of accommodation of new borrowed structures or material, and what are the style repertoires and style range emerging from these changes. My first general remark about the book is that from the first chapter onwards it becomes evident that the focus of the investigation is too vast and

166

Reviews

that one is thus tempted not to designate it as ‘‘focus’’. In the introduction a brief survey of the most significant approaches to Indian sociolinguistics is presented: the range of the recorded investigations is wide and encompasses philological and structural descriptions, the concept of sprachbund, theoretical issues of language contact and language change as well as the cover terms of linguistic variation, bilingualism and multilingualism. The book under review is about all of these issues. This implies that many aspects cannot be analysed with the required depth and that some relevant theoretical frameworks (e.g. Thomason, Sarah G., and Terrence Kaufman. 1988. Language Contact, Creolization, and Genetic Linguistics. Berkeley: University of California Press) are ruled out by the shortness of the volume. In other words, the book is far too short to accommodate all the issues the author would like to. A blatant exemplification of this hindrance is found when the question arises: ‘‘What are the social conditions conducive to the transfer of linguistic features from one language to another?’’ (9, [emphasis mine]). A more detailed study of works on language contact and structural borrowing reveals that the term ‘‘conducive’’ as used by Pandharipande is deceptive. There is no compulsion for linguistic features to be transferred, even when the societal conditions would push in this direction. Rather, there are factors of both extra-linguistic and intra-linguistic nature which may favour the phenomenon of transfer (Thomason and Kaufman 1988; Gardani, Francesco. 2002. Borrowing of inflectional morphemes in language contact. M. A. Thesis, Universita¨t Wien [submitted to press]). Part One revolves around the issue of language pluralism in the MarathiHindi contact setting and dedicates three chapters (Chapter One is the introduction) to, respectively, the formal linguistic outcomes of linguistic convergence (Chapter Two), the sociolinguistic function of that convergence (Chapter Three), and possible motivations and restrictions on language change in the Nagpur area (Chapter Four). Chapter Two is concerned with the linguistic convergence of spoken Marathi and Hindi in the Nagpur area. The borrowing occurrences involving other languages such as English, Persian, and Sanskrit are not taken into account. The chapter is subdivided into two main sections which deal, respectively, with borrowings from Hindi into Marathi, and from Marathi into English. For each language two varieties are specified, namely Pune Marathi (Standard Marathi) and Nagpuri Marathi, on the one hand, and Standard Hindi and Nagpuri Hindi, on the other hand. The data presented in Chapter Two is interesting. It encompasses borrowing of lexical, semantic, phonological, syntactic, and morphological nature. On a negative note, the data is sometime collocated under the wrong heading: The borrowing of the conditional construction (cf. 2.2.3.1) is no syntactic borrowing but rather a semantic loan, which is

Reviews

167

used to express a syntactic construction. The borrowing of the pronoun apan is not of a syntactic nature but an occurrence of morphological borrowing. Chapter Two is the most interesting part of the book because it shows how deeply a language contact setting may exert influence on the structures of the languages involved. Nevertheless, some shortcomings of a conceptual and formal nature arise. First, in most cases the set of the data cannot be traced back to bibliographical references. Nor does it stem from the personal research of the author. Not indicating where the data used stems from is a severe shortcoming in scientific writing. Second, the author does not provide an overview of the overall sociolinguistic situation of the contact setting. This should comprise, whenever available, information about the multilingual society, the number of speakers, the degree of the societal multilingual competence, intermarriage, status as well as prestige of the languages involved. This lack is made up for in Chapter Three. The design of one’s own work pertains, of course, to the free choice of the author. However, it would have been more sensible to present the sociolinguistic setting first and then the concrete cases of borrowing. Third, information about the discreteness of the linguistic systems is missing in Chapter Two as well as in Chapter Three. The notion of discreteness of linguistic systems involved in a language contact context and its tolerability with the concept of borrowing is of great relevance. For discreteness to be postulated, linguistic differences must be recognisable on all levels of the language system. Clearly, borrowing is easier to accomplish between non discrete systems (e.g. dialects) because here the di¤erences are far smaller than between languages. Dialect interference may occur ‘‘both between stable, strongly di¤erentiated dialects and between weakly di¤erentiated dialects through the di¤erential spread (in waves) of particular changes’’ (Thomason and Kaufman 1988: 9). Nonetheless, borrowing is individuated in the most straightforward and the least misleading way when it occurs between discrete systems. Fourth, the genetic a‰liation of the languages involved in the contact situation is not discussed here but, once again, taken into account in Chapter Three only, where the author states simply that ‘‘Hindi and Marathi are genetically related (i.e., Indo-Aryan) languages’’ (42). This is a matter of typological distance. Many linguists, especially Meillet (Meillet, Antoine. 1921. Linguistique historique et linguistique ge´ne´rale. Paris: Champion), believed typological closeness to be the indispensable requirement for grammatical transfer. In contrast, more recent studies (Thomason and Kaufman 1988; Gardani 2002) hold the position that the absence of typological closeness is by no means an obstacle to grammatical transfer. Nevertheless, it holds true that typological congruence favours the introduction of structural elements into the receiving language.

168

Reviews

The last part of Chapter Two is concerned with an explanation of linguistic change in terms of ‘‘grammatical complexity vs. functional transparency’’: in three and a half pages the author touches on most relevant issues in linguistic theory. The first point to be elucidated is that of the types of borrowing that a language may accomplish. Pandharipande claims that ‘‘[m]ost of the borrowings have added new linguistic units [ . . . ] to the borrowing language’’ (36). The really interesting question here would be not whether or not borrowing is of an additive nature, but mostly whether new oppositions are brought into the receiving language. This would beg more ¨ berlegungen zu einer Klassifikation discussion. Breu (Breu, Walter. 1996. U des grammatischen Wandels im Sprachkontakt (am Beispiel slavischer Kontaktfa¨lle). In Sprachtypologie und Universalienforschung (STUF) 49: 21–38) has suggested a very accurate scheme, which accounts for six di¤erent types of e¤ect on the structure of the recipient language. For Breu, contact-induced language change involves the increase and decrease of oppositions in the linguistic system. Moreover, the question whether there is a hierarchical ‘‘borrowing scale’’ (38) has already been addressed by Thomason and Kaufman (1988: 74–76). At times Pandharipande’s terminology is problematic: the entrance through transfer of structural or concrete linguistic units into the receiving language does not necessarily imply that ‘‘language change has complicated the grammatical systems of the languages in contact by adding grammatical constructions’’ (37, [emphasis mine]). This change in language should be addressed as linguistic enrichment rather than as an increase in ‘‘complexity’’. Also the term ‘‘functional’’ which refers to the ‘‘social function’’ of language is problematic because in general linguistics ‘‘functional’’ has a di¤erent meaning. The conclusion of the author that the linguistic changes in Marathi and Hindi have the ‘‘function to establish and maintain the cultural and emotional identity of the people in Central India’’ is the core claim and at the same time the main hypothesis of the book. Chapter Three o¤ers a brief sketch of the history of Marathi-Hindi bilingualism in Nagpur (by the way, the reproduction of geographic maps would haven proven very useful!) as well as an interesting discussion about the social function and distribution of the di¤erent language varieties as codes. The author hypothesises that the choice of a code must also be viewed as a manner of maintaining a specific regional identity (45, 48). In the last chapter of Part One Pandharipande aims at determining restrictions on the processes of linguistic convergence, considering convergence and maintenance as two forces which operate upon languages in contact. A deductive catalogue of constraints is presented. These are seen as ‘‘devices for maintaining the separate identities of the languages in contact’’ (58). Of course, the belief that constraints on mutual borrowing ‘‘prevent the [total] merger of Hindi and

Reviews

169

Marathi into one language’’ (58) is to some extent too strong an assumption. Similarly, it is inappropriate to state that ‘‘[i]n order to maintain their separate linguistic identities it is necessary to maintain two separately identifiable grammatical systems’’ (60) since this would presuppose a very strong and conscious control by speakers on language usage. Nevertheless, postulating the intertwining of codes and identities as a device for maintaining regional and linguistic identity through code choice is by all means plausible. Part Two features two chapters on the process of nativization (as accommodation) within the contact of Marathi with English, Persian, and Sanskrit. In Chapter Five the question of the status of borrowings is tackled and whether or not there is a hierarchy of nativization of borrowings, i.e. the process of assimilation of borrowed material and structures into the receiving language. This is discussed within the context of the borrowings from English, Persian, and Sanskrit into Marathi. The investigation of the data makes clear that the borrowed material is marked and kept distinct from the native Marathi stock. Nevertheless, several degrees of nativization are recognised and a hierarchic ‘‘nativization scale’’ is established. The loans from Sanskrit are more accommodated than those from Persian and English. Furthermore, the English loans are more marked and the Persian ones are more nativized. The explanation that the positioning of the involved languages within the scale is the result of the psycholinguistic and sociological attitude of the Marathi speakers towards the languages cited above is plausible. To corroborate this hypothesis the author mentions the example of villagers who do not treat the English loans di¤erently from the genuine Marathi stock. As English is here viewed as the symbol of modernization, the attitude of these speakers towards it di¤ers from that of the inhabitants of the urban centres. Chapter Six investigates the functional dimension of code-mixing between Marathi-English and Marathi-Sanskrit. The starting point of the discussion is that modernization can be looked at as ‘‘language change which marks the language as ‘modern’, as opposed to ‘traditional’, in the speech community’’ (90). Three major processes are identified within the overall process of modernization of Marathi society and language: ‘‘Englishization’’, as opposed to ‘‘Sanskritization’’ and ‘‘Hindiization’’. Each process is the response or the reaction of the Marathi society (as depicted in the interviews with the editors of two Nagpur daily newspaper) to the process of modernization and looked at as the marker of the multi-identity (respectively international, national, and regional) of the Marathi speakers. On a negative note, I have to observe that section 6.3, on the local varieties and the newspaper register, is just far too long (eleven pages). Furthermore, the claim that the use of dialects or varieties of Marathi better signal the emotional identity of the speakers (101) is obvious.

