Demonstratives in Cross-Linguistic Perspective 9781108424288, 9781108333818, 2017061572

Citation preview

Demonstratives in Cross-Linguistic Perspective

Demonstratives play a crucial role in the acquisition and use of language. Bringing together a team of leading scholars, this detailed study, the first of its kind, explores meaning and use across 15 typologically and geographically unrelated languages to find out what cross-linguistic comparisons and generalizations can be made, and how this might challenge current theory in linguistics, psychology, anthropology and philosophy. Using a shared experimental task, rounded out with studies of natural language use, specialists in each of the languages undertook extensive fieldwork for this comparative study of semantics and usage. An introduction summarizes the shared patterns and divergences in meaning and use that emerge. stephen c. levinson is Emeritus Co-Director of the Max Planck Institute for Psycholinguistics, and Professor of Comparative Linguistics at Radboud University, Nijmegen. His research focuses on language diversity and its implications for theories of human cognition. He is the author of over 300 publications. sarah cutfield is Visiting Fellow, Linguistics, at the Australian National University. Her specialties are descriptive and typological linguistics, sociolinguistics, linguistic anthropology, language documentation, Australian Aboriginal languages, creole languages, language ideology and identity. michael j. dunn has been Professor of General Linguistics at Uppsala University since 2014. His academic background is in language description, linguistic typology and phylogenetics, and his current research focus is on the evolutionary dynamics of language change. n. j. enfield is Professor of Linguistics at the University of Sydney. He is head of a Research Excellence Initiative on The Crisis of Post-Truth Discourse. His research on language, culture, cognition and social life is based on long-term fieldwork in mainland Southeast Asia, especially Laos. se´ rgio meira is a researcher at Museu Paraense Emilio Goeldi, Belém. He specializes in the Cariban and Tupian language families of lowland South America and in the Tiriyó language in particular.

Language, culture and cognition Editor Stephen C. Levinson, Max Planck Institute for Psycholinguistics

This series looks at the role of language in human cognition – language in both its universal, psychological aspects and its variable, cultural aspects. Studies focus on the relation between semantic and conceptual categories and processes, especially as these are illuminated by cross-linguistic and cross-cultural studies, the study of language acquisition and conceptual development, and the study of the relation of speech production and comprehension to other kinds of behaviour in cultural context. Books come principally, though not exclusively, from research associated with the Max Planck Institute for Psycholinguistics in Nijmegen, and in particular the Language and Cognition Group.

1 Jan Nuyts and Eric Pederson (eds.) Language and Conceptualization 2 David McNeill (ed.) Language and Gesture 3 Melissa Bowerman and Stephen C. Levinson (eds.) Language Acquisition and Conceptual Development 4 Gunter Senft (ed.) Systems of Nominal Classification 5 Stephen C. Levinson Space in Language and Cognition 6 Stephen C. Levinson and David Wilkins (eds.) Grammars of Space 7 N. J. Enfield and Tanya Stivers (eds.) Person Reference in Interaction: Linguistic, cultural and social perspectives 8 N. J. Enfield The Anatomy of Meaning: Speech, gesture, and composite utterances 9 Giovanni Bennardo Language, Space, and Social Relationships: A foundational cultural model in Polynesia 10 Paul Kockelman Language, Culture, and Mind: Natural constructions and social kinds 11 Jürgen Bohnemeyer and Eric Pederson Event Representation in Language and Cognition 12 Jan P. de Ruiter (ed.) Questions: Formal, functional and interactional perspectives 13 Jennifer Green Drawn from the Ground: Sound, sign and inscription in Central Australian sand stories 14 Stephen C. Levinson, Sarah Cutfield, Michael J. Dunn, N. J. Enfield and Sérgio Meira (eds.) Demonstratives in Cross-Linguistic Perspective

Demonstratives in Cross-Linguistic Perspective Edited by

Stephen C. Levinson Max Planck Institute for Psycholinguistics, Nijmegen

Sarah Cutfield Australian National University, Canberra

Michael J. Dunn University of Uppsala

N. J. Enfield University of Sydney

Sérgio Meira Museu Paraense Emílio Goeldi, Belém

University Printing House, Cambridge CB2 8BS, United Kingdom One Liberty Plaza, 20th Floor, New York, NY 10006, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia 314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India 79 Anson Road, #06-04/06, Singapore 079906 Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning, and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781108424288 DOI: 10.1017/9781108333818 © Cambridge University Press 2018 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2018 Printed and bound in Great Britain by Clays Ltd, Elcograf S.p.A. A catalogue record for this publication is available from the British Library. Library of Congress Cataloging-in-Publication Data Names: Levinson, Stephen C., editor. | Cutfield, Sarah, editor. | Dunn, Michael, 1969– editor. | Enfield, N. J., 1966– editor. | Meira, Sérgio, editor. Title: Demonstratives in cross-linguistic perspective / edited by Stephen Levinson, Sarah Cutfield, Michael Dunn, N.J. Enfield, Sergio Meira. Description: Cambridge : Cambridge University Press, [2018] | Includes bibliographical references and index. Identifiers: LCCN 2017061572 | ISBN 9781108424288 (hardcover) Subjects: LCSH: Grammar, Comparative and general – Demonstratives. | Grammar, Comparative and general – Deixis. | Language and languages – Crosscultural studies. | Language and languages – Philosophy. | Semantics (Philosophy) Classification: LCC P299.D46 D56 2018 | DDC 415–dc23 LC record available at https://lccn.loc.gov/2017061572 ISBN 978-1-108-42428-8 Hardback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Contents

List of Figures List of Tables List of Contributors Preface 1 Introduction: Demonstratives: Patterns in Diversity stephen c. levinson

page vii x xiii xv 1

2 The Demonstrative Questionnaire: “THIS” and “THAT” in Comparative Perspective david p. wilkins

43

3 Lao Demonstrative Determiners Nii4 and Nan4: An Intensionally Discrete Distinction for Extensionally Analogue Space n. j. enfield

72

4 Dalabon Exophoric Uses of Demonstratives sarah cutfield

90

5 Brazilian Portuguese: Non-contrastive Exophoric Use of Demonstratives in the Spoken Language se´ rgio meira and raquel guirardello-damian

116

6 “See This Sitting One”: Demonstratives and Deictic Classifiers in Goemai birgit hellwig

134

7 Tzeltal: The Demonstrative System penelope brown and stephen c. levinson 8 Yucatec Demonstratives in Interaction: Spontaneous versus Elicited Data ju¨ rgen bohnemeyer 9 Lavukaleve: Exophoric Usage of Demonstratives angela terrill

150

176 206 v

vi

Contents

10 Tiriyó: Non-contrastive Exophoric Uses of Demonstratives se´ rgio meira

222

11 Trumai: Non-contrastive Exophoric Uses of Demonstratives raquel guirardello-damian

242

12 Saliba-Logea: Exophoric Demonstratives anna margetts

257

13 Warao Demonstratives stefanie herrmann

282

14 Chukchi: Non-contrastive Spatial Demonstrative Usage michael j. dunn

303

15 Yélî Dnye: Demonstratives in the Language of Rossel Island, Papua New Guinea stephen c. levinson 16 Tidore: Non-contrastive Demonstratives miriam van staden

318 343

17 The Jahai Multi-term Demonstrative System: What’s Spatial about It? niclas burenhult

361

Index

381

Figures

1.1 1.2 1.3 1.4

1.5 1.6 1.7

1.8

1.9 1.10

3.1

3.2

Idealized model of speaker-anchored radial spatial categories page 7 The distinct uses of demonstratives (after Levinson, 2004) 10 Cardioid shape of proximal space in some languages 20 Extensions of demonstratives in some European languages (in the scenes, S labels speaker, A addressee and X the referent where not clear) 21 Some two-term oppositions in this book 22 Unmarked versus marked oppositions in Yélî Dnye demonstratives 24 Three ways in which ‘medial’ terms may actually be unmarked terms (dotted extensions show areas where a neutral functions like a medial because it is not pre-empted by other terms) 25 Sketch of the core proximity oppositions in nominal demonstratives in the 15 languages (N = Neutral or unmarked, Far is far distal, S&A is a joint speaker-addressee anchor) 26 Sketch of Jahai notion of ‘exterior’ space 28 The nominal systems in this book laid out as a tree of increasing semantic complexity (S-Prox = speaker-based proximal term, S-Dist = speaker-based distal term, S-Far = speaker far-distal term, N = neutral term unmarked for proximity, A-Prox = addressee-based proximal, S&A-Prox/Dist = speaker and addressee based proximal/distal) 36 Two scenes with no contrast between spacing of speaker, addressee, and referent, but with contrast in the status of the interactional space. (Dotted line represents interactionally defined ‘engagement area’) 82 Hypothetical scenario, with speaker (stall attendant) and addressee (customer) on either side of a market stall table, both contained within the space defined by the market stall. Division of here-space between speaker and referent is possible, and

vii

viii

3.3

4.1 4.2

4.3 5.1 5.2 5.3 7.1 7.2 8.1 8.2 8.3 8.4 8.5 10.1 12.1 13.1 14.1 15.1 15.2

List of Figures

given accessibility to addressee of the relevant information, nan4 may be used. Scenario from video recording (Enfield, 2003a), with speaker (stall attendant) at one market table, and addressee (older woman) at another market stall some 15 metres away. While division of here-space between speaker and referent is possible, the information relevant for the speaker’s having made such a division (e.g. due to interactional dynamics) is not accessible to addressee – thus, due to the principle of ‘recipient design’ (see text), nan4 may not be used. Approximate boundaries of languages of the Top End of the Northern Territory (based on Harvey, 2009) Depiction of scene 9, when the speaker’s here-space equates to speaker-addressee shared space and includes the speaker, addressee and the referent Arrangement of speaker and two or three referents in table-top space Areas covered by esse and aquele (speaker and addressee relatively far from each other) Areas covered by esse and aquele (speaker and addressee relatively near) Areas covered by esse and aquele (speaker and addressee relatively near) Demonstratives used for three objects on a table, all within speaker’s reach Demonstrative and directionals in time reference Demonstrative scenes 13 and 16 Demonstrative scenes 9 and 12 Accessibility in the demonstrative scenes Demonstrative scene 11 Anchoring and attention calling in Yucatec spatial deictics Oppositions in Tiriyó third-person pronouns (inanimate forms only for convenience) Circle schema of results presented in Table 12.5 Scenes where no demonstrative pronoun was used The exophoric demonstratives in Chukchi along a proximal–distal cline Demonstratives used (with pronoun n:ii) for single objects on a table Demonstratives used (with pronoun n:ii) for two objects on a table

85

86 93

101 110 127 128 129 165 172 191 191 195 196 198 238 273 289 314 323 323

List of Figures

15.3 Demonstratives used (with pronoun n:ii) for three objects on a table 15.4 When addressee is opposite speaker 15.5 The three dimensions of Yélî Dnye deictic determiners 15.6 Unmarked vs marked oppositions in the speaker-oriented series 15.7 Contrastive use of demonstratives (numerals indicate order of mention) 16.1 Simplistic analysis of ta 16.2 Royal up/down on Tidore. All objects or movements located in the direction of the sultan’s palace are ‘up’ 16.3 Royal up/down on Tidore. All movements and objects towards the port to Ternate are ‘down’ 16.4 Going to the market 16.5 Tidore region 16.6 Use of demonstratives in geographical and interactional space 16.7 Re, ge and ta in interactional space 16.8 A dog trotting off into geographical space 17.1 Scenes 1, 3, 6, and 11: Speaker-anchored exterior acceptable. The added arrow indicates the required position of the referent, if different from that indicated in the questionnaire 17.2 Scenes 2, 4, 5, 10, and 18: Addressee-anchored exterior acceptable. The added arrow indicates the required position of the referent, if different from that indicated in the questionnaire

ix

323 324 325 328 335 346 347 347 349 350 354 356 359

373

374

Tables

1.1 The languages surveyed in this volume page 13 1.2 Numbers of deictic distinctions versus numbers of terms in pronominal/adnominal demonstratives in the languages in this volume 17 1.3 A typical paradigm mismatch 18 1.4 Major types of spatial distinction in the nominal demonstratives of the 15 languages, where 1 = attested 23 1.5 Semantic parameters involved in the ‘design space’ of demonstrative oppositions 35 1.6 Generalizations about demonstrative systems: Hypotheses based on this book 37 2.1 Visibility parameters 54 2.2 Three two-term (‘speaker-anchored’) demonstrative systems compared: Proximal and distal terms in English, Ewe, and Italian 56 2.3 Brazilian Portuguese 57 2.4 Turkish 58 2.5 Japanese 59 3.1 Distinguishing properties of Lao demonstratives 75 3.2 Distribution of the Lao demonstrative determiners nii 4 and nan4 according to native-speaker judgements in response to the questionnaire 78 4.1 Dalabon demonstratives used in non-contrastive exophoric reference 94 4.2 Distribution by scene (Wilkins, 1999a) of the spatially specific Dalabon demonstratives 99 4.3 Contrastive uses of Dalabon demonstratives in table-top space 111 4.4 Spatial pre-emptive heuristics in the Dalabon demonstrative paradigm 113 5.1 Written Portuguese demonstratives (after Cunha and Cintra, 1984: 328) 118

x

List of Tables

5.2 Spoken Brazilian Portuguese demonstratives (based on Pavani, 1987; Castilho, 1993) 5.3 Demonstratives in combination with place adverbs (singular forms only) 5.4 Results of the demonstrative questionnaire (total of six consulted speakers) 5.5 Proposed analysis of the demonstrative system (singular forms only) 6.1 The demonstrative word 6.2 The structure of the noun phrase 6.3 Results of the Demonstrative Questionnaire (five consulted speakers) 6.4 Classifiers and their semantics 6.5 The anaphor and the demonstratives 7.1 Basic oppositions in Tenejapan Tzeltal demonstratives 7.2 Initial and terminal deictic pairings in Tenejapan Tzeltal 7.3 Results of the exophoric elicitation task for Tzeltal 8.1 Synopsis of Yucatecan spatial indexicals (based on Hanks, 1990: 18–19) 8.2 The semantics of the space-deictic determiners and adverbs of Yucatec according to Hanks (1990) 9.1 1st/2nd person personal pronouns 9.2 3rd person demonstrative pronoun ‘he/she/it’ 9.3 3rd person demonstrative pronoun ‘the other one’. The paradigm is defective. 9.4 Demonstrative modifier ‘this’ 9.5 Possible analysis of exophoric demonstratives 9.6 Summary of results of the exophoric demonstratives questionnaire for Lavukaleve (cited in neuter sg forms) 9.7 Functions of exophoric demonstratives 10.1 Tiriyó third-person pronouns (after Meira, 1999a: 154) 10.2 Results of the demonstrative questionnaire (total of four consulted speakers) 10.3 The Tiriyó system of third-person pronouns: A revised version 11.1 Demonstrative pronouns (previous studies) 11.2 Demonstrative modifiers (previous studies) 11.3 Place adverbs and demonstratives (previous studies) 11.4 Results of the elicitation conducted with the Wilkins (1999) Demonstrative Questionnaire 11.5 Demonstrative pronouns 11.6 Demonstrative modifiers 11.7 Place adverbs

xi

120 121 124 131 136 137 140 145 148 153 160 163 180 185 208 208 209 209 212 214 220 224 227 239 244 244 245 249 253 254 254

xii

List of Tables

12.1 Three-way deictic contrast across classes of spatially deictic terms 12.2 Buhutu spatial deictics (adapted from Cooper, 1992: 96) 12.3 Three-way contrast across classes of spatial deictics 12.4 Distribution of =ne and =wa according to discourse mode 12.5 Results of the demonstrative questionnaire (seven speakers) 13.1 Barral’s person-oriented system 13.2 Romero-Figeroa’s (1997) system 13.3 Combinability of demonstrative stems 13.4 Usage of the demonstrative pronouns in the questionnaire (consulted speakers) 13.5 Metalinguistic commentary and ideal definitions given by consultants 13.6 Warao demonstrative forms and definitions 13.7 Language-specific variables for Warao demonstratives 13.8 Warao demonstratives in participant and distance anchoring 14.1 Demonstrative stems 14.2 Demonstrative stems and the corresponding demonstrative adverbs 14.3 The preferred and alternative (bracketed) demonstratives at different ranges from speaker and addressee 15.1 An (inaccurate) first approximation: The core set of spatial demonstratives 15.2 Demonstrative pronouns and corresponding adverbs, according to the first approximation in Table 15.1 15.3 Henderson’s (1995: 46) analysis of the deictics 15.4 Yélî Dnye demonstratives used in the scenes in the Wilkins (1999) Demonstrative Questionnaire 15.5 Another arrangement of Yélî Dnye demonstratives 15.6 Basic collocations with gesture 15.7 Conditions on the deictic center’s attention with respect to referent 15.8 Epistemic certainty 15.9 Adverbial (manner) uses of demonstratives 15.10 Ordered slots in the pre-verbal nucleus 16.1 Demonstratives in Tidore 16.2 Application of scenes from the Wilkins questionnaire 17.1 The Jahai demonstrative system

258 259 260 266 272 285 286 287 290 295 299 299 300 304 306 311 322 322 324 326 327 333 334 334 335 337 344 353 368

Contributors

ju¨ rgen bohnemeyer, Department of Linguistics, University at Buffalo – SUNY, USA penelope brown, Language Acquisition Group, Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands niclas burenhult, Centre for Languages and Literature, Lund University, Sweden sarah cutfield, School of Literature, Languages and Linguistics, Australian National University, Canberra michael j. dunn, Department of Linguistics & Philology, Uppsala University, Sweden n. j. enfield, Faculty of Arts and Social Sciences, Department of Linguistics, The University of Sydney, Australia raquel guirardello-damian, School of Modern Languages, University of Bristol, UK birgit hellwig, Department of Linguistics, University of Cologne, Germany stefanie herrmann, Institut für Linguistik/Germanistik (ILG), Universität Stuttgart, Germany stephen c. levinson, Language and Cognition Group, Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands anna margetts, School of Languages, Literatures, Cultures and Linguistics, Monash University, Melbourne, Australia se´ rgio meira, Museu Paraense Emílio Goeldi, Coordenadoria de Ciências Humanas (CCH), Belém, Brazil miriam van staden, Amsterdam Center for Language and Communication, University of Amsterdam, The Netherlands xiii

xiv

List of Contributors

angela terrill, Department of Cultural Anthropology and Ethnology, Uppsala University, Sweden david p. wilkins, Department of Linguistics, Australian National University, Canberra

Preface

This book has had a long incubation. On the basis of a field task designed by David Wilkins, published here as Chapter 2, successive generations of linguistic fieldworkers at the Max Planck Institute for Psycholinguistics have contributed chapters to this volume. Editorial teams have also changed, as earlier editors moved on to other jobs and projects. The first editorial team consisted of Michael Dunn and Sérgio Meira, Sara Cutfield was then brought in by Nick Enfield both to contribute and to edit the newer chapters, and finally, I myself took over to see the volume through with the help of Edith Sjoerdsma and Ludy Cilissen. I have tried to provide a worthy introduction to what is a unique collection of chapters exploring the semantic and pragmatic typology of this topic, which has played a central role in linguistic and philosophical theory, and is now being picked up as an important issue in the neurosciences. s te p h e n c . l e vi ns o n

xv

1

Introduction: Demonstratives: Patterns in Diversity Stephen C. Levinson1

This book is an in-depth look at demonstratives in 15, nearly all unrelated, languages (see Table 1.1 below). Demonstratives are, of course, the little words (or morphemes) like this and that in English, which serve, in their central function, to pick out a referent in the speech event.2 Demonstratives have played a crucial role in linguistic and philosophical thought, but monographic treatments are rare (see Hanks, 1990; Diessel, 1999), and what we know about them crosslinguistically is limited by the paucity of details found in the average grammar of a language. This volume tries to put this right, by examining demonstratives in depth across the 15 languages. A special feature of this collection of studies is that they have used as part of their analysis precisely the same field instrument, thus providing a tertium comparationis, or a grid for precise comparison, of a kind that has never been utilized before. The studies have each been conducted in the field with multiple participants, providing much richer data than is commonly found in language descriptions. This allows us to offer some generalizations about the underlying distinctions found in demonstrative systems with a new certainty. The authors have gone on to supplement this comparison point with observations of their own, derived from long-term investigation in the field. The volume offers a corrective to a large number of preconceptions found in the linguistic, anthropological, psychological and philosophical literature. This introduction has two major components. A section on preliminaries sketches the state of the art as reflected in the literature from linguistics to brain science. It draws out some of the main issues that make demonstratives so theoretically important, explicates many of the theoretical distinctions that have been made and provides some of the main results from contemporary 1

2

This introduction has benefitted substantially from elaborate notes on the chapters made by Sarah Cutfield, and from a short draft by Michael Dunn. I am grateful to Gunter Senft, Harald Hammarström, David Peeters, Niclas Burenhult, Penelope Brown and others for comments on earlier drafts. In this introduction the term demonstrative refers in the first instance to the pronominal or adnominal forms, while adverbs like here and there are termed demonstrative adverbs. The demonstrative paradigm of any particular language may include additional elements like anaphoric terms.

1

2

Stephen C. Levinson

research. The second section of the introduction turns more centrally to the business of this volume, explaining the central task used to structure the studies. A substantial subsection draws together the findings from the separate studies and offers some strong generalizations that have emerged from this collective exercise. 1

Preliminaries: This Volume in Context

1.1

The Importance of Demonstratives

Demonstratives like this and that are within the top 20 most frequent words in English and are among the most deeply conserved and ancient words in languages (Pagel et al., 2013); indeed, their etymology can rarely be traced (Diessel, 1999, but see Hellwig, this volume; Rosés Labrada, 2015). Demonstratives are also among the earliest words learned by children, and often the first closed-class opposition (Clark, 1978; Tanz, 1980). In acquisition, they follow the earlier use of pointing with which they become associated, pointing marking the initiation of systematic intentional and referential communication, with shared attention focused on a third entity (Tomasello et al., 2005; Liszkowski et al., 2012). The association with pointing makes crystal clear that demonstratives have as one of their most important functions a focusing of joint attention on an object in the environment. This makes them a kind of ideal model system for the study of language use: a single word and gesture can function as a full referring act, with all the complexities of the joint attention, common ground, multimodality and pragmatic integration involved in more complex utterances (Clark et al., 1983). Demonstratives and pointing may also be thought of as an ancient substrate of language, closely allied with animal communication systems which are always concerned with the here and now, showing little of the ‘displacement’ typical of human language (despite which they figure only rarely in discussions of language origins). Demonstratives have also played a critical role in our theory of language. They form part of the deictic field, that is, those expressions in a language that are built for contextual resolution, by reference to the situation of speaking. The deictic field is often divided into the semantic domains of space, time, person, discourse and social dimensions, covering not only demonstratives but also temporal expressions like now and tense, personal pronouns like I and you, anaphoric or related expressions, and honorifics (Levinson, 2004). Each of these domains makes essential reference to the context of utterance – one can’t resolve Now you do this! without seeing who was addressed when with what demonstration. Deixis poses fundamental problems for understanding the semantics of natural languages and consequently has greatly exercised

Demonstratives: Patterns in Diversity

3

philosophers of language. The tendency has been to conceive of semantics as specifying the states of affairs that match descriptions, with the contribution each linguistic expression makes to that specification being systemically explored. But deictic elements like demonstratives clearly do not play a direct role in that specification – instead, they point to dimensions of the context of utterance which have to be imported to complete the description. A great deal of thought in philosophy and formal semantics has gone into conceptualizing the mechanisms involved, with meaning and reference relativized to context (see, e.g., Braun, 2015 for review), so that You made this can be explicated as, say, ‘Anne E. Smith made the indicated painting on or before 7th May 2017’. From the point of view of a cognitive theory of language, this is problematic, because it may not conform to the thought the utterer had in mind when he spoke (perhaps he thought the painting was a photograph and that the maker was called Alice). This dilemma has never been satisfactorily resolved in the theory of semantics (Levinson, 2004). In addition, the paradoxes of selfreference (as in This statement is false) have teased philosophers for two millennia. Deixis is what makes languages special and especially complicated compared to the artificial languages of logic and computation because the incorporation of contextual factors into referential language produces a hybrid system which resists any easy theoretical reduction. Demonstratives have always been taken to be the prototype elements of deixis (the term comes from the Greek for ‘pointing’) and have played a key role in semantic theory. It is therefore surprising how relatively little direct exploration of demonstrative semantics and use has been undertaken in any specific language (with exceptions noted below; see also Weissenborn and Klein, 1982; Senft, 1997; 2004; Levinson and Wilkins, 2006). Another gap has been systematic comparison of demonstrative systems across languages, taking into account the details of how the semantics and usage vary across languages. This volume aims to at least partially fill these gaps. How does one recognize a demonstrative in an unfamiliar language? Largely by function: demonstratives are specialized to refer by exploiting aspects of the context of the speech event to which they direct attention, and hence they often expect a corresponding gesture. It is actually quite difficult to specify the function exactly. Diessel (2006) suggests the function can be specified as (a) indicating the location of a referent relative to the deictic centre, and (b) coordinating speaker and addressee attention on that referent. The two conditions are inter-related: if either (a) or (b) is successful, the other condition is likely to be met, and thus there are two routes to referent identification. Note that although we have a preconceived idea of the core function of demonstratives, namely drawing attention with a gesture to a physical referent in the vicinity of the speech event, the function needs to be generalized to nonmaterial referents and non-gestural usages as in this city or this strange smell,

4

Stephen C. Levinson

where the demonstrative ‘locates’ the referent in the presumptive common ground of mutually assumed entities. Or if you and I are at a restaurant table with a candle in front of us, this candle coordinates attention on a mutually manifest entity (Smith, 1982) without the need for gesture or location information. In many languages there is rich information about the qualities of the referent (e.g. in Goemai, this volume, about the position and support of a referent) which may obviate locational specification. And in many languages, as reviewed below, there may be special forms reserved not for drawing attention to a referent, but for exploiting mutual awareness of it. Even if we take the key prototype of demonstrative function as an expression with a gesture drawing attention to something in the environment, that alone will not suffice to individuate demonstratives, for it turns out that many definite referring expressions can be used in just such a way (as in What a peculiar man said pointing at a man). So demonstratives are also identified by being closedclass items that form small contrastive sets, with distinct properties of morphological combination and distribution in a clause. The form and syntax of demonstratives are of course language specific, and many details of this sort will be found within this book. Thus, it is the combination of distinctive function and a relatively small closed-class set of terms that enables recognition of a demonstrative category cross-linguistically. 1.2

The Form and Syntax of Demonstratives

The place of demonstratives in the pronominal field of a language varies according to whether a language has third person pronouns proper or just uses demonstratives instead, and in the latter case according to whether there are also dedicated anaphoric pronouns (not aligning with the personal pronouns), and so whether demonstratives are also employed for that function. Bhat’s (2004; 2013) typological survey of pronouns suggests that less than half of all languages have third person pronouns clearly modelled on first and second pronouns; and a third of languages base their third person (anaphoric) uses on demonstratives, sometimes derivationally, or by using a specific demonstrative (often the remote one) as an anaphoric pronoun. Diessel (1999) provides a useful overview of the morphology and syntax of demonstratives based on a balanced sample of 85 languages. Demonstratives may be free or bound forms or clitics, they may inflect for case, or agree in gender, with a tendency for adverbial forms to be less bound and less inflected. The traditional division of demonstratives into three subclasses, demonstrative pronouns like this, demonstrative adjectives or adnominals as in this book, and demonstrative adverbs like here (location) or thus (manner) generally holds up cross-linguistically, to which a subclass of presentationals (as in French voilá!) is often added. Diessel (1999) also makes a case for a rarer additional category

Demonstratives: Patterns in Diversity

5

of ‘demonstrative identifiers’ occurring in minimal clauses serving to introduce a referent (glossing, for example, as ‘This is . . . ’). In addition, there are verbs of ‘doing like this’ in many languages (Guérin, 2015). These different subclasses can be distinguished on distributional grounds, pronominals constituting an NP on their own, while adnominal demonstratives combine with a nominal phrase to constitute an NP. In about 70 per cent of languages, the pronominal and adnominal stems are the same in form (Diessel, 2013a), but in the rest they differ. Typically, pronominal and adnominal forms will differ in root form, in inflectional possibilities or syntactic distribution. Nevertheless, in some generative accounts they are both conceived to be determiners regardless – intransitive determiners in the case of demonstrative pronouns, and transitive ones in the case of adnominal ones (Abney, 1987; see Diessel, 1999: 62–71 for discussion). This raises the question as to exactly how demonstratives are to be distinguished from definite articles. Lyons (1977) has suggested that definite articles are just demonstratives unmarked for spatial distinctions, for example. But as we will see, many of the systems described in this book have neutral or unmarked demonstratives which contrast strongly in form and function with definite articles in the language in question (see, e.g., the chapter on Tzeltal). Once again, the crucial difference is semantic – the distinction between an instruction to find the referent in the context of the speech event (demonstratives) versus an instruction to find it in the universe of discourse (definite articles), a distinction often reflected in form and always in function. The demonstratives surveyed in this book have quite different formal properties. Some are affixes as in Chukchi, which, when attached to a pronominal root, make a demonstrative pronoun, and to an adverbial or place root or simply a locative case, an adverbial. Others, as in Tzeltal and Yucatec, are circumclitics with very complex co-occurrence with stems. Some of the demonstrative roots are adverbs, which then derive pronouns, as in Jahai. Many demonstrative adnominals inflect for case, number, gender, animacy and the like, and a wide range of such types can be found in this volume. 1.3

The Semantics of Demonstratives

We have pointed out that deictic expressions in general get their interpretations from the context of the speech event – I refers to the current speaker, now to an interval including the time of speaking and here to a location including the place of speaking (Bühler’s [1934] 1982 origo). In that respect such expressions are variables waiting for further specification from the environment – they are themselves referentially under-specified. The same holds for demonstratives: this or that are instructions to find the referent in the context but give little clue about how to do this – there may be some spatial opposition (a point we will return to), and obviously identification will be

6

Stephen C. Levinson

much helped by a pointing gesture (and occasionally, in languages like Goemai, this volume, where demonstratives contain classifiers based on the referent’s specific properties). In general, then, demonstratives, as with other deictics, work by being semantically general to a point that they invite the recipient to use contextual clues to find a definite interpretation. In I’ve hurt this finger, it is the vacuity of this that directs visual attention to the speaker’s hands. Similarly, This smells bad may refer to whatever the speaker is holding up to her nose, or to the room we just walked into, or whatever is plausible in context. The kind of semantics built into demonstratives is therefore necessarily shallow. How then is the referent recognized by the recipient? Quite largely just because a demonstrative by convention indicates that the speaker warrants that the addressee can find the referent in the context, given whatever semantic properties the particular demonstrative requires the referent to meet and other signals like gaze and gesture. How should the addressee find the referent, given the relative semantic vacuity of the demonstrative? In just the same way that Schelling Games are resolved (Schelling, 1960), by working out what the speaker thinks that the addressee thinks the speaker supposes to be the salient object of attention in the domain (see, e.g., Clark, 1996). This reflexive reasoning is reflected in the activation of the frontal lobes and the attentional and ‘theory of mind’ neural networks during demonstrat§ive use (Peeters, Chu, et al., 2015). But important clues to the identity of the referent lie not only in the demonstrative chosen but also in the contrastive items not selected, and here the principles of semantic and pragmatic oppositions become salient: if that is used, the implication is that some crucial properties for the use of this did not obtain. Those properties may be built into the demonstrative semantics (e.g. this may specify spatial proximity), or they may come about by pragmatic obviation (see section 3 below): for example, that may not actually specify spatial distance but be neutral for distance, but because this was avoided (which may imply spatial proximity), that will pick up the complement of possible referents. In this sort of way, understanding the precise semantics of demonstratives is important for understanding how they function. Where (as mostly) there are contrastive items, these must somehow divide the space of possible referents. One obvious way of doing this is by spatial zones, e.g. distinct radii around the speaker, but spatial zones can also be specified around the addressee, or around the joint space occupied by speaker and addressee. The following figure shows an idealized model of radial categories around the speaker (here the deictic origo, Bühler (1982 [1934])), a model that seems to be presumed in many grammatical descriptions. This volume throws considerable doubt on whether any language actually has a system like this.

Demonstratives: Patterns in Diversity

7

Here There Yonder

Figure 1.1 Idealized model of speaker-anchored radial spatial categories

In addition to horizontal proximity, more rarely the vertical dimension may be used to carve the search space into higher versus lower zones. These will give restricted search domains for the referent. Less obvious ways of narrowing the search domain are also possible. For example, a contrast can be made between things that are already in our joint attention, as judged by gaze, for example, and things that are now in my attention but not yet in yours. Such systems have been described for Japanese and Turkish (see, e.g., Özyürek, 1998; Küntay and Özyürek, 2002, 2006). Another distinction which occurs is between referents visible to the speaker or recipient or both versus those obscured. Many systems sometimes also specify properties of the referent, such as gender or animacy. Building on earlier work by Anderson and Keenan (1985), Fillmore (1997) and others, Diessel (1999) provides a useful overview of the kinds of contrasts that have been reported in grammars of 85 languages that span many major language families. Hanks (2009) provides a more systematic review of the underlying semantical concepts, noting that systems elaborate on variants in the origo or ‘indexical ground’, variants in the mode of access to the referent and various properties of the referent itself. This book supplements earlier work by going into much deeper detail on the actual nature of the contrasts encoded, albeit in a relatively small sample and with restricted methods. Recently, there has been extensive debate on what might be called the ‘spatial bias’ in the description of demonstrative systems (Hanks, 2005; 2011). This (alleged or real) bias is the presumption that spatial distinctions, usually in terms of distance from the speaker, form the primary semantic axis of contrast between demonstrative items. This bias is reflected in descriptive grammars, where without much investigation demonstratives are labelled as ‘proximal’, ‘distal’, ‘medial’ or the like. Earlier surveys and typologies have reflected this bias in the grammatical descriptions (Halliday and Hasan, 1976; Lyons, 1977; Anderson and Keenan, 1985; Diessel, 1999; Dixon, 2003). Hanks (1990; 1992; 2011) points to much more subtle interactional factors

8

Stephen C. Levinson

lying behind some systems, and rather than thinking of the distinctions being organized as constraints on search domains, he emphasizes modes of conceptual access to referents and ways of directing attention – the different modes of perception (vision, audition, etc.), inference and retrieval from memory or discourse. On this analysis, the deictic field naturally extends to anaphora and textual reference (or to endophora as well as exophora), and to an ethnographic understanding of what makes referents salient in a context (Hanks, 1990). Hanks (2011) also notes that the spatial bias has tended to be associated with an egocentric or speaker bias, and that the addressee’s access to the referent can be at least as important – the one must signal but the other must grasp, and some demonstrative systems clearly utilize a distinction about where the addressee’s attention currently is. Hanks argues that the origo or deictic centre, even in what are described as speaker-centric systems, may be much more open, in fact unmarked, and may routinely be addressee transposed, or inclusive of both speaker and addressee (experimental work on Spanish using similar methods to those pioneered in this book substantiates this: Coventry et al., 2008, and Jungbluth, 2003; see also Peeters, Hagoort and Özyürek, 2015 on Dutch). Hanks’ (1990; 2005) work on Yucatec Maya, based on interactional observation, makes a compelling case for dethroning spatial distinctions (but see Bohnemeyer, this volume, on the same language). There is no doubt that an ethnographically grounded interactional perspective, sensitive to social boundaries, is essential to understanding tokens of usage. But Enfield (2003, and this volume), using interactional records, shows that in the case of Lao the same methods substantiate a spatial analysis, albeit a surprising one, in which a spatial sphere of activity or attention is crucial for a distal demonstrative, while the contrastive alternate is not a proximal demonstrative but one unmarked for location. In general, the degree to which spatial coding is crucial seems to be a language-specific matter. But what is clear is that the spatial spheres involved will always be pragmatically elastic. Here, recent work in psychology and the neurosciences is pertinent. Kemmerer (1999) carefully considered the possibility that the proximal/distal distinction so often reported in descriptions of demonstratives maps onto a neurologically determined distinction between peripersonal space and extrapersonal space (a distinction reflected in patients with spatial neglect) but rejects it on the basis that demonstrative use is far more flexible, and many demonstrative systems seem to make a three-way distinction. Coventry et al. (2008), however, show that in both English (a two-term system) and Spanish (a three-term system) usage is indeed sensitive to the peripersonal domain (roughly one’s sphere of reach), and moreover, when one extends the peripersonal reach by giving participants a stick, the proximal domain increases accordingly! In English (but not in

Demonstratives: Patterns in Diversity

9

Dutch, see Peeters, Hagoort and Özyürek, 2015) the association of this with proximity of referent to speaker is strong enough under conditions of joint attention to elicit an N400 (a neural marker of integration difficulties during interpretation) where there is a spatial incongruency (Stevens and Zhang, 2013). Many papers in this volume report a specific notion of proximity which seems to coincide with the reaching zone, and within which proximal demonstratives may be obligatory (see discussion below and Peeters et al., 2014). And much recent experimentation supports a deep conceptual connection between spatial cognition and demonstratives (see, e.g., Coventry et al., 2014; Bonfiglioli et al., 2009). The discussion of the importance of spatial distinctions is, then, by no means over, and the chapters in this book continue the discussion. The number of distinctions made in a demonstrative system has often been used as an organizational framework for description and typology, especially for demonstrative pronouns and adnominals (Frei, 1944; Anderson and Keenan, 1985; Diessel, 1999; Dixon, 2003). Although Anderson and Keenan (1985) describe a number of systems as having four or more degrees of spatial distance (up to seven in the case of Malagasy), Fillmore (1982) and Diessel (1999) are rightly sceptical. Nevertheless, the World Atlas of Language Structures provides some frequencies for the number of spatial distinctions in adnominals, with c. 5 per cent of languages having four or more such distinctions (Diessel, 2013a). Hanks (2011) notes that some complex systems resist a simple radial treatment in distance from ego, requiring instead, for example, notions of laterality or exteriority (areas outside the primary space, as in West Greenlandic; see also Jahai, this volume), or distinct indexical grounds or origos, or the kind of interactional factors mentioned above. Whatever the basis of the distinctions, it is important to note that demonstrative pronouns, adnominals and adverbs may make different numbers of cuts (Hanks, 2011). English makes a binary opposition across all three grammatical sub-domains, but many languages have richer distinctions, whereas Tongan distinguishes four demonstratives in the pronouns and only two in the adnominals (Diessel, 2013a). Where the number of distinctions is the same but the forms are different across all these grammatical classes, we have a special type of lexical organization, where a parallel series of semantic oppositions is pursued across a number of form classes – a semantic template or ‘semplate’ (Levinson and Burenhult, 2009). All these paradigm patterns are of considerable interest, and they appear to be very stable over long periods of time, offering a window on long-range language relationships (ongoing work by Nick Evans and Simon Greenhill). We return below to the semantics of demonstratives in the light of the chapters in this volume.