170

Reviews

Part Three features two chapters about the creative dimension of bi/ multilingualism in Maharashtra, the increase of style repertoire and of style range in creative literature to reflect the socio-cultural identity of Maharashtrian society. The first notion introduced in Chapter Seven is language-mixing which is supported by the relevant theoretical background. Unfortunately, no distinction between the phenomena of code-mixing and code-switching is made. Although in the literature the term ‘‘code-mixing’’ usually refers to individual, rule-less (as opposed to code-switching) occurrences of switching between two or more languages in the speech of a multilingual speaker (Myers-Scotton, Carol and Janice L. Jake. 2001. Explaining aspects of code-switching and their implications. In One Mind, Two Languages: Bilingual Language Processing, Janet Nicol (ed.): 84–116. Malden, MA: Blackwell), here it seems to be used synonymously with the notion of ‘‘contact languages.’’ (Thomason, Sarah G. 2001. Language Contact. An Introduction. Edinburgh: Edinburgh University Press.) Section 7.2 is concerned with cases of individual transfer occurrences of bound morphemes in the speech of uneducated persons. These are referred to as ‘‘borrowings’’ but a thorough analysis of the case studies reveals that they are not. It is well known that contact-induced morphological change is not rare in word formation but exceptional in inflection. The only monograph available on the issue of inflectional borrowing (cf. Gardani) rules out occurrences of individual borrowing. It is true that language change originates ultimately through individual innovations and that multilingual people are the first source of such innovations but both lexical and grammatical transfer is to be assumed to come about only if it overflows the boundaries of the single individual speech and comes to be used by a numerically or socially prominent part of the linguistic networks. Moreover, Gardani amounts to a definition of inflectional borrowing as acknowledged only when inflectional morphemes are attached to native words of the receiving language. This excludes cases like examples (311)–(313) which are mere quotations of foreign forms. Irrespective of the fact that Pandharipande has a good knowledge of the general relevant literature, her discussion of structural borrowing is not up-to-date, since she takes into account neither Thomason & Kaufman’s (1988) theory of language contact nor van Coetsem’s (van Coetsem, Frans. 2000. A General and Unified Theory of the Transmission Process in Language Contact. Heidelberg: Winter) approach to the transfer process. This is very unfortunate because these are seminal studies in the theory of contact linguistics. Once again, the research questions (‘‘whether there are any conditions which determine the types of borrowing possible in a given language’’ and whether or not these conditions are universal or language particular, 130) formulated in 7.4 are conspicuously wider than the scope and the space available to this book.

Reviews

171

The book concludes with a chapter on the range of styles in use by playwrights in modern Marathi drama (after 1950). Following the thread of the book that variation (as language change as well as in repertoire) is functional as to convey identity and socially definable categories (such as intimate vs. formal), the author claims that the development of major styles (the englishized, the sanskritized, and the persianized styles) has to do with their use and sociolinguistic function in di¤erent domains of communication (tradition vs. modernity). Dispensing with matters of form last, a sharp critique has to be levelled against the editorial work around this publication. It is deplorable that a review article has to record as many blemishes as the following: The body of the text is not justified in the foreword nor is it the case in the references. In many cases the interlinear (morpheme) glosses are not aligned (e.g. examples (25), (26), (50), (54), (58), (75), (76), (86), (96), (175), (176), (285b), (288b), (289b), (307). This list is not exhaustive!). Moreover, the glosses in (12)–(14) and (69) are not at all straightforward. Mistakes recur throughout the book: just one example to illustrate: p. 9 ‘‘and’’ is used two times instead of ‘‘the’’. On p. 81, the legend of the table belongs with examples (194)–(199) to follow. A similar case is (32). The most manifest example of this bad editorial work is found in the use of bibliographical references. On pp. 8, 30, 44, 52 (three times), 53 and 95, ‘‘Kachru (1980)’’ is quoted. A quick look at the references reveals that there are two Kachrus, both referred to as Kachru (1980). The author probably refers to Kachru Y. (1980) but even the best disposed reader could not guess which article is quoted! On p. 65 ‘‘Kachru (1975)’’ is cited. Again, there are two. Which one is meant here? In the references themselves the use of the first name is not consistent: in almost all cases it is abbreviated but in Kahane, Henry and Renee (1979) and in Southworth, Franklin (1971) it is not. Moreover, throughout the references there are many notes of a private nature which do not belong in a final product: ‘‘Fishman, J. 1969 (can’t find it on the computer, need more information, cf. p. 103)’’; ‘‘Joshi, A. 1982 (book or paper?)’’; ‘‘Nadkarni, D.S. 1975 (This is the only one dated 1975 found on the computer. Is it the correct reference?)’’; ‘‘Shridar and Shridar [sic: no firstnames!] (need to find the reference)’’; ‘‘Tendulkar, V. 1963 (the book is not on shelf in the library. Need to find place: publisher.)’’; ‘‘Woolford 1980 (publisher?)’’. Overall, the reader gets the impression that the author (a professor at University of Illinois at Urbana-Champaign) is asking her assistants to seek information. Further flaws in what follows: Tagliavini’s 1969 book is titled ‘‘Le origini delle lingue neolatine’’ and not ‘‘neolatin’’. Von Wartburg wrote in 1951 a work on the origin of the Romance ‘‘Vo¨lker’’ and not ‘‘Volker’’ (cf. also p. 89 in the notes). Weinreich (1967) is cited: here again,

172

Reviews

there is a mistake in the spelling as the French word for linguistics is linguistique and not ‘‘linguistigue’’. Weinreich’s opus Languages in contact (1953) has been inserted twice, although the author herself notices that the quoted edition is the edition reprinted in 1968. When citing the edition of 1968, the author omits the plural morphemes in ‘‘finding and problem’’. These are serious shortcomings and the result of careless editorial work. The overall impression left by the book is that it is quite accessible, the single chapters are well linked and at the same time su‰ciently diverse to be of genuine interest to the target audience as they present appealing data and some interesting claims. Unfortunately, many of its assumptions call for much more attention in terms of depth and space. Although the book is a quite useful resource for anyone wanting to explore the terrain of sociolinguistic research in Central India, some conceptual inconsistencies and plenty of deplorable formal shortcomings strongly reduce its scientific value.

Wali Kashi. Marathi. Munich: Lincom Europa 2005. 72pp. þ i–iv. (Languages of the World/Materials 441). ISBN 3 89586 314 9. EUR 37.50 —Reviewed by Leena Kelkar-Stephan, Halle Marathi, one of the major languages of the Indo-Aryan family, is spoken by approx. 96 million people in India (62,481,681 first language speakers according to the Census of India 1991). It is one of the eighteen o‰cial languages of India, with a long literary tradition. Marathi as a foreign language is taught in a few universities outside of India and the numerous works on Marathi grammar, with a few exceptions, are written in Marathi and are thus inaccessible to the non-Marathi speakers and to learners of Marathi as a foreign language. There is in e¤ect a dearth of pedagogical grammars, teaching aids, course books, texts or translation articles, and audio material for Marathi. Kashi Wali’s (hereafter KW) latest book Marathi gives basic information on the phonology of Standard Marathi, its writing system, morphology and syntax. The book contains an introduction (pp. 1–2), which however does not state the aim and scope of the book, followed by four chapters and a bibliography, with no word index. In Chapter 1 (pp. 3–8) KW briefly describes phonological properties of Marathi, such as vowels, consonants and their distribution, syllable structure, suprasegmentals and some of the Sandhi rules in Marathi. In Chapter 2 (pp. 9–36), the sub-section on nominal morphology, KW treats standard topics like gender, number and case in nouns and pronouns. In this chapter, KW also touches upon the distribution of two special reflexives sw taha e

Reviews

173

and ap næ which are said to be special to Marathi in the Indo-Aryan family. However, KW herself has commented extensively on this latter topic in her earlier publications. Sub-sections on adjectives, adverbs, nominal categorizers as well as an account of the verbal morphology round up this chapter. Chapter 3 (pp. 37–64) reviews Marathi syntax, taking a closer look at noun phrases, adjectival phrases, adverbial and postpositional phrases. In this chapter, simple, compound and complex sentences are also discussed, with an interesting account on the grammatical relations in Marathi sentences. Chapter 4 (pp. 65–66) contains a sample text. However, here again the reader is kept in the dark as to the purpose of this text. Bibliography (p. 67–72) includes ‘‘an extensive bibliography including past and present grammars, recent dissertations and scholarly articles in generative linguistic theory.’’ In view of the desideratum of material on Marathi mentioned above, every new work on Marathi grammar is a welcome contribution. KW’s book is a handy summary of older Marathi grammars and linguistic research by herself and by other authors on many topics. e