10

Stephen C. Levinson

1.4

The Uses of Demonstratives

The analysis of demonstratives is much complicated by the fact that they tend to get used for many different functions, some beyond strictly deictic uses, such as tracking referents in discourse. The following diagram (from Levinson, 2004) displays a complex taxonomy of different uses. A first cut can be made between the deictic and non-deictic uses, of which anaphora is the most important: that man has a different role when used to point someone out than it has when embedded in a text like I was introduced to Mr Little: That man was to have a huge influence on my fortunes. The distinction is usually made more coarsely in terms of exophoric (external to the text) versus endophoric (textinternal) uses, but, as Fillmore (1997) pointed out, one needs to distinguish anaphora, where a term simply picks up the same reference as a prior term, from discourse deixis, where a term refers to a chunk of discourse itself, as in Bloop! It sounded like that. There are additional non-deictic uses like the empathetic that goddamn son of a bitch (see the chapter on Dalabon, this volume, and Naruoka, 2006) or the recognitional uses (Himmelmann, 1996) as in Do you remember that wonderful holiday in Morocco? Within the exophoric uses, one needs to distinguish those expressions (gestural uses) that require a gesture (as in This eye hurts) from those that don’t (symbolic uses as in This room is beautiful). Notice that, contrary to some remarks in the literature (e.g., Diessel, 2014), many uses of demonstratives do not require gestures, and much of this book is concerned with those uses. Where gestures occur, the kind of gesture required may vary, from a head nod or lip point, to a marked gaze, to a manual point, or to a demonstration (This finger hurts). The shape of the hand and its orientation can also be Deictic

Exophoric Gestural Contrastive Non-contrastive Symbolic Transposed Discourse deictic

Non-deictic

Anaphoric Anaphoric Cataphoric Empathetic Recognitional

Figure 1.2 The distinct uses of demonstratives (after Levinson, 2004)

Demonstratives: Patterns in Diversity

11

distinctive – The motion in He went that way is more likely to be accompanied by a lax hand than an index finger point. Contrastive uses of demonstratives are also distinctive, as in This mug is larger than that one, versus non-contrastive uses It’s hard to read in this light (see section 2.2.2.3 below). One celebrated property of deixis is its ability to be transposed (Bühler, [1934] 1982), especially in a narrative context, to some other time and place, as in This was the book he needed, John realized, where the origo is not in the situation of utterance but transferred to the narrative event (see also Fillmore, 1997). Many uses of demonstratives are transposed, and part of the uncertainty of analysis may lie in whether a deictic ground (origo or anchor) is basically speaker-centric and transposed to the addressee (This is your glass pointing at the glass nearer to you), or whether it includes both perspectives to start with. Another kind of usage distinction cross-cutting the exophoric uses above is a distinction between a referent already in the addressee’s attention versus a referent that is not yet so. As mentioned at the outset, a key function of demonstratives is to draw the addressee’s attention to an object or event in the immediate environment. Once attention is drawn further reference can often be made with a demonstrative, but sometimes of different form, as is required in Turkish (Küntay and Özyürek, 2006) and some of the languages discussed in this volume (e.g. Jahai, Yucatec, Tzeltal). English demonstrative pronouns like this and that can occur across all these distinctive usage types. But many languages have forms specially reserved for some of these functions. For example, many languages have special anaphoric pronouns of distinctive form. Some demonstratives are specialized to gestural uses, as in yay big in American English, or to manner adverbials glossing ‘thus, (do) like this’. These distinctions are important for the proper study of demonstratives across languages (König, 2012; Umbach and Gust, 2014; van der Auwera and Sahoo, 2015). It is generally agreed that the exophoric uses (and specifically the gestural ones) are primary and prototypical. The reasoning for this is that, first, these are the earliest learned forms and usages in language acquisition. Second, these uses are probably the most frequent in daily interaction (although not of course in text corpora) and tend to be the unmarked usages, the default interpretations. Third, they would seem to be diachronic sources for other grammatical items, and indeed for the anaphoric uses of the same forms. This book is focused on these exophoric uses. 2

The Current Volume

This volume differs from earlier typological surveys of demonstratives in that it is not based on extracting cross-linguistic generalizations from existing grammars or descriptions. It aims instead for a deeper understanding of meaning and usage in a small but widely dispersed sample of languages, so probing for some

12

Stephen C. Levinson

of the key parameters of world-wide variation (a similar but European-based survey can be found in Da Milano, 2005). As already mentioned, the chapters in this volume have a common framework, namely a uniform task that was conducted in the field by each of the contributing authors. The task was designed to get beyond the simple and usually inaccurate ‘proximal versus distal’ kind of accounts of demonstratives. It is also designed to resolve the kinds of controversies about the correct analysis of systems that can be found in the literature, as, for example, the treatment of Turkish as having a speakeranchored proximal versus distal (Kornfilt, 1997) or a speaker-anchored versus addressee-anchored system (Lyons, 1977) or one primarily concerned with attentional factors (Küntay and Özyürek, 2002). The chapters review the nature of demonstratives in 15 languages, drawn from 14 language families or isolates from all the major continents, and many different kinds of social and subsistence systems (there have been speculations that demonstrative systems might be linked to such social variables; see Perkins, 1992). This is a well-dispersed sample, which should reveal any clear universal structure, while indicating directions of linguistic variation. Table 1.1 provides a summary of the languages, in the order in which they are described in this book. The following sections of the introduction describe the task and explore some of the main results that have emerged from the comparative exercise. 2.1

The Task

All the authors in this volume took a particular elicitation instrument to the field designed for the purpose, David Wilkins’ (1999a) ‘Demonstrative questionnaire: “THIS” and “THAT” in comparative perspective’ (included here as Chapter 2). As the title makes clear, this was quite restricted in scope, aimed at the demonstrative pronouns and adnominal demonstratives in particular, although supplemented by other instruments (particularly ‘Eliciting contrastive use of demonstratives for objects within close personal space’ of the same date (1999b)). The title is, however, a bit misleading, as the instrument is not a straightforward questionnaire; rather, it describes a task to be carried out in elicitation sessions with (preferably) five or more native speakers in their natural ecology, where scenes are to be enacted or played out in local settings, improvising the relevant interactions. Different scenes are described (with a numbered sketch as a mnemonic) with different locations of speaker, a single addressee and a referent, and the fieldworker attempts to re-enact the scene with native speakers.3 English sentential frames were given as 3

A similar, and derivative, questionnaire can be found in Da Milano (2005), where the emphasis is on different sentence and speech act frames; it was used in a European survey, where translational equivalents are easier to find.

Demonstratives: Patterns in Diversity

13

Table 1.1 The languages surveyed in this volume Language name

Glottolog reference

Language family

Location

Subsistence type

Researcher

Lao

laoo1244

Tai–Kadai

Laos

Nick Enfield

Dalabon

ngal1292

Gunwinyguan

Brazilian Portuguese

braz1247

IndoEuropean

Arnhem Land, Australia Brazil

Rice growing, market economy Hunter-gatherer

Goemai Tzeltal

goem1240 tzel1254

Afro-Asiatic Mayan

Nigeria Mexico

Horticulturalists Slash and burn cultivators, market economy

Yucatec Maya

yuca1254

Mayan

Mexico

Lavukaleve Tiriyó

Lavu1241 trio1238

Isolate Cariban

Trumai

trum1247

Isolate

Solomons Brazil, Surinam Xingu, Brazil

Slash and burn cultivators, market economy Horticulturalists Horticulturalists

Saliba

sali1295

Austronesian

Warao

Wara1303

isolate

Chukchi

chuk1273

Yélî Dnye

yele1255

ChukotkoKamchatkan Isolate

Tidore

tido1248

Jahai

jeha1242

North Halmaheran (West Papuan) Austro-Asiatic

Papua New Guinea Venezuela, Guyana S Siberia, Russia Papua New Guinea Tidore island, Indonesia Malaysia

Urban and rural industrial state

Horticulturalists

Fishers, horticulturalists Fishers, horticulturalists Hunter-gatherers, reindeer herders Fishers, horticulturalists Horticulturalists, traders Hunter-gatherers

Sarah Cutfield

Sérgio Meira & Raquel GuirardelloDamian Birgit Hellwig Penelope Brown, Stephen Levinson Jürgen Bohnemeyer Angela Terrill Sérgio Meira Raquel GuirardelloDamian Anna Margetts Stefanie Herrmann Michael Dunn Stephen Levinson Miriam van Staden Niclas Burenhult

examples to the fieldworker, but this was not a translation exercise – the aim was to contrive a situation in which the native speaker could role-play and volunteer the relevant demonstrative. Investigators were also encouraged to keep a lookout for naturally occurring instantiations of the same or similar arrangements. Scene 6, for example, reads:

14

Stephen C. Levinson

The referent is just beside speaker (within easy reach), on side away from addressee: The object is difficult, if not impossible for addressee to see. ‘I’ve just finished reading _ book.’ ‘Do you want to borrow __book?’ • • • •

Does it make a difference if Addr knows the object is there vs doesn’t know? Does it make a difference if object has been mentioned before? Must the speaker point? What if object was more visible?

Each of the 25 scenes thus invokes many parameters, attempting to control not just spatial parameters but also the nature of the setting (bounded or unbounded), the relative locations of speaker, addressee and referent and sometimes a bystander, the current attentional focus of speaker and addressee, whether the referent has been mentioned earlier in the interaction, the presence or absence of accompanying gesture and so forth. The scenes vary the distances between speaker, addressee and referent on an implicit seven-point scale (from speaker’s body part, referent touching body part, reaching distance, a couple of steps away, tens versus hundreds of metres away, to horizon). The scenes are not presented pictorially to consultants, the figures being merely an aid for the fieldworker; rather, as mentioned, they are acted out. The fieldworker therefore has to ‘localize’ the scenes, finding a local instantiation that closely matches the described situation in its pertinent aspects and attempting to act it out on a corresponding scale. By getting multiple consultants to do the same task, one is able to factor out personal differences, or differences in construal of the scene that might otherwise be misleading. The task was particularly focused on exploring different indexical grounds (e.g. speaker or addressee as origo), possible distance distinctions, the role of visibility or access to the referent, together with the role of gesture and attention. To be precise, the kinds of demonstrative pronoun use in focus here are exophoric (+/− gestural), non-contrastive uses (the contrastive use being explored in the other instrument just mentioned). The task was reasonably demanding to run, as it involved putting participants in different locations, sometimes at rather great remove from speaker or referent, and could best be done with an assistant, taking some hours to run through. Such a task is of course limited, but it in fact yields information of a kind difficult to get in any other way. Missing from the task, for example, was any probing for ‘such’ or ‘like this’ demonstratives (König, 2012). Fieldworkers were urged to amplify the information by using additional tasks, noting informal usage around them and examining videotapes of natural interaction – as Hanks (2011), Jungbluth (2003) and Enfield (2003) have emphasized, natural interactional usage is critical to understanding these systems (and in fact information on ‘like this’ demonstratives will be found in, for example, the Tzeltal chapter). But one excellent payoff from administering the task is much

Demonstratives: Patterns in Diversity

15

greater clarity about the relevant demonstrative systems than in preceding accounts. For example, the Venezuelan indigenous language Warao, described by Herrmann in this volume, has previously been described as a personoriented system, or as a two-way speaker-proximal versus distal system, while in fact it turns out to basically introduce a three-way contrast within speaker’s reach, outside speaker’s reach and way beyond. In the case of Yucatec, the account here by Bohnemeyer based on the task yields a different account from the well-known analysis by Hanks (1990, 2005) based on ethnographic observation, specifically showing that an addressee-anchoring is doubtful. In a number of other languages, the authors themselves had held different views before administering the task (see, e.g., the chapters on Trumai, Lavukaleve, Jahai, Tidore or Tzeltal). The task was constructed subsequent to the publication of Hanks’ (1990) interactional account of Yucatec demonstratives based on dense ethnographic observation, and attempted to capture some of the variables he had noted. However, interactional analyses yield complementary data, and it is interesting to compare descriptions obtained by both means: sometimes the accounts converge (cf. Enfield, 2003 and this volume), but sometimes they diverge (cf. the accounts of Yucatec given by Hanks, 1990, and Bohnemeyer, 2012 and this volume). 2.2

Some Results

The first impression of the demonstrative systems here sampled from the world’s languages is one of overwhelming diversity. They differ in their formal exponents – whether they are affixes, clitics or free forms, whether they are fundamentally nominal, adverbial or otherwise. They differ in whether they form a neat paradigm, for example whether the demonstrative pronouns make the same number and nature of distinctions as the adverbs, or otherwise. They differ in their semantics – whether, for example, they are centred on speaker, addressee or both, or whether they delimit definitive spatial extensions or something more complex. They differ in their range of usage, for example whether the same form is used for exophoric and anaphoric reference or whether dedicated anaphoric forms form part of the paradigm; or whether the forms require a gesture or one is merely optional; or whether they presume shared attention on a referent or create it. What, however, they share in common is a core function – namely the identification of a referent (or place) in the shared environment of the speech event by appeal to a Schelling-like presumption that recognition can be achieved (see Clark, 1996: 62f.), aided by any gestural (and gaze) indications and semantic specification inherent in the forms.

16

Stephen C. Levinson

2.2.1

Form and Function

Since Frei (1944), the typology of demonstratives has often been laid out in terms of the number of opposing terms employed – thus two-term, three-term and multi-term systems (see, e.g. Anderson and Keenan, 1985; Diessel, 1999). But this is misleading for a number of reasons. First, there is confusion about whether distinctions made in the adverbs should enter the count – Diessel (1999) considers that they should and so denies the existence of one-term demonstrative systems, even though his native German has only one clear demonstrative pronoun. Incidentally, in this introduction we will use the term demonstrative to refer to pronominal or adnominal forms, specifying ‘demonstrative adverb’ when that is what is meant. Second, there is unclarity about whether the count is the number of terms in the paradigm (and if so, whether it includes or excludes dedicated anaphorics), or the number of underlying semantic distinctions. For example, Diessel (2013b) counts what he takes to be just the semantically encoded distance distinctions in 234 languages and finds that 54 per cent of languages make two such distinctions, 38 per cent make three such distinctions, 3 per cent make four distinctions and 2 per cent make five distinctions. This information is based on grammars, which tend to use a ‘proximal’/‘distal’ terminology by default rather than by careful examination – as we have seen, authors of grammars who have used the demonstrative task often find themselves to have been wrong. The semantic distinctions of relevance may be restricted to spatial distance only or include additional deictic factors (e.g. attention, visibility, direction) or include properties of the referent (e.g. number/gender/animacy) and more exotic distinctions. For example, Goemai is only a two-term system if one abstracts out the deictic prefixes from the pronominal forms – if not, it is a system with 36 pronominal forms, made up of combinations of the nine positional classifier roots, two numbers and the two demonstrative forms. Clearly, for languages where the demonstratives are primarily affixes or clitics (e.g. Goemai, Tzeltal, Yucatec in this book), the forms will tend to multiply according to which items they are attached to in order to form demonstrative pronouns. In many languages, e.g. Brazilian Portuguese, pronominals typically occur with adverbials (which in this case make more deictic distinctions than the pronominals) on the pattern of ‘this here’ – there are eight such phrasal combinations, plus six independent forms, so 14 in total (see also Jungbluth and Da Milano, 2015). Table 1.2 shows, for the languages in this book, how these different ways of counting do not align. The numbers should be taken with caution, since what is a ‘deictic feature’ is contestable, and not all authors spell out all the possibilities, concentrating on forms central to their analyses. Dedicated anaphoric terms within the demonstrative paradigm are here treated separately, and

Demonstratives: Patterns in Diversity

17

Table 1.2 Numbers of deictic distinctions versus numbers of terms in pronominal/adnominal demonstratives in the languages in this volume

Language

Main deictic semantic distinctions in nominals

Distinct demonstrative pronominal forms

Non-deictic semantic distinctions in pronominals

Lao Goemai

2 2

2 36

Yucatec Tzeltal Warao

2 3 3

over 18 over 22 11

Brazilian Portuguese

3

Saliba Trumai

3 3

6 (14 including Adverbial combinations) 3 18

0 Number (2), Posture (9) Various Various Location, Existence, Number, Subordination Gender

Tiriyó

4

18

Dalabon Chukchi Lavukaleve

4 4 4

5 4 (* cases) 28

Yélî Tidore Jahai

5 7 9

5 7 9

0 Gender, Number Animacy, Number 0 4 Number, Gender 0 0

Additional dedicated anaphoric demonstrative morpheme(s)*

Adverbial forms (in brackets additional dedicated anaphoric forms)

0 1

2 2 (+1)

1 0 1

4 3 (+1) 3

1

4

1 +?

3

4

4

+ ? +

2 4 * 3 cases ?

1 ? 0

4 (+1) 7 (verbal) 9

* ‘+’ indicates that there are at least some dedicated anaphorics; ‘?’ marks uncertainty in the description

systems with composite pronouns resist easy classification. The table compares nominal to adverbial forms, but the pronominals are in many cases formed from an adnominal in combination with, for example, a semantically general nominal. Note from the table that many systems have large numbers of forms generated by number, gender and other distinctions, so the mean number of pronominal forms is around 13 for this sample. The Tzeltal and Yucatec cases are harder to classify, since they have multiple forms compounded from different deictic elements.

18

Stephen C. Levinson

Table 1.2 makes clear that there is no simple relation between the number of forms and the number of deictic distinctions. Clearly the added information about number, gender, animacy and the like – what Diessel (1999) calls qualities of the referent – may help to identify the referent and proliferate forms. The Goemai posture information, specifying whether the referent is hanging, standing, sitting or lying, is clearly very helpful in this regard and is in a sense spatial and classificatory, but not deictic. These referent properties are one way to aid recognition of the referent. But there are many others. Let us discount the specific semantic specifications of properties of the referent and consider just purely deictic properties – parameters that make essential reference to the properties of the speech event rather than the referent. These are, on a first approximation, the properties enumerated in the first column of Table 1.2. An interesting question is what are these, and to what extent do they follow a predictable typology of distinctions – for example, the radial distances presumed in the sources used by Diessel (2013b). We discuss these issues in section 2.2.2 below. Paradigmatic Oppositions If we set aside the formal distinctions to do with properties or qualities of the referent (number, gender, animacy, positional orientation, etc.) and just look at the deictic distinctions, we can compare the numbers of distinctions (regardless of semantic basis) in pronominal versus adverbial demonstratives. What we see here (comparing columns 2 and 6 in Table 1.2) is that in some cases (Chukchi, Yélî, Jahai, Goemai, Warao, Saliba, Trumai, Lavukaleve) there seem to be the same number of pronominals as adverbials, often because the one set is derived from the other. Thus Jahai makes nine oppositions in each field. In other cases though (Lao, Yucatek, Tzeltal, Brazilian Portuguese), there is a mismatch, and this can play a significant role in usage patterns. For example, Table 1.3 shows a relatively frequent type of mismatch. Where this kind of formal paradigm mismatch exists, it is very common for the pronominals to be reinforced with an adverbial since the adverbials add

Table 1.3 A typical paradigm mismatch (‘proximal’ and ‘distal’ are here merely indicative labels – the content may be various) Pronominals

Adverbials

‘proximal’ ‘distal’

‘proximal’ ‘distal’ ‘far distal’

Demonstratives: Patterns in Diversity

19

information – indeed, in Brazilian Portuguese usages such as esse acqui ‘this here’ are the norm, and single pronominals the exception (thus although the language has just two demonstratives, in combination with four demonstrative adverbs six distinctions are actually made). A similar pattern is found in Tzeltal with a binary opposition reinforced by a ternary one, as in tey a mene, ‘that over there’, although adverbial reinforcement is less required. Lao also has the same abstract paradigm pattern, but the adverbs do not seem to play the same supporting role (see Enfield, 2003 and this volume). There are other ways in which adverbial distinctions may outnumber nominal or adnominal ones, for example where there are two sets of adverbials making orthogonal distinctions (as in Hanks’ 1990 account of Yucatec; see Bohnemeyer, this volume). The reverse pattern, where there are more pronominal forms than adverbial ones, appears to be rarer. Dalabon (this volume) may appear to be a case of this sort, but in fact there are additional adverbial forms not yet fully researched (Sarah Cutfield, p.c.).

2.2.2

Semantic Distinctions

2.2.2.1 ‘Distance’: Proximity and Origin

The literature has made much of ‘distance’ distinctions in demonstratives, from Frei (1944), through Anderson and Keenan (1985) to Diessel (2013b). What is actually discussed, though, is proximity to speaker or addressee or both, and thus both proximity and origo (indexical ground). The majority of the systems in this book are entirely speaker anchored, but some have both speakeranchored and addressee-anchored terms (e.g. Trumai, Saliba, Yélî Dnye, Jahai), and others only speaker-and-addressee-centred terms (Goemai and Brazilian Portuguese), an anchorage that other languages have only in the distal forms (e.g. Trumai, Saliba). One of the few plausible implicational ‘universal’ hypotheses emerging from this book is that the possession of an addresseeanchored term implies the possession of a speaker-based one. The chapters in this book show that proximity is an elastic notion, and according to each language, it has different extents depending on multiple pragmatic factors, well described, for example, in the chapter on Lao (see also Enfield, 2003). There is no doubt that a notion of current interactional space is involved, and hence the location of the addressee makes a difference to the shape of the space even in systems that have no addressee-anchored form. Some recurrent factors noted in the chapters are as follows: (a) pointing enlarges the proximal space considerably (see, e.g., chapters on Lavukaleve and Tzeltal); (b) even within speaker-anchored systems, the relatively distal location of the addressee may expand the notion of proximal; (c) areas within reach of speaker and addressee tend to be treated specially (e.g., a speakeranchored or addressee-anchored proximal may then be strongly pre-emptive

20

Stephen C. Levinson S

Figure 1.3 Cardioid shape of proximal space in some languages

over a neutral or unmarked term); (d) when a referent is proximal but right behind the speaker, a speaker-anchored proximal term often cannot be used, suggesting a cardioid shape to the proximal space (see, e.g., Warao, Tiriyó, Goemai, Brazilian Portuguese, Tidore and Jahai chapters, but note this is not so for Lavukaleve or Yucatec); (e) contrastive usage often neutralizes proximity. In addition to these specific factors, many social interactional factors intervene to help define a ‘near space’ or interactional zone, from a distal space outside that zone. We may begin by demonstrating the utility of the questionnaire for addressing these issues on familiar European languages (they are not included in the volume here, with the exception of Brazilian Portuguese, since there is already significant literature available). The studies were done by colleagues at the Max Planck Institute for Psycholinguistics.4 Taking a selection of eight scenes in the questionnaire which map roughly onto a spatial distance metric from the speaker’s point of view, we can illustrate the different extensional range of a ‘proximal’ and ‘distal’ deictic in five European languages (English, Dutch, Italian, Russian and Brazilian Portuguese) – see Figure 1.4. In this array of languages, only Italian looks anything like the textbook or grammar book system of a neat carving of space into a speaker-anchored proximal versus distal sphere (see Bonfiglioli et al., 2009). A couple of observations can immediately be made: (1) these languages have contrasting extensions for their ‘proximal’ versus ‘distal’ demonstratives. Interestingly even closely related English and Dutch show small differences. (2) Many of the systems show significant overlap in the use of the two terms: English and Dutch permit the distal demonstrative just about everywhere (dotted lines indicate possible if less-preferred use), while in contrast Russian permits the proximal just about everywhere. (3) Brazilian Portuguese usage shows discontinuities over this speaker-anchored distance measure, with a proximal use just where the referent is far from the speaker but close to the addressee. This points to some sort of addressee anchoring (see also Meira, 2003; this volume). This shows immediately the importance of what Hanks (2011) called the 4

We thank David Wilkins for English and Italian, Mariet Verhelst and others for Dutch, Michael Dunn for Russian, Sérgio Meira for Brazilian Portuguese (see Chapter 10).

Demonstratives: Patterns in Diversity 1

6 S

21

19

9

A

S

S

12 A

S

13 A

S A

16

24

S

S A A

English

this that

Dutch

deze (dit) die (dat)

Italian

questo

Russian

ètot

A S

quello tot Brazilian esse Portuguese

aquele

Figure 1.4 Extensions of demonstratives in some European languages (in the scenes, S labels speaker, A addressee and X the referent where not clear)

indexical ground, or the anchorage of the distance (or other) measures involved. These extensional patterns are clues to the underlying intensions or abstract meanings. We will argue below, for example, that where a term may be used indiscriminately over a wide range of scenes, it is likely to be an unmarked (semantically general) term in a privative opposition. So, for example, the fact that English that can be used for referents close to the speaker suggests that it is actually unmarked for distance but picks up its distal meanings by pragmatic opposition to a marked proximal item, this.5 The mechanism here is a wellunderstood application of Gricean conversational principles, which enjoin the use of a more specific term if it applies – if the more specific term is not used, that conversationally implicates that it doesn’t apply (Levinson, 2000). Russian then would be the inverse case, with an unmarked proximal, and a more specific distal which tends to restrict the meaning of the unmarked term. These conversational inferences are, like all conversational implicatures, defeasible, so in circumstances in which it is not misleading the less specific term may still be used where the specific would otherwise be expected. This yields some of the flexibility that makes the precise description of demonstratives rather hard. Let us now pursue the other two-term pronominal/adnominal systems which are investigated in this book (we will see that this categorization is misleading in a moment). If we compare the systems of Lao (Tai-Kadai family, South-East Asia) and Goemai (West Chadic, Nigeria), we see they look different yet again 5

As can be expected, there are different analyses of the English demonstratives, with some presuming this is the semantically unmarked part of the opposition (e.g., Dixon, 2003: 81) – see discussion in Enfield (this volume).

22

Stephen C. Levinson 1

6 S S

19

9

A

S

12 A

S

13 A

SA

16

24

S

S A A

Lao

nii 4 nan 4

Goemai

n´d'én´ nòe n´d'énáng

Yucatec

–a’ –o’

AS

Figure 1.5 Some two-term oppositions in this book

(Figure 1.5). In Lao, both the proximal and the distal have large potential extensions, with just the closest and most distant scenes requiring the proximal and distal respectively. Again this suggests a system with pragmatic pre-emption, but one has to turn to much further detail to extract a clear analysis of the intrinsic meaning of the terms. In fact, for Lao the underlying semantics seems to be that the distal form nan4 carries the marked semantic content ‘referent not here’, while nii4 is unmarked, picking up the residue, and can refer to all but the most distant referents. Goemai, on the other hand, appears to have both marked proximal and distal terms, but they are both from a shared speaker and addressee anchorage – thus something can be ‘near’ to a distant addressee and still get the proximal term (as in scene 16 in Figure 1.5 below). Interestingly, this kind of ‘we-thinking’ is now recognized to play a role in other kinds of joint activity (Galotti and Frith, 2013), and especially in reference, viewed as a joint activity (Bangerter and Clark, 2003; Peeters and Özyürek, 2016). Yucatec seems to offer the inverse to Lao – that is to say, the distal (or ‘nonimmediate’) form seems to be unmarked and the proximal marked (cf. Russian and English above), a pattern that seems to be the more general. We turn now to consider a bit more systematically the distinctions found in the 15 languages in this book. Table 1.4 tabulates the major types of spatial distinctions. It should be noted at once that two languages (Jahai, Tidore) have important spatial distinctions of another kind, namely directional systems making reference to, inter alia, geographical features. These will be discussed later. Setting these extra-spatial distinctions aside, the systems differ on three main dimensions: (i) the origo or anchorage (Hanks’ 2009 ‘indexical ground’): terms may be anchored on the speaker, the addressee or both. (ii) zones around the anchors – here glossed proximal, distal or far distal. (iii) the possession of a spatially neutral or unmarked term that contrasts with the terms which have a clear anchorage and a clear zone of application.

Demonstratives: Patterns in Diversity

23

1

1 1 1

1 1

1 1 1 1

1

Terms on other dimensions epistemic attention

1

1 1 1 1 1

NEUTRAL term

Speaker +Addressee-anchored DIST

Speaker +Addressee-anchored PROX

Addressee-anchored DIST

1 1 1

Addressee-anchored PROX

1 1 1 1

Speaker-anchored FAR DIST

Speaker-anchored DIST

Yélî Dnye Chukchi Lavukaleve Saliba Lao Yucatec Tiriyó Trumai Tzeltal Portuguese Warao Goemai Tidore Jahai Dalabon

Speaker-anchored PROX

Table 1.4 Major types of spatial distinction in the nominal demonstratives of the 15 languages, where 1=attested

1 1 1 1

attention new/old, auditory

1

1 1

1

1

1

1

1 1 1

1 1

1 1

1

geocentric geocentric identifiable vs unfamiliar

As an example, consider Yélî Dnye, which has speaker-anchored proximal and distal terms, an addressee-anchored term and an unmarked or neutral term (and a dedicated anaphoric term). Figure 1.6 may help to explain such a system: the diagram illustrates that referents very close to S, those significantly distant from S and those very close to A all have preferred terms, but that the unmarked or neutral form has general extension, so particularly picks up the middle ground where the other terms do not apply. The importance of these neutral terms is one of the major discoveries of this volume, occurring in half of the languages. It is also possible that the distal terms in Saliba, Tzeltal and Dalabon may better be analysed as neutral terms, because they have very wide extensions overlapping with the other terms, in which case this would be a majority encoding strategy. In the case of Tiriyó,

24

Stephen C. Levinson Kî (unmarked/neutral)

Ala (Prox to S)

Mu (Dist to S)

Ye (Prox to A)

Figure 1.6 Unmarked versus marked oppositions in Yélî Dnye demonstratives

special contrastive uses suggest that the term in question is an unmarked distal, whereas in Lavukaleve it is a true neutral unmarked for proximity (Meira and Terrill, 2005). The analysis, sketched above, is that by Grice’s maxim of Quantity, a speaker should use the most informative expression which applies. Thus in Yélî Dnye, if the object is in reach of the speaker, or conversely of the addressee, or alternatively in the far distance, then those more specific terms should be used; otherwise the unmarked term may be used. This explains the wide range of potential usage of the unmarked term in this language, as in the others described here, it being not strictly incorrect but potentially misleading to use it where the more specific terms apply. A possible problem for this analysis (discussed by Bohnemeyer, this volume) is that, especially for speaker-proximal or far-distal referents, it sometimes seems impossible to replace them with the neutral term. Here a comparison with temporal deixis may be helpful: if tomorrow is Friday, then it would be very odd to refer to tomorrow as ‘Friday’, because ‘tomorrow’ strongly preempts it. Deictic pre-emption is peculiarly strong. Similarly, the first person pronoun pre-empts the use of one’s own name – it would be odd in normal circumstances for Anne to say “Anne thinks that a better solution is Anne’s”. Nevertheless, in the right circumstances, for example answering the phone, Joe can say “Joe speaking”, so there is no grammatical or semantic impediment, rather a usage pre-emption. On the basis of the findings in this volume, there are grounds to be suspicious of reports of ‘medial’ terms – that is to say, systems with a three-way distance contrast. Instead, it seems likely that many of these are actually two terms with clear codings for proximal and distal, and then a third term that is unmarked, or relatively unmarked. There are at least three ways this may come about, shown in Figure 1.7, and the volume clearly attests two of them. With this preamble, we can show the similarities and differences between the systems by using the kinds of diagrams in Figures 1.6 and 1.7 above to sketch the nature of the spatial and anchorage oppositions in the main nominal

Demonstratives: Patterns in Diversity

25

(a) e.g. Lavukaleve (unmarked/neutral)

(Prox to S)

(Dist to S)

(b) e.g. Tiriyó, Tidore (unmarked distal)

(Prox to S)

(Dist to S)

(c) Unattested, but the two-term Lao system suggests this could perhaps occur (unmarked/neutral)

(Prox to S)

(Dist to S)

Figure 1.7 Three ways in which ‘medial’ terms may actually be unmarked terms (dotted extensions show areas where a neutral functions like a medial because it is not pre-empted by other terms)

demonstratives (setting aside for the moment the absolute dimensions in Tidore as well as the intrinsic dimensions in Jahai – the diagram shows what Burenhult calls the ‘accessibility’ demonstratives for Jahai). Figure 1.8 presents these spatial oppositions abstracted out of the many additional dimensions that structure these demonstrative sets. Clearly there are many different kinds of systems here, in fact only the pairs Chukchi and Warao, Goemai and Brazilian Portuguese, and Tidore and Lavukaleve (and possibly Saliba and Trumai) appear to make very similar oppositions – with the upshot that there are 11 or 12 distinct systems in the sample of 15 languages. Nevertheless, Figure 1.8 makes clear some striking regularities across the systems. Nearly all systems have a speaker-based, or speaker-and-addressee-based, proximal – the exception here is Lao. Neutral or unmarked terms are common, contrary to Anderson and Keenan (1985) and Diessel (1999: 38). Hence there are grounds for suspicion that most if not all terms that might be described as ‘medial’ in descriptive grammars are in fact neutral or unmarked distal, and pick up their medial usage through privative opposition with proximal or far-distal terms (see Meira and Terrill, 2005).

Stephen C. Levinson

26

Yélî Dnye S-Prox

Goemai

Tiriyó S-Dist

S-Prox

S&A-Prox

S-Dist S-Far

A-Prox

S&A Dist

N Chukchi S-Prox

Brazilian Portuguese

Warao

S-Dist

S-Prox

S-Dist

S&A-Prox S-Far

S-Far N

N

Saliba

S&A Dist

Tidore

Tzeltal S-Prox

S-Prox

S-Prox

S-Far N

A-Prox

N (A-Prox or S-Dist)

S&A Dist (N?) (Also Absolute direction system)

Trumai

Yucatec

Lavukaleve S-Prox

S-Prox

S-Prox

S-Far A-Prox

N

N S&A Dist Dalabon

Lao S-Dist

S-Prox

Jahai S-Dist

S-Prox A-Prox

N

S-Dist (or N?)

A-Dist (N) (Also Absolute & Intrinsic direction system)

Figure 1.8 Sketch of the core proximity oppositions in nominal demonstratives in the 15 languages (N = neutral or unmarked, Far is far distal, S&A is a joint speaker-addressee anchor)

2.2.2.2 Other Kinds of Spatial Information in Deictic Systems: Direction So far we have been considering spatial distinctions which effectively, by coding or implicature, specify search-domains that are at distances notionally or in idealized form radial from speaker or addressee or both (note though the kind of ‘deformation’ of such radial areas in, for example, the cardioid pattern in Figure 1.3). On the grounds that these terms do not normally specify angular or directional information, Levinson (1996; 2003) proposed that demonstrative systems should not be considered spatial frames of reference, that is, a system of cognitive coordinates capable of specifying a direction. He was particularly keen to point out that what had been called ‘the deictic frame of reference’ was

Demonstratives: Patterns in Diversity

27

actually a more complex system involving projected origos and angles, and not necessarily involving the deictic centre, which he called ‘the relative system’ (involving in English expressions like left, right, front, back). A good sample of cross-linguistic analyses of frames of reference in languages can be found in Levinson and Wilkins (2006). Frames of reference have been of particular interest to the psychology of space, and Levinson (2003) distinguished three kinds of basis for spatial reckoning which may be reflected in languages: (i) the intrinsic frame of reference which uses the designated facets of ground objects (as in ‘in front of St Peters in Rome’, the ‘intrinsic frame of reference’), (ii) the relative frame of reference which maps ego’s or alter’s own coordinates (left, right, front, back) on the environment (as in ‘he left his glasses to the left of the phone’), (iii) the absolute frame of reference which maps geographic or abstract external coordinates onto local arrays, as in ‘I left the mattock north of the tree’. Since then a number of objections have been made to the demotion of a deictic frame of reference (Diessel, 1999, 2014), on the basis, for example, that (a) demonstratives are deictic, mostly speaker anchored, mostly with spatial distance coding, and clearly a fundamental type of expression across languages, and (b) they are normally accompanied by a gesture providing the angular information. With regards to the second point, the chapters in this book make clear that pointing is by no means essential or general with demonstratives, and indeed the gestural uses are often distinctively different from the non-gestural ones – for example, proximity is typically enlarged by pointing. Perhaps a stronger point is that the demonstrative terms themselves sometimes encode angular information, usually of a geocentric or ‘absolute’ sort. Burenhult (2008) proposes a typology of angular information encoded in demonstratives which seems to be of two kinds. The first, noted early on by grammarians, is that some languages encode absolute (geocentric) direction, or array-external directions, the most common being information about whether the referent is above or below the deictic centre. Although grammars are rarely explicit about this, a word or affix meaning ‘that up there’ may indicate that the referent is literally higher than the deictic centre, or that it lies on a notional gradient (upriver/downriver, uphill/downhill) which actually delivers an angle on the horizontal. Alternatively, demonstratives may actually directly make reference to absolute coordinates, although these are usually glossed in terms of geographical gradients. These phenomena were already noted in Anderson and Keenan (1985) and before. But Burenhult (2008) adds that demonstrative systems may also make use of an intrinsic or array-internal frame of reference. These ‘absolute’ and ‘intrinsic’ directionalities are attested in this volume by two languages which offer perhaps the most elaborate systems here reviewed:

28

Stephen C. Levinson

Tidore and Jahai. Tidore has a core system with a proximal, a distal and far distal, pretty much equivalent to Lavukaleve (see Figure 1.8 above). But superimposed on this, Tidore has two geographically anchored axes – an ‘up/down’ axis, and a ‘seaward/landward’ axis. The interpretation of the ‘up/down’ axis is particularly complex, since it can be driven by verticality, lay of the land, direction of the royal palace, sea currents, etc. These absolute or geocentric axes constrain the distal term, so by pragmatic presumption producing a four-way differentiation between speaker-proximal, distal, far-distal and geographic space where geocentric coordinates kick in. The Jahai system illustrates the possibility of having demonstratives encoding not only externally anchored terms in an absolute frame of reference but array-internal organization in an intrinsic frame of reference. But the core of the system has speaker-based versus addressee-based opposition, crossed with an accessibility parameter, yielding four central terms. Superimposed on this is an absolute or geographical external ‘up/down’ opposition, which can have either a vertical interpretation or a horizontal interpretation in terms of upriver/downriver direction. Then, in addition, there is an intrinsic or array-internal mode of calculating direction, which Burenhult (2008; this volume) calls ‘exterior’, which distinguishes referents on the side of the participant away from the shared interaction space, as sketched below in Figure 1.9 (but see Burenhult, this volume). Hanks (2011) points to West Greenlandic as possibly making similar distinctions. Burenhult’s analysis may well apply to other systems less well described, but this awaits further research; for example, it appears to apply to the usage conditions for the proximal in Tidore, because the zone of proximity there extends further out on the side away from the addressee. Burenhult (2008: 113–114) suggests that some of the motion demonstratives that have been reported elsewhere which generalize to or from alignment demonstratives (as in Nunggubuyu) may also involve an intrinsic frame of reference. Burenhult also notes as a generalization that demonstratives in the relative frame of reference, e.g. meaning ‘that to the left’ do not seem to exist

S

A

S’s Exterior

Figure 1.9 Sketch of Jahai notion of ‘exterior’ space

A’s Exterior

Demonstratives: Patterns in Diversity

29

(although colloquial Tamil in usage allows a left/right bias to the interpretation of proximal/distal forms: Levinson, 2003: 108).

2.2.2.3 Non-spatial Distinctions in Demonstratives The usage of all of the systems described involves many non-spatial criteria, including whether the referent has been mentioned before, is in the current addressee’s focus of attention, is visible, is manipulable and so forth – many of these factors operate to constrain the relevant notions of proximity without being in any way encoded in the semantics of the terms (see, e.g., Chapter 3 on Lao). Clearly systems that mark gender, number and other properties of the referent (as in Trumai, Tiriyó, Goemai and Lavukaleve) involve coding of nonspatial attributes, but these are non-deictic in character. Some languages (like Tariana (Aikenvald, 1994) or Kilivila (Senft, 1996; 2004)) combine classifiers with demonstratives, and what is represented in this book is an unusual obligatory combination in Goemai of demonstrative and a positional classifier element, which specifies whether the referent is sitting, standing, hanging or lying. The positional information is spatial, but not deictic, as it does not depend on speaker or addressee’s location. Nevertheless, there are important deictic non-spatial properties that get coded directly in the semantics of demonstratives and don’t just constrain normal usage, and these parameters are briefly reviewed in the paragraphs that follow. Attention As mentioned at the outset, part of the root function of demonstratives seems to be to manipulate attention (Diessel, 2006). Earlier studies of Turkish and Japanese have suggested that attentional parameters may be built into the semantics of demonstratives: a re-analysis of Turkish demonstratives for example shows a distinction between şu (not yet in joint attention) and two other forms (bu and o) that presume joint attention and distinguish proximity (Küntay and Özyürek, 2002). Elaborating some suggestions by Diessel (2006), Burenhult (this volume) considers demonstratives to be ‘attention managers’ and makes a useful distinction between attention-drawers (with a subtype attention-correctors) and attention-confirmers, which in Jahai require different forms (Burenhult, 2003). The forms in question belong to one of three dimensions that structure the nine-way demonstrative oppositions in Jahai, a dimension Burenhult calls ‘accessibility’, a dimension we return to below. In Goemai, the main function of demonstratives is to draw attention to, and identify, the referent; this contrasts with Trumai, where demonstratives can be used only if attention is already on the referent or perhaps more generally on the interactive space. Chukchi has particles or clitics that prefix to the demonstratives forming terms which serve specifically to draw the addressee’s attention

30

Stephen C. Levinson

to the referent, and consequently normally occur with a point. Tzeltal has a proximal component (in-) that has a presentational quality, presumes visibility and draws attention to the referent. Similarly, Yucatec uses the presentational prefix on the demonstratives in order to draw the addressee’s attention to the referent. Saliba tends to use demonstrative adverbs to draw attention to referents. In contrast, the Yélî Dnye speaker- and addressee-proximals both presuppose mutual attention is already on the referent – hence the unmarked neutral must be used to draw attention to the referent. Tiriyó offers two forms of the proximal, one for drawing attention to new referents, and one that presupposes that attention is already drawn (it is thus typically an anaphoric). In sum, attentional distinctions are often built into demonstrative systems, and more generally, even in familiar languages, seem to play a role in their usage (see, e.g., Piwek et al., 2008; Peeters et al., 2014 on Dutch). Visibility Many of the chapters refer to the relevance of the visibility or otherwise of the referent. However, in the cases reviewed, visibility seems to be more a criterion constraining use than a part of the demonstrative semantics. In Chukchi an invisible referent tends to force a periphrastic construction (glossing, e.g., ‘which is there’) instead of a demonstrative, and a similar constraint is found in Goemai and Warao. In Tzeltal, there are constraints on the proximal forms, which seem to require visibility of referent. Saliba has a pragmatic condition on the visibility of referents – if they are not visible, they are presumed to have been mentioned or previously established, in which case the anaphoric element preempts the demonstratives; alternatively, a periphrastic form with the demonstrative adverbs is employed. In Yucatec, a presentative form of the deictic is used for calling attention to a referent in the immediate visual field, whereas when the referent is not in that field, an adverbial form (glossing ‘that there’) is used. Other chapters show that what on earlier accounts was thought to be a visibility distinction in fact turns out to have a more evidential character, to do with indirect or non-visual evidence (see the chapters on Yélî Dnye, Tiriyó and Jahai), as discussed immediately below. Mode of Apprehension, Evidentiality: An Epistemic Dimension Jahai has a demonstrative reserved for referents recognized only through their emissions, by sound or smell or inward pain. At first sight an ‘invisible’ term, it in fact cannot be used for referents directly touched as well as those not seen. In a similar way, Yélî Dnye has a term wu that on earlier accounts was reserved for invisible referents, but on close analysis refers to referents indirectly inferable: it cannot be used for invisible things whose location is clearly known. A typical use is ‘What’s that (wu)?’ when hearing a sound of unknown origin. Exactly the same kind of usage is reported from Tiriyó for the demonstrative previously thought to be reserved for the invisible, but which in fact

Demonstratives: Patterns in Diversity

31

refers to an audible but invisible referent – a non-audible invisible referent cannot be so referred to. In the dialect of Yucatec studied by Hanks (1990), there is a terminal particle used for referents audible but not visible; this is absent from the Yucatec studied by Bohnemeyer (this volume), but usage of the other deictics is also governed by this kind of evidential dimension – if one cannot see the referent but knows it is there, the distal demonstrative is paired with the distal adverb (te’l-o’). Accessibility A number of the chapters describe a criterion for proximity that can best be characterized as reachability, corresponding closely to the peripersonal space of neurology. For example, in Yélî Dnye the speaker- and addressee-proximals both have strong conditions on their use: if the referent is in the hand or within easy grasp, these proximals must be used in normal circumstances, strongly pre-empting the unmarked neutral term. In Saliba a referent that when sitting cross-legged is reachable is referred to with the proximal, but when one stretches the legs out forwards and it becomes not easily reachable, the proximal is no longer preferred. In Yucatec, the speaker-anchored proximal is used whenever the referent is in reach, and mostly not when the referent is out of reach. Reaching is also reported as the relevant criterion for the Lavukaleve and Tidore proximals. We have already seen how this accessibility criterion may create a cardioid pattern around speaker or addressee, where things behind one cannot easily be reached without turning around, although this is curiously variable across languages (e.g., in Yucatec the proximal can still be used for referents behind one). In some of the languages where this pattern arises, it seems to be an outcome of another constraint, namely that the referent should be visible, as in Warao. But in others it seems to be a genuine constraint based on reaching without effort. Thus Brazilian Portuguese makes an accessibility distinction in the distal compound (pronominal-adverbial) forms, between places that are easy versus hard to access. Notice that these findings are in line with recent experimental work on European demonstratives (Peeters et al., 2014); Coventry et al. (2008) found that increasing subjects’ reach with a stick increased the scope of the proximal. Jahai also has terms that could roughly be glossed as speaker- versus addressee-anchored proximal/distal oppositions, but Burenhult argues that the fine detail of usage argues for an analysis in terms of accessibility: e.g. if a speaker-accessible form can be used for a referent X, but an obstacle is placed between X and the speaker, the accessible form is no longer felicitous. Thus reachability and approachability play a crucial role, but the notion of accessibility is also conceptual and refers to whether a referent is or is not in the common ground or attention of speaker and addressee. Dalabon appears to have specialized demonstratives that make this distinction, which are alternatives to a proximal/distal set that is also in the inventory.

32

Stephen C. Levinson

Contrastive Use The central task used to generate the findings presented in this volume specifically avoided contrastive uses of the demonstratives (i.e. contrastive as in This one not that one). Nevertheless remarks on contrastive uses will be found throughout the volume. In addition, a number of the authors ran the extra task, specifically designed to explore contrastive uses, as in ‘Eliciting contrastive use of demonstratives for objects within close personal space’ by David Wilkins (1999b). One of the findings from this volume is that although there do not often seem to be distinct contrastive forms (but see, e.g., Schapper and San Roque, 2011 on Bunaq), contrastive uses are generally different from noncontrastive uses and rarely follow the English pattern where in contrastive contexts first reference uses the proximal and second reference uses the distal demonstrative as in First take this pill, then that one (Fillmore, 1997). One pattern is neutralization of proximity. Thus Chukchi speakers tended to revert to the same proximal form if A is contrasted with B in table-top space, as did Lavukaleve and Yucatec speakers; Dalabon speakers also prefer this but may deploy their special ‘common ground’ form for a distal referent in table-top space if pointing is disallowed. Another pattern is that order of mention is irrelevant, and proximity determines the forms used, as in Yélî Dnye; the same holds in Saliba unless the referents are equidistant from the speaker, in which case only in the contrastive use is the proximal contrasted to the distal for equidistant referents. The study of contrastive use can throw considerable light on the meanings of terms: true neutrals resist contrastive use, while distalneutrals allow such usage (Meira and Terrill, 2005). Gesture and Pointing The demonstrative task aimed to gather information about both gestural and non-gestural uses. One emerging result of this volume is that a pointing gesture extends the proximal zone, in some cases indefinitely. This is true in, e.g., Tzeltal, Saliba, Yucatec, Tiriyó, Lavukaleve and Dalabon. It seems that the gesture often has a presentational quality and this is what extends the proximal zone: thus Tzeltal gesture forces use of the proximal or presentative form, which then has extensive range right up to the horizon. Similarly, in Lavukaleve the presentative form of the demonstrative requires a point, and pointing occurs specifically with the presentative form. In Tiriyó, too, the gesture forces use of a proximal form for close referents and shrinks the space so that middle-distance referents can get the proximal. In a number of languages, the proximal forms require a gesture. Yélî Dnye proximals must be accompanied by a pointing gesture, or at least gaze, unless the referent is actually being manipulated. In Yucatec, the proximal forms must be accompanied by a gesture. In Saliba, pointing is associated with the speakerbased proximal rather than the addressee-based one, and pointing clearly

Demonstratives: Patterns in Diversity

33

extends the range of this proximal. In other languages like Warao, apparently all the exophoric demonstratives must be accompanied by a gesture, and if one’s hands are full, a dedicated anaphoric (or established referent) form must be used instead. In Tiriyó, touching or close pointing requires the proximal demonstrative. The discussion so far has been loose about ‘gesture’, although as far as we know pointing with the index finger is universally available (although supplemented by other bodily indications in many other cultures). Some have suggested that the function of the demonstrative is to draw attention to the pointing, which then directs attention to the referent (Bangerter, 2004). Alternatively one might claim that it is the gaze that actually does the work, the gesture drawing attention in turn to the gaze (see Enfield, 2001 on lip-pointing). Again, here we find cultural diversity. Some of the chapters provide specific information on the form of pointing associated with the demonstratives. Chukchi uses the index finger point, Tzeltal the index point for locations and the flat palm down for directions, Dalabon, Lao, Tiriyó and Yélî Dnye also offering lip- or head- and eye-pointing. In some languages, especially those that use an absolute frame of reference elsewhere in their spatial system, pointing may not be accompanied by gaze in the same direction (Levinson, 2003). Although lip- or head-pointing may be occasioned by occupation of the hands, when the hands are free it may have a special role as a demoted gesture (suggesting, for example, that the referent is obvious, or we shouldn’t point at the referent out of respect; see Enfield, 2001). In some cultures there are systematic taboos, for example on pointing with the left hand (Kita and Essegbey, 2001). More systematic information on this issue would be useful in future work on demonstratives (see Wilkins et al., 2007 for suggestions about data collection, and Kita, 2003). An interesting possibility is that gesture becomes less prominent in systems with more information encoded in the semantics of demonstratives, for example with directional, animacy, gender, number and classifier information. This would be in line with recent experimental findings that show longer, slower pointings when these are carrying the main directional information (Peeters, Chu, et al., 2015; see also Bangerter, 2004), but there are no more than hints in this direction from the chapters in this book. Dedicated Anaphorics in Demonstrative Paradigms This volume is focused on exophoric reference where demonstratives indicate referents in the speech situation, because this usage seems to be primary (e.g. learned first in childhood) and is the usage that has intrigued philosophers, psychologists and linguists. But many demonstrative systems use the same forms in exophoric (contextual) and endophoric reference (or anaphora sensu lato). In English, for example, ‘that X’ (as in ‘That woman you mentioned’) can refer backwards (anaphorically sensu stricto) to a previously mentioned

34

Stephen C. Levinson

referent, while ‘this X’ is likely to be forward looking or cataphoric (as in “This strange guy came up to me and asked for a dime”). English has third person pronouns, but many languages (e.g. some Australian ones, but not Dalabon in this volume) instead use demonstratives in this function. Where there is use both environmentally and textually, the two usages may not influence one another. But where a language provides a whole different set of forms for anaphoric reference, they may obviate the use of demonstratives in an endophoric function, as in Goemai, Lavukaleve or Yélî Dnye. Consider, for example, Yélî Dnye, as described in this volume: you can point at a referent and refer to it with, say, a speaker-centric demonstrative ala ‘this’, but the very next reference to that object even when accompanied by a point must use the dedicated anaphoric form (yi), even though this is still a deictic as well as an anaphoric usage. In this way the possession of a dedicated anaphoric system may make a very real difference to the use of demonstratives, and this obviation needs to be mentioned in any good description. In Warao, for example, the exophoric forms have an obligatory accompanying pointing gesture, while the anaphoric form picks up already established referents or those where no gesture can be made (because, for example, the hands are otherwise occupied). In Tiriyó, two forms of the proximal seem in fact to be specialized, one for drawing attention to a referent, and another that presumes prior attention through earlier reference. 3

Conclusions: Some Generalizations about Demonstrative Systems

One of the clear messages of this volume is that current typologies of demonstratives are inadequate, and we are far from being able to formulate tight ‘universals’ of demonstratives. What we are now in a position to do, however, is to lay out the semantic parameters of the design space for demonstratives – that is, the dimensions on which they vary. These dimensions are summarized in Table 1.5, where these parameters can be partially cross-combined, so that, for example, one can have an addressee-anchored proximal, but not apparently an addressee distal. Where constraints are known to be likely, they are listed below in Table 1.6, as possible generalizations. The reason that a proper typology, as a tree of possibilities, currently escapes us is that the constraints on the co-occurrence of these features are largely unknown. Obviously, where there are only two terms, the possibilities are limited, but for three or more terms the parameters in Table 1.5 offer many possibilities. Perhaps it makes sense to ask the following question: if cultural evolution is going to make a two-opposition system, or a three- or four- or more opposition system, which semantic features are likely to be selected at each increment of oppositions or terms? Given the remarks above we should dispense with the

Demonstratives: Patterns in Diversity

35

Table 1.5 Semantic parameters involved in the ‘design space’ of demonstrative oppositions Parameters Origin Proximity Distality Unmarked Access Attention Absolute direction Intrinsic direction (Burenhult, 2008) Evidential Gesture Dedicated anaphorics Properties of referent Emotional affect

Values speaker reach unmarked distal neutral accessible, no obstacle drawing vertical single participant ground non-visible with point nominal number, gender positive affect to referent

addressee interaction area far distal unmarked distal Inaccessible, obstacle

speaker & addressee beyond interaction area

presupposing horizontal joint participant ground audible, smellable without point adverbial animacy negative affect to referent

correcting

indirectly ascertained other kinesic index classifier category, etc. speaker–addressee relationship

counting of terms, the formal exponents of difference, and concentrate rather on the basic semantic oppositions that are made, selecting from the parameters in Table 1.5. Here the distribution of features that have been described in this volume may be illuminating when laid out as a tree of incremental possibilities from top to bottom, as in Figure 1.10. It is tempting to interpret the tree in diachronic light, as the order in which oppositions might be added, but as Diessel (1999) points out, the historical sources of demonstratives are usually opaque. One should note that there appears to be considerable freedom in whether, having made one distinction using feature X to yield terms X1 and X2, feature Z is used to make a subdivision of say X1, or whether Z may outrank X, and X be used to yield subtypes of Z. For example, suppose we take the ‘attention-drawing versus attention-confirming’ parameter: in Turkish, this parameter would be the first split in a decision tree, with a secondary proximity distinction between the referents within joint attention (Küntay and Özyürek, 2002); but in Tiriyó one must first decide if the item is proximal, and then whether attention is already on the item or not. We have arranged the papers in this book in line with the complexity order in Figure 1.10. Thus the book starts with chapters on Lao and Dalabon, both systems with speaker-based proximity systems, but in the case of Lao on a negative proximity base – a ‘not here’ term. We then proceed to systems anchored on both speaker and addressee, Brazilian Portuguese and Goemai –

36

Stephen C. Levinson

Two-opposition systems

Three-opposition systems

Four-opposition systems

S-Prox N

S-Prox S-Dist

N S-Dist

S-Prox A-Prox

Tzeltal, Yucatec

Dalabon

Lao

Unattested

S-Dist

S-Far

S&A-Dist

Lavukaleve

Tiriyó

Trumai, Saliba

S-Far

S&A-Prox S&A-Dist Braz. Portuguese, Goemai

A-Prox

Warao, Yélî Dnye Chukchi

Multi-opposition systems

Up/Down Sea/Land

Up / Down S-Exterior A-Exterior

Tidore

Jahai

Figure 1.10 The nominal systems in this book laid out as a tree of increasing semantic complexity (S-Prox = speaker-based proximal term, S-Dist = speaker-based distal term, S-Far = speaker far-distal term, N = neutral term unmarked for proximity, A-Prox = addressee-based proximal, S&A-Prox/Dist = speaker and addressee based proximal/distal)

Goemai introducing the complexities of a referent classifier system. Chapters on Tzeltal and Yucatec complete the coverage of systems based primarily on a binary opposition – since their formal exponents are many and complex, they introduce additional semantic parameters like presentational elements. We then proceed to systems that introduce a third semantic element, starting with speaker-anchored systems in Lavukaleve and Tiriyó which differ in the semantic nature of their third term, and subsequently to chapters on systems where the third element is anchored on both speaker and addressee. Next come systems that introduce yet a further semantic element over and above a speaker-based three-element system – in Warao and Chukchi this involves the introduction of a speaker-based far-distal, while in Yélî Dnye it involves the introduction of an addressee-based term. Finally we come to systems that make multiple oppositions, and here we see yet again that such systems can be built on different bases: in Tidore, absolute directional terms are overlaid on a speaker-anchored core system, while in Jahai absolute and intrinsic directional terms are overlaid

Demonstratives: Patterns in Diversity

37

Table 1.6 Generalizations about demonstrative systems: Hypotheses based on this book 1. 2. 3. 4. 5.

6.

7.

8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

There is no addressee-anchored demonstrative without a corresponding speaker-anchored one. It is rare, but possible (as shown by Lao), to have no speaker-based proximal. There are no addressee-anchored distal demonstratives. Speaker- and addressee-origos are not the only kind; speaker and addressee anchoring can be the only or main form of origo. A demonstrative system should for functional reasons have wide spatial coverage of the environment of the speech event (although some spatial discontinuities may result from attentional/visibility/reachability criteria, as in Fig. 1.3). It follows that a two-term system with only a speaker-proximal and addressee-proximal two-term system is unlikely to be found, since coverage would be too partial. Genuine medials (mid-distance) demonstratives may be hard to find, since they mostly seem to amount to unmarked neutrals or unmarked distals in privative opposition to proximals and fardistals. Four or more distinctions will never be distance-based alone, but always involve additional parameters like switches of anchorage, evidential distinctions or frame of reference distinctions. All demonstrative systems involve some notion of proximity, even if (as in the case of Lao) it is negative. Spatial zones invoked by demonstratives can be expected to be continuous rather than discontinuous spaces. Proximity notions are only metric and relatively inflexible in some proximal demonstratives, where reaching space of speaker, addressee or both may be involved. Proximal notions often involve reaching criteria and thus may often not extend behind the origo or anchorage point (speaker or addressee). Demonstrative systems may involve both the absolute and the intrinsic frames of reference, but never the relative frame. Terms described as reserved for ‘invisible’ referents in grammars are more likely to have an evidential basis, being reserved for audible referents, or referents indirectly ascertained. If a proximal allows concurrent pointing, pointing always extends the range of a speakeranchored proximal term. Possession of a dedicated anaphoric pronoun or adverb in the paradigm will preclude use of the exophoric demonstratives in anaphoric function. If there are more adverbial distinctions than nominal ones, adverbs and nominal demonstratives are more likely to occur as compound forms. Contrastive uses of demonstratives tend to neutralize proximity distinctions. Across languages, as the semantics of demonstratives becomes richer (with coding of direction and referent properties, for example), the importance and frequency of gesture declines.

on a basic opposition between speaker-based and addressee-based terms. The book illustrates very clearly how languages are largely free to pick and choose from the parameters in Table 1.5. No reader of this volume can fail to be impressed by the diversity of the 15 systems reviewed in this book. They vary fundamentally in formal type, in the

38

Stephen C. Levinson

semantic distinctions they make, in the pragmatic variables they are sensitive to and the functions they perform. Even concentrating just on the distance or proximity distinctions, we have seen there are nearly as many types (12) as there are languages reviewed in this book (15). Nevertheless, in the spirit of universals research, one may hazard the generalizations in Table 1.6. To conclude this introduction, demonstrative systems remain, as mentioned at the outset, a key ‘model system’ for the study of language in relation to its use. At first sight they seem quite simple, little signals of referential intention, but because they articulate the complexities of mutual awareness, of referent identification in the full three-dimensional sensorium of context and face-toface interaction, they can rapidly be seen to offer insights into the enormous complexity of the triadic attention (between speaker, addressee and referent) that nine-month-old infants are already beginning to glimpse (Tomasello et al., 2007). This makes demonstratives not only an interesting model system for cross-linguistic comparison, as in this book, but also a perfect laboratory for the study of language acquisition and the cognitive and neural underpinnings that make it possible (see, e.g., Peeters, Hagoort and Özyürek, 2015; Redcay and Saxe, 2013; Peeters et al., under review). One can be confident that future research will exploit the rich communicational infrastructure that crosslinguistic variation in demonstratives reveals. References Abney, S. (1987). The English noun phrase in its sentential aspect. PhD dissertation, MIT. Aikhenvald, A. (1994). Classifiers in Tariana. Anthropological Linguistics, 36(4), 407–465. Anderson, S. & Keenan, E. (1985). Deixis. In T. Shopen, ed., Language typology and syntactic description. Vol. III. Cambridge: Cambridge University Press, pp. 259–308. Bangerter, A. (2004). Using pointing and describing to achieve joint focus of attention in dialogue. Psychological Science, 15(6), 415–419. Bangerter, A. & Clark, H. H. (2003) Navigating joint projects with dialogue. Cognitive Science, 27, 195–225. Bhat, D. N. S. (2004). Pronouns. Oxford: Oxford University Press. (2013). Third person pronouns and demonstratives. In M. S. Dryer & M. Haspelmath, eds., The world atlas of language structures online. Leipzig: Max Planck Institute for Evolutionary Anthropology. (Available online at http://wals.info/chapter/43, July 20 July 2016.) Bohnemeyer, J. (2012). Yucatec demonstratives in interaction: Spontaneous vs. elicited data. In A. C. Schalley, ed., Practical theories and empirical practice. Amsterdam/ New York: John Benjamins, pp. 99–128.

Demonstratives: Patterns in Diversity

39

Bonfiglioli, C., Finocchiaro, C., Gesierich, B., Rositani, F. & Vescovi, M. (2009). A kinematic approach to the conceptual representations of this and that. Cognition, 111(2), 270–274. Braun, D. (2015). Indexicals. In E. N. Zalta, ed., The Stanford encyclopedia of philosophy. Spring edn. (Available online at http://plato.stanford.edu/archives/spr2015/ entries/indexicals/.) Bühler, K. (1982 [1934]). The deictic field of language and deictic words. In R. Jarvella & W. Klein, eds., Speech, place and action. Amsterdam: John Wiley, pp. 9–30. (Translated from Bühler, Karl. 1934. Sprachtheorie: Die Darstellungsfunktion der Sprache. Jena, Germany: Fischer). Burenhult, N. (2003). Attention, accessibility, and the addressee: The case of the Jahai demonstrative ton. Pragmatics, 13, 363–379. (2008). Spatial coordinate systems in demonstrative meaning. Linguistic Typology, 12, 99–142. Clark, E. V. (1978). From gesture to word: On the natural history of deixis in language acquisition. In J. Bruner & A. Garton, eds., Human growth and development. Oxford: Clarendon Press, pp. 85–120. Clark, H. H. (1996). Using language. Cambridge/New York: Cambridge University Press. Clark, H. H., Schreuder, R. & Buttrick, S. (1983). Common ground and the understanding of demonstrative reference. Journal of Verbal Learning and Verbal Behaviour, 22, 245–258. Coventry, K. R., Valdés, B., Castillo, A. & Guijarro-Fuentes, P. (2008). Language within your reach: near-far perceptual space and spatial demonstratives. Cognition, 108 (3), 889–895. Coventry, K. R., Griffiths, D. & Hamilton, C. J. (2014). Spatial demonstratives and perceptual space: Describing and remembering object location. Cognitive psychology, 69, 46–70. Da Milano, F. (2005). La deissi spaziale nelle lingue d’Europa. Milan: Angeli. Diessel, H. (1999). Demonstratives: Form, function, and grammaticalization. Amsterdam: John Benjamins. (2006). Demonstratives, joint attention, and the emergence of grammar. Cognitive Linguistics, 17(4), 463–489. (2013a). Pronominal and adnominal demonstratives. In M. S. Dryer & M. Haspelmath, eds., The world atlas of language structures online. Leipzig: Max Planck Institute for Evolutionary Anthropology. (Available online at http:// wals.info/chapter/42, accessed 15 January 2016.) (2013b). Distance contrasts in demonstratives. In M. S. Dryer & M. Haspelmath, eds., The world atlas of language structures online. Leipzig: Max Planck Institute for Evolutionary Anthropology. (Available online at http://wals.info/chapter/41, accessed 16 January 2016.) (2014). Demonstratives, frames of reference, and semantic universals of space. Language and Linguistics Compass, 8(3), 116–132. Dixon, R. M. W. (2003). Demonstratives: A cross-linguistic typology. Studies in Language, 27(1), 61–112. Enfield, N. J. (2001). ‘Lip-pointing’: A discussion of form and function with reference to data from Laos. Gesture, 1(2), 185–211. doi:10.1075/gest.1.2.06enf.

40

Stephen C. Levinson

(2003). Demonstratives in space and interaction: Data from Lao speakers and implications for semantic analysis. Language, 79(1): 82–117. Fillmore, C. J. (1982). Towards a descriptive framework for spatial deixis. In R. J. Jarvella & W. Klein, eds., Speech, place and action: Studies in deixis and related topics. New York: Wiley, pp. 31–60. (1997). Lectures on deixis. Stanford, CA: CSLI Publications. Frei, H. (1944). Systèmes de deictiques. Acta Linguistica, 4, 111–129. Gallotti, M. & Frith, C. D. (2013). Social cognition in the We-mode. Trends in Cognitive Sciences, 17(4), 160–165. Guérin, V. (2015). Demonstrative verbs: A typology of verbal manner deixis. Linguistic Typology, 19(2), 141–199. Halliday, M. A. K. & Hasan, R. (1976). Cohesion in English. London: Longman. Hanks, W. F. (1990). Referential practice: Language and lived space among the Maya. Chicago: University of Chicago Press. (1992). The indexical ground of deictic reference. In A. Duranti & C. Goodwin, eds., Rethinking context: Language as an interactive phenomenon. Cambridge: Cambridge University Press, pp. 43–76. (2005). Explorations in the deictic field. Current Anthropology, 46(2), 191–220. (2009). Fieldwork on deixis. Journal of Pragmatics, 41, 10–24. (2011). Deixis and indexicality. In W. Bublitz & N. R. Norrick, eds., Handbook of pragmatics. Berlin: Mouton de Gruyter, pp. 315–346. Himmelmann, N. (1996). Demonstratives in narrative discourse: A taxonomy of universal uses. In B. Fox, ed., Studies in anaphora. Amsterdam: John Benjamins, pp. 205–254. Jungbluth, K. (2003). Deictics in the conversational dyad: Findings in Spanish and some cross-linguistic outlines. In F. Lenz, ed., Deictic conceptualisation of space, time, and person. Amsterdam: John Benjamins, pp. 13–40. Jungbluth, K. & Da Milano, F., eds. (2015). Manual of Deixis in Romance Languages. Berlin: de Gruyter. Kemmerer, D. (1999). ‘Near’ and ‘far’ in language and perception. Cognition, 73, 35–63. Kita, S. (2003). Pointing: Where language, culture, and cognition meet. Oxford: Psychology Press. Kita, S. & Essegbey, J. (2001). Pointing left in Ghana: How a taboo on the use of the left hand influences gestural practice. Gesture, 1, 73–95. König, E. (2012). Le role des deictiques de manière dans le cadre d’une typologie de la deixis. Bulletin de la Société de Linguistique de Paris, 107(1), 11–42. Kornfilt, J. (1997). Turkish. London: Routledge. Küntay, A. C. & Özyürek, A. (2002). Joint attention and the development of the use of demonstrative pronouns in Turkish. In B. Skarabela, S. Fish & A. H. J. Do, eds., Proceedings of the 26th Annual Boston University Conference on Language Development. Somerville, MA: Cascadilla Press, pp. 336–347. Küntay, A. C. & Özyürek, A. (2006). Learning to use demonstratives in conversation: What do language specific strategies in Turkish reveal? Journal of Child Language, 33(2), 303–320. doi:10.1017/S0305000906007380.