Bernhard Hurch, ed., Studies on Reduplication. 2005. Berlin: Mouton de Gruyter. 640 pages. Hardbound. Price: 119,63 Euros —Reviewed by Lionel Wee, National University of Singapore Studies on Reduplication is intended to serve as a state-of-the-art volume, reflecting ‘‘a series of di¤erent positions, di¤erent problems, di¤erent perspectives on one broad problem—reduplication’’ (Hurch: 9). There is no doubt that the volume serves its intended purpose admirably. It contains discussions of reduplication from a remarkably varied list of languages (including Acehnese, Chechen, Madurese, and Thao), a range of language types (including sign language, child language, and pidgins and creoles), and provides analyses that are couched within di¤erent theoretical assumptions (though there is the not unexpected influence of Optimality Theory in a number of contributions). This extremely rich and comprehensive volume is therefore of great value to scholars interested in reduplication. This would, I assume, include a not insignificant number of scholars interested in South Asian languages, since these languages are particularly rich in reduplicative morphology. As this review is intended for a publication specializing in South Asian languages, let me first comment on the general organization of the volume before focusing more extensively on those papers that deal specifically with reduplication in South Asian languages. The volume contains 24

174

Reviews

papers (excluding the Preface and Introduction). While one can glean some kind of order in the way the papers are organized—particularly the later papers—it is a pity that, editorially, a greater e¤ort was not made to impose some kind of thematic coherence on the volume. For example, Hurch clearly recognizes that there are papers that deal with similar issues and attempts to group these together. Thus, the contributions by Dressler et al. (Reduplication in child language), Leroy and Morgenstern (Reduplication before age two), and Sofu (Acquisition of reduplication in Turkish) are placed sequentially because they all deal with children’s acquisition of reduplication. Likewise, the contributions by Bakker and Parkvall (Reduplication in Pidgins and Creoles), Kouwenberg and LaCharite´ (Less is more: Evidence from diminutive reduplication in Caribbean Creole languages), and Abraham (Intensity and diminution triggered by reduplicating morphology: Janus-faced iconicity) are grouped together because they deal with reduplication in pidgins and creoles. As are those by Pfau and Steinbach (Backward and sideward reduplication in German Sign Language), and Wilbur (A reanalysis of reduplication in American Sign Language) because they discuss reduplication in sign languages. But the principle for organizing the earlier papers appears to be far less transparent. True, the first three papers (Rubin’s Reduplication: Form, function, and distribution; Gil’s From repetition to reduplication in Riau Indonesian; and Inkelas’ Morphological Doubling Theory: Evidence for morphological doubling in reduplication) can all be said to deal with a host of foundational issues (Respectively: What kinds of reduplication are there? How can we distinguish reduplication from repetition? Should reduplication be based on phonological identity—as has been influentially presumed—or should it involve morphosemantic identity instead?). But if this is the organizing principle, then it is not clear why Singh’s Reduplication in Modern Hindi and the theory of reduplication (which comes much later) does not immediately follow, since this, too, raises a number of substantive issues about the nature of reduplication. Instead, we have a number of papers that arguably form a coherent category in that they deal with a range of very specific reduplicative phenomena. For example, Downing’s The Emergence of the Marked: Tone in some African reduplicative systems provides an analysis of tonal asymmetries. McLaughlin’s Reduplication and consonant mutation in the Northern Atlantic languages focuses on backcopying, and treats reduplication as the juxtaposition of stems rather than deriving the reduplicant from the base. Nelson’s Wrong side reduplication is epiphenomenal: Evidence from Yoruba argues that cases where the reduplicant seems to appear on the opposite end of the base in fact do not exist. Shaw’s Non-adjacency in Reduplication provides an explanation of adjacency violations by making a number of interesting assumptions about the nature of the base: that it

Reviews

175

can be characterized in either morphological or prosodic terms, and that it is a derivative concept rather than a grammatical category unique to reduplication. Finally, Urbanczyk’s Enhancing contrast in reduplication highlights various strategies for disambiguating the grammatical functions associated with particular types of reduplication. There is also a series of papers that have apparently been placed together because they all adopt a diachronic orientation towards reduplication, often with a special focus on specific languages or language families (Shibasaki’s On the grammaticalization of verbal reduplication in Japanese; Haugen’s Reduplicative allomorphy and language prehistory in Uto-Aztecan; Rose’s Reduplication in Tupi-Guarani languages: Going into opposite directions). Again, if this is what motivates the grouping, then it is not clear why Kulikov’s contribution (Reduplication in the Vedic verb: Indo-European inheritance, analogy and iconicity) is not placed together with these. Finally, there are papers that appear to be concerned primarily with the size of the reduplicant. Just how big or how small should we expect the reduplicant to be? El Zarka’s On the borderline of reduplication: Gemination and other consonant doubling in Arabic morphology and Maas’ Syntactic reduplication in Arabic explore di¤erent directions on this issue. The former seeks to include gemination, thus arguing that the reduplicant can be as small as a segment, while the latter looks to include syntactic constructions, arguing that syntactic reduplication in Arabic languages is a grammatical device for modifying the semantics of the verbal predicate. This concern about the size of the reduplicant makes it natural for Keane’s Phrasal reduplication and dual description to be placed together with these two papers, but instead, it comes much earlier in the volume, preceding Singh’s contribution. I would not want to push this point about the volume’s organization too hard, however. It is a ‘‘small-small’’ blemish on what is otherwise a wonderfully broad and informative collection. But some attempt at thematic coherence would have created an opportunity for Hurch to engage in a more critical discussion of the various issues that scholars interested in reduplication are grappling with today, and this would have also allowed him the opportunity to (i) suggest how the issues could be prioritized; (ii) explore the extent to which some of the issues transcend theoretical boundaries; and (iii) suggest additional issues or directions of his own. As it stands, the Introduction concisely summarizes the contents of the individual contributions, occasionally noting recurring themes, but avoids providing anything like a more thought-provoking discussion. I cannot help but feel that a more provocative Introduction would have given the reader a much broader and nuanced overview of the field (accompanied, of course, by the appropriate caveats that actual contributions are

176

Reviews

often poly-thematic and hence may cross-cut thematic boundaries, and that the themes themselves are neither necessarily exhaustive nor immune to criticism—caveats that, I remind the reader, also apply to my own remarks about the organization of the volume). I suspect that Hurch refrained from a more critical intervention because he saw himself primarily as a facilitator, and wanted to allow the various papers to speak for themselves. This is admirable and modest, but I would have very much welcomed hearing what Hurch himself has to say about the current state of studies on reduplication, particularly since (the Introduction aside) the volume does not contain a specific contribution from him. Let me now attend to papers that deal explicitly with South Asian languages, beginning with Keane’s contribution. Keane is to be lauded for her willingness to address the broad range of reduplicative phenomena across a number of South Asian languages, drawing on Bengali, Hindi, Kannada and Tamil. Her discussion of these data directs our attention to a number of interesting issues. One, are there in fact dissimilatory strategies that function to keep certain patterns of reduplication distinct from others (p. 240)? Two, how context-dependent is the meaning that is signaled by a given pattern of reduplication (p. 241)? Three, and this relates to the preceding question, how stable are the intuitions of native speakers vis-a`-vis reduplication data, when the patterns are only ever found in highly colloquial speech (p. 241)? These questions arise in the context of Keane’s discussion of South Asian echo reduplication, most of which convey meanings that may be glossed as ‘and the like’ or ‘and so forth’, as seen in her Tamil data, where the base (which includes a derivational su‰x) has been reduplicated. The reduplicant itself uses a fixed /ki/ to replace the initial CV sequence of the base (p. 247). [cantoocam-aaka kintoocam-aaka] joy-fully echo ‘joyfully and so forth’ Intriguingly, there are cases where the verb and its indirect object also appear to be reduplicated (p. 248). [kumaarukku ku
utteen-~~u kimaarukku ku
utteen-~~u] poy Kumar. DAT give.PAST.1S-QUOTE echo lie collaatee say.NEG.IMPER ‘Don’t lie that you gave (it) to Kumar or some such nonsense.’ As Keane points out (p. 251), echo reduplication appears to operate over syntactic as well as morphological structures, and this has implications for

Reviews

177

any attempt to locate reduplication within some component of the grammar such as the lexicon. Keane’s own solution is couched within a framework of declarative phonology, but as far as I can see, while declarative phonology contains a number of useful devices (default inheritance and unification working in conjunction with structural descriptions of the base) that allow the echo reduplication to be formally described, it falls short of actually explaining the data, a point conceded by the author herself (p. 256). In this regard, Singh’s paper, which deals with data from Modern Hindi, is uncompromising in its quest for explanations about the nature of reduplication. Singh begins by pointing out that reduplication patterns which insist on replacing the initial consonant of the base with a fixed consonant are extremely prevalent in South Asian languages. This, he argues, has implications for the question of whether partial reduplication (either historically or synchronically) can always be assumed to derive from full reduplication (p. 265). Another important question arises from an examination of examples like those below (p. 266; p. 272). The first example obviously bears the kind of semantics that Keane glosses as ‘and the like’ in her paper. Singh’s own preference is to dub this the ‘Et Cetera’ interpretation. Crucially, while both examples seem to utilize the ‘Et Cetera’ principle, only the first example involves reduplication. bhagavaan bhagavaanvagavaan ‘God’ ‘crazy things like God’ chaaya paani chaayapaani ‘tea’ ‘water’ ‘refreshments’ Singh’s point is that accommodating both examples under the same analysis requires abandoning the assumption that there exists ‘‘an identifiable, definable kind of morphology that should be called reduplication’’ (p. 275). Of course, it may be asserted that the semantics are in fact di¤erent so that the ‘Et Cetera’ principle applies in the case of ‘crazy things like God’ but not ‘refreshments’. Another argument might be that even if the same semantics is conceded to apply, there is no reason to conclude from this that the morphological processes involved are similar. These issues hinge on a finer-grained semantic analysis of the ‘Et Cetera’ principle and the theory of morphology that we wish to buy into. Be that as it may, it is clear that the assumption that reduplication constitutes a distinct morphological process has for some time now begun to acquire the status of an axiom. So, whether or not we accept the specific theoretical framework that Singh works within (Whole Word Morphology), Singh is certainly right to remind us that the assumption is one that has too easily been taken for granted, and as such, warrants critical revisiting.