Demonstratives: Patterns in Diversity

41

Levinson, S. C. (1996). Frames of reference and Molyneux’s question: Cross-linguistic evidence. In P. Bloom, M. Peterson, L. Nadel & M. Garrett, eds., Language and space. Cambridge, MA: MIT Press, pp. 109–169. (2000). Presumptive meanings: The theory of generalised conversational implicatures. Cambridge, MA: MIT Press. (2003). Space in language and cognition. Cambridge: Cambridge University Press. (2004). Deixis. In L. Horn, ed., The handbook of pragmatics. Oxford: Blackwell, pp. 97–121. Levinson, S. C. & Burenhult, N. (2009). Semplates: A new concept in lexical semantics? Language, 85, 153–174. doi:10.1353/lan.0.0090. Levinson, S. C. & Wilkins, D. (2006). Grammars of space. Cambridge: Cambridge University Press. Liszkowski, U., Brown, P., Callaghan, T., Takada, A. & De Vos, C. (2012). A prelinguistic gestural universal of human communication. Cognitive Science, 36, 698–713. doi:10.1111/j.1551–6709.2011.01228.x. Lyons, J. (1977). Semantics. Vol. II. Cambridge: Cambridge University Press. Meira, S. (2003). Addressee effects in demonstrative systems: The cases of Tiriyó and Brazilian Portuguese. In F. Lenz, ed., Deictic conceptualisation of space, time, and person. Amsterdam: John Benjamins, pp. 3–12. Meira, S. & Terrill, A. (2005). Contrasting contrastive demonstratives in Tiriyó and Lavukaleve. Linguistics, 43(6), 1131–1152. Naruoka, K. (2006). The Interactional functions of the Japanese demonstratives in interaction. Pragmatics, 16(4), 475–512. Özyürek, A. (1998). An analysis of the basic meaning of Turkish demonstratives in faceto-face conversational interaction. In S. Santi, I. Guaitella, C. Cave & G. Konopczynski, eds., Oralité et gestualité: Communication multimodale, interaction. Paris: L’Harmattan, pp. 609–614. Pagel, M., Atkinson, Q. D., Calude, A. S. & Meade, A. (2013). Ultraconserved words point to deep language ancestry across Eurasia. PNAS, 110, 8471–8476. Peeters, D. & Özyürek, A. (2016). This and that revisited: A social and multimodal approach to spatial demonstratives. Frontiers in Psychology, 7, article 22 (February 2016). Peeters, D., Chu, M., Holler, J., Hagoort, P. & Özyürek, A. (2015). Electrophysiological and kinematic correlates of communicative intent in the planning and production of pointing gestures and speech. Journal of Cognitive Neuroscience, 27(12), 2352–2368. doi:10.1162/jocn_a_00865. Peeters, D., Hagoort, P. & Özyürek, A. (2015). Electrophysiological evidence for the role of shared space in online comprehension of spatial demonstratives. Cognition, 136, 64–84. Peeters, D., Snijders, T. M., Hagoort, P. & Özyürek, A. (under review). The neural integration of pointing gestures and speech in a visual context: An fMRI study. Ms. Peeters, D., Zeynep, A. & Özyürek, A. (2014). The interplay between joint attention, physical proximity, and pointing gesture in demonstrative choice. In Proceedings of the 36th annual meeting of the Cognitive Science Society. Austin, TX: Cognitive Science Society, pp. 1144–1149. Perkins, R. D. (1992). Deixis, grammar and culture. Amsterdam/Philadelphia: John Benjamins.

42

Stephen C. Levinson

Piwek, P., Beun, R.-J. & Cremers, A. (2008). ‘Proximal’ and ‘distal’ in language and cognition: Evidence from deictic demonstratives in Dutch. Journal of Pragmatics, 40, 694–718. Redcay, E. & Saxe, R. (2013). Do you see what I see? The neural bases of joint attention. In H. S. Terrace & J. Metcalfe, eds., Agency & joint attention. New York: Oxford University Press, pp. 216–237. Rosés Labrada, J. E. (2015). Grammaticalization of lexical elements as deictics: Evidence from Sálibani. Talk to MPI Nijmegen Grammar Group, 27 August 2015. Schapper, A. & San Roque, L. (2011). Demonstratives and non-embedded nominalisations in three Papuan languages of the Timor-Alor-Pantar family. Studies in Language, 35(2), 380–408. Schelling, T. C. (1960). The strategy of conflict. Cambridge, MA: Harvard University Press. Senft, G. (1996). Classificatory particles in Kilivila. New York: Oxford University Press. ed. (1997). Referring to space: Studies in Austronesian and Papuan languages. Oxford: Clarendon Press. ed. (2004). Deixis and demonstratives in Oceanic languages. Canberra: Pacific Linguistics. Smith, N., ed. (1982). Mutual knowledge. London: Academic Press. Stevens, J. & Zhang, Y. (2013). Relative distance and gaze in the use of entity-referring spatial demonstratives: An event-related potential study. Journal of Neurolinguistics, 26(1), 31–45. Tanz, C. (1980). Studies in the acquisition of deictic terms. Cambridge: Cambridge University Press. Tomasello, M., Carpenter, M., Call, J., Behne, T. & Moll, H. (2005). Understanding and sharing intentions: The origins of cultural cognition. Behavioral & Brain Sciences, 28(5), 675–691. Tomasello, M., Carpenter, M. & Liszkowski, U. (2007). A new look at infant pointing. Child Development, 78, 705–722. Umbach, C. & Gust, H. (2014). Similarity demonstratives. Lingua 149, 74–93. Van der Auwera, J. & Sahoo, K. (2015). On comparative concepts and descriptive categories, such as they are. Acta Linguistica Hafniensa, 47(2), 136–173. Weissenborn, J. & Klein, W., eds. (1982). Here and there: Cross-linguistic studies on deixis and demonstration. Amsterdam: John Benjamins. Wilkins, D. P. (1999a; this volume). The 1999 demonstrative questionnaire: ‘This’ and ‘that’ in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24. (1999b). Eliciting contrastive use of demonstratives for objects within close personal space (all objects well within arm’s reach). In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 25–28. (Available online at http://fieldmanuals.mpi.nl/volumes/1999/elicitingcontrastive-demonstratives-personal-space/.) Wilkins, D. P., Kita, S. & Enfield, N. J. (2007). ‘Ethnography of pointing’: Field worker’s guide. In Asifa Majid, ed., Field manual volume 10. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 89–95. (Available online at http://field manuals.mpi.nl/volumes/2007/ethnography-of-pointing/.)

2

The Demonstrative Questionnaire: “THIS” and “THAT” in Comparative Perspective David P. Wilkins1

1

Purpose

This “questionnaire” is an elicitation tool which is meant to help a researcher begin to identify the (extensional) range of use of (some of the) basic spatial demonstrative terms in their research language. It attempts to correct some of the shortcomings of previous tools that we have developed for the exploration of demonstratives. Previous tools were both too narrow and too open-ended: they focused too narrowly on contrastive use of demonstratives for objects in table-top space, and they did not constrain situations sufficiently to allow for rigorous crosslanguage comparison. Prior trialing of the current elicitation tool suggests that it will allow us to design a similarity space for the extensional comparison of (some) demonstrative terms cross-linguistically, much in the same way that we were able to compare the application of “COME” and “GO” verbs across languages. It is important to note, at the outset, that this elicitation tool has NOT been designed to cover all the relevant distinctions that are known to exist within the demonstrative systems of the world’s languages. Instead, it has concentrated on those parameters within systems which, cross-linguistically, appear to be the most common. Thus, it has been designed to help differentiate and compare: (i) speaker-anchored versus addressee-anchored versus speaker&addresseeanchored versus other-anchored terms (ii) distance distinctions (up to at least four degrees of distance distinction from speaker) (iii) distinctions of visibility versus non-visibility.

1

Numerous people have contributed to the initial design and/or subsequent improvements of this questionnaire. These include: Felix Ameka, Michael Dunn, Jürgen Bohnemeyer, James Essegbey, Raquel Guirardello, Birgit Hellwig, Sotaro Kita, Steve Levinson, Anna Margetts, Asli Özyürek, Angela Terrill, and Barbara Villanova.

43

44

David P. Wilkins

Further, if employed as intended, it should help the researcher assess the roles played by gesture, addressee knowledge and attention, and different domains of object access in guiding the selection and use of demonstrative terms. 2

Design and Use

This elicitation tool centers around a set of 25 diagrammed scenes in which a speaker is referring to a single object (non-contrastively) within the context depicted. These are NOT stimuli to be shown to language consultants. They are scenes to help you organize your own elicitation tasks and to help you keep track of relevant parameters and oppositions to test. You, as researcher, are meant to understand the intention of the diagrams, and then decide the best way to get descriptions of the depicted scene. It is most preferable if you recreate the scene at the appropriate scale. You are also advised to keep track of natural demonstrative usage and see which, if any, of the 25 scene types they appear to represent. Next to each diagram is a description of the main features of the scene. Different manipulations of the same scene are also listed. Typical sentential frames in English are given as an example, but these are only intended as a guide to the intended distinctions. To Guarantee Comparability (a) Remember that each scene deals with reference to a single unique object; (b) Avoid contrastive reference. (c) We are interested primarily in expressions which involve demonstrative pronouns or demonstrative adjectives. You should record full utterances with these terms. (Note, while the same elicitation tool can also be used for demonstrative adverb usage, it would be preferred that you use it for elicitation of demonstrative pronouns/adjectives first and later for demonstrative adverbs. Of course, in some languages it is common to use both demonstrative pronouns/adjectives and adverbs, but you should let that come naturally. Focus on object reference, not place reference.) (d) You should be aware that in many of the scenes, the nature of the referent object is not specified. Since variations in size of an object can drastically affect demonstrative selection, you are asked to keep referent objects within a size range somewhere between a book at one extreme through to a chair at the other extreme. The object is preferably an inanimate object that could be picked up and held in two hands. (e) The consultant’s default or preferred description for a scenario may not in fact use a demonstrative. That’s fine, we’re interested to know this.

The Demonstrative Questionnaire

45

However, the researcher should probe whether one or more of the demonstrative terms could be used. (f) In many instances below, there are subsidiary questions dealing with pointing. Of course pointing conventions differ from culture to culture, and Discussion note #2 in the appendix to this chapter, as well as Chapter 4 of this manual (see also Wilkins and Pederson, 1999), may help you sort out the different details of this. Usually, however, cultures employ a more explicit convention beyond mere orienting – i.e. they use head pointing or lip pointing or manual pointing. A question like “Is pointing obligatory?” here means is a pointing convention beyond mere gaze orienting (or body stance) obligatory. Of course, where there’s more than one convention it would be interesting to know which ones would typically be selected. Method of Recording There is no strict recommendation here. Although the elicitation can be done at one sitting, it may be best to do parts of it as specific contexts suggest themselves. While one would ideally like to get everything on videotape, especially given the importance of accompanying indexical gestures, it may sometimes be more practical to take pen and paper notes. However, make sure to be as explicit as possible as to context, response, and accompanying gestures. Number of Consultants Minimally 3, preferably 5, and ideally 10. NOTE: It will help to read Discussion Notes #1 and #2 to understand the logic behind the choice of the following scenes, and subsidiary questions. These two discussion notes appear as an appendix to this task. 3 1.

The 25 Demonstrative Scenes Speaker (Spkr) points to own body part. In this case one of his/her teeth. “ _____tooth hurts.” “The ball hit me on _____ tooth.” • Does close pointing versus touching make a difference? • Does it make a difference if Addressee (Addr) already has attention on tooth versus attention being drawn? [In some languages teeth are more alienable body parts, so you may also want to try fingers, hands, shoulders.]

S

46 2.

3.

4.

5.

David P. Wilkins Spkr points to Addr’s body part. In this case one of Addr’s teeth. “Did you know _____tooth is chipped?” “Your right,_____tooth is yellow.” • Does close pointing versus touching make a difference? • Does it make a difference if Addr already has attention on tooth versus attention being drawn? [In some cultures, index finger pointing at someone else is impolite. Check whether there is any natural form of indexical reference for this situation.]

S

Spkr notices a movable object in contact with his/ her body. In this case, a bug on his/her shoulder. “ _____ bug is bothering me.” • Does it make a difference if Spkr’s attention has just gone to bug, or has been on it for a while? • Does it make a difference if Addr already has attention on bug versus attention being drawn?

Spkr points to movable object in contact with Addr’s body. In this case a bug on Addr’s shoulder. “Look at _____bug on your shoulder.” “What kind of bug is _____?” • Does degree of closeness of point to referent make a difference? • Does it make a difference if Addr already has attention on bug versus attention being drawn? Spkr references movable object in contact with Addr’s body, but without using a manual point? [Might use gaze or head point or lip point.] “Look at _____bug on your shoulder.” “What kind of bug is _____?” • Does it make a difference if Addr already has attention on bug versus attention being drawn?

A

S

S

A

S

A

The Demonstrative Questionnaire 6.

The referent is just beside Spkr (within easy reach), on side away from addressee. The object is difficult, if not impossible for Addr to see. “I’ve just finished reading_____book.” “Do you want to borrow _____book?” • Does it make a difference if Addr knows the object is there versus doesn’t know? • Does it make a difference if object has been mentioned before? Must Spkr point? • What if object was more visible?

7.

The referent is just in front of Spkr, and visible to Addr (but not within Addr’s reach). “I’ve just finished reading _____book.” “Do you want to borrow _____book?” “Have you read _____book?” • Does it make a difference if Addr already has attention on object versus attention being drawn? • Must Spkr point?

8.

The referent is in between Spkr and Addr and equidistant from both (and within arm’s reach of both). “Is _____your book/radio?” “I like _____book/radio.” “Do you want to borrow _____book?” • Does it make a difference if Addr already has attention on object versus attention being drawn? • Must Spkr point? • Does ownership of object make a difference?

9.

The referent is just in front of Addr, and visible to Spkr (but not within Spkr’s reach). “Is _____your book/radio?” “I like _____ book/radio.” “Do you want to borrow _____book?” • Does it make a difference if Addr already has attention on object versus attention being drawn? • Must Spkr point?

47

S

S

A

A

S

A

S

A

48

David P. Wilkins

10. The referent is just beside Addr (within easy reach), on side away from Spkr. The object is difficult, if not impossible for Spkr to see, but Spkr knows where object is. “Is _____your book/radio?” “I like _____book/radio.” “Do you want to borrow_____book?” • Does it make a difference if Addr already has attention on object versus attention being drawn? • Must Spkr point? • What if object was more visible?

S

11. Referent object is just behind the Spkr. The Addr is at some distance away but can readily see object (although it is well out of arm’s reach). The Spkr knows where the object is, even if she/ he cannot see it. The Spkr never turns to look at the object. “Is _____your book/radio?” “I like _____book/radio.” “Do you want to borrow _____book?” • Does it make a difference if the Spkr points or not? • Must Spkr point? • Does it make a difference if object has been mentioned before? • Does it make a difference if Addr already has attention on object versus attention being drawn? 12. Referent object is equidistant from Spkr and Addr, in front of (and between) them. It is easily visible to both. To get the object each would only have to walk about five paces. “Is _____your book/radio?” “I like _____book/radio.” “Do you want to borrow _____book?” • Does it make a difference if Addr already has attention on object versus attention being drawn? • Must Spkr point? • Does it make a difference if object has been mentioned before?

A

A

S

S

A

The Demonstrative Questionnaire 13. Spkr and Addr are sitting next to each other at one end of a large cleared space. The area of the space is about the size of a football field. There is another person at the other end of the space, and the referent is in front of this person, visible to both Spkr and Addr. “ _____ball/radio is a good one.” “I wonder where he got _____ball/radio” • Does it make a difference if Addr already has attention on object versus attention being drawn? • Must Spkr point? • Does it make a difference if object has been mentioned before? 14. Spkr and Addr are sitting next to each other at one end of a large cleared space. The area of the space is about the size of a football field. There is another person at the other end of the space. The referent is right at the center of the space (equidistant from Spkr/Addr and other). “ _____ball/radio is a good one.” “I wonder if _____ball/radio is his” • Does it make a difference if Addr already has attention on object versus attention being drawn? • Must Spkr point? • Does it make a difference if object has been mentioned before? 15. Spkr and Addr are sitting next to each other at one end of a large cleared space. The area of the space is about the size of a football field. There is another person at the other end of the space facing away from Spkr/Addr and the referent is in front of him. The referent is not visible to Spkr/ Addr, but the Spkr knows about object and its location. “ _____ball/radio is a good one.” “I wonder if _____ball/radio is his” “Did you see _____ball/radio he has?” • Does it make a difference if Addr knows the object is there versus doesn’t know? • Does it make a difference if object has been mentioned before? • Does it make a difference if Spkr does not know of existence of specific object, but conjectures existence from action of other (“He’s really getting stuck into _____thing.”). • Is pointing natural in this situation?

49 S A

S A

S A

50

David P. Wilkins

16. Spkr is sitting at one end of a large cleared space, and Addr is sitting at the other. The space is about the size of a football field. The Spkr has to shout to the Addr. The referent is in front of the Addr, and visible to speaker. “ _____ball/radio is a good one.” “Is _____ball/radio yours?” • Does it make a difference if Addr already has attention on object versus attention being drawn? • Is pointing natural? • Does it make a difference if object has been mentioned before?

S

A

17. Spkr is sitting at one end of a large cleared space, and Addr is sitting at the other. The space is about the size of a football field. The Spkr has to shout to the Addr. The referent is in the center of the space, equidistant from Spkr and Addr. “ _____ball/radio is a good one.” “Is _____ball/radio yours?” • Does it make a difference if Addr already has attention on object versus attention being drawn? • Is pointing natural? • Does it make a difference if object has been mentioned before? 18. Spkr is sitting at one end of a large cleared space, and Addr is sitting at the other. The space is about the size of a football field. The Spkr has to shout to the Addr. The Addr is facing away from Spkr and the referent is in front of him. The referent is not visible to Spkr, but the Spkr knows about object and its location. “ _____ ball/radio is a good one.” “Is _____ball/radio yours?” • Is pointing still natural? • Does it make a difference if object has been mentioned before? • Does it make a difference if Spkr does not know of existence of specific object, but conjectures existence from action of Addr? (“What’s _____thing your playing with?”).

S

A

S

A

The Demonstrative Questionnaire

51

19. Spkr is standing outside a home looking in through window. Addr is at other end of room away from window. Referent is near window and visible to Spkr (and Addr). [So object is physically closer to Spkr than Addr.] “Is _____your book/radio?” “I like _____book/radio.” • Does it make a difference if the Spkr points or not? Must Spkr point? • Does it make a difference if object has been mentioned before? • Does it make a difference if Addr already has attention on object versus attention being drawn? 20. Spkr and Addr are inside a house looking out of (open) door. They are near the doorway. The referent is just outside of door (near it). The referent is easily reached by both Addr and speaker (and equidistant from both). “I like _____book/radio.” “Who’s book/radio is _____?” • Does it make a difference if the Spkr points or not? Must Spkr point? • Does it make a difference if object has been mentioned before? • Does it make a difference if Addr already has attention on object versus attention being drawn? • Does term change with change in closeness of Spkr/Addr to door? Closeness of object to door? 21. Spkr and Addr are inside a house looking out of (open) door. They are near the doorway. The referent is a few meters away (next to a large immovable object). The object is technically closer (and in line) with Spkr [i.e. “on the Spkr’s side of the house”] “I like _____book/radio.” “Who’s book/radio is _____?” • Does it make a difference if the Spkr points or not? • Must Spkr point? • Does it make a difference if object has been mentioned before? • Does it make a difference if Addr already has attention on object versus attention being drawn?

S A

A

A

S

S

52

David P. Wilkins

22. Spkr is inside a house looking out of open door. Addr is sitting outside at a distance (a few meters away). Referent is just outside the door (outside, but physically closer to Spkr). “Is _____your book/radio?” “I like _____book/radio.” • Does it make a difference if the Spkr points or not? Must Spkr point? • Does it make a difference if object has been mentioned before? • Does it make a difference if Addr already has attention on object versus attention being drawn?

A

S

23. Addr is inside a house looking out of open door. Spkr is sitting outside at a distance (a few meters away). Referent is just outside the door (outside, but physically closer to Addr). “Is _____your book/radio?” “I like _____book/radio.” • Does it make a difference if the Spkr points or not? Must Spkr point? • Does it make a difference if object has been mentioned before? • Does it make a difference if Addr already has attention on object versus attention being drawn? 24. Large-scale geographic space. Spkr and Addr next to one another looking out across a river into some hills (several kilometers away). Spkr is pointing to referent which is visible up in the hills. “I’ve climbed to _____black rock.” “Have you been to _____cave?” “See _____bicycle.” • Does it make a difference if Addr already has attention on object versus attention being drawn? • Does it make a difference if object has been mentioned before? 25. Large-scale geographic space. Spkr and Addr next to one another looking out across a river into some hills (several kilometers away). Spkr is pointing to referent which is not visible because it’s in the hills on the other side. “I’ve climbed over to _____black rock.” “Have you been to _____cave?” “Your father made _____statue.” • Does it make a difference if Addr knows the object is there versus doesn’t know? • Does it make a difference if object has been mentioned before? • Must Spkr point?

S

A

A

A

S

S

The Demonstrative Questionnaire

4

53

Some Comments on Intended Parameters of Distinction Encoded in These Scenes

It is impossible to present a simple grid of distinctions within which each of these scenes can be fitted. However, the following comments may help the researcher to better understand some of the parameters involved. • Although there is no fixed order in which scenes should be presented, they have been ordered for the convenience of the researcher. Basically they pass through the various physically and socially determined “distance domains” that are given in Discussion Note #1 in the appendix to this chapter. That is, the scenes move from Personal Space through Interactional Space through Home Range Space through Large-Scale (Geographic) Space. • If we only consider speaker-based distance, then we can identify as many as seven distinct distances at which objects are located: object is body part (scene 1); object is in contact with body (scene 3); object is within arm’s reach of speaker (scenes 6, 7, 8); object is within easy access (a few steps from speaker) (scenes 9, 12, 20, 22); object tens of meters away (scenes 14, 17, 21, 23), object 100+ meters away (scenes 13, 15, 16, 18), and object some kilometers away (scenes 24–25). • A very similar scale also holds if we consider only addressee-based distance (i.e., as many as seven distinct distances obtain): object is body part (scene 2); object is in contact with addressee’s body (scene 4); object is within arm’s reach of addressee (8, 9, 10, 16, 18); object is within easy access (a few steps from addressee) (scenes 7, 12, 20); object tens of meters away (scenes 14, 17), object 100+ meters away (scenes 13, 15), and object some kilometers away (scenes 24–25). • Several scenes help test what happens when objects are equidistant from both speaker and addressee. These are scene 8 (object in arm’s reach of both); scenes 12 and 20 (object a few steps away from both); scenes 14 and 17 (object tens of meters from both); scenes 13 and 15 (object 100+ meters from both); and scenes 24 and 25 (object kilometers away from both). If a term is only anchored at addressee or at speaker, then we might find some hesitation or variability in the employment of terms in some of these scenes. However, if a term is anchored at both speaker and addressee, there may be no problems (unless another parameter gets in the way). • The visibility parameters in these scenes can be tabulated as in Table 2.1 (+ = visible; – = not visible; and ± = this is a factor that should be manipulated). • The main intentions of scenes 19 through 23 have more to do with established boundaries in lived space than they do with containment or buildings. That is, in some field sites walled rooms, doors, or windows may be odd, but this need not preclude relevant investigation. It is often the case that sleeping areas or cooking areas or camping areas have understood “social” boundaries to them

54

David P. Wilkins

Table 2.1 Visibility parameters

Scene #1 Scene #2 Scene #3 Scene #4 Scene #5 Scene #6 Scene #7 Scene #8 Scene #9 Scene #10 Scene #11 Scene #12 Scene #13 Scene #14 Scene #15 Scene #16 Scene #17 Scene #18 Scene #19 Scene #20 Scene #21 Scene #22 Scene #23 Scene #24 Scene #25

Visible to Speaker

Visible to Addressee

− + + + + + + + + ± − + + + − + + − + + + + + + −

+ − + + + ± + + + + + + + + − + + + + + + + + + −

such that one person can be within the (social) confines of one area and someone else outside those bounds. These scenes are meant to test whether such boundaries make any difference to the application of demonstratives. Can they, for instance, override actual distance relations? That is to say, in a scene like 19, is it more relevant for demonstrative selection that the object is physically within the “social” living area of the addressee, or is it more important that it is physically closer to the speaker, who is outside that living area? • Similar issues of accessibility are available with respect to scene 11, but in this case it is visual accessibility that is relevant. That is to say, will demonstrative selection be determined by the fact that the object is close to (but not visible) to the speaker, or by the fact that it is visibly accessible to the addressee although not close to the addressee, or . . . ?

The Demonstrative Questionnaire

55

• The issue of whether terms apply that mean “close to third person” are taken up in scenes 13 and 15. It is of interest to know if these contrast with scene 21, for instance, in which an object is located with respect to an inanimate entity (a tree), or scenes 24 and 25. • We are working on the presumption that the application of demonstrative terms will manifest prototype effects. As such we suspect that various scenes will be considered central members for a term in one language but a peripheral member for a term in another language. Thus, for example, scene 17 is a perfect example of the use of the mid-distant (speakeranchored) demonstrative in Arrernte (yanhe) but is a problematic (peripheral) case of the use of the Brazilian Portuguese term which refers to an object proximal to Speaker and Addressee (esse). Similarly, languages which do not have a “near addressee” demonstrative, and only have speaker-anchored terms, appear to treat scenes 16 and 13 (and 15 and 18) in an identical manner, whereas languages which do have a “near addressee” demonstrative, regularly apply this demonstrative to scene 16 (and 18) and use a different one for scene 13 (and 15). In short, the scenes have been selected both to test the boundaries between common systems and to test for prototype effects within common systems. 5

Organizing the Data for Analysis

As a first guide to organizing the data for analysis and comparative purposes, we provide the four tables on the following pages. These are based on very preliminary data collection but should give an idea of how systems of different types will treat these scenes. Thus, the first table (Table 2.2) shows how three different languages (English, Ewe, and Italian) with a “simple” speaker-based two-term system (proximal and distal) compare in their use of these terms. This table contrasts nicely with that for Brazilian Portuguese (Table 2.3). In Brazilian Portuguese there is also a basic two-term system, but both terms are anchored on Speaker&Addressee. Thus we see a distinction in how “speaker-based” terms carve up the scenes as opposed to “speaker&addresseebased” terms. Also of interest in the Brazilian Portuguese data is the further specification which the deictic adverbials provide. These adverbials are differently anchored from the demonstratives, and we see an addressee-anchored proximal adverbial and an addressee-anchored proximal adverbial as well as two adverbials which mean ‘away from speaker&addressee, but not far’ and ‘far away from speaker&addressee’. The third and fourth tables contrast two distinct three-term systems, Turkish (Table 2.4) and Japanese (Table 2.5). These two systems have often been treated in identical descriptive terms, but these tables help confirm the results of Kita and Özyürek which show these to be two very different systems. In particular, while Japanese has a true

(7)

(1)

A

A

(22)

S

(8)

(6)

(3)

S

A

A

S

(11)

A

Ewe: ‘prox’ only

S

S

(20)

S

A

Ewe: ‘prox’ preferred, ‘dist’ possible

A

Ewe: ‘prox’ or ‘dist’ equally acceptable

A

S

(23)

(17)

(21)

S

A

Ewe: ‘dist’ preferred, ‘prox’ may be possible

(24)

(16)

Ewe: ‘dist’ only

S

A

Centrally Distal

A S

S A

(13)

Italian: ‘dist’ only

English: ‘dist’ only

(14)

S A

A

(10)

S

S

A

(12)

A

A

S

S

(9)

(5)

A

English: all speakers prefer ‘dist’ but some can also use ‘prox’

S

Italian: ‘dist’ preferred but ‘prox’ possible

(proximal / distal cline)

(19)

S

English: ‘prox’ or ‘dist’ equally acceptable

A

(4)

(2)

A

Italian: ‘prox’ preferred but ‘dist’ possible

English: all speakers prefer ‘prox’ but some can also use ‘dist’

S

S

Italian: ‘prox’ only

English: ‘prox’ only

Centrally Proximal

S

S

Table 2.2 Three two-term (‘speaker-anchored’) demonstrative systems compared: Proximal and distal terms in English, Ewe, and Italian

S

(11)

A

esse N aqui esse aqui

S

(6)

esse N (+ aqui) esse (+ aqui)

A

A

(19)

S

(7)

(1)

S

(22)

A

S

S

A

S

S

Table 2.3 Brazilian Portuguese

A

S

S

A

(16)

S

(10)

(5)

esse N (+ aí) esse (+ aí)

A

A

esse N aí esse aí

A

ESSE

(9)

(4)

S

(23)

S

A

(20)

A

S

S

S

esse N esse

A

(8)

(3)

(2)

A

(17)

S

A

esse N aí esse N aí + (?esse aí) esse aí + (too distant and not (?aquele N aí) visible) [MOST DIFFICULT]

(18)

A

S

A

(21)

S

(14)

S A

(12)

A S

A S

(25)

(24)

(13)

S A

aquele N lá aquele lá

AQUELE

A

aquele N (+ alí) aquele (+ alí)

S

(8)

A

S

S

(7)

(6)

(3)

(1)

A

A

S

S

(17)

S

A

BU

A

(22)

S

S

SU

(19)

S

(4)

A

A

BU BU /SU BU /SU BU /SU SU / ??BU no pointing, so choice depends role of attention state of addr’s [relevant that close it is in both on attention less clear for initial attention scene involves spkr & addr bu = addr’s att’n use of bu ??? doesn’t matter pointing] attention space already on [su still used to for choice of su = addr hasn’t [?? su] object (with or draw addr’s either form noticed obj without att’n] before pointing) ?bu = maybe if su = addr’s att’n already est’d in drawn to obj the discourse

S

Table 2.4 Turkish

A

A

(21)

S

(20)

S

(12)

(11) A

A S

S

(14)

SA

(9)

(2) A

A

AS

A

(24)

(23)

S

O

S

(18)

A

S

(16)

S

(13)

(15)

SA

SA

(5)

A

A

AS

S

(25)

(10)

A

SU / BU / O SU / O SU /O O O O su, bu and o all choice varies attention does no pointing and no pointing and with pointing with pointing; with attention not affect both spkr & attention but addr’s attention may and nature of choice [pointing addr have doesn’t matter attention or may not be point used with both attention doesn’t matter relevant su = draw forms for all shared on [non-visibility depending on addr’s att’n to scenes] object relevant: in both scene (relevant obj scenes speaker for 12, o = shared att’n cannot see maybe 11) & no point at object] 14, distant point in 2

S

S

S

S

KO

(7)

(6)

(3)

(1)

A

S

A

S

S

(11)

(8)

(2)

A

A

A

A

A

(22)

S

(20)

S

(19)

S A

KO or SO different factors affect choice in each case (e.g. ± pointing; ± contact; ±ownership)

S

S

Table 2.5 Japanese

(9)

A

SO (??KO)

S

S

S

S

A

(10)

(5)

(4) A

A

SO

S

(23)

A

(18)

A

S

(17)

S

(16)

S

A

A

A

(21)

S

(12)

A

A (or attentiondrawing SO)

S

A S

A

(24)

(15)

S A

(14)

S A

(13)

S A

(25)

A (but with further specification and constraints on type of phrase that follows)

A S

60

David P. Wilkins

addressee-anchored proximal form, Turkish instead has a form which functions to draw addressee’s attention but is not tied spatially to either addressee or speaker. Examine these tables closely to give you an idea of the types of distinctions that are possible, and once you have your own data, try to similarly tabulate the scenes. Remember, since this task was NOT designed to capture all the distinctions possible within a system, you cannot expect to map all the terms you encounter onto this table. Moreover, there is a danger that for larger systems, this form of tabulation may give initially false generalizations over the use of a term, since the relevant parameters for very large systems are not systematically covered. As such, we would be obliged for suggestions as to future scenes you think we should include to improve this questionnaire.

Appendix Demonstrative Discussion Notes #1 Subject: “Distance Scale” [Domains of (tangible) accessibility in terms of what one would have to do to come in contact with an object that has been referred to – “experientially based”.]

In the deixis literature, and in grammatical descriptions of demonstrative systems, explanations of how descriptive terms like proximal, medial, and distal are realized in practice are often made either in terms of “experientially based” notions of ease of (tangible) access, or in terms of rough (Westernbased) measures. For instance, in discussing prototypical usage of the Proximate, Immediate, and Distant demonstrative roots of Nunggubuyu, Heath (1984: 269–270) writes: The Immediate is also used for locations conceptualised as being within easy access (not necessarily closer to addressee than to speaker), for example when speaker and addressee are sitting together and speaker indicates an object a few metres away . . . Again, if speaker and addressee are sitting together, Immediate would be used for something a few feet away, but ordinarily the Distant would be used to refer to something more than about 20 metres away.

Similarly, for Korean, Ho-min Sohn (1994: 295) writes: The i-series [close to speaker] . . . is used to refer to something close to or contacting with the speaker. Relativity is observed in the fact that the speaker must use the i-series to indicate something (e.g., a bug), say, on the addressee’s shoulder, if he is touching it or closely pointing to it. The ku-series [close to addressee] is used to refer to something

The Demonstrative Questionnaire

61

relatively close to or contacting with the addressee. The ce-series [away from speaker and addressee] is used to refer to something close to neither. If something is located equidistantly between the speaker and addressee, however, the speaker may use i if it is within a couple of meters from him, and ku or ce if it is over several meters away.