178

Reviews

Kulikov’s paper begins with a succinct, yet detailed summary of the reduplication patterns found in Vedic verbs. In Vedic, reduplication is found in the perfect, the present, the aorist, the desiderative, and two types of intensive (p. 431). One of the most interesting aspects of the paper comes from Kulikov’s discussion of the origin of the reduplication vowel in the Vedic presents. The options are either a or i, and citing the work of Kortlandt (1987 Archaic ablaut patterns in the Vedic verb. In Festschrift for Henry Hoenigswald: On the Occasion of his Seventieth Birthday, George Cardona, and Norman H. Zide (eds.), 219–223. Tu¨bingen: Narr.), Kulikov (p. 437– 438) suggests quite convincingly that both vowels are in fact traceable to the full and zero grades of the reduplication syllable in Proto-Indo-European. Pulling together arguments about the tendency towards consonant cluster simplification (especially in geminates), the functional motivation to preserve the formal character of the reduplication pattern, and the predisposition to generalize the choice of reduplication vowel, Kulikov leads us to the hypothesis that i began life as an epenthetic vowel in the zero grade reduplications, and subsequently spread over to the full grade as well. This hypothesis nicely dovetails with the synchronic observation that while i is undoubtedly highly prevalent in the Vedic verbs, there are still occasionally alternations with forms that show traces of the original a. Thus, whatever one’s theoretical bent or areal interest, this excellent volume fully succeeds in conveying the range of intriguing issues surrounding the study of reduplication. Reading through this volume, one not only gets a palpable sense of how far linguists have come in understanding reduplication, but more importantly, a clearer idea of how much further we have to go, and the issues we need to grapple with, if we are to ever get there.

Paul St-Pierre and Prafulla C. Kar (eds.). 2005. In Translation: Reflections, Re-fractions, Transformations. Delhi: Pencraft International. 286 pp. —Reviewed by Carmen Wickramagamage, University of Peradeniya Let me first confess to what I am (as well as to what I am not) before I o¤er my considered opinion on the book that I have been asked to review. I cannot claim to be a Translation Studies scholar with an established reputation in the field of TS. I have not participated enough in the debates that characterize the field of Translation Studies to claim access to this title. My entry into the field of Translation Studies is very recent, almost accidental. When I happened upon Translation Studies, I did so strangely enough as a teacher—and one of the coordinators—of the Program in Translation Studies that we run at the University of Peradeniya, Sri Lanka. At the

Reviews

179

University of Peradeniya, we have discovered Translation Studies rather late. Some might even say that Sri Lanka too has discovered Translation Studies rather late, that is, its social and political uses, although one would have thought that considering the rather brutal and intractable ethnic conflict and civil war that mar the socio-politicalscape of Sri Lanka today we would have turned to Translation sooner as one way to resolve the dispute by peaceful means. The net result of course of this failure to o‰cially acknowledge Translation as a felt national need has meant that to date Translation Studies mostly does not attract as either students or practitioners the best there is in the field in terms of bi-lingual and bi-cultural competence. After all, who would want to make a professional commitment to a field where the pay is poor, jobs are scarce, and recognition nil? So, while a Professor of Linguistics once told me that we should be prepared to work with students who possess 112 languages, what we must work with in Sri Lanka are those with two 12 languages (that is, if we are lucky!). My approach to the book therefore is as a teacher of Translation Studies, as someone who has constantly to grapple with the issue of how to teach, and what to teach, when we teach Translation Studies in the given sociopolitical and economic climate of Sri Lanka today. I am looking in other words for new directions in pedagogy, fresh inputs into the academic content, and insightful reformulations of the disciplinary parameters of Translation Studies, etc., when I read anything having to do with Translation— all with an eye to becoming a better teacher. The Program in Translation Studies, University of Peradeniya, aims at turning out competent translators who in the world outside of academia will be called upon to engage with the problems that arise in connection with the translation of actual texts—mostly ordinary, pragmatic texts of every day. So the question for me always is what makes a professional translator? What ingredients should go into her/his production? I am always looking for answers to these questions. In Translation is in many ways a refreshing read. Its essays o¤er not just ‘‘reflections’’ on but ‘‘refractions and transformations’’ of matters familiar to many in Translation Studies. Among the most exciting refractions, for instance, are a series of essays that in their di¤erent ways destabilize or problematize that opposition familiar to those in Translation Studies between Source and Target—Languages and Texts. More often than not, we assume that a relationship of symmetry obtains between the Source and Target Languages—hence, the quest for the holy grail of ‘‘equivalence.’’ Not so to these scholars. Take ‘‘Language and Translation’’ by Anthony Lewis, for instance. This essay asks us to get away from that understanding of the relationship between Source and Target Languages as one of ‘‘symmetry [hence notions such as equivalence and faithfulness]’’ (23) and to

180

Reviews

think of translation as ‘‘the transformation of a communicative act by one community (or an individual from that community) into a meaningful communicative act for another community [or an individual in that community]’’ (23). Sukanta Chaudhuri goes further and argues for the liberation of the translation from the debt that links it to the source text, which ascribes a ‘‘lesser status’’ or relationship of dependence to the target text visa`-vis the ‘‘original’’ in the eyes of readers, critics and legislators. To him, translation enables the possibility of reincarnation for the Source Text. The translator then is a reader but who unlike other readers ‘‘leaves a record of his reading in an externally embodied version of the work’’ (75). The relationship of dependence, in other words, does not say it all. If the translated text is a ‘‘mere’’ reproduction in other words of the original, how can there be more than one translation of the same text? Multiple translations attest to the status of the translation as always a creative interpretation. And this redefinition of the Target Text or Translation as a ‘‘creative transposition’’ (as Roman Jakobson put it, cf. Venuti, Lawrence (ed.). 2000 The Translation Studies Reader. New York & London: Routledge: 113– 118), or as a ‘‘creative interpretation’’ as Chaudhuri puts it, could provide the impetus for a revisionist definition of translation in the realm of law. Salah Basalamah foregrounds the legal implications in a global climate that shows a growing concern with intellectual property rights—a concern that may have an adverse e¤ect on translation activity. (Indeed, this is already happening in Sri Lanka where publishers are leery of publishing translations without the necessary paperwork regarding copyright clearance, etc., in place after Sri Lanka placed its signature on the agreement a couple of years ago.) He challenges the International Copyright Law (Trade-Related aspects of Intellectual Property Rights, or TRIPS) that presents translations as derivations and hence subject to the relevant clauses of TRIPS. He does not see translations as mere reproductions but as ‘‘one possible reiteration in another context, emphasizing its di¤era´nce with the original and o¤ering it survival, a new life, a new identity’’ (103). In other words, translation ensures the after-life, as Walter Benjamin would say, of the original. But while Benjamin, according to Basalamah, maintains a strict opposition between the original and translation, he sees a translation as both an original and not an original, occupying in fact an in-between state between original and derivation: ‘‘Is it ‘original’ because the original text is still perceptible through the lines of the translation, or is it ‘original’ because its contribution to the survival of the original text is such that the latter is indebted to the translation?’’ (104). This is a refreshing insight indeed into the relationship that obtains between the source and target texts in translation. To him, the texts exist then in ‘‘a relationship not of filiation, but of dialogue and interaction, which characterize relations between texts

Reviews

181

in general’’ (104). This line of argument and thought is brought to a climax as it were by Alexis Nouss. He too denounces the conventional terminology of ‘‘source’’ and ‘‘target’’ cultures/languages/texts as ‘‘essentialist and homogenizing.’’ To him, ‘‘the sub-versive function of translation destabilizes notions of appropriation, or of internal or external cultural values. It undoes the borders designating works as ‘foreign;’ it puts into question the notion of the ‘original,’ through which anteriority is easily transformed into authority’’ (223). He agrees with Basalamah that translation is thus always between sameness and di¤erence. A translation proclaims, according to Nouss, that ‘‘this can be said in this language and in that other one as well’’ (225). So in translation, ‘‘what comes after gives a reality and existence to what comes before, but where also it would never have come after if there had not been a before . . .’’ (226). Metissage then is the appropriate term for the relationship that obtains between texts in the translation situation. He thus proposes ‘‘translated between’’ for translated from/into because translation does not, should not, eliminate the distance or interval between languages. It must make that interval, that uncertainty of meaning, evident (227). Such deconstructivist interventions into long-established and upheld ‘‘truths’’ in the field are always refreshing. And yet one cannot help but wonder how the new approach might help mount a defense against charges of copyright infringement in a court of law. After all, the fact remains that what comes after in the case of translation has a relationship of a sort with what went before. Moreover, a strictly legalistic view-point could also assert that what came after would never have existed if not for what went before while what went before would have existed (in many cases) even without what came after although perhaps not in the same ways as those provided by translation. Another set of essays that took my fancy engaged with the phenomenon of Translation in a post-colonial context. These essays, in their di¤erent ways, invite us to critically reflect on what we do when we translate within a post-colonial context. Two essays, by Paul St-Pierre and Saji Mathew, situate their critiques within the specific context of India although what they assert is of equal relevance to other post-colonial multi-lingual contexts. Is translation, ask St-Pierre and Mathew, enabling the hegemonic spread of or the global dominance of English? What does globalization actually mean for the language politics of India? Mathew is not happy about the preponderance of translations into English from the other languages. All roads in translation it appears are leading to English with a pan-Indian linguistic community and culture created via English. But translating into English from the so-called ‘‘regional’’ languages will not bring about pan-Indian unity or identity when only 4% of the population understand English, says Mathew. Instead, ‘‘the ‘regional self ’ of the textual space is fallaciously

182

Reviews

transmuted into a ‘national self ’ and the original locationality in heterographic space ignored’’ (156). But how does one by-pass English or ignore its pseudo-national status, and turn it into just another ‘‘Indian’’ language? For Mathew, both parity across and understanding between language groups can be promoted and fostered via ‘‘Indian literatures in translation,’’ not ‘‘Indian literature in translation [into English].’’ So, translate into as many Indian languages as possible, says Mathew. I was not however sure how feasible his solution was to this unsatisfactory state of a¤airs that obtain when it comes to the translation scene in India. Two questions arise: are there competent translators to undertake translations between regional languages, say Malayalam and Oriya, for instance? Is it financially feasible? It is not easy after all today to escape the call of the ‘‘filthy lucre.’’ Mathew proposes ‘‘indirect translations as solution to the problem of finding competent translators for certain language combinations.’’ But isn’t such a translation then a ‘second-hand’ product at best, at a second remove from the original, with all of the problems for translation that that entail? And would there be an adequate readership for such translations? After all, in this day and age, very few publishing houses would want to invest in a project that may not show positive returns. The problem that Paul St-Pierre speaks of in the specific context of Orissa is related to the issue of the pan-Indian dominance of English that Mathew speaks about: the emergence of a non-Oriya speaking Oriya community who depend on translations from Oriya into English in order to remain in touch with their culture and language. The literate and a¿uent urban middleclass in Oriya is increasingly anglophile, St-Pierre seems to say. It is the translations from Oriya into English that enable them to keep a foot in both worlds. In the post-colonial world, we have come full circle. We are back to creating a class of ‘‘brown sahibs’’ who are ‘‘local’’ in skin-color only, as Macaulay so presciently put it in his (in)famous Minute on Indian Education in 1835. This class of Indians who are English in everything but color (does English have a color any more?) can now assuage their guilty conscience by reading Oriya texts in translation. Such is the ‘‘violent translation’’ that they have undergone as a result of European/British imperialism. The pity is that we have not recovered from this ‘‘epistemic violence’’ (Gayatri Spivak’s memorable phrase) but are sliding more and more within its grip as English takes on a new role as the seemingly neutral and natural medium of global communication. In Sri Lanka, too, we see this happening, with the absurd speed at which both the State and the populace embrace English-medium education, thus foregoing the opportunity to create bi-lingual or tri-lingual speakers via the education system. Translations into English enable them to assuage their conscience and to convince themselves that they are still in touch with their ‘‘local’’ languages and cultures while to

Reviews

183

all intents and purposes they have donned the cap of English. Traduttore traditore (qtd. in Jakobson: 118). In the era of globalization, translation again takes on the role of traitor, returning us to debates that we thought we had left behind with independence from colonial rule. The issue of language politics, or parity between languages, in the era of globalization, which e¤ectively means the global dominance of English, then, goes beyond the narrowly conceived notion of equivalence in conventional translation studies. What does equivalence mean when it comes to English and, say, Sinhala (a language spoken in Sri Lanka) where the huge maw of English seems ever ready to swallow the latter in the name of supplying terminology to handle new cognitive domains? The ever-closer involvement of English in the life of Sinhala, the alacrity with which Sinhala speakers take to ‘‘things’’ English, makes one fearful for the survival of the latter in the years to come. Will translation help preserve and retain the identity of Sinhala as a separate language? And, as usual, Gayatri Spivak, in her inimitable fashion and style, has something to contribute to this debate. Calling European imperialism the ‘‘founding crime of the world we live in today’’ (243), Spivak draws our attention to the issue of parity, or equivalence, not just in terms of ‘‘cultures’’ and ‘‘languages’’ as they are conventionally understood but rather in terms of ‘‘standard’’ or mainstream languages and cultures versus dialects and subaltern communities. ‘‘The toughest problem is translation from idiom to standard . . . The speaker of some form of standard Bengali cannot hear the self-motivated subaltern Bengali unless organized by politically correct editing, which is equivalent to succor from above’’ (248). As Spivak correctly points out, the attainment of social justice through translation entails the non-elimination of di¤erence, or the safeguarding of equivalence, at this level too. And for Spivak the issue is not so much translation into English from other Indian languages but the retention of the subaltern di¤erence as the texts get translated into English or any other Indian language. In other words, the Bengali spoken by the subaltern should not be standardized so as to eliminate the di¤erence between subaltern and elite. The subaltern must not only speak but must be heard in terms of her di¤erence: ‘‘it is this act of hearing-torespond that may be called the imperative to translate’’ (247). Spivak dismisses those critics who have rejected the crucial place accorded translation in Spivak’s project to make the subaltern speak or be heard. ‘‘The subaltern can speak, does speak, and is heard, in her own native language,’’ say her critics (247). To this charge, Spivak poses the ‘‘hearing-to-respond’’ dimension to translation—clearly by those whose ‘‘hearing’’ and listening matter in the corridors of power. However, I would like to pose the following question to Spivak: doesn’t this speaking-listening situation imagined and imaged by her merely confirm existing power relations without any attempt

184

Reviews

at inserting a critical wedge that might gesture towards the eventual dismantling of those power relations? In other words, what next, after the subaltern has spoken and the elite have heard what she has to say? Is it again business as usual? How will it lead to the elimination of the inequalities that subtend her di¤erence from the elite? Or is that hearing the crucial first step in the process of dismantling? For Probal Dasgupta, in ‘‘Tra‰cking in Words,’’ on the other hand, the imperialist agenda is an inescapable fact of the whole enterprise of translation. He coins the terms ‘‘Fundamentalism’’ and ‘‘Foundationalism’’ to describe the two major impulses that helped spread the activity of translation across the globe. ‘‘Fundamentalism’’ describes mainly the proselytizing imperative, whether it is preaching the virtues of Christianity via the Bible or that of European civilization via its classical texts. The relationship between the Source Text/Culture and the Target Text/Culture here was hierarchical with the target language in the position of a lowly or unworthy vehicle via which the host/recipient culture becomes beneficiary to the superior point of view o¤ered in the source text/culture. One would have thought ‘‘foundationalism’’ would turn out to be its opposite since it describes the spread of a ‘‘techno-scientific’’ mindset and ideology. But, as Dasgupta points out, both betray the ‘‘missionary’’ impulse. ‘‘Through the e¤orts of a neutral everybody coming from a generalized everywhere, a coherent body of valid knowledge already accumulating in the industrialized societies will grow into a universal scientific foundation for all human knowing and striving’’ (44). Knowledge itself comes to be identified with the ‘‘west,’’ constructed and promoted at the expense of other knowledges and other points of view. Whether explicitly religious or overtly secular, both promote the western worldview and attempt at converting (translating?) the rest of the world’s population to that cast of mind. Both enable the hegemony of the ‘‘western’’ civilization and its stranglehold upon the world’s non-western peoples. So, should we then disavow translation in order to resist this hegemonic spread of western thinking via translation? Or, and this is my view, should we use translation to oppose/destabilize western thought forms and social structures with those from other parts of the world? We may be tra‰cking in words but words are all we have to resist the global dominance of English and the ‘‘gospel’’ of occidental worth and merit that it carries. And, so, I now return to the question with which I began this essay: what has all of this got to do with translation pedagogy, or with the production of competent translators, which is our primary objective as teachers in and coordinators of the Translation Studies Program within a context such as Sri Lanka? Many of the debates that the above-cited essays participate in—are they relevant or extraneous to the job of training quality translators who as translators in state institutions might be mostly translating

Reviews

185

government documents and circulars in English or Sinhala into Sinhala and Tamil or English and Tamil? (Some of course would be feeding the global demand for information and entertainment through translations of technoscientific and literary texts.) Would it be useful to them to know the reformulation of the relationship between Source and Target texts into one of dialogue from that of filiation? Will they be confronted with the need, or perhaps given the opportunity, to resist the global hegemony of English in the translation industry as a whole? Or will they have to deal with the heteroglossic and heterographic aspects to language that require them to use translation as a vehicle for subaltern voices? Rajendra Singh, in his contribution to this volume, would answer with a resounding ‘‘no.’’ To him, they are all extraneous issues that distract one from what ought to be the real focus of Translation Studies. I would hesitate at such a definite ‘‘no,’’ and would say both ‘‘yes’’ and ‘‘no.’’ Indeed, in their normal bureaucratic lives and selves, they would neither be expected, nor rewarded for, foregrounding those complex theoretical perspectives. Working under time and other administrative constraints, theirs is not to ask the question why. However, in both their personal lives and in their everyday translation activity, they might get the opportunity to do justice to the local languages vis-a`-vis English by translating directly between Sinhala and Tamil, for instance, rather than going through English. They might be able to engage in a more nuanced translation that actually gestures towards a subaltern voice and perspective without eliminating the dialectal variation by a more standard version. They might refrain from making the Target Language into some lowly vehicle for the reception of a superior language content and point of view. They might go on to translate documents that foreground a ‘‘native’’ or ‘‘local’’ point of view that significantly di¤ers from but o¤ers a worthwhile alternative viewpoint to the hegemonic western point of view on how the world turns. So, indeed, the competent translator that the Translation Studies Program of the University of Peradeniya imagines cannot but benefit from the debates—reflections, refractions, transformations—outlined above. Yet something else is necessary: a good knowledge of the two language systems. Indeed, given the dearth of competent bi-linguals who would take up translation as a profession in Sri Lanka due to inadequate recognition by society and state, training in language becomes even more acute. One might call this a language-based approach to translation studies. This is the approach that In Translation intentionally and otherwise eschews and for which Rajendra Singh stands like the proverbial voice in the wilderness. Several essays in the book make it a point to mention a ‘‘turn’’ away from language to a turn towards culture. So Daniel Simeoni fires the first salvo: ‘‘everyone is into culture now’’ (6). Why? Because ‘‘translations are textual artifacts whose functions are inscribed in the particular social culture that