The purpose of this note is to present a systematization of the more common of the “explanatory” observations that are used to characterize abstract “distance distinctions”. In particular, since many observations are in terms of both physically and socially determined “distance domains” which describe what a speech act participant would have to do to gain tangible access to the referent, this short note will deal exclusively with such domains. A critical assumption that has been made in this endeavor, is that such a “domain scale” may be orthogonal to features like visibility and referent size (although, of course, these may all combine in a natural and typical way). These “domain scales” can be related to such commonly used scale notions as “small-scale” space, “medium-scale” space, and “large-scale geographic” space. A first approximation to a heuristically useful “domain scale” is given in Table A.1. The “question-set” at the end of this document will give a clearer picture of what is intended by the nine basic distinctions in this table. For the moment, let’s assume they are self-explanatory and go on to discuss why such a systematization may be useful. A frustrating thing about reading descriptions of demonstrative systems is that range of application is often poorly discussed, and more often than not left to characterizations like that presented for Nunggubuyu and Korean above. One of the most common things said about the application of terms is that they are “relative”, rather than absolute measures. This is especially common when discussing distance-dimensions. However, the nature of the relative application is rarely spelled out – are all demonstrative terms available in all access domains? or are they spread “relatively” across access domains? or . . . ? The problem seems to be most critical with “distal” or “not near” notions, which are often characterized as being ‘very far away’, ‘many meters away’, ‘hard to reach or unreachable’, and thereby suggesting that they are NOT usable for items within personal or interactional space. However, it is rarely clear whether such characterizations are of semantic features, prototype applications, or stereotypical cases, and these must surely be distinguished. An illustration from Arrernte may help. Here I will consider the application of four of the seven spatial deictic demonstratives: nhenhe ‘proximal to speaker’, yanhe ‘medial from speaker’, nhakwe ‘distal from speaker’, and alertakwenhe ‘distal to both speaker and addressee’. I would also claim that the use of the proximal, medial, and distaI forms (i.e., the first three terms) is relative. However, while it appears that nhenhe ‘proximal’ can be used to refer to an object in all nine domains, in the first two domains (“body parts” and

BODY parts

in arm’s REACH

within SOCIAL SPACE of conversation

PERSONAL SPACE INTERACTIONAL SPACE SMALL-SCALE SPACE

CONTACT with body

within REACH OF OTHER

USED SPACE [space covered by daily travels]

HOME RANGE MEDIUM-SCALE SPACE

IMMEDIATE LIVED SPACE [home base]

MANY DAYS’ WALK major journey [beyond horizon] GEOGRAPHIC (MAPPED) SPACE LARGE-SCALE SPACE

DAY’S WALK minor journey [within horizon]

Table A.1 Domains of access which appear regularly, but in a non-systematic fashion, in explanations of how terms like proximal, medial, and distal are to be understood

The Demonstrative Questionnaire

63

“contact”) the proximal form is the only demonstrative that is possible. Otherwise, it appears that three (relevant) objects within each of the other seven domains can be contrastively described using all three terms, as long as (speaker-based) relative distance within the domain is met. Of course, the proximal, medial, and distal (speaker-based) terms are regularly used to refer to objects distributed across domains. In fact, any time one asks an Arrernte consultant to choose things in the visible environment to exemplify how one would use nhenhe versus yanhe versus nhakwe, they regularly choose similar objects (say rocks or bushes or people), and then use nhenhe for an object within the interactional space, yanhe for an object much further away, but within home-range, and nhakwe for an object that is at the farthest end of visible space, typically just within the horizon. I myself have never encountered a consultant who chooses three objects that are all within interactional space, for instance, to contrast the “ideal” usage of the terms, even though all three terms can certainly be applied for objects within the local social interactional context that are arrayed at different distances from the speaker. (Of course, if the same question about exemplary usage was made within a confined windowless room, one may well get an array constrained to being within the one (interactional) domain chosen.) The point being, “idealized” usages of demonstrative terms which emphasize their contrastive differences are, in Arrernte, arrayed across items located in different domains of ‘access’, but this says little about their relative nature or their actual semantic characterization. Finally, let’s consider the term alertakwenhe which is anchored at ‘speaker and addressee’. This term is only ever used with pointing and refers to a visible object. Crucially, this term seems to be much more “domain-constrained” than the other three terms discussed. The referent must be at an absolutely far distance from both speaker and addressee – not easy to reach, and often hard to see though visible. In examining its uses, it only appears to be used for objects that occur in the last three domains (if visible). These observations are diagrammed in Figure A.1. As the figure should make abundantly clear, one needs to distinguish ranges of actual application from ranges of ideal application, as well as distinguishing the domains in which terms can actually be used contrastively versus domains where only a subset of terms is available. We may be able to accomplish this, in part, by being clearer about “distance/domain scaling”. I should point out that speakers’ own folk definitions of demonstrative terms may make reference to similar experiential access domains, but in the absence of exemplification with “pointing” may choose different domain “idealisations”. For instance, folk definitions given for the Warlpiri proximal form nyampu, emphasize contact with speakers and translate as ‘Nyampu is for nearby – like if you can touch it with your hand’ or ‘Nyampu, that is close up, like what we hold in our hand’. I do not have similar data for Arrernte, but

64 BODY

David P. Wilkins CONTACT REACH

REACH OF SOCIAL OTHER SPACE

IMMEDIATE USED LIVED SPACE SPACE

DAY’S WALK

MANY DAYS WALK

nhenhe (proximal) yanhe (medial) nhakwe (distal) alertakwenhe (distal from spkr&adr +vis) “ideal” nhenhe when asked to contrast terms

“ideal” yanhe when asked to contrast terms

“ideal” nhakwe when asked to contrast terms

Figure A.1 Distinguishing the ranges of application of Arrernte demonstrative terms from “idealized” application (showing explicitly the regions in which terms enter into contrast)

would not be surprised if the “decontextualized” idealizations of folk definitions focused on different types of “domain access” than “contextualized” contrasts where object pointing is involved. Some Questions That Might Be Considered With Respect to the Distance/Domain Scale • Body Parts and Places on Body: What demonstratives are used when: • referring to one’s own body parts, or spots on the body • contrasting two similar body parts (eyes), or spots on the body • when referring again to a part touched/held and referred to by interlocutor (e.g., when a doctor holds arm and says “does it hurt on this spot” and one answers “yes, it hurts on___spot”) • when referring to movements of feeling from one part of body to another • when talking about body parts that one can visibly access versus ones that are not visually accessible • when talking about body parts as visually accessed versus felt (proprioceptively accessed) (“Those feet of mine look so ugly”; “These feet of mine feel so painful”) • Things in contact with the body: What demonstratives are used when: • referring to objects in fixed contact with body: e.g., dirt on face; clothing; etc. • referring to movable/transferrable objects in contact with body: • bug (“This beetle won’t leave me alone”/“See this fly on my feet”) • something in one’s hand: just being held versus being presented/offered • referring to something one is standing/sitting/lying on (this bed is uncomfortable)

The Demonstrative Questionnaire

•

•

•

•

65

• referring to something on the interlocutor’s body which the speaker is touching with hand/finger (e.g., arm, clothing, bug) (i.e., contacting something also in contact with addressee) • referring to some independent object that part of the body is lightly up against or touching (or something touching the body) (“This tree I just brushed past has rough bark”) Things within easy arm’s reach (without getting up or taking any steps or turning significantly): What demonstratives are used when: • referring to a movable versus non-movable object in easy reach (of speaker as origo, addressee as origo) • contrasting more than one object within easy arm reach • where objects extend in away axis / across axis / upward / downward Things out of easy arm’s reach, but within easy arm’s reach of another participant in the interaction: What demonstratives are used when: • asking another person in the conversation to pass you something that is in their immediate reach • referring to something that you have to get up and walk a few steps to • referring to a non-movable object in the vicinity of another participant • contrasting an object in speaker’s reachable space with an object in addressee’s reachable space Things that fall within the shared social space set up by the interlocutors’ positions. (Need to keep track of “shape” of shared space): What demonstratives are used when: • referring to objects within space of 2 interactants, 3 interactants, 10 spread interactants, etc. • referring to objects in social space that are not easily reached by any participant • referring to objects in a social space created by two participants side-byside, two participants face-to-face, three participants in a triangle; five particpants in a circle • referring to objects which are in between speaker and interlocutor and reachable by both • contrasting objects at various positions within shared social space Things within the immediate lived (home) space (The place where “family” gathers together on a daily basis – for eating / cooking / sleeping): What demonstratives are used when: • referring to objects at the distant periphery of “home” space • referring to objects each of which falls in a different subdomain of home space • contrasting objects spread from speaker and addressee at center of home space to periphery of home space

66

David P. Wilkins

• when speaking to people across divides of home space versus being within the same domain of home space (independent of actual distances involved) • Things in the space which is used on a daily basis, including all the places usually traveled to from home-base and back within the day (village area / larger camp / water source / toiletry spot / agricultural or hunting area / market / workplace / etc.). What demonstratives are used when: • referring to objects (visible versus non-visible) at a place within the daily home-range, but which is outside of the “home-base” • contrasting objects (of appropriate scale) arrayed from speaker at home center through to the periphery of the home-range • Things that are just a day’s journey (by foot) away, or which are considered to be minor journeys in the region, involving a night’s stay somewhere else (at the edge of visible accessible space / within the horizon space / next village or two over / part of the accessible network of paths outside of the home range / etc.). What demonstratives are used when: • referring to the destination place • referring to different resting spots or landmarks along the way • contrasting different spots / landmarks extended along the projected travel path • referring to the visible distant part versus the non-visible distant part of such a journey path • Things in the space of a major journey which cannot be accomplished in one day, and require two or more sleeps en route. Usually a journey where the travelers will spend several days at the end point (visiting far-flung family relations / exploring new regions / going to a major town for work or trade / etc.). What demonstratives are used when: • referring to the destination place • referring to different resting spots or landmarks along the way; including places where one will sleep en route • contrasting different spots/landmarks extended along the projected travel path • contrasting different spots/landmarks near the destination

Demonstrative Discussion Notes #2 Subject: Pointing, Touching, Presenting – The Relation of Nonlinguistic Indexical Accompaniment to Demonstrative Choice While authors like Bühler ([1934] 1982), Fillmore (1982; 1997), Kaplan (1989), and Hanks (1992; 1996) have acknowledged, and emphasized, the significance of accompanying gestures in “true” acts of spatial demonstrative deixis, there has been little systematic exploration of the role and nature of such

The Demonstrative Questionnaire

67

accompaniments. It is now clear that one of the failings of our earlier questionnaires and tasks for exploring demonstrative reference is that they did not pay sufficient attention to variation in the types of indicative acts that accompany demonstrative terms. Bühler has called such acts “guidelines” and notes that “any deictic word without such guidelines is running blind to its meaning.” The purpose of this note is to highlight one important facet of this problem; one that has significant ramifications for intra- and inter-language comparability. Various elicitation tasks have now led me to the observation (which probably should have been obvious much earlier) that the linguistic choice of demonstrative is not fully independent of the choice of accompanying indexical act (i.e., the choice of “guideline”). That is to say, it often happens that in the same physical context, referring to the same object (at the same “proximity” and “scale”), different non-linguistic indexical acts correspond systematically with differences in demonstrative choice. Let’s consider several examples. In elicitations with Arrernte speakers, where three objects (cups) were evenly spaced across the width of a table (0.7 m) on the axis away from the speaker, there was variable treatment of the object furthest from the speaker. When speakers referred to the furthest object with NO accompanying manual point, but were gazing in that direction or were using a “lip point” or a “head nod” toward the object, they regularly used the distal demonstrative form nhakwe. However, when they used a fully extended, index finger point, the same speakers systematically used the medial demonstrative yanhe to refer to the same object. In very rough terms, the manual point occurs with a demonstrative term that suggests the object site is closer to the speaker, whereas the use of body orienting (without manual point) occurs with a demonstrative term which suggests the object site is further from the speaker. Similarly, let’s consider one Japanese speaker’s view of the following two scenes: (a)

(b) A S

A S

68

David P. Wilkins

In scene (a) the speaker is pointing to one of the addressee’s teeth. In scene (b), the speaker has her back to an object that she knows is right behind her, and the addressee, who is at a little distance from both speaker and object, is looking at the object that is behind the speaker. For both scenes, a ko-series term (“speaker proximal”) or a so-series term (“addressee proximal”) can be used to refer to the object, but different factors relating to the accompanying indexical act correspond to the preference of one term or the other. With respect to scene (a), the closer the pointing finger of the speaker comes to the tooth of the addressee, the better the use of the ko-series, and the worse the use of the soseries. At the point where the speaker’s finger touches the tooth, the ko-series is the only possible choice (i.e., use of the so-series becomes impossible). Thus, although the tooth is clearly part of the addressee (although a relatively more alienable part than the ear, for instance), the addressee-based term becomes impossible if the speaker’s point is touching it, and only the speaker-based term is possible. In contrast, the description of scene (b) changes according to whether the speaker is pointing or not pointing (in the situation under consideration, the speaker never turns toward the object). If the speaker does not point to the object, but merely refers to it (e.g. “Do you want to borrow _ book behind me?”), then the so-series is considered the natural choice, and use of the ko-series is decidedly odd. The use of the addressee-based term suggests that the fact that the object is in the (gaze) attention of the addressee, rather than the speaker, is more important than the fact that it is significantly nearer the speaker than the addressee. However, if the speaker makes a manual point to the area behind himself, then it is the ko-series that is considered the natural choice, and the so-series is odd. For the next example, we turn to an elicitation task in which two “sticky notes” (with a different geometrical figure on each) are stuck face down onto a small note pad. The note pad is horizontal and essentially held within the lap of the speaker, and the two squares of paper on it are aligned on the away axis. So, while both “objects” are well within arm’s reach of the speaker, technically one square of paper is physically closer to the body of the speaker, and the other is physically further away. The purpose of the elicitation is to get a contrastive response in which both objects are referred to one after the other (e.g., “this is the circle and that is the square”), and the starting point for the contrast is varied by certain manipulations (i.e., speakers sometimes start by talking about the one closest to them first and the furthest second, and at other times the order of description is the other way around). Whether they are simply pointing to each of the paper squares, or touching each of them, English speakers can freely say “this is X” and “that is Y” (largely independent of where they start, or technical distance). However, consultants from a number of other languages (e.g., Ewe, Italian, Persian) regularly show a different response pattern depending upon whether they are pointing without touching or pointing by touching. In Ewe, for

The Demonstrative Questionnaire

69

example, if the speaker is closely pointing, but not touching, the closer piece of paper can be referred to with the proximal form ke (or sia), and the further piece of paper can be referred to by the distal term kemi (or má). However, as soon as the speaker touches the paper (no matter how many contrasts are involved, or how far the stretch is), it must be referred to using the proximal form, and any contrast is made in other ways (i.e., you must say the equivalent of “this is X and this is Y” when touching each object in turn). Thus, indication by touching seems to have a special status in demonstrative reference, such that if something is touched it is brought into the proximal/personal sphere of the speaker. Technically, the non-touching point is no further (in the horizontal dimension) from the object of reference than the indication by touch, but touch versus non-touch certainly affects demonstrative choice. Finally, let’s consider an example from Dutch. Kirsner (1993) has observed that a speaker’s references to clothing they are wearing are more natural using the proximal form deze rather than the distal form die. He gives the following example (1): (1)

Vind je dat deze / ?die trui bij het jasje past? Do you think that this / ?that sweater goes with the jacket?

The one exception he noted to this generalization has to do with shoes, which can be referred to proximally or distally, as in example (2). (2)

Vind je dat deze / die schoenen bij het jasje passen? Do you think that these / those shoes go with the jacket?

Follow-up elicitation with Dutch speakers confirms these observations. However, what Kirsner failed to note is that in a sentence like (2) the choice of deze versus die appears to be consistently associated with different indexical acts (different “guidelining”). In modeling imagined uses of each of the two variants of example (2), Dutch consultants (apparently unconsciously) selected distinct non-linguistic acts to accompany the difference in demonstrative selection. When using the proximal form, speakers will simply either look down at their shoes or lift one of their feet up a little (but there is no manual pointing). By contrast, when using the distal form of the sentence, speakers regularly chose to point to their shoes. In this case then (within the personal sphere), it appears that non-manual orienting acts are associated with a demonstrative term that suggests the object site is “closer” to the speaker, whereas the manual point is associated with a demonstrative term that suggests an object site that is “further” from the speaker. In other words, we have a pattern of usage that is somewhat the reverse of the Arrernte example we began with. (I suspect, however, this is less a difference between languages and referential practice than a difference between reference to objects within personal space, as opposed to reference to objects within a more extended interactional space.)

70

David P. Wilkins

These examples should be sufficient to justify my main point, which is that choice of demonstrative and choice of accompanying indicative act often vary with respect to one another in a systematic fashion, and this variation is regularly independent of any of the physical determinants of the context itself. That is, the same extralinguistic context can be described by different utterances, each of which combines a different lexical demonstrative with a different non-linguistic act. (I am not, of course, intending to imply that each communicative act is under the same construal of the situation. Quite the contrary, I presume each communicative act suggests a different construal, but not one that is going to be obvious from the mere physical aspects of the extralinguistic context – a point which Asli has regularly stressed.) In a very real sense, we are talking about differences in “referential practice”, to borrow a term from Hanks (although with slightly different implications). Hanks (1996) himself, in talking about the Mayan terminal deictics aʔ ‘immediate (proximity)’ and oʔ ‘non-immediate (proximity)’ has noted that they regularly cooccur with different gesture patterns. He writes (1996: 250): All aʔ forms are associated with high-focus gestures, such as extending the referent in the hand, touching or pointing to it with directed gaze, all of which imply that the Spkr. is in a relation of contiguity with the object. ... All oʔ forms are associated with relatively less focal gestures, such as a vague toss of the hand or a less ostentatiously directed point. In many cases, there is no gesture at all. These are forms used to make references to objects in the Adr.’s zone or in the common ground.

Such observations are critical to helping us understand the use and semantics of the linguistic forms. However, Hanks goes on to show how these combinations of gesture and linguistic form regularly apply to quite distinct contexts (prototypical contexts of use), but he does not indicate whether they may both apply within the one physical context, referring to the same object, in order to convey distinct construals of the situation. This is not meant to be a criticism, but an observation which may help to clarify distinct aims and questions. What are the cooccurrence patterns of demonstrative form and indexical act? Are different pairings of linguistic and non-linguistic demonstrative associated with demonstrably distinct contexts? Can different pairings be used to make reference to the same object within the one context? If so, does the selection of one pairing versus another entail a different construal of the situation? If so, how can we demonstrate that? To conclude, if we don’t pay attention to the pairing of different linguistic forms with different non-linguistic indexical accompaniments, we run the risk of a false or confused analysis of the use and semantics of the linguistic forms. At the very least, a continuing catalogue of examples like the one I’ve started

The Demonstrative Questionnaire

71

above would be useful. Ideally, researchers will be more sensitive to such variations, and will start to identify and describe the nature, semantics, and use of the different forms of indexical accompaniment. Already we have seen that relevant dimensions of variation appear to be: (a) manual pointing versus no pointing; (b) manual pointing versus other means of orienting; (c) pointing without touching versus pointing with touching; (d) distance at which the point is held in relation to the object; and (e) whether object is in the visual attention of the addressee but not of the speaker. What other means and distinctions are available to prevent “deictic terms running blind to their meaning”? References Bühler, K. [1934] (1982). The deictic field of language and deictic words. In R. J. Jarvella & W. Klein, eds., Speech, place and action: Studies in deixis and related topics. Amsterdam: John Wiley, pp. 9–30. (Trans. from Bühler, K. 1934. Sprachtheorie: Die Darstellungsfunktion der Sprache. Jena, Germany: Fischer). Fillmore, C. J. (1982). Towards a descriptive framework for spatial deixis. In R. J. Jarvella & W. Klein, eds., Speech, place and action: Studies in deixis and related topics. New York: John Wiley, pp. 31–60. (1997). Lectures on deixis. Stanford, CA: CSLI Publications. Hanks, W. F. (1992). The indexical ground of deictic reference. In A. Duranti & C. Goodwin, eds., Rethinking context: Language as an interactive phenomenon. Cambridge: Cambridge University Press, pp. 43–76. (1996). Language form and communicative practices. In J. Gumperz, & S. Levinson, eds., Rethinking linguistic relativity. Cambridge: Cambridge University Press, pp. 232–270. Heath, J. (1984). Functional grammar of Nunggubuyu. Canberra: Australian Institute of Aboriginal Studies Ho-min Sohn (1994). Korean. London/New York: Routledge. Kaplan, D. (1989). Demonstratives. In J. Almog, J. Perry, & H. Wettstein, eds., Themes from Kaplan. Oxford: Oxford University Press, pp. 481–563. Kirsner, R. S. (1993). From meaning to message in two theories: Cognitive and Saussurean views of the Modern Dutch demonstratives. Conceptualizations and Mental Processing in Language, 3, 81–114. Wilkins, D. P. & Pederson, E. (1999). Ethnography of pointing questionnaire. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 41–44. (Available at http://fieldmanuals.mpi.nl /volumes/1999/ethnography-of-pointing-questionnaire/.)

3

Lao Demonstrative Determiners Nii4 and Nan4: An Intensionally Discrete Distinction for Extensionally Analogue Space N. J. Enfield1

In this chapter I describe the two-term system of demonstrative determiners in Lao, a Southwestern Tai language spoken by some 15 million people in Laos, Thailand, and Cambodia. This chapter can be thought of as a sister of Enfield (2003a), a study of the system based on observational data much of which was collected on video. Here I shall present results of the Wilkins (1999) Demonstrative Questionnaire (Chapter 2, this volume; see Introduction for background) which, in combination with the observational findings reported in Enfield (2003a), support the following analysis. The demonstrative determiner nii4 is semantically neutral, with no spatial meaning (whether of ‘place’ or ‘distance’) encoded. Nan4 is semantically marked, specifying that the referent is something ‘not here’. This is a spatial meaning, but one of absolute location with respect to a conceptually defined zone rather than scalable distance from a point. Many previous studies (summarised, for example, in Anderson and Keenan, 1985; Diessel, 1999; and Dixon, 2003) report that demonstratives can encode distinctions in metric distance from speaker to referent, physical accessibility to referent, visibility of referent, and attention on referent, among other things. In the Lao case, these factors are indeed crucial to demonstrative selection in certain contexts, yet they are not encoded in the semantics of the demonstratives. Factors of distance and physical accessibility and dynamics of the interaction can determine the conceived extension of spatial perimeters, and thereby affect demonstrative selection. Visibility and attention can affect demonstrative selection through interaction with the special mechanisms of demonstrative reference as a joint achievement by speaker and addressee together.

1

I gratefully acknowledge the support of the Max Planck Society and the European Research Council (ERC grant HSSLU). This chapter has benefitted from comments on an earlier draft by Jürgen Bohnemeyer, Sérgio Meira, and Sarah Cutfield.

72

Lao Demonstratives

73

To establish these claims, it is necessary to begin with some preliminaries on the mechanisms by which interlocutors make spatial deictic reference, after which I give some facts about the larger demonstrative system of Lao and its context within the grammar. I then go on to describe results of the questionnaire, with reference to other observational data. The chapter concludes with a brief summary. 1

Mechanisms of Spatial Deictic Reference

Deictic expressions assume a centre or origo (Bühler, [1934] 1982), in which core semantic parameters of deixis are defined. Bühler ([1934] 1982: 13) suggests that ‘i’, ‘here’, and ‘now’ are simple elements inherent in the origo. To these may be added the addressee ‘you’ and a simple demonstrative meaning which I gloss as ‘dem’ (Enfield, 2003a, 2009). ‘Dem’ is the bare demonstrative meaning, neutral with respect to the domain it applies to. If a lexical item realises this meaning alone (as would an unmarked demonstrative), then it will be available for use in any of the various demonstrative domains (exophoric, anaphoric, discourse, recognitional; Himmelmann, 1996). The element dem will be a component in definitions of more complex demonstratives, as well as related elements such as articles. Let us consider now the mechanisms by which speakers and addressees make deictic reference, concentrating first on the only inherently spatial core deictic element, namely ‘here’ (Wierzbicka, 1996; Goddard and Wierzbicka, 2002). The extensional perimeters of ‘here’ are neither fixed nor semantically specified (as is the case for the extensional perimeters of ‘now’ in time). Semantically, ‘here’ is a place. If I say “The thing is here”, I tell you where it is (i.e. in which place), but not how far away it is – the thing could be millimetres or light years away, and in both cases still be ‘here’.2 This does not mean that distance is irrelevant to the encoding and interpretation of ‘here’ in real contexts. Referents such as one’s own front tooth or a bug on one’s shoulder (see column 1 of Table 3.2, below) are so close to the origo that for them to be encoded as ‘not here’ is unlikely if possible at all. However, more than metric distance comes into the interpretation. In using an intensionally discrete yet extensionally elastic basic deictic element like ‘here’ for speakers to successfully communicate a specific extensional interpretation (e.g. “here in this room” as opposed to “here in this house”), they must be confident that their interlocutors will be able to converge, in the context, on the same interpretation. The extension of the perimeter 2

While the notion of ‘near’ involves a sense of scale (e.g. it can be modified by ‘very’) ‘here’ does not. Further, ‘near’ has a gradable opposite ‘far’ (which cannot be defined as ‘not near’). Thus, while ‘not near’ is not necessarily ‘far’, ‘not here’ is necessarily ‘not here’.

74

N. J. Enfield

intended in the context is not determined by spatial semantics, but by an exclusively interactional aspect of pragmatics called the Schelling Game (Schelling, 1960). In a Schelling Game, players have a ‘coordination problem’ and face the challenge of having to converge on a mutually beneficial solution, yet without the possibility of discussing the problem (nor of reading each other’s minds). Instead, they must rely on assumptions about shared common ground, including what is shared by virtue of being perceptually salient to players in the situational context and what is shared by virtue of common personal experience and/or common cultural membership (Clark, 1996). The use of demonstratives is a striking example of a Schelling Game phenomenon in language (Clark et al., 1983). The simple demonstrative meaning ‘dem’ provides no information as to how a referent object may be located, unlike say a demonstrative with extra semantics marking some specification of distance (e.g. that the referent is something ‘not here’) or attention (e.g. that the referent is something ‘you are not attending to now’). When a demonstrative with the bare meaning ‘dem’ is used, to identify the referent object, the addressee must rely on semiotic resources beyond linguistic semantics, such as indexical or ‘presentational’ gestures (pointing and placing; Clark, 2003) or some other perceptual signal (for instance, the visual prominence of certain candidate referent objects; cf. Clark et al., 1983). When speakers use a demonstrative meaning simply ‘dem’, they assume that the intended referent is one whose identity should be already clear to, or readily determined by, an addressee. Actually determining the identity of the referent is a collaborative and contextdependent achievement (Clark, 1996). Mediating the use and interpretation of deictic and other expressions is the principle of ‘recipient design’, that a speaker ‘designs his utterance in such a way that he has good reason to believe that the addressees can readily and uniquely compute what he meant on the basis of the utterance along with the rest of their common ground’ (Clark et al., 1983: 246; cf. Sacks and Schegloff, 1979). Addressees thus do not expect speakers to say things whose interpretations will require information that speakers do not believe addressees have access to. I assume that all demonstrative elements share ‘dem’ as a core semantic component, and if a language has a truly unmarked demonstrative, then ‘dem’ alone will be its meaning – such a demonstrative will not be restricted in use to any one domain (e.g. ‘exophoric’). In demonstrative systems with more than one term, one term will be unmarked (containing only the core element ‘dem’), and further terms will contain the core element ‘dem’ plus additional semantic information which narrows down the referential search domain (in different ways for different demonstratives). This means that the semantics of different terms in multi-term systems are not equivalent in complexity. Thus, in the Lao system described here, one term has the bare meaning ‘dem’, while its

Lao Demonstratives

75

contrasting term has this meaning plus a specification of ‘where’ the referent is. The standard assumption for two-term systems is that the contrasting semantic specifications express a symmetrical contrast of ‘distance’ (cf., e.g., Lyons, 1977; Anderson and Keenan, 1985; Diessel, 1999). In the terms I am working with here, the two terms would thus mean ‘dem+far’ and ‘dem+near’, respectively.3 In the case of Lao, however, this is inaccurate for two reasons. First, the two terms are not symmetrical in semantic marking (rather, one demonstrative is marked with respect to the other), and second, the semantic value marked is not one of distance (‘near’ versus ‘far’) but of place. In other words, the marked demonstrative says something about where the referent is, not about how far it is. How this discrete intensional specification maps onto analogue extensional space is determined only in context. 2

Lao Demonstratives and the Two-term Determiner System Nii4 versus Nan4

2.1

The Larger Set of Lao Demonstratives

There are five demonstrative elements in Lao, and only two – nii4 and nan4 – are generally used as nominal modifiers (i.e. in simple noun phrases such as ‘this chair’). The remaining three are locative adverbials. None of the demonstratives can freely head noun phrases, and as such are distinct from nominal elements such as pronouns. Table 3.1 shows the set of demonstratives, with a number of distinguishing properties: Table 3.1 Distinguishing properties of Lao demonstratives

Form

Gloss

1. Can appear freely as direct nominal modifier?

nii4 nan4 han5 phii4

‘this’ ‘that’ ‘there’ ‘here’

+ + − −

3

2. Can appear independently as agent or theme NP?

3. Can be compl. of maa2 ‘come’?

4. Can be compl. of paj3 ‘go’?

5. Can appear independently as locative compl. of juu1 ‘be.at’?

− − − −

+ − − +

− − + −

+ − + +

As noted already, ‘far’ and ‘near’ are not true binary opposites – a system with one element positively specified for ‘far’ and another for ‘near’ would be unable to cope with elements which are neither ‘near’ nor ‘far’. This problem does not arise for the ‘here’ versus ‘not here’ analysis suggested here.

76

N. J. Enfield

This chapter concentrates on the two demonstrative determiners, nii4 and nan4. The adverbials phun4 and han5 occasionally occur in what appears to be nominal-modifying function, but which is comparable to English expressions like the book there – i.e. where the adverbial meaning remains. The adverbials are not in paradigmatic opposition to uses of nii4 and nan4 (and accordingly, the questionnaire never elicited any demonstrative other than nii4 or nan4).4 Non-spatial uses of the demonstratives include discourse deixis, recognitional deixis (Himmelmann, 1996; Enfield, 2003b), and frequent uses as de-stressed particles, marking discourse status of participants or propositions they refer to. The demonstrative determiners also occur in anaphoric expressions, but never by themselves (i.e. never without being part of a noun phrase headed by some nominal). These other uses are not considered in this chapter (but see Enfield, 2007: 97–104) – note, however, that in the case of the unmarked demonstrative nii4, I do not consider any of the other uses (except for the de-stressed discourse particle functions) to be extensions or to involve additional semantic content. 2.2

Semantics of the Two-term System

The Lao demonstrative determiner system for spatial reference contains two terms, whose values are, at first glance, distal and proximal. The extent of previous analysis of the semantic distinction between nii4 and nan4 has been simply to gloss the two as ‘this’ and ‘that’, respectively (e.g. Wright, 1994: 40). However, nii4 and nan4 are not semantically symmetrical –nor do they literally encode distinctions of distance (see Enfield, 2003a for full explication of this point; presented in summary form in this section). The ‘proximal’ nii4 is semantically unmarked, encoding nothing more than a basic demonstrative meaning (expressed here as ‘dem’), while nan4 is semantically marked, sharing with nii4 the same basic demonstrative meaning ‘dem’ but in addition also specifying that the referent is something ‘not here’. (1) (a) nii4 = ‘D E M ’ (b) nan4 = ‘D E M + NOT- H E R E ’

The two demonstrative determiners are paradigmatically related (that is, in certain grammatical contexts, they are interchangeable) and form an entailment scale (Levinson, 2000: 79). This means that the meaning of one is wholly contained within the meaning of the other, having consequences with respect to the fact that they are available as alternatives in precisely the same grammatical 4

I am unable to comment in any detail on the semantics of the adverbials – they are of course of great interest, with an extra dimension provided by phun4 ‘yonder’. However, they are beyond the scope of the discussion in this chapter (see Enfield, 2007: 97–101).

Lao Demonstratives

77

context. The ‘stronger’ or more marked of the two – nan4 – encodes the meaning that the referent is something ‘not here’, while the ‘weaker’ or less marked – nii4 – may implicate (but does not entail) the converse, namely that the referent is ‘here’. The notion of a multi-term demonstrative system as consisting of a semantically unmarked term which gives rise to inferences in paradigmatic opposition to semantically marked alternatives was, as far as I am aware, first suggested by Wierzbicka (1980: 27)5 and is explored in many other contributions to the present volume (see introductory chapter for discussion). 3

Results

Responses by informants to the 25 situations illustrated in the Demonstratives Questionnaire are provided in Table 3.2, which shows the scenes distributed across five columns.6 The headings of columns 1–5 in Table 3.2 explain the rationale for the groupings given. The situations illustrated in columns 1 and 5 elicited discrete judgements, which can be interpreted to mean that in the column 5 situations, the referent must be regarded as ‘not here’ (and thus is referred to using nan4), while in the column 1 situations, the referent cannot be regarded as ‘not here’ (and is referred to, by default, using nii4). Consultants’ response to the scenes shown in remaining columns 2–4 – more than two-thirds of the total scenes – was that either nii4 or nan4 could be used (while in most cases one or another was preferred), which means that what determines the preference for nii4 or nan4 in a given case is not depicted in the illustrations. It is not that the semantic distinction between nii4 and nan4 is vague – rather, I maintain that it is absolutely discrete. The apparent vagueness arises from the fact that the extension of the notion ‘here’ assumed by the speaker may change from moment to moment. As a complement to more subtle observations from natural interaction, however, the results of the questionnaire become more informative. 3.1

‘Distal’ Nan4: A Semantic Distinction of Place, Not Distance

The critical factor in a speaker’s decision to use nan4 and encode a referent as ‘not here’ is the conceived extension of their here-space (Enfield, 2003a). The spatial extension of one’s here-space is not inherently defined but rather is determined by the interpretation of space as having places within it. And, 5

6

The semantic explication of a ‘distal’ form as ‘not here’ also originates with Wierzbicka (1980: 27), with reference to English. She regards English this as the unmarked member of the English pair (as does Dixon, 2003: 81), while others have suggested the opposite (Halliday and Hasan, 1976: 59). I conducted the questionnaire in full with six consultants, while many others were consulted opportunistically whenever situations arose which corresponded to scenes in the questionnaire.

S

S

S

(11)

(7)

(6)

(3)

(1)

1. nii 4 only

A

A

A

S

S

S

S

S

S

(8)

(5)

(4)

(2)

A

A

A

A

2. nii 4 preferred, nan 4 possible

S

A

(23)

A

(22)

S

(19)

S

(17)

S

A

A

S

A

S

S

(20)

A

AS

(24)

(21)

S

(16)

S

(14)

(13)

A

A S

S

(25)

(18)

A

(15)

S A

S A

5. nan 4 strongly preferred (nii 4 extremely marginal)

(10)

A

S A

A

S

4. nan 4 preferred, nii 4 possible

(12)

(9)

A

3. nii 4 / nan 4 equally likely

Table 3.2 Distribution of the Lao demonstrative determiners nii4 and nan4 according to native-speaker judgements in response to the questionnaire

Lao Demonstratives

79

crucially, a speaker’s encoding of a given interpretation is mediated by their expectation that the addressee can converge on the intended interpretation. In situated interactions of the kind illustrated in the 25 scenes, an important factor contributing to interlocutors’ potential construals of here-space perimeters is the status of the engagement areas of the speaker and addressee, i.e. the conceived extensions in space of their current manual and attentional engagements (Goffman, 1963: chapter 9; Scheflen, 1976; Kendon, 1977: chapter 5). Also important are the logical possibilities of speakers’ and addressees’ engagement areas overlapping with each other, and/or including the referent object. Manual, interactional and attentional engagements of interlocutors can jointly determine active zones in interactional space (cf. Meira, this volume) which acquire enough salience to determine here-space perimeters mutually apparent to speaker and addressee. When a Lao speaker chooses to refer to something in the physical space as ‘nan4 thing’, he or she is assuming that there is a here-space perimeter salient enough to the addressee to be understood as the one outside of which the referent is to be found. The scenes in column 5 show the referent well and truly ‘not here’ from the speaker’s point of view. In all three scenes (15, 18, and 25), one important point is that the speaker is ‘gravitationally anchored’ by another interlocutor (the addressee in two cases) – that is, the speaker and another person in the scene are right next to each other, allowing them to combine in spatial ‘weight’, encouraging a conceived spatial zone perimeter to gravitate closer to them and create a tighter and more local here-space. This overrides any conceivable engagement area here-space created by the interaction of speaker and addressee (e.g. in scene 18). In the column 4 scenes, it is also clear why a speaker would most likely choose to encode the referent as something ‘not here’. In three cases (scenes 13, 21, and 24), it seems natural to construe speaker and addressee as sharing a here-space, and in each case the referent is outside that space (as for 15 and 25 in column 5). In scenes 10 and 16, on the other hand, the referent is saliently in the addressee’s here-space, and selection of nan4 reveals the speaker’s construal that speaker and addressee do not share the same here-space. This is clearest in scene 16, where the presence of a third participant (next to the speaker), by the gravitational effect just mentioned, helps to make the speaker’s here-space more locally confined and thereby puts the distant referent in a place ‘not here’. Considerations of politeness can also encourage speakers to explicitly mark their non-assumption of sharing a here-space with their addressee (as a speaker might do, for example, by selecting nan4 in scene 10). In the column 3 scenes, interlocutors are as likely as not to interpret the referent as being ‘not here’. In the column 2 situations, the most obvious conceived here-space of the speaker contains the referent, so nan4 becomes an unlikely (but nevertheless still possible) choice. In real situations

80

N. J. Enfield

corresponding to the scenes in columns 2 and 3, what is it that pushes people to make the choice of using nan4, if both forms are possible? As emphasised already, what matters is whether or not the referent is considered by the speaker to be outside his or her here-space and that this is apparent enough to the addressee for them to be able to make the right interpretation. 3.2

Contextual Factors in Selecting between the Two Demonstratives

There are three important contextual factors which can determine a sufficiently salient here-space perimeter: first, real physical barriers in space; second, semi-physical barriers in space determined by culturally conventional ‘closures’, such as doorways or other household borders, or posture of interlocutors (Goffman, 1963; Scheflen, 1976; Hanks, 1990); third, non-physical perimeters created by dynamics of the personal interaction (Kendon, 1977; Enfield, 2003a). Let us consider these in turn. The first important pragmatic parameter contributing to construal of a herespace is that of physical access. I have observed a number of examples in which ‘distal’ nan4 is used for referents which are close to the speaker but out of reach due to the presence of some obstacle. One example collected on video involved a speaker using nan4 to refer to a water outlet on a wall less than 2 metres away – a distance which in itself would by no means exclude nii4 – but where the water outlet itself was out of physical reach of the speaker, since a washing machine and an interlocutor were blocking the narrow path between speaker and referent object (for detailed description, see Enfield, 2003a). In this case, nan4 does not semantically encode ‘lack of physical access to the referent’. The lack of access is a pragmatic factor encouraging the speaker’s judgement that the referent is something ‘not here’. In another example (observed), a customer in a shop was selecting from among a number of small items inside a glass cabinet, and her addressee (the shopkeeper) was behind the cabinet, with full manual access to the items. The speaker identified the item she wanted, and pointed to it, using nan4, despite the item being in full physical view and being within inches of the speaker’s hand (cf. scene 4). In a third case (observed), two people – one much taller in stature than the other – were at an open closet door, looking for some towels. The shorter person was directing the taller person to select particular towels which were on the highest shelf of the cupboard, out of the shorter person’s physical reach. The shorter person repeatedly used nan4 for each reference. The towels being out of her physical reach (in a manner especially obvious to the addressee) resulted in their being saliently interpretable as ‘not here’ from the speaker’s perspective, and they were encoded accordingly by nan4. A second determinant of here-space perimeters is the presence of a kind of ‘barrier’ which does not literally restrict physical or other access but creates

Lao Demonstratives

81

spatial closure by recognised cultural conventions (Goffman, 1963: chapter 12; Scheflen, 1976; Hanks, 1990). Scenes 19–23 clearly aim to probe this kind of distinction, but interestingly in the Lao case, the doorway dividing the house interior from exterior in scenes 19, 22, and 23 did not strongly encourage conception of a corresponding here-space perimeter (which would have resulted in a judgement of nan4 as preferred). The reason for this is that Lao speakers do not generally observe doorways, thin or low walls, etc. as determinants of spatial closure as is conventional in some cultures (e.g. urban USA: Goffman, 1963; Scheflen, 1976). Lao social interaction, whether in house, village, school, rice field or market place, often has a diffuse character, and (rather unlike northern European cultures, for example) anyone within earshot of a conversational interaction can more or less openly claim involvement in it (i.e. one does not have to pretend that one is not listening to what is being said). Thus, the house interior/exterior border between speaker and addressee in scenes 19, 22, and 23 does not strongly affect Lao speakers’ sense of space being divided up. That the referent in scenes 20 and 21 is more likely to be encoded by nan4 is not due to there being a house interior-exterior border between speaker and referent, but rather due to the gravitational closure created by speaker and addressee being very close to each other and away from the referent. Note, however, that there certainly are cases in which culturally established perimeters between ‘places’ can determine here-space extensions relevant to the selection of demonstratives – see section 4 below for discussion of an example in which conventional enclosure of market stalls as distinct places is crucial in demonstrative selection. A third source of here-space perimeter enclosure is the interactionally established engagement area, i.e. the conceived fluid ‘shell’ which encloses areas of attentional and manual engagement (Kendon, 1977; Enfield, 2003a). While engagement area perimeters are often determined by joint attention or manual interaction, such perimeters may also enclose single interlocutors due to ‘auto-engagements’. In one case (recorded on video and described in detail in Enfield, 2003a), a merchant who is standing less than a metre from a customer uses nan4 to refer to an item on the market table immediately between them, within close physical reach – given this spatial layout, speakers would normally prefer nii4, but in this case the speaker was attentionally and manually preoccupied with wrapping one of the customer’s previous purchases. At the moment of speaking, ‘here’ was restricted to where her manual autoengagement was, encompassing only herself and the goods in her hands. In order to test the idea that interactionally established spatial closures are relevant determinants of here-spaces (and thus of the selection of nan4 as opposed to nii4), I devised the following two scenes. Both scenes had the same spatial configuration of speaker, addressee, and referent, where the referent was located close to the addressee. In one scene, the speaker and the addressee were

82

N. J. Enfield

A

S

(a) Addressee engaged in interaction (a ball game) with speaker, focused on referent (a ball): nii 4 preferred.

A

S

(b) Addressee engaged in interaction (a ball game) focused on referent, speaker not involved: nan 4 preferred.

Figure 3.1 Two scenes with no contrast between spacing of speaker, addressee, and referent, but with contrast in the status of the interactional space. (Dotted line represents interactionally defined ‘engagement area’)

together engaged in an activity that was focused on the referent. In the other scene, the addressee was not involved – instead, a third person was engaged with the addressee on an activity that focused on the referent. In Figure 3.1a, speaker and addressee are focused on a joint activity of playing with a ball. The dotted line surrounding them represents the perimeter of their mutually salient engagement area. When consultants were asked to judge whether nii4 or nan4 was better in the frame ‘Is _ ball yours?’, they all agreed that while both were possible, nii4 was preferred. In Figure 3.1b, on the

Lao Demonstratives

83

other hand, the addressee is engaged with a third person, and the speaker is outside of this engagement, and (thus) outside the addressee’s here-space. The referent is saliently ‘here’ for the addressee, and, in addition, the speaker is saliently not sharing that here-space at the relevant moment – thus, the preferred choice of demonstrative is nan4, encoding the idea that the referent is, for the speaker, ‘not here’. 3.3

Unmarked Nii4: When Nan4 is not Applicable

In the scenes shown in columns 1 and 2 of Table 3.2, the choice of nii4 is preferred or obligatory because in the situations illustrated it is difficult or impossible to regard the referent as ‘not here’ (especially, for example, when the speaker is touching or holding the referent object). To use nan4 in these cases, encoding that the referent was something ‘not here’ would most likely result in failure to arrive at joint reference by speaker and addressee. This may be overridden, however, at least in the column 2 cases. I have already discussed in the previous section how contextual factors can give rise to ‘distal’ nan4 being used when ‘proximal’ nii4 would be expected. While the column 4 scenes are likely to be encoded with nan4, as discussed above, there is sometimes no need to encode the idea that the referent is something ‘not here’, despite there being quite a distance between the speaker and the referent. I have witnessed on many occasions speakers using nii4 to refer to things quite far away, sometimes very far away. As already mentioned, social interaction in a Lao village is often diffuse, where speakers and addressees are spaced widely apart, and in such cases everything in the visible space is naturally regarded as ‘here’ (unless a more proximal factor calls for a momentary conceptual division of the space). In one example (recorded on video and discussed in more detail in Enfield, 2003a), several people in diffuse conversation are distributed along a 30-metre section of an irrigation channel, when someone walks along the channel on the opposite side pulling two buffaloes by a rope. Someone else observes that one of the buffaloes is pregnant, and when I (holding the camera) ask for clarification as to which one is being referred to, one of the speakers points at the far one, walking away now some 30 metres or more in the distance, saying ‘nii4 one’. The more diffuse nature of the interaction resulted in a larger and more spread-out here-space, encompassing a number of people. The difference between using nii4 and nan4 in scenes such as those in column 4 can depend on the relative diffuseness of the interaction, as consultants confirm. The claim that nii4 is semantically neutral with regard to the encoding of spatial information may lead one to expect it to fully overlap with uses of the marked spatial term nan4. But, as shown in column 5 of Table 3.2, there are

84

N. J. Enfield

situations in which informants regard nii4 as extremely marginal if possible at all (scenes 15, 18, and 25). The explanation is simply that the contextual conditions in those scenes so strongly suggest that the referent is ‘not here’ (as discussed in section 3.1, above), that nii4 is pragmatically, rather than semantically, inadmissible. This is a complication for the logic of an entailment scale analysis, since it should be possible for an element whose semantics are wholly contained in those of another to be used in all the same contexts. However, in real usage, pragmatic effects can be strong, determining people’s judgements about what is possible at all. 4

Influence of Addressee Location Arising from Schelling Game Pragmatics

The Lao demonstrative system is speaker-oriented in semantic terms, since both the basic demonstrative meaning ‘dem’ and the spatial deictic locus ‘here’ are inherently centred in the deictic origo, the locus of the speaker ‘i’ (Bühler, [1982] 1934). But the location of the addressee can make a crucial difference to speakers’ selection of one or the other demonstrative, in particular when it affects the addressee’s access to relevant information about the scene. Recall that demonstratives pose coordination problems which are solved by Schelling Game type interactions (section 1, above). A marked demonstrative, by virtue of being marked, encodes information additional to the basic demonstrative meaning ‘dem’, giving the addressee the extra demand of having to assess the status of that specification in the context (a demand which is absent in the case of the unmarked term). This predicts, via the principle of recipient design (Schegloff, 1972; Sacks and Schegloff, 1979/2007; Clark, 1996), that when the relevant information is not available to the addressee, the marked demonstrative will not be used, even if it would otherwise have been appropriate. Furthermore, the unmarked demonstrative will be used instead. Information crucial to interpreting the marked term nan4 may sometimes not be mutually available (i.e. it may sometimes not be obvious to both speaker and addressee that the information is available to both). One situation in which this is the case is when the relevant part of the setting (in particular, the exact location of the referent object) is out of the visual field or current attention of the addressee. An example comes from a video recording of a marketplace exchange in Vientiane (for details see Enfield, 2003a). The typical urban Lao market place is a large open area containing many adjoining market stalls of similar size, with goods laid out openly on stall tables. While stalls often have nothing to physically partition them from each other, there is nevertheless some degree of conventional enclosure dividing them (Goffman, 1963: 151), such that they each may be regarded as self-contained places. In this example, a customer

Lao Demonstratives Perimeter of market stall as ‘place’ — if this defined current HEREspace, speaker would use nii 4.

85

SPKR Possible HERE-space division within market stall place — if this were active, speaker would use nan 4 to refer to the referent object, encoding the idea that it is construed as ‘not here’.

REFERENT ADDR

Figure 3.2 Hypothetical scenario, with speaker (stall attendant) and addressee (customer) on either side of a market stall table, both contained within the space defined by the market stall. Division of here-space between speaker and referent is possible, and given accessibility to addressee of the relevant information, nan4 may be used.

approaches a stall, picks up a cabbage from the table, and asks the stall attendant standing across the table ‘How much is nii4 one?’ (a typical use of nii4 as ‘proximal’). Now it is the stall attendant’s turn to reply and tell the customer the price of the cabbage. In semi-structured field elicitation sessions (see Enfield, 2003a), I have described this scene to consultants and asked them which demonstrative would be appropriate in responding to the customer at this point. They say that the stall attendant could use either nii4 or nan4. (The spatial layout is as in Fig. 3.2, similar to scenes [9] and [10] in columns 3 and 4 of Table 3.2.) With speaker, addressee, and referent located fully within the conventional enclosure of a single place – the market stall space – there is no clearly primary division of the space: the speaker has the option to regard the whole space as ‘here’ (including both speaker and addressee), or to subdivide that space by construing a here-space perimeter between herself and the referent (which is in the customer’s hand). Now, returning to our example, what in fact happens next is that the stall attendant does not know the price of the cabbage, and has to turn around and call out to the real owner of the stall, an older woman who is standing some 15 metres behind and away, engaged in conversation with a neighbouring stallowner. While this older woman has visual access to the referent object, she has not been monitoring the interactional dynamics within the speaker’s immediate area (the market stall), so she is not aware on what basis a here-space perimeter could be construed by the speaker as subdividing the market stall space itself. The extension of here-space most saliently available to the addressee coincides with the speaker’s ‘place where I am now’ – namely, the market stall as a whole (see Figure 3.3).

86

N. J. Enfield

SPKR

ADDR

REFERENT

HERE-space division most salient for addressee-and-speaker, given that addressee has not been involved in or monitoring the interaction within the market stall place where the speaker is.

Figure 3.3 Scenario from video recording (Enfield, 2003a), with speaker (stall attendant) at one market table, and addressee (older woman) at another market stall some 15 metres away. While division of here-space between speaker and referent is possible, the information relevant for the speaker’s having made such a division (e.g. due to interactional dynamics) is not accessible to addressee – thus, due to the principle of ‘recipient design’ (see text), nan4 may not be used.

The speaker (the stall attendant) thus has no choice but to use nii4 to refer to the cabbage (in the customer’s hand), since to have used nan4, given her interlocutor’s most likely construal of the speaker’s here-space, would have led to the interpretation that the referent was ‘not at this market stall’. That nan4 was impossible here is confirmed by consultants – but in another context with the same spatial layout (i.e. referent some 2 metres away from speaker and in another person’s hand), nan4 would certainly be possible. Further examples provide additional evidence. In one video-recorded example, the addressee is seated on a kitchen floor engaged in food preparation, and a knife sits on the floor, unknown to her, just behind her back. The speaker is walking around the kitchen, and when she spots the knife, she says ‘Is nii4 knife the one we took to the rice fields?’. At the moment of speech, the knife was physically closer to the addressee than to the speaker (similar to scenes [9] and [10]), yet consultants agree that ‘distal’ nan4 would have been highly inappropriate if possible at all. This is because the addressee is known not to have access to relevant information which she would need in order to figure out why the speaker had encoded it as something ‘not here’. By contrast, nii4, being semantically unmarked with respect to spatial distinctions, makes no such demand. Because the speaker knows this, and because both interlocutors assume a maxim of recipient design, by which addressees expect utterances to

Lao Demonstratives

87

have been designed for them such that they will not have to refer to information they do not have access to, then marked nan4 is ruled out in this case.7 In sum, the fact that demonstrative reference is essentially a kind of Schelling Game can make the addressee a pivotal player in a speaker’s selection between demonstratives, even where there is no semantically marked addressee-anchoring in the demonstratives themselves. 5

Conclusion

The two-term demonstrative determiner system of Lao appears at first glance to encode a simple spatial distinction ‘proximal’ versus ‘distal’. Upon closer inspection, however, it emerges that while both terms share the basic demonstrative meaning ‘dem’, neither term semantically encodes the idea that the referent is at a certain distance (i.e. ‘far’ versus ‘near’), and only one encodes the idea that the referent is in a certain spatial location (i.e. nan4 with the semantic component ‘not here’). No spatial distinction of any kind is semantically encoded by the unmarked demonstrative nii4. A range of contextdependent factors, including metric distance to the referent, visibility of the referent, physical access to the referent, attention on the referent, location of the addressee, demands for politeness between interlocutors, culturally conventional spatial closures, and interactional activation of engagement zone perimeters, can all affect the selection of demonstratives in consistent and principled ways, yet without any of them being encoded in the semantics. There is no reason to doubt that factors such as these could be semantically encoded by demonstratives in other languages, as has been widely reported, but data of the kind presented in this chapter suggest that such reports must be verified with much more sensitive investigation than has been the standard to date. The approach we have taken here is to use multiple methods to get at the meaning and function of this system. In an earlier study (Enfield, 2003a), data from video recordings of repeated in situ uses of Lao demonstratives provided initial evidence for the analysis presented here. The Demonstratives Questionnaire method, foregrounded in this book, not only provided new evidence in support of the same analysis but also helped to sharpen our statements about where the distinctions in the system lie. This is especially true in that, as this chapter has demonstrated, the questionnaire method allows us to design scenarios to test language-specific hypotheses raised by naturalistic observation of demonstrative usage. 7

This suggests an additional scene to be included in the questionnaire – namely, a situation in which the referent is close to the addressee, within the speaker’s sight, but out of the addressee’s sight (i.e. scene 11 with the ‘spkr’ and ‘addr’ roles reversed).

88

N. J. Enfield

The analysis offered here attributes a pivotal role to the conception of places in space. Extensional facts about spatial settings of discourse are not encoded in linguistic meaning, but they do help to determine interlocutors’ construals of the meaning that physical space can inherit. Demonstratives bring in the interactional component of deixis, often adding spatial semantics as well, all the while presuming intensionally discrete places within extensionally analogue space. References Anderson, S. R. & Keenan, E. L. (1985). Deixis. In T. Shopen, ed., Language typology and syntactic description. Vol. III. Cambridge: Cambridge University Press, pp. 259–308. Bühler, K. ([1934] 1982). The deictic field of language and deictic words. In R. J. Jarvella & W. Klein, eds., Speech, place, and action. New York: John Wiley and Sons, pp. 9–30. (Translated excerpt from Sprachtheorie: Die Darstellungsfunktion der Sprache, 1934.) Clark, H. H. (1996). Using language. Cambridge: Cambridge University Press. (2003). Pointing and placing. In S. Kita, ed., Pointing: Where language, culture, and cognition meet. Mahwah, NJ: Erlbaum, pp. 243–268. Clark, H. H., Schreuder, R. & Buttrick, S. (1983). Common ground and the understanding of demonstrative reference. Journal of verbal learning and verbal behaviour, 22, 245–258. Diessel, H. (1999). Demonstratives: Form, function, and grammaticalisation. Amsterdam/Philadelphia: John Benjamins. Dixon, R. M. W. (2003). Demonstratives: A cross-linguistic typology. Studies in Language, 27(1), 61–112. Enfield, N. J. (2003a). Demonstratives in space and interaction: Data from Lao speakers and implications for semantic analysis. Language, 79(1), 82–117. (2003b). The definition of WHAT-d’you-call-it: Semantics and pragmatics of ‘recognitional deixis’. Journal of Pragmatics, 35(1), 101–117 (2007). A grammar of Lao. Berlin: Mouton de Gruyter. (2009). The anatomy of meaning: Speech, gesture, and composite utterances. Cambridge: Cambridge University Press. Goddard, C. & Wierzbicka, A., eds. (2002). Meaning and universal grammar. Amsterdam: John Benjamins. Goffman, E. (1963). Behaviour in public places. New York: The Free Press. Halliday, M. A. K. & Hasan, R. (1976). Cohesion in English. London: Longman. Hanks, W. F. (1990). Referential practice: Language and lived space among the Maya. Chicago: University of Chicago Press. Himmelmann, N. P. (1996). Demonstratives in narrative discourse: A taxonomy of universal uses. In B. Fox, ed., Studies in anaphora. Amsterdam: John Benjamins, pp. 205–254. Kendon, A. (1977). Studies in the behaviour of social interaction. Bloomington: Indiana University / Lisse: Peter de Ridder.

Lao Demonstratives

89

Levinson, S. C. (2000). Presumptive meanings: The theory of generalized conversational implicature. Cambridge, MA/London: MIT Press. Lyons, J. (1977). Semantics. Cambridge: Cambridge University Press. Sacks, H. & Schegloff, E. A. (1979/2007). Two preferences in the organization of reference to persons in conversation and their interaction. In N. J. Enfield & T. Stivers, eds., Person reference in interaction: Linguistic, cultural, and social perspectives. Cambridge: Cambridge University Press. (Originally published in Everyday language: Studies in ethnomethodology, ed. G. Psathas, 1979. New York: Irvington, pp. 15–21.) Scheflen, A. E. (1976). Human territories: How we behave in space-time. Englewood Cliffs, NJ: Prentice-Hall. (Written with Norman Ashcraft.) Schegloff, E. A. (1972). Notes on a conversational practice: Formulating place. In D. Sudnow, ed., Studies in social interaction. New York: The Free Press, pp. 75–119. Schelling, T. C. (1960). The strategy of conflict. Cambridge, MA: Harvard University Press. Wierzbicka, A. (1980). Lingua mentalis. Sydney: Academic Press. (1996). Semantics: Primes and universals. Oxford: Oxford University Press. Wilkins, D. P. (1999; this volume). The 1999 demonstrative questionnaire: ‘This’ and ‘that’ in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24. Wright, P. S. (1994). A Lao grammar for language learners. Bangkok: Thammasat University. (Special issue of Journal of Language and Linguistics.)

4

Dalabon Exophoric Uses of Demonstratives Sarah Cutfield1

1

Introduction

In this chapter I describe the four demonstratives used in exophoric reference in Dalabon, a non-Pama-Nyungan language of Northern Australia. To summarize the main findings of this chapter in advance, the Dalabon demonstrative paradigm divides into spatial and non-spatial sets, both of which are used in non-contrastive exophoric reference. The data reveal a strong preference to index a referent as being located in the speaker’s herespace, and to use the non-spatial demonstratives with distal referents, instead of the semantically distal form. These practices generate two types of implicature (scalar and conversational), which result in the non-spatial demonstratives gaining implied distal meanings. Demonstrative theorists (e.g. Halliday and Hasan, 1976; Lyons, 1977; Fillmore, 1982; Anderson and Keenan, 1985; Diessel, 1999) have argued that spatial reference is primary with respect to other demonstrative reference (e.g. discourse, recognitional). The Dalabon data suggest that it is not spatial reference which is primary per se, but rather, referring to a referent via the speaker’s here-space which is primary. This chapter also describes a typologically under-reported phenomenon of using a non-spatial demonstrative to index a referent as located in interactional ‘no-man’s-land’. The analysis presented here is developed from data collected using the Demonstratives Questionnaire (Wilkins, 1999a), and supplemented by data from contrastive use elicitations and in-situ observation. Central to interpreting language in context is the recognition that language use involves the speaker and addressee working together to solve the co-ordination

1

This analysis is a revised version of work presented in my Ph.D. and was developed with the financial support of an ELDP Individual Graduate Studentship fieldwork grant and a postdoctoral fellowship at the Max Planck Institute for Psycholinguistics, funded by the HUSSLE project. I am grateful to the Dalabon speakers whose data is presented in this article, to my Ph. D. supervisors Heather Bowe and Anna Margetts for their guidance, to Nick Enfield and Gunter Senft for their comments on an earlier version of this chapter. I alone am responsible for any remaining imperfections in this analysis.

90

Dalabon Demonstratives

91

problem that is communication (Schelling, 1960; Lewis, 1969; Clark, 1996). Demonstratives are one referential device used to solve communication problems, as they index referents in situational space, or in the realm of the interlocutor’s common ground. In order to describe the use of the Dalabon demonstratives in situational space, I will first define some terms which are employed here to describe how interlocutors make spatial deictic reference. The here-space (Enfield, 2003: 89) is a ‘conceptually defined area’ meaning ‘where I am now’ (cf. Wierzbicka, 1996). The here-space is specified for location, but not distance. Its boundaries are largely determined by social interaction, and indexed by the use of deictics such as demonstratives. Physical obstacles and social boundaries can influence the location of the here-space boundary. An engagement area (Enfield, 2003) can also contribute to the construction of a social boundary which may then be interpreted as a here-space boundary. An interlocutor’s engagement area is ‘the place which is, at moment t, the conceived site of a person’s currently dominant manual and attentional engagement’ (Enfield, 2003: 89). For example, a speaker engaged in rolling tobacco into a cigarette can be understood to have an engagement area which involves themselves, the tobacco and cigarette paper. A referent outside this engagement area – though still proximal to the speaker – may be referred to with a non-proximal demonstrative on account of the speaker recognizing a here-space boundary between their engagement area and the referent. The speaker-addressee shared space refers to the speaker and addressee’s shared interactional space, which may also create a here-space boundary around them. Depending on the arrangement of the speaker and addressee and the distance between them, their shared space could be a tight circle, or a distant oblong. In this discussion, the space outside the speaker’s here-space is called their there-space. It is bordered only by the here-space and extends infinitely outwards in every direction away from the speaker’s here-space. The four demonstratives surveyed divide into spatial and non-spatial sets. The spatial demonstratives nunda and djakih are the here-space and there-space forms, respectively. The non-spatial demonstratives kanh ‘that (identifiable)’ and nunh ‘that (unfamiliar)’ are used to indicate that a referent can more readily be recognized according to its accessibility. In section 2, I provide a brief typological and sociolinguistic overview of Dalabon, before introducing the demonstrative paradigm in section 3. Section 4 details the questionnaire methodology and results. Data from contrastive and emotionally deictic uses of the demonstratives is considered in section 5. Heuristics and implicatures arising from the analysis are presented in section 6, before the final discussion and conclusions in section 7.

92

Sarah Cutfield

2

The Language and Its Speakers

Dalabon (Gunwinyguan, non-Pama-Nyungan) is a severely endangered language from southwestern Arnhem Land, Northern Australia (see Figure 4.1).2 It is closely related to the Bininj Kunwok dialect chain, which neighbours Dalabon territory to the north. Other neighbouring languages Rembarrnga (east), Ngalakan (south-east) and Jawoyn (south-west) are also members of the Gunwinyguan language family. There remain perhaps three fluent speakers of Dalabon, who live in different communities. On the whole, Dalabon people have shifted to speaking Kriol in southern communities and Bininj Kunwok in the north. Dalabon is a polysynthetic language, with grammatical relations and the tense/aspect/mood of the clause indexed on the verb. 3

Description of Form Classes and Previous Claims

In this study, I focus on the set of demonstratives in Dalabon which have nominal distribution (i.e. they can occur as the head of a phrase, or as a modifier of a head) and are used in non-contrastive exophoric reference to entities. Therefore, deictics with only adverbial distribution are excluded (e.g. nidjarra ‘here’, kanihdja ‘there’). The nominal demonstratives3 in Dalabon function in all four of Diessel’s (1999) demonstrative distributions (pronominal, adnominal, adverbial, identificational).4 Demonstratives refer only to entities in the pronominal, adnominal and identificational distributions, and as such, only these distributions are included in the analysis.5 The Dalabon demonstratives attested in non-contrastive exophoric use are listed in Table 4.1.6 They are presented together with definitions which preview the analysis developed in section 4, as well as previous definitions from the Dalabon dictionary (Evans et al., 2004).7 Dalabon demonstratives are classified here as adnominal when they either immediately precede or follow their semantic head; bear no role-marking morphology; and occur in the same phrase contour as their semantic head 2

3 4

5 6 7

This map is based on the contributions of many researchers, which were collated by Mark Harvey, School of Humanities & Social Science, University of Newcastle, Callaghan NSW 2308, Australia ([email protected]). Henceforth ‘demonstratives’. Some of the Dalabon demonstratives are also attested in the grammaticalized functions of relative pronoun, complementizer, conjunction and interjection. See Cutfield (2012: 135–153) for further discussion and examples. See Cutfield (2012: 127–130) for discussion of the adverbial uses of Dalabon demonstratives. There is a contrastive form nihdja ‘that contrastive (distal)’, which is not attested in noncontrastive use. It is discussed in section 5.1. Evans (p.c. 2003) does not suggest that Evans et al. (2004) present conclusive definitions for the Dalabon demonstratives. Evans’ belief that the Dalabon demonstratives could be considerably further investigated led to the study presented in Cutfield (2012; this volume), and elsewhere.

Figure 4.1 Approximate boundaries of languages of the Top End of the Northern Territory (based on Harvey, 2009)

Sarah Cutfield

94

Table 4.1 Dalabon demonstratives used in non-contrastive exophoric reference Demonstrative

Gloss

Definition

Evans et al. (2004)

nunda djakih kanh~kanunh* nunh

D.here D.there D.ID D.UNF

‘this (in the here-space)’ ‘that (in the there-space)’ ‘that (identifiable)’ ‘that (unfamiliar)’

‘this (presentative)’ ‘that there (with you)’ ‘that’ ‘this (presentative)’, this one’

* These are unstressed and stressed variants of the same form. I use kanh throughout the discussion.

(including across intonation breaks).8 The pre-nominal demonstrative in example (1) serves to illustrate their function in this position as deictic modifiers. The post-nominal demonstratives function as emphatic indexicals (as in: that very one) and are often translated as ‘now’ colloquially. In example (1) the speaker is demonstrating how to dye string to two people of Bulanjdjan subsection (i.e. share the same kinship set). (1)9,10

46:22 nunda yarl-no nga-h-lng-burlhkeyhwo-n de-h– D.here string-NO 1sg-H-SEQ-get.out-PRES 2opp-H-(trunc) ‘this string I’m taking out you two –.’ 46:24 nurra-h-na-n Bulanjdjan burrkunh / 2du>O-H-look-PRES subsection.name two (rise) ‘you two look you two Bulanjdjan.’ (MT, DWOL_Rough_Cuts_1)

8 9

10

These are typically indicated by a hesitation pause, with no restart in the following intonation phrase. The examples are presented one intonation phrase per line. The speaker’s initials are given in brackets at the end of the example, followed by the source media file name. The start timecode for each intonation phrase is given to the left of the line of text. Abbreviations used: – Morpheme boundary, = Clitic boundary (intra-word), – Truncation (interruption, hesitation), Continuing intonation contour, ? Appeal contour, \ Final falling pitch contour, / Final rising pitch contour, = Lengthening (of final segment), Uncertain hearing, X Indecipherable syllable, [ ] Concurrent gesture, > [subject] acting upon [object], 1 1st person (exclusive), 12 1st person (inclusive), 2 2nd person, 3 3rd person, APP Applicative, BEN Benefactive (applicative), DAT Dative, D(EM) Demonstrative, D.here Here-space demonstrative, D.there There-space demonstrative, D.ID Identifiable demonstrative, dis Disharmonic, du Dual, D.UNF Unfamiliar demonstrative, ERG ergative, FEM Feminine (nominal prefix), FUT Future, GEN Genitive, H Realis/Assertive marker, INST Instrumental, INTER Interrogative, IRR Irrealis, LH Left Hand, LOC Locative, MASC Masculine (nominal prefix), NEG Negative, NMZR Nominalizer, -NO -no nominal suffix, O Object, PCUST Past customary, PI Past imperfective, pl Plural, PP Past perfective, PRES Present, R Realis, REDUP Reduplication, REL relative pronoun, RH Right Hand, RR Reflexive / reciprocal, sg Singular, S Subject, SEQ Sequential, SUB Subordinate, trunc Truncated (word or intonation phrase).

Dalabon Demonstratives

95

Demonstratives function as the head of a phrase (pronominal) when they occur in an intonation phrase without a co-referential nominal (2); are noncontiguous with or apposed to a co-referential nominal; bear role-marking morphology (3); or are set apart from a co-referential nominal by prosody (i.e. dislocated). Example (2) is from a personal narrative: the speaker’s aunt is announcing that the speaker has reached marriageable age. Example (3) is an elicited sentence referring to a snake bite. (2) 13:06 djongok-ngan ka-h-yi-ninj ngale nunda– aunt-1sgPOSS 3-H-say-PI alright D.here(trunc) ‘my aunt said “alright this one –”’. 13:08 nunda ka-lng-rarri-minj D.here 3-SEQ-grow.up-PI ’this one has grown up.’ (MT, Tukumba_Oral_History) (3) 21:20 kanh-yih ka-h-ba-nginj D.ID-ERG 3sg-H-bite-PI ‘that one bit me.’ (LB, SC071119N60_BES_01&02LBQB)

Identificational demonstratives occur in copula and non-verbal clauses (Diessel, 1999). In Dalabon, demonstratives occur in two kinds of non-verbal clause: a copula clause and a nominal predicate. I refer to these as identificational copula demonstrative and identificational predicate demonstrative respectively. The identificational copula demonstrative typically features only the demonstrative and the nominal phrase it indexes, in that order, as in (4). Essentially, the demonstrative is the pronominal subject of the copula clause. The identificational predicate demonstrative has pronominal prefixes characteristic of a predicate (5). Both identificational constructions effectively have copula or equative semantics, i.e. ‘this/that is X’, and typically function to present the referent. (4) 32:45 nunda yabok-ngan djadj-no D.here elder. sister-1sg.POSS digging.stick-3sg.POSS ‘This is my sister’s digging stick.’ (LB, SC071119N60_BES_01&02LBQB) (5) pp. 22 ka-h-nunda yarl-no 3sg-H-D.here string-NO ‘this is the string’ (LB, MK, JG, SCNotebook #1, 24/4/08)

Demonstratives are distinguished from free pronouns on formal and semantic criteria. The free pronouns function only as the head of a nominal phrase and may take role-marking morphology (6). Free pronouns are usually the sole

96

Sarah Cutfield

member of the nominal phrase, but they are occasionally modified by demonstratives (7), (8). Pronominal demonstratives may be used to refer to humans (2) or other animates (3); however, they do not encode person or number semantics.11 (6) 30:52 da-h-yinmowo-ninj kardu njing-yih 2sg>3sg-H-tell-PI maybe 2sg-ERG ‘you must have told her.’ (MT, MPI_DEMQ_MT) (7) 49:34 mm kanunh ngey Mm D.ID 1sg ‘Yes, it was me (lit. that me).’ (MKB, MPI_DEMQ_MK) (8) 20:35 mah nunh njing ? INTER D.UNF 2sg (appeal) ‘What about you? (lit. what about that you?)’ (LB, MPI_DEMQ_LB_01)

Possessive pronominal suffixes are not in morphosyntactic competition with demonstratives and can co-occur (9, 10). However, in the data collected for this study, speakers often made reference to a possessed referent present in the speech situation without the use of a demonstrative (i.e. N-POSS). These tokens were typically accompanied by a deictic gesture. (9) 1:56:22 djakih eyeglass-ngu D.there glasses-2sgPOSS ‘those glasses of yours.’ (LB, MPI_DEMQ_LB_02) (10) 1:55:39 nunh eyeglass-ngu nunda D.UNF glasses-2sgPOSS here12 ‘those glasses of yours here.’ (LB, MPI_DEMQ_LB_02)

Place adverbs are employed to provide specific information about a referent’s location, e.g. (10). These do not compete with demonstratives functionally, as the exophoric demonstratives in Dalabon are not specific with respect to the position of referents. Speakers frequently used adverbs in the Demonstrative Questionnaire. Example (11) demonstrates that even when the question prompt 11

12

Dixon (1980: 276–277, 357) notes that in many Pama-Nyungan Australian languages demonstratives function as third person pronouns (particularly singular). The Dalabon examples (7) and (8) above of demonstratives modifying first and second person pronouns indicate that these forms cannot be understood as having third person reference. The here-space demonstrative, nunda, is functioning as a place adverb in this example.