186

Reviews

hosts them’’ (5). Michael Cronin defines the ‘‘cultural turn’’ in the following manner: ‘‘translations [are] no longer to be seen as free-floating aesthetic artifacts generated by ahistorical figures in a timeless synchronicity of language but as works produced by historical figures in diachronic time.’’ And Alexis Nouss is adamant: ‘‘Translation is a cultural fact, not a linguistic fact’’ (222). For Harish Trivedi, the shift from the ‘‘linguistic turn to the cultural turn’’ means the following: that ‘‘texts were constructed not primarily of language but in fact of culture’’ (253). ‘‘Thus, the translation of a literary text became a translation not between two languages, or a somewhat mechanical sounding act of linguistic ‘substitution’ as Catford had put it but rather a more complex negotiation between two cultures’’ (254). Even Anthony R. Lewis’s description of translation as the transfer of a communicative act in one language into another de-privileges language at the expense of culture (23). Is this mere coincidence or a principled standpoint that the book’s editors embrace? It must be a conscious choice because Singh makes it a point to thank the editors for providing him space to articulate a diametrically opposed point of view to that adopted and consciously promulgated by a number of contributors. Even without that voiced avowal, the predilection towards ‘‘culture’’ is clear for, except for Singh, no one articulates or promotes a linguistic approach to translation studies. But the question I have is the following: is this Culture v Language division and debate tenable in the field of translation studies and practice? What exactly do we mean by Language and Culture in the debate within Translation Studies? Is Language here seen as some pure linguistic structure outside time and space, that is, immune to pressure from the community that uses it within a particular spatio-temporal frame? If not, why must we choose one over the other? It is this issue that I wish to address in the following paragraphs. I wish to begin with the famous ‘‘gavagai’’ example of Quine (cf. Venuti 2000: 94–112) that he utilized (unfortunately!) to prove the impossibility of translation. For the few readers who may not know what this story is about: the anthropologist encounters the hunter in a forest clearing. A hare darts across the clearing. The hunter gestures towards the hare and says ‘‘gavagai.’’ What did he say? Quine used this example to argue for the impossibility of translation. But the question for me is the following: what impedes communication (or makes it impossible according to Quine) in this classic encounter? Is it the absence of language or culture? At an obvious level, the hunter and anthropologist do not share a language. The listener may decipher ‘‘gavagai,’’ that is, figure out whether it is word, phrase or sentence, or whether it is noun, verb or adjective, only if s/he knows the linguistic structures, rhythms and patterns of that language. In other words, we cannot discard or disregard the strictly linguistic dimension to language.

Reviews

187

But on the other hand, would sharing a language necessarily enable communication? The hunter and anthropologist may at some future date share a language without the anthropologist always being able to fully comprehend the meaning-systems of the hunter due to cultural di¤erences. This is indeed the case in many post-colonial settings where the native and the colonizer share a language but are not always able to e¤ectively communicate due to the absence of a shared cultural base. The case of English in its former colonies is one of the best examples in support of my claim. Let me begin with a couple of real life examples from the courses that I have been part of as both student and teacher. Faced with the sentence John fired Mary, my fellow student does not see a di¤erence, or the di¤erence, between firing and firing at. They think both mean the same: that John shot Mary. Another example: An American film sub-titled in Sinhala translates graveyard shift as the patrol that the police undertake around the graveyard/cemetery. Or when the narrator of a Discovery Channel program proclaims that the baby is out cold, to indicate the fact that he is fast asleep, the sub-titling reads that the baby is out in the cold. What is entailed in these oh-so-common misunderstandings and hence mistranslations? Of course, these are not the intentional misreadings of the subversive postcolonial that Homi Bhabha or Rafael would celebrate. They are genuine mistakes having to do with knowledge gaps among speakers of one language (English) with many use(r)s. In the first example, is it absence of familiarity with American culture that impedes reading or the absence of familiarity with American English? After all, isn’t it the dropped preposition in one and its presence in the other that accounts for the di¤erence in meaning in John fired Mary or he is out cold? On the other hand, one could argue that no amount of mastery over the grammar or the purely linguistic structures of American English would make a translator read the graveyard shift correctly. It is a usage that is firmly located in culture and to read which a knowledge of the American idiom is necessary. Similarly, that ubiquitous phenomenon spawned by the American TV generation and culture: the couch potato. No amount of dictionaries will illuminate that phrase for you, unless one is looking up a dictionary on American idiomatic usage! And Judith Lavoie’s essay, ‘‘Mark Twain versus William Little Hughes,’’ from the volume In Translation, amply demonstrates how futile an exercise it is to separate the study of language from culture. According to Lavoie, Little Hughes’ translation of Mark Twain’s Huckleberry Finn into French can be used to foreground the extent to which Twain deliberately dispensed with and paid scant regard to structures of social (racial and class) hierarchy, embedded in and reflected in the linguistic usage, so germane to his times. By introducing address terms that connote hierarchy and subordination into a text in which they are significant by their absence, Little Hughes’

188

Reviews

translation then underlines in fact the democratic vision of Twain even as it betrays it by failing to incorporate it in his translation. (After all, most readers of the translation will not have the benefit of the original to compare the translation with.) A translation, thus Lavoie asserts, can reveal the ‘‘structure and signification of the original novel’’ (88). And indeed these status-signifying markers such as ‘‘massa’’ that according to Lavoie Little Hughes introduces into the French translation of Huck Finn are linguistic markers—except that they are linguistic markers that are steeped in and that connote a particular socio-cultural set-up and worldview. So, as a Teacher of Translation who must constantly grapple with the question of what the students must learn and what the competent translator must know, I would state the following: the present division between Language and Culture in Translation Studies is both artificial and untenable. Translation Studies must embrace both language and culture if it is to be true to its mission of producing a) translations that make sense and read well and b) competent translators who will deliver the above. So, while I would not go to the extent of saying with Singh that ‘‘the heart of Translation Studies is language’’ (60), I would say that he has captured in his characteristically terse yet pithy style the relevance of language to Translation Studies: it is the ‘‘study of language mediated by practical concerns’’ (64), or language in use. While Catford (cf. Venuti 2000: 141–148) might be accused of, hence dismissed for, introducing a sterile study of language, or ‘‘shifts’’ as he called them, that says little or nothing of how language makes meaning in context, he should not be turned into the proverbial straw-man to castigate a group of linguists who advocate an ‘‘applied linguistics’’ perspective in Translation Studies. To me, applied linguistics is language in use or language as it is mediated by culture. And culture to me in this context is a particular worldview as it manifests itself in language. And as many philosophers of language have asserted, while at times it is language that determines and shapes our worldview—that is, what we can say and how we can say it—at other times, it is the worldview that determines the shape of the language. There is no natural language that I know which is not integral to, as well as closely integrated with, the aspirations and worldview of a particular culture/community. If that is the case, can a study of language ignore the study of the culture that uses that language? And this is especially the case in translation where one studies what Anthony R. Lewis calls a ‘‘communicative act’’ (23) in one language only in order to reincarnate it as a meaningful ‘‘communicative act’’ for a di¤erent community of speakers in another language. I am not however a purist. Rajendra Singh is, at least when it comes to this particular issue. To him, Translation Studies is in search of a discipline of its own. So it is important to him that Translation Studies should become

Reviews

189

a legitimate discipline by identifying and acknowledging Applied Linguistics as its proper disciplinary base (58). That is, Translation Studies should settle once and for all in its rightful disciplinary home. The study of language should be its focus, all else at best peripheral. ‘‘The problem seems to be that whereas translation has to do with texts and the languages which these texts has to follow the norms of, Translation Theory has increasingly taken on questions that deal only with the institutional matrices the product of translation has its genesis in and finds its way into’’ (59). I am not so concerned with the mixing in current Translation Studies of what Singh might call ‘‘legitimate’’ issues and concerns with their opposite. I celebrate the hybridity that characterize Translation Studies, that enables it like the Octopus to reach out its tentacles and bring within its grasp anything and everything that has even a marginal relevance to Translation Studies. To me, that mixed bag of concerns enriches, rather than dilutes, the new dispensation that goes by the name of Translation Studies for what translation requires in some ways is the non-specialist. What I mean by ‘‘non-specialist’’ here is someone who has some knowledge of almost every discipline under the sun and who will judiciously select from that knowledge to solve the translation problems that come her/his way. So I would like to resort to a biblical metaphor to conclude my reflections on the fascinating collection of essays that In Translation is. If multiple tongues (and multiple cultures?) were God’s punishment to humans for their act of presumption in coming together via one language to build the Tower of Babel to reach heaven, translation has given us the bridge with which to cross the chasms (or to scale the towers) that separate us humans one from the other. The current debate within Translation Studies, Language or Culture (or Culture or Language), should not grow into another chasm that impedes the study and transformation of the babble that surrounds us into meaningful ‘‘communicative acts,’’ which is Translation.