Dalabon Demonstratives

97

is designed to elicit a demonstrative (‘which?’ not ‘where?’), speakers may choose to use adverbs to index and individuate a referent. (11) 06:57 SC: kumarruh mah nunh bark+wanjkih INTER INTER D.UNF watergrass+like ‘Which one is purple (lit. resembles watergrass)?’ 06.59 MT: nahda walum ka-h-yu that.way south 3sg-H-lie.PRES ‘It’s (the one) lying that way to the south’ 07.01 MT: lirri-kah side-LOC ‘on the side’ (N60_WEE_BALLS-MT)

4

Experimental Results

4.1

Methodology

I conducted the Questionnaire (Wilkins, 1999a; Chapter 2, this volume) with four different speakers. For each of the sessions, I directed the speaker and the Dalabon-speaking addressee into the various positions by scene, as depicted in the questionnaire, and asked them to improvise interactions involving mentions of the referent. Three of the speakers were elderly master-speakers of Dalabon. This presented some methodological challenges I expect are common when attempting to conduct stimuli-response elicitation in the endangered language context (see Cutfield, 2009; 2012: 219–222; also Dunn, this volume). Given the small population of surviving Dalabon speakers, it was not possible to conduct the questionnaire with an ‘ideal’ sample of 10 speakers. Also, given the age of the speakers and the design of the task, it was difficult to conduct the questionnaire under strict experimental conditions, controlling for factors such as drawing the addressee’s attention to the referent for the first time versus subsequent mention, and concurrent pointing. The elderly speakers did not always have the patience for the amount of repetition required in the questionnaire, or for numerous questions about the use of other possible demonstrative forms. This methodology resulted in some inconsistency in the testing for the use of each demonstrative form in each scene. There is also a lack of elicited information on speaker preference for a particular demonstrative in a particular scene. Other gaps resulting from this methodology include a lack of negative evidence of the unacceptability of certain demonstrative forms in particular scenes; and inconsistent testing for the influence of concurrent

Sarah Cutfield

98

gestures (or lack thereof), ownership and givenness of referent, and addressee attention on the referent.13 To compensate for these ‘gaps’, I later conducted the questionnaire with a younger speaker (LB) who was more comfortable with the elicitation-style questions I needed to ask. In recording the questionnaire with LB, I was able to ask in every scene about the use of alternative demonstrative forms, controlling for pointing. As a result, the data and analysis is somewhat biased towards the results from one individual (most of the notes on speaker-preferred forms, negative evidence and comparative use of gesture/ownership/givenness, etc. come from LB’s data) and is supported by corresponding evidence from the older speakers. 4.2

Summary Findings

Each of the nominal demonstratives investigated here can be used to make noncontrastive exophoric reference to entities present in the speech situation. It appears nunda is used to refer to referents in the speaker’s here-space, as well as to make first mention of referents in the here-space, and that djakih is used to refer to referents in the speaker’s there-space. The remaining forms are non-spatially specific: kanh is used to indicate to the addressee that the speaker expects the referent to be identifiable to them, while nunh is used to index a referent as being somehow problematic for the addressee to identify. There is no evident proximal/distal (or medial) symmetry in the exophoric use of the demonstratives. Pointing gestures can co-occur with each of the demonstrative forms and influence the choice of demonstrative. The deictic parameters of discourse status, visibility and ownership of the referent are also found to influence the speaker’s choice of demonstrative in exophoric reference. In almost every scene, speakers could use all four of the demonstratives to refer to the referent. This is taken as an indication that the spatial arrangements of speaker, addressee and referent alone do not predetermine the choice of demonstrative in Dalabon. The non-spatially specific forms in particular had a very wide distribution by scene, with kanh ‘that (identifiable)’ being used in every scene. I argue in section 4.3 below that kanh and nunh ‘that (unfamiliar)’ do not encode location. As such, I have excluded these forms from the comparison of distribution by scene presented in Table 4.2, as this table sets out to compare the use of demonstrative forms with respect to location of the 13

David Wilkins (p.c. 2006) notes that the Demonstrative Questionnaire need not be conducted as a strict experimental task per se, but rather as a checklist for tokens occurring in data collected for other purposes. However, in order to have a data set which most closely resembles the scenes in the questionnaire, and investigates each of the deictic parameters in it, it is convenient to proceed through enacting the scenes with speakers, eliciting demonstrative tokens as well as possible.

S

S

S

(8)

(7)

(6)

(1)

A

Column 1

*djakih (1), (6)

A

A

S

A

A

(21)

S

(20)

S

(13)

S A

S

(9)

(4)

(3)

(2)

A

A

S

A

nunda

S

S

S

(14)

SA

(12)

(11)

(10)

A

A

A

(19)

S

(17)

S

(16)

S

djakih

Column 2

A

A

A

A S

A S

S

A S

(25)

(24)

(23)

A

(22)

Choice determined by perception and salience of here-space / there-space boundary

S

S

Table 4.2 Distribution by scene (Wilkins, 1999a) of the spatially specific Dalabon demonstratives

(5)

*nunda (18)

(18)

A

S

A

Column 3 S

*nunda (15)

Neither form attested

(15)

S A

Column 4

100

Sarah Cutfield

referent. Table 4.2 shows the overlapping and complementary tokens of nunda ‘this (in the here-space)’ and djakih ‘that (in the there-space)’, grouped by scene. Column 1 groups scenes for which only nunda was used. Column 2 groups scenes in which either nunda or djakih could be used. Column 3 groups scenes in which only djakih was used. Column 4 shows the scene in which neither nunda nor djakih was used (scene 15). Starred forms of the demonstratives indicate speakers gave negative evidence for the use of these demonstratives in particular scenes. The data in Table 4.2 show that nunda ‘this (in the here-space)’ has greater distribution than djakih ‘that (in the there-space)’. It is also interesting to note that nunda cannot be used in scenes 15 and 18 where the referent is located at some distance and is not visible (cf. scene 16, where it is visible); and in scene 5, where the referent is located with the (nearby) addressee but is not pointed at (cf. scene 4). These observations are discussed in more depth per the parameters of location (section 4.3), pointing (section 4.4) and visibility (section 4.6). 4.3

Location of the Referent

The two spatial demonstratives encode speaker-anchored locational semantics. The location of the referent is processed through interactional realities such as the speaker’s here-space and engagement area. Nunda indexes a referent as located in the speaker’s here-space and djakih indexes a referent as being located in the speaker’s there-space. These spatial demonstratives do not encode location of the referent relative to the addressee or a third person, geographical features such as height or up/down-hill/river, nor do they locate their referent via absolute reference (e.g. compass directions).

4.3.1

Referents in the Here-space Referred to with Nunda

The data discussed in this section suggest that nunda is used to index a referent as inside the speaker’s here-space. Situational factors that contribute to the speaker’s construction of their here-space include proximity of the referent to the speaker, the speaker’s engagement area, speaker-addressee shared space and physical obstacles in the speech situation. For referents very close to the speaker (as in scenes 1, 3 and 6), there was a clear preference for using nunda to refer to them. Nunda was also used in situations where a proximal referent was equidistant from speaker and addressee, or closer to the addressee (i.e. was ‘at’ the addressee) as long as it could also be construed as being in the here-space of the speaker (e.g. scenes 2, 4, 7, 8, 9, 10). In such instances, the use of nunda with an addressee-located referent is understood as indexing the here-space as including the speaker, the addressee and the referent. A representation of this is given in Figure 4.2. This suggests that when the speaker and addressee have

Dalabon Demonstratives

S

101

A

Figure 4.2 Depiction of scene 9, when the speaker’s here-space equates to speaker-addressee shared space and includes the speaker, addressee and the referent.

a shared space (Enfield, this volume), the speaker may equate this with their here-space. Accordingly, the speaker may assume that the addressee will be able to recognize referents inside their shared space as also being located within the speaker’s here-space. A referent located inside the speaker’s engagement area is also recognized as being within their here-space. This accounts for obvious examples such as scene 8, where the referent is located between the speaker and addressee: their interaction creates an engagement area, and speakers typically use nunda to index this referent. Speakers could also use nunda in more complex situations, such as scene 19. This scene depicts the speaker looking through a window at the addressee inside a room, with the referent also inside the room, close to the window. While the physical boundary presented by the window may be perceived as a boundary to the speaker’s here-space, the speaker is talking to the addressee inside the room, and as such, their engagement area can be understood as the area between them. In data elicited according to scene 22, LB is rolling a cigarette with both hands and is focused on the task. This activity creates an engagement area for LB. She is sitting just inside a doorway, with a referent located just outside the doorway, and the addressee sitting a few metres away outside the door. LB’s interaction with the addressee presents an additional engagement area. Despite being engaged with a different object, and the possible boundary presented by the doorway, LB refers to the referent (a ringing mobile phone) with nunda. In so doing, she indexes it as inside her here-space. It appears that either the proximity of the referent to LB (compared to the addressee) or LB’s interactional engagement area with her addressee is of greater salience to LB at the time of utterance than possible here-space boundaries such as could be

102

Sarah Cutfield

construed from her engagement in an activity, or from the physical boundaries of the doorway. This example demonstrates that engagement areas and physical boundaries do not predetermine the boundaries of a speaker’s here-space, as speakers may find other parameters more salient at the moment of referring (e.g. proximity of the referent, alternative engagement areas). That is, a speaker’s engagement area may contribute to their construction of the boundaries of their here-space, but a speaker having an engagement area does not ipso facto generate a here-space boundary. In these situations just described, speakers prefer to use nunda, but other demonstratives were also used. This suggests that while certain situational factors lead speakers to index a referent as being located in their here-space, these factors do not a priori determine the use of nunda. Additionally, speakers could use nunda to refer to more distally located referents (e.g. scenes 13, 24). This is understood as an effect of speakers extending their here-space. Extensions of the here-space may be influenced by concurrent gesture and focus on the referent, which are discussed in sections 4.4 and 4.5 respectively.

4.3.2

Referents in the There-space Referred to with Djakih

The tokens of djakih in the data suggest that this form is used to index a referent as being located in the speaker’s there-space.14 The data collected for this study present a need to revise Evans et al.’s (2004) definition of djakih as indexing the referent as addressee-located. Some tokens of djakih were indeed used to refer to an object in close proximity to the addressee, such that djakih could be understood as ‘that in the addressee’s personal space’, as in scene 5. However, djakih could also be used in scenes 12 and 14, which depict the referent at equal distance from the addressee and speaker. Understanding the use of djakih as indexing the referent as in the speaker’s there-space (which may or may not include the addressee) allows for a monosemous definition which captures the practical usage of this term. For the ‘there-space’ to exist, a boundary to the speaker’s here-space has to be identifiable. Factors which contribute to the speaker recognizing a boundary to their here-space include the referent being non-proximal to the speaker, the boundary of the speaker’s engagement area, physical boundaries in the speech situation and the addressee’s personal space. Speakers reported that djakih cannot be used to refer to something intimately proximal to the speaker (i.e. in the speaker’s here-space), such as their tooth or eye as in scene 1. Instead, speakers used djakih to refer to non-proximal referents, as in scenes 12, 16, 17 and 18. Proximity of the referent was not a determining factor, however, as speaker-proximal referents located clearly 14

This form is commonly used adverbially to mean ‘there’, though such tokens are excluded from the analysis presented in this chapter.

Dalabon Demonstratives

103

inside the addressee’s personal space could also be indexed with djakih, as in scenes 2 and 4. In scene 4, the speaker points at an insect on the addressee’s shoulder. Speakers could use either nunda or djakih in this scene, which is taken to reflect that the speaker can choose whether or not to interpret the addressee’s personal space as a here-space boundary. However, when the spatial arrangement remains the same but the speaker does not point at the referent (scene 5), speakers could no longer use nunda. For the Dalabon data, this is taken as evidence that the addressee’s personal space strongly encourages the perception of a boundary to the speaker’s here-space, but that they can override this boundary with the use of a pointing gesture to extend their here-space to include a referent in the addressee’s personal space. As I only conducted the questionnaire with speaker-addressee pairs who enjoy familiar social relationships,15 I cannot comment on whether a speaker would be more likely to recognize the addressee’s personal space with an addressee with whom they observe greater social distance.16 Data from scene 19 provided evidence that both physical boundaries and the speaker’s engagement area present potential here-space/there-space boundaries. The physical obstacle of the wall between the speaker and the addressee may be perceived as a here-space/there-space boundary, which may lead the speaker to index the referent as being in their there-space. Alternatively, the physical proximity of the referent to the speaker, and the referent being recognized as inside the engagement area created by the interlocutor’s conversation (as reported in section 4.3.1) may lead the speaker to index it as inside their here-space. The speaker must choose between recognizing a physical boundary as a here-space/there-space boundary, or the referent’s relative closeness to them sufficient as criteria for including it in their here-space. LB reported that she could not use djakih in this scene when she and the addressee were having an established conversation (i.e. had a shared engagement area). It is understood that when the speaker and addressee share an engagement area, the engagement area functions as shared here-space, and that referents inside that engagement area will typically be indexed as inside the here-space. In order to demonstrate how she could use djakih in scene 19, LB stepped away from the window slightly, pretended she was engaged in doing something else, and that her utterance with the token of djakih was her first exchange with the addressee. By stepping away from the window, the physical boundary of the wall becomes 15 16

E.g. husband and wife, classificatory sisters (who are also friends), classificatory mother and daughter (who are also friends). Culturally proscribed relationships (e.g. son-in-law and mother-in-law, adult opposite-sex siblings) and their concomitant respect behaviours (e.g. auxiliary registers, speaking through a third interlocutor, complete avoidance) are well described for Aboriginal communication (see, e.g., Heath et al., 1982).

104

Sarah Cutfield

more salient. LB’s creation of an alternative engagement area also demarcates her exclusive here-space, which did not include the referent or the addressee.

4.3.3

Non-spatial Demonstratives

As mentioned in section 4.2, the demonstratives kanh ‘that (identifiable)’ and nunh ‘that (unfamiliar)’ have very wide exophoric distribution. Both forms can be used in nearly every scene, which suggests they are semantically general with respect to location. They can be used to refer to proximal (scene 1) or distal (scenes 15, 24) objects, or to referents whose location is not evident or relevant (scene 11). Their uses to index a referent as ‘identifiable’ or as ‘unfamiliar’ are discussed in section 4.5. In situational reference, these nonspatial forms are in fact used more commonly to refer to distal referents than the there-space demonstrative, djakih. Additionally, while these non-spatial forms can be used to refer to proximal referents, speakers have a strong preference to index proximal referents with the here-space demonstrative nunda. As a result, the use of the non-spatial demonstratives gives rise to an implicature that the referent is distal from the speaker. This implicature can be defeased, as illustrated by (elicited) examples of speakers using the non-spatial demonstratives to refer to intimately proximal referents (e.g. examples 10, 12). The unfamiliar form, nunh, is also used when the location of the referent relative to the speaker’s here-space/there-space boundary is ambiguous (e.g. in scenes 8–10). This form can be used to index the referent’s location as ‘noman’s-land’, and its use allows speakers to avoid having to solve the coordination problem of committing to indexing the referent’s location with respect to their here-space boundary. This is discussed further in sections 4.5–4.7. 4.4

Deictic Gestures

The prototypical pointing gesture is described by Kita (2003: 1) as: ‘a communicative body movement that projects a vector from a body part. This vector indicates a certain direction, location or object.’ After McNeill (1992), I use the term deictic gesture as a hypernym to refer to those gestures Kita defines as pointing gestures. This includes lip points, eyebrow points, gaze, etc. I reserve the term pointing for manual points only. Concurrent deictic gestures were frequently attested in the questionnaire data (81 per cent of demonstrative tokens). The Dalabon data show that when a deictic gesture affects the dynamics of the interactional space, then it affects the choice of demonstrative. There is a strong preference for using deictic gestures with here-space referents, and perhaps by implication, a greater likelihood of constructing a referent as in the here-space when making a concurrent deictic gesture, and in particular, a point.

Dalabon Demonstratives

105

Pointing at a referent appears to have the effect of extending the speaker’s herespace in the direction of that referent. For example, scenes 4 and 5 both depict the speaker referring to a referent, e.g. a fly, on the addressee’s shoulder, while the speaker and addressee are close to one another. In scene 4 the speaker is pointing to the fly with her finger, whereas in scene 5 the speaker is not pointing but instead gazing at the referent. Speakers used djakih, kanh and nunh in both scenes, but the here-space demonstrative nunda was only used in scene 4, with a manual point. There was also a clear preference for using nunda in scene 4, over the other demonstratives. This supports the view that concurrent pointing contributes to the referent being indexed in the speaker’s here-space. Concordantly, there is also a correlation between lack of manual pointing and the use of non-spatial demonstratives: the non-spatial demonstratives kanh and nunh tend to occur when there is no deictic gesture, or when the deictic gesture is not a manual point (e.g. scene 5 as just discussed). This suggests that the nonspatial demonstratives are more presupposing. These observations are demonstrated in example (12) below. This data from scene 1 features a change from no deictic gesture and the unfamiliar demonstrative nunh to the here-space demonstrative nunda when the speaker restarts the utterance and adds a manual point. (12) 14:05 nunh mumu-ngan kanh – D.UNF eye-1sgPOSS REL (trunc) ‘that eye of mine which –’ 14:06 [nunda da-h-nana-n] D. here 2sg>3-H-see.REDUP-PRES ‘this one you’re looking at.’ [LH is pursed (thumb, index finger and middle finger tips touching) and raised to left eye, and held there. Stroke: dah] (LB, MPI_DEMQ_LB_01)

The Dalabon data suggest that speakers can use deictic gestures (and, pointing in particular) to override potential here-space boundaries, as presented by the addressee’s personal space (e.g. scenes 4, 5, 8–10), physical boundaries (scene 19) or distance (scene 14, 25). That is, the act of pointing extends the speaker’s here-space. In this way, deictic gestures affect the interactional space, and affect the choice of demonstrative. There is also a strong correlation between use of a deictic gesture and the first mention of a referent. LB reports that it is not possible to make first mention of a proximal referent with nunda (as in scene 2), without a manual point. When given a constraint of not manually pointing at the addressee’s eye in this scene, LB claims to prefer to use the non-spatial kanh instead, with concurrent gaze and optional head nod. However, shortly after this discussion, she used nunda with accompanying eye gaze in scene 2 to refer to the now-

106

Sarah Cutfield

established referent. This example – while an exception to the pattern of speakers using manual points with the here-space demonstrative – also demonstrates that manual pointing is not obligatory with every use of nunda. 4.5

Discourse Status

The questionnaire data on discourse status in the speech situation is less straightforward to interpret compared to the locational and gestural data. In the experimental set-up, the speaker and addressee both knew of the referent’s existence and location. Additionally, the referent was the focus of many prior elicitations. As such, it was not always apparent whether a given demonstrative token was to be interpreted as a first or subsequent mention, or whether it was functioning to draw the addressee’s attention to the referent. Having identified some methodological challenges in obtaining relevant data, it was nonetheless possible in some two-thirds of cases to classify first versus subsequent mentions, and to observe some general tendencies in the data. The here-space demonstrative nunda is typically used to make first mention of a (proximal) referent, while the identifiable demonstrative kanh is typically used to make reference to established or otherwise identifiable referents. While these tendencies can be observed, in fact, both nunda and kanh can be used for either first or subsequent reference. This is understood in terms of indexing versus grounding. When speakers use nunda in either first and/or subsequent mention, the speaker is drawing attention to (indexing) the referent’s location in the here-space, for example the speaker repeatedly using nunda while talking about and handling a referent. The use of kanh indicates that the speaker assumes the referent is grounded in the discourse, either by means of having been already mentioned, or by virtue of being evident in the speech situation. That is, ‘common ground’ (rather than ‘mentionedness’ per se) seems to influence the choice of demonstrative: the here-space demonstrative indicates the referent’s location is in the common ground, while the identifiable demonstrative assumes the referent is in the common ground, but is not specific by which means. By contrast, the ‘unfamiliar’ demonstrative nunh is used when the referent is not grounded in the discourse, or, is not in the ‘common ground’. This lack of grounding may come about by one of two means: either as a result of an unsuccessful first mention with nunda or kanh; or as a result of the location of the referent not being self-evident. The first case is demonstrated in example (13), with data from scene 14. The speaker makes first mention of the referent with kanh and a manual point. In so doing, she attempts to construe the referent as grounded. The addressee indicates this attempt has failed, prompting the speaker to again refer, this time with nunh.

Dalabon Demonstratives

107

(13) 47:49 LB: [barnada-h-na-n kanh mobile phone kanihdja INTER 2sg>3-H-see- PRES D.ID mobile.phone there bulhdjarn] in.the.middle ‘Can you see that mobile phone there in the middle?’ [Raises extended right arm and points with right index finger at referent. Stroke: kanh, held till end of utterance] 47:52 ND: [kahke mak nga-h-n-iyan nga-h-kabburla-no] nothing NEG 1sg>3-H-see-FUT 1sg-H-blind-NO ‘No, I can’t see it, my eyesight is poor (lit. I’m blind).’ [Gazes in direction of LB’s point] 48:01 LB: [nunh-bay] D.UNF-?towards ‘that one.’ [Raises extended right arm and points with right index finger at referent. Stroke: nunh, held past end of utterance] (LB, MPI_DEMQ_LB_03)

The second case is illustrated by instances of speakers using nunh when the location of the referent relative to the here-space/there-space boundary is ambiguous. When speakers used a manual point in scene 9, they typically used nunda to index the referent, as the referent was unambiguously constructed as being inside their here-space. When directed not to use a point in this scene, the location of the referent relative to a here-space/there-space boundary was no longer obvious to the speaker. Speakers reported not being able to use the spatially specific demonstratives nunda or djakih in this scenario. The referent is in plain view, and proximal to both interlocutors, so it is not the case that the speaker estimates that the addressee won’t be able to recognize the referent. Instead, the referent appears to be located in ‘no-man’s-land’ with respect to interactional space, which precludes the use of spatially specific demonstratives. By using nunh in such cases, speakers are indexing the referent as ‘not readily grounded in interactional space’. This is explored further in section 4.7. 4.6

Visibility

Visibility is not a semantically encoded feature of the Dalabon demonstrative paradigm, with each of the demonstratives being used to refer to the non-visible referent in scene 11. However, demonstratives are not a preferred strategy when referring to non-visible referents, with speakers typically using adverbs or bare nouns in the first instance. Nonetheless, the data generated by the questionnaire does offer evidence of the use of the demonstratives with non-visible referents.

108

Sarah Cutfield

When referring to non-visible things, speakers choose between drawing attention to the referent’s location and appealing to the addressee being able to identify the referent without knowing its location. When speakers did use a demonstrative to refer to a non-visible referent, there was a preference to employ the latter strategy, using the identifiable demonstrative, kanh. This practice of not drawing attention to the non-visible referent’s location has culturally specific politeness implications, for the addressee’s positive and negative face.17 Making requests for objects owned by the addressee is usually done with the identifiable demonstrative kanh. This is simultaneously a positive and negative politeness strategy in Dalabon. The use of kanh indexes a relationship of solidarity between the speaker and addressee: ‘I am assuming we have enough common ground for you to be able to identify the intended referent, and because we share common ground, we are close enough that I can ask you for this thing.’ This appeals to the addressee’s positive face wants. In order to appeal to the addressee’s negative face wants to not be imposed upon, speakers can use the identifiable demonstrative as part of a culturally appropriate strategy to deny an off-record request for access to the addressee’s (non-visible) possession. In using the identifiable demonstrative to not specifically locate the referent, the speaker is allowing the addressee to pretend to ‘fail’ to recognize the intended referent (and thereby deny the request, without having to deliver an explicit ‘no’, which would not be culturally appropriate). Drawing attention to the location of a non-visible referent (e.g. scene 11) is most commonly done with nunda ‘this in the here-space’ or nunh ‘that (unfamiliar)’. Both of the spatial demonstratives nunda and djakih were used to locate the non-visible referent with concurrent deictic gestures: nunda with a manual point (common); djakih with an open- handed deictic gesture (infrequent). Nunda is used for non-visible referents whose (proximal) physical location is highly salient; e.g. in naturally occurring data, LB fumbles for an asthma inhaler which is in her handbag and out of sight, all the while referring to it with nunda. This form could also be used with distal and non-visible referents as in scene 25, but not the non-visible referent in scenes 15 and 18. By contrast, the unfamiliar demonstrative nunh seems to be used to draw attention to a difficult-to-locate non-visible referent. QB uses nunh to make first mention of the referent in example (14) which is located immediately behind the addressee’s back. This example comes from scene 11, with the interlocutor roles reversed. (14) 1:42:38 manjh-kenh mah nunh – INTER INTER D.UNF (trunc) ‘What’s that –.’ 17

This is discussed further in section 4.7.

Dalabon Demonstratives

109

1:42:39 dolku-ngu-kah back-2sgPOSS-LOC ‘at your back?’ (QB, MPI_DEMQ_LB_02)

In this situation, the speaker and the addressee have a co-ordination problem in establishing the referent in the discourse. The addressee cannot see the referent, nor does she have a clear view of the speaker, so QB is unable to refer to the referent with a spatial demonstrative or a deictic gesture. The speaker is indicating she is not familiar with the referent (what’s that?), and as such, as she is unable to provide identifying information to the addressee. Therefore, she cannot use the identifiable demonstrative to appeal to the addressee’s ability to identify the referent. This is compounded by her not being able to point to the referent and accounts for her use of the adverbial phrase ‘at your back’ to locate the referent. In this instance of reference to a non-visible referent, nunh is used to signal to the addressee that the interlocutors have a co-ordination problem in locating and grounding the referent. 4.7

Ownership

Ownership of a referent is not a semantically encoded feature of demonstrative semantics in Dalabon, nor is possessive morphology in formal competition with the demonstratives (as seen in line 14:05 of example 12). However, ownership of the referent can be shown to affect the speaker’s conceptualization of their herespace, which in turn affects their choice of demonstrative. In scenes which present the addressee’s personal space as a potential here-space boundary (scenes 8–10), the here-space demonstrative nunda is used when the speaker owns the referent, while the unfamiliar demonstrative nunh is used when the addressee owns the referent. That is, when the speaker owns the referent, they can ignore the addressee’s personal space as a potential boundary to their here-space; when the addressee owns the referent, the speaker is more likely to index the referent as ‘problematic to index with respect to my here-space boundary’. This phenomenon is particularly apparent in examples where the ownership of the referent is being contested by the speaker. It is important to first point out that in cases where ownership of the referent is not known or relevant, speakers refer as per the referent’s location (as described in section 4.3), which may also be influenced by concurrent deictic gestures (as described in section 4.4). When claiming that the referent in scenes 8–10 belongs to them, speakers are more likely to use the here-space demonstrative and a concurrent point. That is, owning a referent makes speakers more likely to construe it as being located inside their here-space. When the addressee is the owner of the referent, the unfamiliar demonstrative can be used to construe the referent as problematic with

110

Sarah Cutfield

respect to being inside or outside the speaker’s here-space – and ultimately avoids the issue. Speakers also used the unfamiliar demonstrative nunh when questioning (rather than asserting or contesting) ownership of the referent, as in example (15) from scene 9 data. By using the unfamiliar demonstrative in this example, the speaker is avoiding presupposing both (a) that the referent is hers (which she could do by using the here-space demonstrative) and (b) that the referent is definitely not hers (which she could do by using the there-space demonstrative djakih to index it as outside her here-space). (15) 01:09:41 [nunh aykladj ngey-kun] D.UNF glasses 1sg-DAT ‘Are those glasses mine?’ [gazes at referent] (LB, MPI_DEMQ_LB_02)

5

Additional Data

In this section, I consider additional data on the interactional uses of Dalabon demonstratives, in particular: contrastive uses and emotionally deictic uses. 5.1

Contrastive Uses

In addition to the Demonstratives Questionnaire (Wilkins, 1999a), I conducted contrastive elicitations with Dalabon speakers (both of my own design, and those designed by other researchers, e.g. Pederson and Wilkins, 1996; Levinson, 1999; Wilkins, 1999b). In these elicitations, two or three objects were placed in front of the speaker in table-top space scale, as depicted schematically in Figure 4.3. Table 4.3 shows the typical referring patterns for two referents (Sets 1–2) and three referents (Sets 3–5), when reference is first made to Referent #1, then proceeds to Referent #2 (then #3). The sets are listed in order of ‘naturalness’ and observed speaker preference, i.e. Set 1 was consistently produced as the first token, without any prompting to control for demonstrative choice or concurrent gesture. Set 2 was typically produced when speakers were asked

Figure 4.3 Arrangement of speaker and two or three referents in table-top space

Dalabon Demonstratives

111

Table 4.3 Contrastive uses of Dalabon demonstratives in table-top space

Set 1 Set 2 Set 3 Set 4 Set 5

Pointing

Referent #1

Referent #2

Referent #3

+ manual point − manual point + manual point − manual point − manual point

nunda nunda nunda nunda nunda

nunda kanh nunda nunda kanh

nunda kanh nihdja

Note: These data are not exhaustive, but rather, reflective of spontaneously used and most common patterns in this data set. For example, both djakih and nunh are attested in the contrastive data, but usually only in response to a request for the speaker to use these forms. A more thorough analysis of the full data set remains a goal for future research.

to not point with their hands. The same applies for Sets 3 and 4, with three referents. Set 5 was sometimes given as an alternative (and less preferred option) to Set 4 and also features the use of the distal contrastive nihdja. All of the tokens with no manual point featured some other deictic gesture at the intended referent, in particular lip-pointing, gaze and raised eyebrows. These were uniformly emphatically employed. A preliminary analysis of these summary contrastive data suggests that the here-space form nunda is the preferred demonstrative for reference to any referent in table-top space scale – regardless of contrast with another referent; that a concurrent deictic gesture is obligatory to disambiguate referents; and that a manual point is the preferred deictic gesture type. Additionally, when speakers are unable to use manual points to disambiguate contrasting referents, they employ verbal strategies, with the identifiable demonstrative kanh being used to refer to the less-proximal member(s) of the set. The contrastive demonstrative nihdja ‘that (distal contrastive)’ only occasionally appears and is used to refer to the most-distal member of a set. That speakers prefer to use nunda to refer to all referents in this distance scale, and that they do not readily employ the there-space demonstrative djakih to index distal referents in the set suggests that none of the referents are located far enough away from the speaker to be considered located in the speaker’s ‘there-space’. The use of the non-spatial kanh and semantically contrastive nihdja instead in such cases also suggests the speaker is using strategies which avoid indexing the referent as distal. Additionally, the use of kanh and nihdja with the lack of a manual point suggests that the speaker can less readily conceive of the more distal members of the set as being located inside their here-space. That is, the contrastive data supports observations made in section 4.4 that manual points help extend the speaker’s here-space, or, conception of what is ‘proximal’ to them (cf. Enfield, 2003: 111–113).

112

Sarah Cutfield

5.2

Emotional Deixis

Cutfield (2012) shows that the Dalabon demonstratives are also used to index the speaker’s stance towards the referent and/or their interlocutor. These emotionally deictic18 uses can be understood as pragmatic effects of their intensional semantics, as described in section 4. The here-space demonstrative nunda is used to index the speaker’s affinity with or positive affect towards the referent; the unfamiliar demonstrative nunh indexes the speaker’s lack of affinity with or negative affect towards the referent; while the identifiable demonstrative kanh indexes a relationship of solidarity between the interlocutors. Examples of these uses can be seen in section 4.7, where speakers use nunda to index their ownership of (= ‘affinity with’) the referent; nunh to index their lack of ownership of (= ‘affinity with’) the referent; and kanh to index their solidarity with the addressee, as part of a positive politeness strategy to ask to use the addressee’s possession. 6

Heuristics and Implicatures in the Paradigm

In this discussion, I have identified two presumptive heuristics (Levinson, 2001) which account for much of the demonstrative usage surveyed here. Each of these heuristics gives rise to an implicature that a referent referred to with a non-spatial demonstratives is non-proximal to the speaker. Heuristic #1: Of the demonstrative forms which can be used to refer to speaker-proximal referents, use the here-space form whenever the speaker’s here-space is readily construed. It is most informative to index a referent as being in the here-space. Implicature #1: This means the here-space demonstrative pre-empts other demonstratives which can be used with proximal referents (i.e. non-spatial demonstratives). Consequently, the non-spatial demonstratives are rarely used with speaker-proximal referents. This is scalar implicature: the non-spatial demonstratives are unmarked with respect to location, and their use is preempted with referents in the speaker’s here-space by the marked (and more informative) here-space demonstrative. As a result of this implicature, the nonspatial demonstratives gain the implied meaning ‘not-here-space’. Heuristic #2: Whenever a non-proximal referent can be co-ordinated to nonspatially, use a non-spatial demonstrative: it is most informative to identify a non-proximal referent by its identifiable/unfamiliar status, instead of its distal location.

18

See Lakoff (1974) for a discussion of emotional deixis in English, and Stivers (2007: 85–89, 95) for a Conversation Analysis of English that as an alternative recognitional used to disassociate the referent from the speaker and addressee.

Dalabon Demonstratives

113

Table 4.4 Spatial pre-emptive heuristics in the Dalabon demonstrative paradigm Location

Pre-empts

Possible

HERE THERE

Here-space Non-spatial

Non-spatial There-space

Implicature #2: This gives rise to a conversational implicature, where frequent use of the non-spatially-specific forms with distal referents can implicate that any referent referred to with a non-spatial demonstrative is distally located. In practice, these heuristics and their resultant implicatures mean that the non-spatial demonstratives are rarely used with proximal referents, and that they are more commonly used with distal referents than the there-space demonstrative. Both of these implicatures are defeasible, however, as the nonspatial demonstratives may be used to refer to speaker-proximal referents. Table 4.4 presents a summary of which demonstrative form(s) either preempts others or is possible according to the referent’s location. An additional heuristic is that when a referent’s location is indeterminate (e.g. it is not visible, or its location with respect to the speaker’s here-space/ there-space boundary is ambiguous), use a non-spatial demonstrative. It is typically not possible to use a spatial demonstrative in these circumstances. In such cases, the non-spatial demonstratives are not pre-empting the use of a spatial demonstrative per se, nor are there implicatures arising from this use. 7

Discussion and Conclusions

This analysis identifies four Dalabon demonstratives which are used in noncontrastive exophoric reference: two forms have spatial semantics, two are non-spatial. The spatial forms index the referent as being located in the speaker’s here-space or there-space respectively. The spatial arrangement of the speech situation does not predetermine whether a referent can be understood as ‘here/proximal’ or ‘there/distal’. The questionnaire data show that the speaker’s here-space is created and influenced by multiple (often cooccurring) situational factors: proximity of the referent to the speaker and its position relative to the addressee; concurrent manual pointing; discourse status; attention-direction; referent visibility and ownership. There is a clear preference in the Dalabon data to construe a referent as being in the speaker’s herespace when the interactional and spatial specifics of the speech situation make this feasible. Additional contrastive and emotionally deictic data further demonstrate speakers’ preference to use the here-space demonstrative when

114

Sarah Cutfield

these contexts permit. This suggests that for Dalabon speakers, locating the referent inside the speaker’s here-space is the primary means to refer to it – it is the most informative way to achieve reference. With non-proximal referents, however, their distal location is apparently not a very informative means of achieving reference. Dalabon speakers prefer to use the non-spatial demonstratives over the there-space demonstrative in such cases. This suggests that as a referent’s location is less readily recognized as ‘here’, the less informative location becomes as a deictic parameter. Instead, the perceived cognitive accessibility of the referent to the addressee (i.e. identifiable, unfamiliar) now becomes the most informative means of achieving reference. In Dalabon demonstrative usage, the here-space and there-space are not equally employed in exophoric reference, i.e. they are not symmetrically opposed. These observations that the here-space is the preferred means through which situationally present referents are referenced in Dalabon and that non-spatial means are preferred to the there-space for non-proximal referents suggest that it is here-space reference which is primary, and not spatial reference more generally. This is a more nuanced description of ‘primary demonstrative reference’ than is typically found in existing space-as-primary claims (see Halliday and Hasan, 1976; Lyons, 1977; Fillmore, 1982; Anderson and Keenan, 1985), and it needs further investigation and corroboration with data on other languages. References Anderson, S. R. & Keenan, E. L. (1985). Deixis. In T. Shopen, ed., Language typology and syntactic description. Cambridge: Cambridge University Press, pp. 259–308. Clark, H. H. (1996). Using language. Cambridge/New York: Cambridge University Press. Cutfield, S. (2009). Bias, elicitation and endangered language description. Paper presented at the International Conference on Language Documentation and Conservation, Honolulu. Cutfield, S. (2012). Demonstratives in Dalabon: A language of south-western Arnhem Land. Doctoral thesis. School of Languages, Cultures and Linguistics, Monash University, Melbourne. Diessel, H. (1999). Demonstratives: form, function and grammaticalization. Amsterdam: John Benjamins. Divon, R. M. W. (1980). The Languages of Australia. Cambridge: Cambridge University Press. Enfield, N. J. (2003). Demonstratives in space and Interaction: Data from Lao speakers and implications for semantic analysis. Language, 79(1), 82–117. Evans, N., Merlan, F. & Tukumba, M. (2004). A first dictionary of Dalabon. Maningrida: Bawinanga Aboriginal Corporation.