C

Dialogue

Language policies and lesser-known languages in India* PROBAL DASGUPTA

1. Introduction In this intervention I make a case against treating lesser-known languages as points on a scale that ranges from the neediest communities to the least needy. My argument refers to factors that become salient during the transition from a modern order of nation-states to an unsettled dynamism involving heterogeneous spaces. Under the exigencies and anxieties of globalization, the experiences become traumatic and call for urgent e¤orts to formulate and address what are experienced as vital issues on the ground. For clarity, I briefly present first the view I oppose, calling it ARSA, the Aid Recipient Spectrum Approach. I then outline the alternative I advocate, ILCEA, the Inter-Local Community Empowerment Approach. Once these have been introduced, the Language Policies in India section situates the problem with respect to India; the ARSA section discusses how the consensus both in India and elsewhere has swung towards ARSA; and the final section focuses on ILCEA, considers reasons for wishing to turn the tide and proposes ways of bringing this about. ARSA takes a cluster of aid donors as given. It seeks to draw up a list of recipients deserving appropriately targeted aid. On this view, a community deserves aid at di¤erent levels for di¤erent dimensions. If the assistance is to empower the community at the level of its language, then the targeting must be tailored to the needs associated with specific empowerment goals. These are to be measured in terms of language vitality indicators. The communities that lie the furthest from empowerment must be prioritized for maximal assistance. According to this logic, middle-ranking communities also need aid, but will have to wait in the queue, and must be persuaded not to sulk. ILCEA takes the view that two communities can deal with each other on a basis that becomes mutually empowering if the process brings out their agency in ways that respect both symmetry and di¤erence. Such a dialogue cannot be validly framed from above by a supposedly rational system

194

Probal Dasgupta

instructing each member of every dyad to behave properly. Welfare is maximized by communities in an interlocal dialogue that strengthens each community’s sense of agency. On this approach, the most natural dialogue partner for a community speaking a lesser-known language is the immediately dominant regional language constituency. When confidence comes to be seen as the main resource, economic issues appear in a di¤erent light because of a revision in the conceptualization of what shall be counted as goods. Relevant economic models must closely hug the interlocal ground and cannot be predicted from generalized patterns. It follows that there can be no single global spectrum from those most urgently deserving aid to those least urgently deserving it. Nor can one usefully visualize any single process characterizable as linguistic aid from the global power centres to the local power peripheries. On the contrary, given globalization, particular dialogic dyads become more crucial than ever before. The next section prepares the ground for a debate between ILCEA, the viewpoint I advocate, and ARSA, the prevalent viewpoint I oppose. 2. Language Policies in India A few stereotypical points are frequently made when journalists or ordinary citizens comment on how language policy in independent India has been dealing with lesser-known languages, an issue generally reduced to a tribal question. The republic’s first rulers, Bad News, inherited white colonial attitudes towards tribals. But Nehru, Good News, imported a variant of Soviet language policy into the Indian political system. The problems of all tribal languages are similar enough that the constitution’s guaranteed umbrella reservation of educational opportunities and jobs for the scheduled castes and the scheduled tribes (SCs and STs) can o¤er uniformly useful support to them all. The beliefs just listed are (I am citing yet another stereotype now, this is not yet my voice!) contested in the space of a neoliberal cynicism that has enabled the upper caste Hindu elite to detach many Dalits from the SC/ST coalition, interpellating Dalits as Hindus and convincing them and others that a technologically advanced country must drive tribals out of their traditional habitats for the public good. Now, all these stereotypes, though they are di¤erent from each other, are embedded in a general discourse about language and literacy that takes the direction of progress and the homogeneity of technology for granted. It is di‰cult to know where to begin if one wishes to move out of the non-debates of the small talk that consistently stayed at this uninformed level and never took o¤ as a public discourse. At the same time, we can hardly a¤ord the arrogant luxury of scholarly contempt directed at the thoughtless crowd and the vapid journalism it elicits. Public policy always

Language policies and lesser-known languages in India

195

has to be fashioned through a process that enables the population’s initial beliefs to collide with each other and with the contributions of experts including linguists. That some of these initial beliefs are called stereotypes by some of their critics is only a factor facilitating this arranged collision process. One idea might be to try to formulate contrary thoughts critical of the stereotypes presented. Trying this approach, let us negate the Bad News claim and say that the point of post-independence policy was not to harm tribal languages but to ensure due attention to the majority language of each region. The temporary marginalization of tribal languages was an unintended side e¤ect and did not indicate any colonial mind-set inherited by the republic’s first rulers. Negating the Good News claim, let us attribute the regional language policy not to Nehru’s fascination for the Soviet Union but to the traditional pressure of the linguistic state agenda that the Indian National Congress had been stampeded into after the resistance to Curzon’s partition of Bengal, Potti Sriramulu’s fast unto death, and a historical cascade triggered by these and related factors. Are steps like these going to move us out of the stereotypical space? The sad truth is that such moves, which are in fact sometimes made, remain equally embedded in a general discourse about language and literacy that takes the direction of progress and the homogeneity of progressinducing technology for granted. This general discourse takes spoken language to be the lowest rung of a ladder of ascent. Then there are tools. Writing is the first tool. Digitalization deploys the latest generation of tools. Societies have to climb the ladder to make it to the highest rung. Those at the top of the ladder are the best placed to help those who are still climbing. Help takes the form of transferring technology and the ability to use it. Tribal cultures have remained stuck in preliterate methods of transmitting values and norms. Linking them to the literate world is the highest priority for Indian language policy as far as lesser-known languages are concerned. One notices immediately that this general set of assumptions, summarized in the ladder metaphor, is compatible with the stereotypical proposition that in 1947 Indian power moved into the hands of brown whites who inherited the colonial habit of oppressing tribals. The same ladder metaphor, however, is equally compatible with the simple-minded reversal of the stereotype, a reversal that says a true nationalist leadership began by empowering a Telugu speaking Andhra Pradesh, a Marathi speaking Maharashtra and so forth without prejudice to the claims of tribals and other marginal communities less well-placed to benefit at once from national independence. Negating the stereotypes as they stand only has the e¤ect of producing the appearance of a debate. The ladder metaphor remains intact through this entire process.

196

Probal Dasgupta

The way out of stereotypes, then, cannot be to simply negate them, as such a move would leave the underlying assumptions undisturbed. This would also happen if we were to succumb to the temptation of focusing too hard either on the socialist model Nehru is supposed to have imported from the Soviet Union’s experience or on neo-liberal alternatives. The kinds of rethinking thrown up by India’s 1991 liberalization rest on the assumption that the market’s economic logic is systematically distinct from the excessive administration of commanding heights socialism. Real di¤erences do exist, to be sure. But even a moment’s reflection shows that the market economy shares with socialism exactly the same prejudices about the intrinsic value of technology, the identifiability of a single direction of progress, the misery of the tribals, and the designation of writing as a first tool and of digitalization as a state of the art toolkit. While socialism and liberalism may deliver di¤erently, their classical shopping lists look identical when language policy is at stake. Independent India’s language policy evolved along with its interpretation of the constitutional notions of the states and of the unitary Indian republic consisting of them. The overall agenda was focused on development, and the fact that that period’s model was socialistic quite possibly did not matter. The republic was to deliver development. In most parts of the country this delivery would occur through the intermediate organization of the state. This arrangement was viewed as politically meaningful in regions with a literate and competent elite capable of running a responsible legislature. Again, this kind of viewing seems to be done in comparable ways by socialists and liberals. Regions with this desirable trait were organized into linguistic states over the 50s and 60s. By the 70s, the dominant regional language in each state had attained the hegemony in education and cultural life that had been intended, and provincial politics rearticulated itself around this pan-Indian fact. The other side of the picture is that it was understood, from the very start, that some areas of the country lacked the type of elite the republic had faith in. The union territories were salient examples. The republic administered them directly, working for a future when prototypical linguistic states would emerge there. Some of these regions, in the northeast, were tribal. But the presence of tribals in Orissa, Andhra Pradesh, Maharashtra, West Bengal, Bihar exemplifies the impossibility of correlating this process with tribality. In the other direction, so does the union territory of Goa, Daman and Diu. Goa’s emergence as a state was connected to the process of the republic convincing itself that Konkani was a distinct language and not a dialect of Marathi. One might argue that nothing follows from Goa. The republic does continue to see Goa as a special case, given the Portuguese past, the tourism

Language policies and lesser-known languages in India

197

focus making multilingualism necessary, and other factors that are rapidly adduced whenever the issue is raised. But the number of special cases is embarrassingly large. The ideal type of a mainstream consisting of distinct monolingual elites hegemonic in prototypical states has always been a nonstarter. Punjab was such a special case that it had to split in the 60s and went into crisis in the 80s. Assam was so special that it unleashed a crisis of a di¤erent type in the same decade. The proverbial wealth of many Punjabis and the endowments for Sikh studies professorships in metropolitan countries may make it hard to see that Punjabi, compared to Hindi, is a lesser-known language. But it is widely recognized that languages with as small and underdeveloped a population as Asamiya are indeed lesserknown languages in any sense of the term. Thus, we cannot a¤ord to confine our discussion to tribal languages. However, in order to locate our issues on a map that has been shaped by the salience of tribal languages, we do need to take a look at what their prioritization has been like. The Indian republic’s explicit policy has arranged separate, protected and partial cultivation for tribal languages. The prototypical case has tribal children receiving primary schooling in their mother tongue, once the republic has sent the linguists over, produced the books and so on. The republic expects the output of primary schooling to make a gradual and pain-minimized transition to secondary education in the dominant regional language of that part of India, or, if there is a vacuum, in Hindi or English. This background is what renders intelligible the otherwise surprising fact that neither Santali nor any variant of Naga is the o‰cial language of any Indian state. The Indian republic has spent considerable money to develop certain tribal languages, giving them a priority that Asamiya or Konkani were denied. But the project was not going to be pushed to the point of using a tribal language as the medium of instruction at university departments of history, political science, or sociology. The system has taken the view that there are no resources to pursue such utopian projects and that the tribals themselves do not want developmentally irrelevant decorations. Consequently lexical development at the tertiary educational level has not been on the agenda for tribal languages. Thus on the one hand tribal languages were not going to be integrated into the world system of mutually translatable modern languages. On the other hand, the Indian republic evinced no interest in the indigenous knowledge resources embedded in the vocabulary systems of tribal languages. The fact that under pressure on this front from international initiatives a layer of such talk is now being hastily added to the discourse on tribal welfare merely betokens the ad hoc willingness of the republic’s functionaries to

198

Probal Dasgupta

say whatever is likely to fetch funding. The republic’s actual policy has been to treat tribal cultural and linguistic spaces as frozen in time and as incapable of cognitively dynamic interaction with modernity. Only tribals as individuals can be rescued from cognitive and developmental stasis, the Indian republic has believed, by weaning them from their primordial identity and inserting them into the provincial mainstream. Of course I say ‘‘provincial’’ advisedly, hoping the reader will pay some attention to the oxymoronic sound of the phrase ‘‘provincial mainstream’’. For this is indeed the fate that the republic of India has fashioned for its rescued ex-tribals. The Indian republic certainly did not intend this fate for Konkanis or Asamiyas. Goa and Assam were provinces in their own right with low populations and weak profile languages, twin misfortunes they were expected to deal with as they grew. But a vaguely marked but clearly felt delimitation between an Us and a Them left the republic feeling reasonably sure that Konkanis and Asamiyas belonged with Us who give developmental aid to the unlucky bottom rung. The republic was sure that neither Konkanis nor Asamiyas were to be equated with tribals, with Them who take such aid. I quite understand and sympathize with an Indian republic that has thought like this. It had no choice, steeped as it was in the culture of languages. The question for our times is whether an Indian democracy, engaged with a culture of conversations, will be literate enough to recognize and undo the canonization game that made the old rules of our policy what they were. 3. ARSA It is time to recognize that the game the Indian republic has been playing is a variant of the Aid Recipient Spectrum Approach, ARSA. The Indian variant breaks the set of languages up into at least two subspectra (Them tribal languages and Us others), with further nuances not specified in the summary so far. But the distinctive traits of ARSA manifest themselves with some clarity. Every language is measured in terms of a uniform notion of progress or development. Some languages, at the tribal lowest rung of the ladder, have to be left to their sentiment-only status, without any hope of serious knowledge production occurring in those languages. The welfare enterprise requires that the speakers of those languages be o¤ered a reasonable rescue as individuals. The republic meets this requirement. It gives them graded access to languages that enable them to participate in the world of industrial production and mobility, and in the world of the knowledge that corresponds to and reproduces this economy. There is salvation for the individuals. For their languages, nothing.

Language policies and lesser-known languages in India

199

The ARSA approach as I read it stems from the classical imperative of defining a possible knowledge in terms of a canon-based pedagogy. On that construal of knowledge systems, a society can sponsor a cognitive progression for its children only if they can be taught to move up an age-graded ladder of knowledgeability. Suppose the highest rung on the ladder consists of being able to read, write and understand serious texts corresponding to the industrial state of the art, in principle. The task of producing such texts is interwoven with that of educating young adults who can keep producing them. Such activities are sustainable if the pedagogic interface with their juniors and the economic interface with their industrialized superiors are both sustainable. If we take the industrial state of the art as a given, planetary index that drives the assessment of any locality’s performance, then the tribal language’s race for modernity is over before it has even started. Discouraged for valid reasons, the speakers must pin their hopes on some aspect of their lives other than the language they have been unlucky enough to be born into. This is how things are bound to look to a republic that intends to perform at levels that significant others cannot argue with. India in 1947 was such a republic and could be forgiven for not taking other approaches seriously. Today, it has started performing, and can a¤ord at last to reconsider what performance has meant, what it is going to mean, in and for India. In 1947, the republic was locked into a defiant conversation with British and American achievers, and the point was to prove that Indians could speak the language of industrial achievement. At that stage, a language was what its literary elite could make of it. This limited the number and diversity of the languages India could run its systems with. For the literary elite of each language would have to control its pedagogy; this was the shape of India’s educational options in 1947. Independent India’s budding literary elites were not capable of much more than controlling the provinces. They could run nationalist school systems, show the children how to emote as obedient citizens, and hold out the hope that some of these citizens would one day perform in the world arena through English, India’s window on the world. The rules of such emoting and such obedience were in place where nationalism had mobilized the provincial middle class. In languages where the freedom struggle had not fashioned these rules, India would not be able to invent a nationalist secondary schooling now. The republic could only o¤er a graded transition from tribality to one of the provincial modes of access to Indianness. Here is one way to unpack the logic that prevented independent Indian nationalism from developing secondary schooling, and more, in tribal languages. The main point is what this analysis implies about just why the tribal child must move from primary schooling in Santali to secondary schooling

200

Probal Dasgupta

in (say) Bangla. The logic says that this is because it is only in Bangla, not in Santali, that pre-independent India has fashioned a literary canon in which it can recognize the aspirations of national independence. That canon, in turn, represents an expressive subnational domesticity. This domesticity manages one, national, interface with the Anglophone industrial world that future Indian performance will achieve continuity with. It manages another, provincial, interface with the hinterlands (the women, the peasants, other marginals, the children) from which the future performers (prototypically the upper caste males designated for hegemony) would be selected through educational opportunity narrowing. In other words, it is the canonical uniqueness of English-carried industriality outside India that defines the nation in terms of where its future achievements are headed. Reflecting the big verticality of this relation between India and metropolitan modernity, the canonical literary languages within India construct little verticalities. Tribal languages have to fit into these structures or perish as languages, or so it appeared to independent India’s early strategists. We have to understand that the vocabulary available to them gave them little choice but to think along roughly these lines. 4. ILCEA When we have finished sympathizing with the constricted thinking of the young republic half a century ago, we can start acting our age and playing the role of a mature democracy. The analysis just given suggests that we got locked into the code approach to language precisely because of the point we were trying to prove in the face of derision from significant others, such as the ex-colonial masters and their American successors. That is also why we invested so heavily in English, to complete our accession to modernity. But by now we have made that point. Properly learnt, English itself leads us now to the next step. Through this major language we sought global outreach, the hallmark of modernity. The postmodern transition now under way means that we will now need more ‘‘inreach’’ if we wish to keep growing as a nation. If we want India to be a strong country we must see that the nation will need the muscle tone and synergy that make strength sustainable. We cannot a¤ord to leave a large part of our work force feeling sorry for itself because of the destruction of recent achievements. For globalization has left many cases of newfound linguistic pride shattered. Just as speakers of an Asamiya or a Konkani begin to get ready to join the race, we all suddenly notice that the new race is one that nobody in our neighbourhood can ever hope to win if old definitions of victory continue to prevail. If we leave those definitions intact, many of us will wallow

Language policies and lesser-known languages in India

201

in self-pity, hardly a satisfactory outcome. On this new planet of more widely distributed humility, reconfiguring possibilities of coalition and even friendship becomes a real option. I have no recipe for ensuring that the language code pride that some communities had just begun to achieve can be regained under a globalized sky. I can only suggest that we move out of the ARSA game that viewed languages as codes, into the new game, here called ILCEA, which sees them as sites of discourse. This transition does not involve abandoning the very category of assistance. All communities need to help each other visualize and manage their discourses as dynamic sites of learning by interconnected persons. Once such management leads to synergy and tangible success, all participants will feel genuinely pleased. Out of this pool of cooperating talent, new forms of pride will emerge and, with them, new conceptions of what it is for humans to be languaged beings. The ILCEA vision grows out of actual vectors of the present. Provincial Indians today are less insecure than they used to be. If a sociology teacher freely peppers her classroom performance in Bangla with English words, her Bengali students do not feel threatened, provided that the syntactic frames are in Bangla. The 1950s need for committee-created Sanskrit coinages masquerading as Bangla technical terms is no longer a socially real imperative. Secondary and tertiary education in many major Indian languages can proceed without massive corpus planning exercises. At such a moment, we can a¤ord to consider such education in minor Indian languages as well on the basis of large-scale borrowing. With this radical but evident background point established, let us imagine a concrete case now. Consider the population of a village V that speaks Santali and has local knowledge of Bhojpuri. By default, Santali is always already imbricated in the Hindi nexus and the English nexus. The Santali discourse at V, call it Santali-V, is situated in a niche of discursive ecology schematized as follows at a first approximation: >

Bhojpuri Santali-V Hindi