Dalabon Demonstratives

115

Fillmore, C. J. (1982). Towards a descriptive framework for spatial deixis. In R. J. Jarvella & W. Klein, eds., Speech, place and action. Norwich: John Wiley & Sons, pp. 31–60. Halliday, M. A. K. & Hasan, R. (1976). Cohesion in English. London: Longman. Harvey, M. (2009). Non-Pama-Nyungan languages: land-language association at colonization. Digital maps and information about non-Pama-Nyungan languages held in the AIATSIS Australian Indigenous Languages Collection (ASEDA no 0802). Canberra: Australian Institute of Aboriginal and Torres Strait Islander Studies. Heath, J., Merlan, F. & Rumsey, A., eds. (1982). The languages of kinship in Aboriginal Australia. Oceania Linguistic Monographs No. 24. Sydney: University of Sydney Press. Kita, S. (2003). Pointing: Where language, culture and cognition meet. Mawah, NJ: Lawrence Erlbaum Associates. Lakoff, R. (1974). Remarks on this and that. In M. W. LaGaly et al., eds., Papers from the tenth regional meeting of the Chicago Linguistic Society. Chicago: Chicago Linguistics Society, pp. 345–356. Levinson, S. C. (1999). Deixis and demonstratives. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 29–40. (2001). Maxim. In A. Duranti, ed., Key terms in language and culture. London: Blackwell, pp. 139–42. Lewis, D. (1969). Convention: A philosophical study. Cambridge, MA: Harvard University Press. Lyons, J. (1977). Semantics. Vol. II (2 vols.). Cambridge: Cambridge University Press. McNeill, D. (1992). Hand and mind: What gestures reveal about thought. Chicago: University of Chicago Press. Pederson, E. & Wilkins, D. P. (1996). A cross-linguistic questionnaire on ‘demonstratives’. In S. C. Levinson, ed., Manual for the 1996 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–11. Schelling, T. C. (1960). The strategy of conflict. Cambridge, MA: Harvard University Press. Stivers, T. (2007). Alternative recognitionals in person reference. In N. J. Enfield & Tanya S., eds., Person reference in interaction: Linguistic, cultural, and social perspectives. Cambridge: Cambridge University Press. pp. 73–96. Wierzbicka, A. (1996). Semantic primes and universals. Oxford/New York: Oxford University Press. Wilkins, D. P. (1999a; this volume). The 1999 demonstrative questionnaire: ‘This’ and ‘that’ in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24. (1999b). Eliciting contrastive use of demonstratives for objects within close personal space (all objects well within arm’s reach). In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 25–28.

5

Brazilian Portuguese: Non-contrastive Exophoric Use of Demonstratives in the Spoken Language Sérgio Meira and Raquel Guirardello-Damian

1

Introduction

Brazilian Portuguese is a Romance language spoken by approximately 200 million people in Brazil, South America. Although still mutually intelligible with European Portuguese (spoken by 10 million people in Portugal), these two varieties, after several centuries of independent evolution, have developed a number of specific features (de Carvalho, 1989; Cruz-Ferreira, 1998; Moraes, 1998; Batoréo, 2000; Naro and Scherre, 2000; 2007; Whitlam, 2011). Portuguese has already received considerable attention from linguists; there are several works on Portuguese grammar in general, among which Cunha and Cintra (1984), Mira Mateus (1989), Hutchinson and Lloyd ([1996] 2003), Bechara (2001), Perini (2002), and (for Brazilian Portuguese) Whitlam (2011). Most studies concentrate on the literary variety of Portuguese, leaving the spoken varieties relatively unstudied. This situation has led to an increasing number of published works on, for example, spoken Brazilian Portuguese (see, among others, the Gramática do Português Falado [Castilho, 1991; 1993] series from the University of Campinas). There are several striking differences between written and spoken Brazilian Portuguese. A well-known example is the replacement of thirdperson accusative pronouns by their nominative counterparts (e.g., Eu vi ele [lit. ‘I saw he’], instead of Eu o vi [lit. ‘I him saw’] ‘I saw him’); as a result of that, third-person pronominal subjects and objects are nowadays distinguished only on the basis of word order (e.g., Ele viu ela ‘He saw her’; Ela viu ele ‘She saw him’). Another striking divergence between the written and spoken varieties of Brazilian Portuguese concerns the demonstrative system, which is the main goal of the present work. In the written language, there are three basic terms: este ‘near speaker’, esse ‘near addressee’, and aquele ‘far from speaker and 116

Brazilian Portuguese Demonstratives

117

addressee’. In the spoken language, however, only two terms are in use (esse and aquele), but this two-term system can be expanded by means of combinations with place adverbs (aqui ‘near speaker’, aí ‘near addressee’, ali and lá ‘far from speaker and addressee’), generating the forms esse aqui, esse aí, aquele ali and aquele lá (and, marginally, esse ali; see sections 3 and 4). In view of the number and complexity of the forms, it is not clear whether they should be analyzed as a two-, a four-, or even a six-term system. Moreover, there is no detailed description of their use and semantics in the available literature, especially in contrast with the forms of the written language. The aim of the present chapter is to address the question of how to analyze the form and meaning of Brazilian Portuguese demonstratives by examining their non-contrastive exophoric use with the help of a standard questionnaire (Wilkins, this volume) developed and used for comparative and typological purposes at the Max Planck Institute for Psycholinguistics. The basic conclusion is that the demonstrative system of spoken Brazilian Portuguese is actually richer than that of its written standard counterpart, with the combined forms (esse aqui, esse aí, etc.) adding up to a six-term system, all usable in exophoric contexts. The system is still speaker-and-addressee based in the sense that the position of both speaker and addressee are important for demonstrative choice, but not in the same way that it was for the original system: rather than an opposition between three domains (“near speaker”, “near addressee”, “far from speaker and addressee”), the modern system actually opposes a “near speaker and addressee area” to a “far from speaker and addressee area”, each further subdivided into two sub-areas (“near speaker”, “near addressee”, “far from both speaker and addressee but readily accessible”, “far from both speaker and addressee and not readily accessible”). In section 2, the formal aspects of the demonstrative system are described in detail, with an overview of the extant literature. In section 3, the methodology of the standard demonstrative questionnaire is introduced, and the results obtained from native speakers of Brazilian Portuguese (including the two authors) are presented and discussed. In section 4, the consequences of the results for the analysis of the demonstrative system are discussed. 2

The Demonstratives

2.1

The System of Written Portuguese

In traditional Portuguese grammars (e.g., Cunha and Cintra, 1984), the demonstrative system is described as having variable and non-variable forms. The variable forms present gender and number distinctions (with the plural forms receiving the suffix -s). The system is speaker-addressee-based, i.e., there is a form to refer to entities that are close to the speaker, another for

118

Sérgio Meira and Raquel Guirardello-Damian

Table 5.1 Written Portuguese demonstratives (after Cunha and Cintra 1984: 328) Variable

this (near speaker) this (near addressee) that (far from both)

Masculine

Feminine

Non-variable

este(s) esse(s) aquele(s)

esta(s) essa(s) aquela(s)

isto isso aquilo

entities close to the addressee, and a third form for entities that are far from both (see Table 5.1). The variable forms can be used as modifiers or pronouns. They agree in gender and number with the noun they modify (when used as modifiers) or with the noun they replace (when used pronominally). They can refer to both animate and inanimate entities: (1) o tio a tia o carro a bola os carros as bolas

-

esse tio essa tia esse carro essa bola esses carros essas bolas

-

esse essa esse essa esses essas

‘the (masc., sg.) uncle – this uncle – this one’ ‘the (fem., sg.) aunt – this aunt – this one’ ‘the (masc., sg.) car – this car – this one’ ‘the (fem., sg.) ball – this ball – this one’ ‘the (masc., pl.) cars – these cars – these ones’ ‘the (fem., pl.) balls – these balls – these ones’

The non-variable forms are used only as pronouns, referring exclusively to inanimate or abstract entities (‘this thing’, ‘this idea’).1 They refer to an object that the speaker cannot or will not identify. They are thus normally used in questions about the identity or categorial membership of the referent: (2) O que é isso? ‘What is this (thing)?’

The variable forms, in contrast, are not very felicitous in such questions. They cannot be used alone (as in 3a), since they indicate that something about the referent’s categorial membership is known (‘this (kind of N)’). Even when modifying a noun, they are somewhat marked in such questions and tend to indicate doubt on the true nature of the referent: ex. (3b) can be used if the intention of the letter is not clear (is it an invitation, a threat, an order?). The combination of a variable form with generic nouns like coisa ‘thing’ (3b) comes close to invariable forms in meaning. 1

If a non-variable demonstrative is used to refer to a human being, the utterance has a pejorative tone. Ex. (2) could be used, for instance, to express contempt for the physical and/or moral condition of a human referent.

Brazilian Portuguese Demonstratives

119

(3) a. *O que é esse/essa? ‘What is this?’ b. O que é essa carta? ‘What is this letter?’ c. O que é essa coisa na sua mão? ‘What is this thing in your hand?’

The demonstratives can also be employed as anaphoric pronouns, referring to participants mentioned in previous discourse. For example: (4) Pedro e Ana cheg-aram. Peter and Anne arrive-3Pl.Pfctv.Past Esta traz-ia um bolo, this(fem) bring-3Sg.Impfctv.Past a(masc) cake e aquele carreg-ava uma garrafa de vinho. and that(masc) carry-3Sg.Impfctv.Past a(fem) bottle of wine ‘Peter and Anne have arrived. This one (she) was bringing a cake, and that one (he) was carrying a bottle of wine.’

Notice that the anaphoric uses of (written, standard) Portuguese demonstratives pattern differently from their exophoric uses: as Câmara Jr. (1971) points out, in discourse the distinction is between reference to a point of the text that is close and reference to a point that is distant. In other words, in discourse there are two “distances” only, near and far. The demonstrative system of written Portuguese adjusts to this fact, working now as if it were a two-term system.2 Thus, in anaphoric uses, we have este/a (or esse/a) to refer to a close point of the text, and aquele/a for the reference to a distant point. 2.2

The System of Spoken Brazilian Portuguese

In spoken Brazilian Portuguese there are also variable and non-variable forms, with distinction of gender and number. However, the demonstratives of spoken Portuguese behave as a two-term system, expressing a distinction between reference to close entities and reference to distant entities. The resulting system, shown in Table 5.2, is equivalent to that in Table 5.1 if the first two rows (‘near speaker’ and ‘near addressee’) had collapsed into one. For some speakers, the erstwhile ‘near speaker’ (este/a) and ‘near addressee’ (esse/a) forms are in free variation; for others (including the authors), the ‘near addressee’ forms are now the norm (Rodrigues, 1978; Pavani, 1987; Castilho, 1993). Thus, in spoken Brazilian Portuguese, esse objeto ‘this object’ usually means that the object is close to the speaker, but it can also mean that it is close to the 2

“a importância da função anafórica está na circunstância de que nela não vigora em português o sistema tripartido. Há em essência uma dicotomia entre ponto próximo e ponto distante no contexto básico. A oposição é então a rigor entre êste e aquêle, com o aparecimento de êsse também, para o ponto próximo, como variante livre” (Câmara Jr., 1971: 328; the old spelling êsse, êste, aquêle with a circumflex was kept unchanged).

120

Sérgio Meira and Raquel Guirardello-Damian

Table 5.2 Spoken Brazilian Portuguese demonstratives (based on Pavani, 1987; Castilho, 1993) Variable

this (near) that (far)

Masculine

Feminine

Non-variable

este(s) / esse(s) aquele(s)

esta(s) / essa(s) aquela(s)

isto / isso aquilo

addressee. When the speaker wants to make clear that the referred object is the one close to the addressee and not to him/her, s/he combines a demonstrative with a place adverb. Portuguese is described as having many place adverbs (Cunha and Cintra, 1984: 539), among which the following four are especially frequent in combinations with demonstratives: aqui ‘here (where the speaker is)’; aí ‘there (where the addressee is)’; ali ‘there (far from speaker and addressee; relatively small distance)’; lá ‘there (far from speaker and addressee; relatively big distance)’. The association of the demonstrative esse with a place adverb allows the speaker to refer to an object with precision: esse objeto aqui ‘this(masc.) object here’ clearly refers to an object near the speaker, while esse objeto aí ‘this(masc.) object there’ points to an object that is necessarily near the addressee. Speakers often employ two kinds of combination: Dem + N + Place Adverb (e.g., essa xícara aqui ‘this(fem.) cup here’), or Dem + Place Adverb (e.g., essa aqui ‘this (one) here’). The combination of demonstratives with place adverbs is observed in spoken Brazilian Portuguese not only for the ‘near’ forms, but also for the ‘far’ ones, as can be seen in Table 5.3 (based on the authors’ native-speaker intuitions prior to the research described in section 3). In sum, the loss of contrast between the demonstratives este and esse (and their feminine and non-variable counterparts) in their exophoric use resulted in a change from a three-term system to a two-term system in spoken Brazilian Portuguese. Câmara Jr. (1971) presents a hypothesis to account for this loss of contrast, according to which the following factors are important: (a) Syntactic: the anaphoric uses of the demonstratives might have interfered in their exophoric uses. As already mentioned, the anaphoric uses are based in a binary opposition: “near” and “far” in discourse. This principle observed in the discursive “space” might have influenced the other uses of the demonstratives, i.e., reference to entities in physical/concrete spaces. (b) Phonological: the forms este and esse are phonologically very similar. It is easy to replace one with the other.

Brazilian Portuguese Demonstratives

121

Table 5.3 Demonstratives in combination with place adverbs (singular forms only) ESSE + place adverbs close to Speaker M (var.) F (var.) Non-var.

esse aqui essa aqui isso aqui

AQUELE + place adverbs

close to Addressee

± far from Sp and Add

far from Sp and Add

esse aí essa aí isso aí

aquele ali aquela ali aquilo ali

aquele lá aquela lá aquilo lá

(c) Semantic: three-term demonstrative systems would be too complex. A twoterm system is simpler and agrees with the pattern observed in other western languages, which is the expression of a “near-far” opposition. In other words, Câmara Jr. proposes that these three kinds of factors – syntactic, phonological, and semantic – may have led to the change that happened in the demonstrative system of spoken Brazilian Portuguese (written Portuguese, being more conservative, still preserves the original system). This hypothesis, however, has serious problems. First of all, characterizing three-term systems as “too complex” is hazardous at best, given the existence of much more complicated systems in other languages (see, e.g., Miyaoka, 1996: 348 for Central Alaskan Yupik); it might have been better to characterize reduction as a widespread tendency within the Romance family (attested, e.g., in French, Italian, Catalan, and Romanian). The phonological argument lacks generality: forms which are equally similar (e.g., casta ‘caste’ and caça ‘(hunting) game’, presta ‘it works’ and pressa ‘haste’) show no sign of becoming homophonous.3 Even the syntactic argument seems unlikely: the contrastive anaphoric uses of esse and aquele mentioned by Câmara Jr. (and described in section 2.1 above) belong to the written standard, and are clearly not obeyed in spoken Brazilian Portuguese (in fact, most Brazilian Portuguese speakers have their first contact with this usage in high school grammar classes).4 If this usage is not typical of 3

4

The only other instance of this st > ss change in the spoken language that we are aware of is the current tendency, in certain dialects, to pronounce the second-person past subjunctive endings -aste, -este, -iste as -asse, -esse, -isse (i.e. tu cantasse, comesse, partisse ‘might you sing, might you eat, might you leave’ instead of cantaste, comeste, partiste). This phenomenon, however, is restricted to verb endings (i.e. words like costa ‘coast’, pista ‘clue, path’, fresta ‘crack’, gaste ‘s/he spends (subjunctive)’, basta ‘enough!’, rosto ‘face’, pasto ‘pasture’, infesta ‘it infects’, arrasta ‘s/he/it drags’, etc. do not have any ss variants). Furthermore, there apparently are dialects in which this phenomenon fails to happen, but in which esse still prevails over este. This is not to say that esse and aquele cannot be used anaphorically in spoken Brazilian Portuguese; they are simply not used to refer back to two participants mentioned in discourse (in which esse and aquele would be roughly equivalent to ‘the latter’ and ‘the former’). Impressionistically, the most frequent anaphoric distinction between esse and aquele seems to

122

Sérgio Meira and Raquel Guirardello-Damian

the spoken language, it is hard to understand how it could have the effect that Câmara Jr. attributes to it. A further implicit assumption in Câmara Jr., and one which most researchers would probably not consider problematic, is that the present demonstrative system of spoken Brazilian Portuguese is based purely on the position of the speaker (a “near-far” or “here-there” opposition, the exophoric equivalent of the contrastive anaphoric uses mentioned by Câmara Jr.). However, as will be seen in the remainder of this work, actual usage is more complicated than that: the position of the addressee still matters for demonstrative choice. 3

Investigating the Exophoric of Demonstratives in Spoken Brazilian Portuguese

3.1

Using the Demonstrative Questionnaire

Other studies of Brazilian Portuguese demonstratives concentrate either on the written language (Carvalho, 1976) or on anaphoric uses (Castilho, 1993). In this study, we report on non-contrastive exophoric uses of demonstratives in spoken Brazilian Portuguese, based on results obtained with the help of Wilkins’ Demonstrative Questionnaire (this volume). Elicitation was carried out in 1999 with the help of six speakers, including the two authors. All speakers were in the 24–55 age range at the time of the data collection; three of them are female and three are male; two of them are from Pernambuco (Northeastern Brazil), two are from Goiás (Central Brazil), and two are from São Paulo (Southeastern Brazil). For each scene in the questionnaire, a real-life situation was presented to the speakers, with the help of visual props (i.e., real objects were placed so as to match the target scene with respect to both speaker and addressee). In every case, there was only one object, so as to avoid contrastive demonstrative uses. To avoid the influence of the written language, all speakers were requested to “forget prescriptive grammar” and give spontaneous answers, as if they were dealing with each scene as a part of their everyday life. For each scene, and for each speaker, a description of the scene was first given by one of the authors; the situation was then set up (the visual props were placed according to the description of the situation), and the speaker interacted with one of the authors (the same who described the scene), acting out the scene as closely as possible.5

5

be ‘that which we are talking about now’ (esse) vs. ‘that which was mentioned in previous conversations, or that you know from personal experience or world knowledge’ (aquele). The investigator leading the interaction – one of the authors – would, for instance, place a coin on the speaker’s shoulder and describe a situation in which an insect was crawling there (with the coin standing for the insect) and the speaker wanted to attract the addressee’s attention to it. The investigator would then pretend to be doing something else – e.g. reading a book – and the

Brazilian Portuguese Demonstratives

123

The speakers’ reactions, judgments, and comments were recorded (in notebooks, for three speakers; on mini-disc, for one speaker) and compared. The authors, also native speakers of spoken Brazilian Portuguese, interacted with each other, alternating their roles (not in one sitting, but on different occasions, separated by several weeks); their reactions and judgments were recorded in notebooks. The final classification of all resulting reactions is shown in Table 5.4.6 3.2

Discussing the Results of the Demonstrative Questionnaire

The data summarized in Table 5.4 suggest the following remarks: (1) For certain scenes, demonstratives and place adverbs are not optimal. Scene 11, in which the speaker refers to an object situated behind his back, was considered “difficult” by all speakers. Although they all agree in using the demonstrative esse, sometimes with the place adverb aqui, they are clearly dissatisfied with this solution, unless obvious pointing or a descriptive phrase (e.g., atrás de mim ‘behind me’) co-occurs. (2) There are no scenes for which simple demonstratives, without accompanying place adverbs, were unanimously considered the best choice. In all cases in which a simple demonstrative was used, demonstratives with place adverbs were at least equally good; in most cases, it was the first spontaneous reaction. (3) The position of both speaker and addressee is relevant for the choice of demonstratives (esse, aquele) and place adverbs (aqui, aí), but not in the same way. The place adverbs aqui and aí clearly correlate with proximity to the speaker and proximity to the addressee, respectively, as can be seen by inspecting the first four columns of Table 5.4. Esse and aquele correlate with distance from both speaker and addressee, taken together as a unit: referents located far from this unit (as in the two last columns of Table 5.4) are described with aquele, while referents close to this unit are described with esse. Note that the distance between speaker and addressee is not relevant, as long as the referent is close to one of them: for scenes 16 and 18, in which the referent was close to the addressee, esse was used, although the distance between them (and consequently between speaker and referent) was as big as the distance between speaker and referent in

6

speaker would then (as s/he had been told to do) try to attract the investigator’s attention by saying something like “look at this bug on my shoulder”. There are, of course, certain problems with this kind of approach. First, although the speakers did try to produce natural reactions, there is always the danger, inherent to elicitation, even when staged, that their answers did not accurately reflect their actual usage. Second, since every individual has his or her own personal way of expressing levels of doubt and certainty, it was necessary to interpret their reactions to each situation, which introduced an inevitable amount of subjectivity. Despite these shortcomings, the results seem consistent enough to be significant.

A

(19)

A

Written Portuguese este sometimes mentioned. When esse (without aqui) is accepted, it is considered better if accompanied by pointing of some kind (2 spkr).

(6)

(3)

Demonstratives are not enough; at least pointing is necessary, often also descriptive PP (‘this one behind me’).

S

(7)

S

A

esse aqui better than or equivalent to esse (6 sp.)

(11)

S

S

esse aqui (6 sp.), esse (1 sp.) + additional clarification

S

A

(1)

S

(8)

(4)

(2)

A

A

A

S

A

(23)

A

(22)

S

esse

Touching makes aqui obligatory. If the referent is closer to speaker, aqui is preferred; if it is closer to addressee, aí becomes better.

esse aqui or esse aí depending on distance or touching (5 sp.); esse also possible (5 sp.)

S

S

S

S

(10)

A

(9)

(5) A

A

(18)

A

S

(16)

S

A

Two speakers accepted (but did not prefer) the combination esse ali for (9) and (10). All 6 sp. preferred esse aí for (10).

esse aí better than or equivalent to esse (6 spkr)

S

S

Table 5.4 Results of the demonstrative questionnaire (total of six consulted speakers)

A

(20)

S

(17)

S

(12)

A

A

For (20), 5 sp. had only esse, 1 also aquele; in (12), 4 sp. had esse and aquele, 2 only aquele; in (17), 5 spkrs had aquele and esse, 1 only esse (+PP).

(?) esse (aqui, aí) (?) aquele (ali, lá)

S

S

aquele

For (25), 4 sp. preferred to add further information (‘the ones over there, on the other side of the mountain’).

For (15), 2 sp. found that descriptive PPs would help. Lá was more frequent in (15), while ali was more frequent in (13) and (21).

(25)

(24)

aquele lá better than aquele ali (6 sp.); aquele possible (2 sp.)

(21)

A S

A S

aquele ali / lá better than or equivalent to aquele (6 sp.)

A

(15)

S A

(13)

S A

Brazilian Portuguese Demonstratives

125

scenes 13 and 15, for which aquele was used (since both speaker and addressee were far from the referent). In this respect, it is interesting to point out that scene 17 represents a difficult case for the system: the referent is far from the speaker and also far from the addressee, but it is also located in the middle of the speaker–addressee axis and so not clearly away from the speaker–addressee unit. (4) Pointing and touching were relevant for deciding between the place adverbs aqui and aí, but not between the demonstratives esse and aquele. For instance, in scene (2), touching (and, for two of the speakers, closely pointing to) the addressee’s body part would make the use of the addresseeanchored adverb aí impossible; pointing from a certain distance, on the other hand, would make the use of the speaker-anchored adverb aqui impossible. The same is true, mutatis mutandis, for scenes 4, 8, 22, and 23 (the third column in Table 5.4). In other words, physical contact (or near contact) with the speaker brings the referent into the speaker’s space, thus rendering impossible the use of addressee-anchored terms.7 (5) There were no scenes for which only one of the ‘distal’ place adverbs ali and lá could be used. Unlike aqui and aí, which correlate with speaker and addressee location respectively, ali and lá were not matched onto anything specific. The difference between them seems to depend on construal: speakers unanimously affirm that ali “feels closer” than lá. They seem to spontaneously suggest contrastive situations (one speaker remarked that, if there were two distant objects, she would refer to the closer one as aquele ali and to the farther one as aquele lá). Speakers very often mentioned, while discussing the situations for which they chose aquele ali, that aquele lá would have been acceptable if the object were “a bit farther away”. Since considerably more distant objects (e.g., in scene 13, as compared to scene 12) could still be described with aquele ali, it is at least possible that the speakers had a contrastive situation in mind. (6) The combination esse ali occurred marginally. Three speakers considered this combination impossible; a fourth one considered it strange, but maybe acceptable; the two last speakers, however, considered it possible, and one of them declared that it was relatively frequent in her own usage. It is possible that this variation is actually dialectal, since the speakers who accepted esse ali are from Central Brazil, while the speakers who considered esse ali odd are from other areas of the country. The Central Brazil speakers accepted esse ali for scenes 9 and 10; one of them considered it 7

We would in principle expect touching or close pointing to also render the distal aquele impossible and the proximal esse obligatory. However, for close pointing or touching to be possible, the referent would have to be sufficiently close to the speaker, which would make esse the best choice anyway. It is not clear whether a situation in which these two factors – proximity to the speaker and touching/close pointing – could be teased apart is at all imaginable.

126

Sérgio Meira and Raquel Guirardello-Damian

possible (though not optimal) for scene 17, while the other one was unsure. A first hypothesis is that esse ali is used for referents that are still within the speaker-and-addressee shared space, but perhaps not so close to speaker and/or addressee for aqui or aí to be the obvious choices (see the end of the next section). Further research on this topic remains necessary.

4

Towards a New Semantic Analysis of Demonstratives and Place Adverbs

The discussion of the results of the Demonstrative Questionnaire in the previous section has certain important consequences for the semantic analysis of the demonstratives of spoken Brazilian Portuguese. (1) The use of demonstratives (instead of descriptive phrases or clauses) pragmatically implicates a ‘normal’situation. The most “normal” situation for demonstrative use is apparently when the referent is located somewhere in the visual field of speaker and addressee, most probably in the direction of the speaker’s gaze. That “normality” is pragmatically implicated rather than asserted can be seen from the fact that demonstratives can still be used in scenes like 11, though not so felicitously. For other scenes in which the referent was not, or hardly, visible, speakers also tended to consider adding further information “more adequate” (e.g., scenes 15, 25). (2) Exophoricity is distributed across the demonstratives (esse, aquele) and the place adverbs (aqui, aí, ali, lá). As was said in remark (2) in section 3.2, there were no clear cases of exophoric use for which simple demonstratives without place adverbs were the optimal choice. Observations from native speakers suggest that the presence of place adverbs actively favors the exophoric interpretation (e.g., two speakers independently remarked that using only esse or aquele would be compatible with an anaphoric interpretation – ‘the one I mentioned before’ – but that this would not be possible if aqui, aí, ali, or lá were added). Consider, for instance, an abstract noun like ilusão ‘illusion’. Anaphoric uses of demonstratives are possible and quite frequent (essa ilusão ‘this illusion [which we are talking about]’, aquela ilusão ‘that illusion [which I had mentioned in some other occasion and am now reminding you of]’). The co-occurrence of a place adverb (e.g., essa ilusão aqui ‘this illusion here’ or aquela ilusão lá ‘that illusion there’), however, sounds rather strange, unless the speaker is somehow attempting a reification of the specific illusion s/he is talking about (s/he may, for instance, be referring negatively to a book that defends a certain viewpoint, or to people in a protest march – both cases of entities that can be situated in space). That is to say, the exophoric “force” of the simple demonstratives esse and

Brazilian Portuguese Demonstratives

127

aquele, although present (since they can be used alone in exophoric contexts, as the answers to the demonstrative questionnaire show), is “weak” enough that the addition of a place adverb is seen as a natural step: a demonstrative plus place adverb combination is a more felicitous or more helpful way of referring to a physical object. (3) The demonstratives constitute a speaker-and-addressee-centered system. The demonstrative esse is used when the referent is close to either the speaker (e.g., scenes 1, 3, 6) or the addressee (e.g., scenes 4, 9, 16, 18). The importance of the addressee becomes clear when one compares scenes 16 and 18 with scenes 13 and 15: the referent is, in all these scenes, equally distant from the speaker but not from the addressee. Thus, the proximity to the addressee in scenes 16 and 18 made esse the only possible choice, while aquele must be used for 13 and 15, in which the referent is distant from the addressee (and thus the speaker). Thus, esse treats the speaker and the addressee as forming one unit, a speaker-and-addressee “common zone”, in which the referent is placed, while aquele situates the referent in the complementary zone, far from the speaker-and-addressee unit. A system of this kind has, to the best of our knowledge, never been reported in the literature. In this respect, it is interesting to point out that scene 17 represents a case in which this system breaks down: the speaker and the addressee are far away from each other, and both are far away from the referent, which is situated halfway between them. This scene constituted a problem for most speakers: it elicited the highest levels of doubt and variation. The occurrence of both esse and aquele in this context, often from the same speaker, suggests that there was doubt as to whether the area between speaker and addressee (the “doubt area” in Figure 5.1) should be considered to belong to the “common zone” or not when they are far away. Speakers apparently have no consistent preferences. Scenes in which the speaker and addressee are clearly closer to each other lead to a clearer connected “common area”, without an intervening

Speaker ESSE

Doubt Area

Addressee ESSE

AQUELE

Figure 5.1 Areas covered by esse and aquele (speaker and addressee relatively far from each other)

128

Sérgio Meira and Raquel Guirardello-Damian

Speaker

Addressee ESSE

AQUELE

Figure 5.2 Areas covered by esse and aquele (speaker and addressee relatively near)

“no-man’s-land”: thus, scenes 6, 22, and 23 elicit no variation in demonstrative choice. This is illustrated in Figure 5.2. The other scenes for which there was variation – 12 and 20 – seem to have more to do with uncertainty as to where the frontier between the speaker–addressee common zone (or esse area) and the complementary aquele area lies. (4) The place adverbs also form a speaker-and-addressee-centered system, but without a “common area”. The adverb aqui clearly indicates proximity to speaker, while aí indicates proximity to addressee (compare aqui scenes 1, 3, 6, 7, 19 with aí scenes 5, 9, 10, 16, 18, and with “doubt” scenes – either aqui or aí, depending on perceived closeness to speaker or to addressee – 2, 4, 8, 22, 23). One can see that aqui and aí partition the speaker–addressee “common area” into a speaker-centered area and an addressee-centered area, which agrees with the remark, made by most speakers, that aqui and aí can co-occur only with esse. The other two place adverbs, ali and lá, are likewise restricted, according to most speakers, to co-occurring with aquele. Although they can also be viewed as partitioning the aquele area into sub-areas, they are not really parallel to aqui and aí. This can already be seen in the answers to the demonstrative questionnaire, in which ali and lá are much more interchangeable: unlike aqui and aí, there were no scenes in which only ali, or only lá, was acceptable. The preference for lá (without, however, excluding ali) in scenes 24 and 25 suggests that ali marks ‘closer’ referents, while lá indicates ‘more distant’ ones. This is compatible with comments by all speakers to the effect that distance is the main difference between ali and lá: ali is ‘closer’, lá is ‘farther away’ (cf. also collocations with the adverbs longe ‘far’ and logo ‘soon, next’: only lá longe ‘far away’ and logo ali ‘nearby, not far from here’ are well formed, while *ali longe and *logo lá are clearly impossible).

Brazilian Portuguese Demonstratives

129

ESSE lá

ali Speaker

Addressee

aqui

aí (?)

AQUELE

Figure 5.3 Areas covered by esse and aquele (speaker and addressee relatively near)

These differences are illustrated in Figure 5.3. Notice that the broken line separating the ali–lá areas marks this opposition as different from the aqui–aí opposition.8 Figure 5.3 provides us with a convenient framework for looking at the exceptional cases: the sequence esse ali, mentioned in the sixth remark of section 3.2 as possibly defining a different dialect, perhaps like the ‘positive anymore’ dialects of American English (see, e.g., Labov, 1991; Murray, 1993; Shields, 1997).9 The occurrences in answers to the demonstrative questionnaire 8

9

One might want to argue that there is still a parallelism between the aqui–aí and the ali–lá oppositions, in that the first element of both pairs marks closer proximity to the speaker, while the second marks more distance from the speaker. Thus, both aqui and ali would be “proximals”, while aí and lá would be “distals”, the main difference being simply that they are restricted to different areas (i.e., aqui is proximal in the esse or speaker-and-addressee area, while ali is proximal in the complementary aquele area). This hypothesis, although very appealing, is problematic: the ‘distal’ term aí is clearly anchored on the addressee, since it cannot be used for referents that are within the speaker’s area, but ‘farther away’. For instance, if a speaker is confronted with two obviously close objects, one of which is closer to him/her than the other, but neither of which is closer to the addressee (the speaker might be looking at two pieces of paper on top of a table, while the addressee is on the other side of the room), s/he simply would not be able to distinguish them by using aqui and aí. If used in this scene, aí would have to mean ‘close to addressee’, i.e., on the other side of the room. The use of ali and lá to distinguish two objects situated in the aquele zone (i.e., far from both speaker and addressee) is not constrained in this way. In other words, the distinction between aqui and ali is based on the opposition between the two ‘poles’ on which these terms are anchored, the speaker and the addressee; for ali and lá, there may be one ‘pole’ for ali, the speaker-and-addressee common area (but note that the referent cannot be “too” close to this area, or else aqui or aí will become more adequate), but there is clearly no such ‘pole’ for lá. One speaker offered a possible context for the (unexpected) sequence aquele aqui. If someone wants to refer to an object (e.g., a suitcase) which had been mentioned at some earlier point in time (‘that suitcase’) and which is now, unexpectedly to the addressee, close to the speaker, s/he

130

Sérgio Meira and Raquel Guirardello-Damian

were few: two speakers accepted (but did not prefer) esse ali for scenes 9 and 10, and one accepted (but again did not prefer) it for scene 17; these are all scenes in which the referent is in the speaker-and-addressee (or esse) area but relatively far from the speaker. Although the data are too scant for any principled claims to be made, a first suggestion is that the speakers who allow the sequence esse ali are using it to “map” the zone in the speaker-andaddressee (‘esse’) area that is not clearly within the aqui or the aí areas (i.e., the region marked with a question mark in Figure 5.3). 5

Conclusion

The analysis that was proposed and argued for in the previous section can be summarized as follows: • Spoken Brazilian Portuguese has a speaker-and-addressee-centered system of exophoric reference. • The two demonstratives esse and aquele locate the referent in two zones: the ‘speaker-and-addressee’ shared space for esse, and the complementary ‘away from speaker-and-addressee’ space for aquele. • These two spaces are further partitioned in two “subspaces”: (1) a speakercentered space, defined by aqui, and (2) an addressee-based space, defined by aí, a ‘nearby space’ (i.e., not far from the speaker-and-addressee space) defined by ali, and a ‘far’ space defined by lá.10 • It is not felicitous to use demonstratives in the “doubt area”, and perhaps in other areas as well (e.g., behind the speaker’s back).

10

could use this sequence (e.g., pega aquela mala aqui! ‘Get that suitcase here!’). The consulted speaker is clearly using aquele with an anaphoric sense and aqui with a demonstrative sense. The other speakers found this combination at least puzzling, and perhaps even uninterpretable; “repair measures” were suggested (e.g., inserting a long pause between aquela mala and aqui), but without much success. This sequence is clearly not conventionalized; it seems to be, at best, an instance of creative usage from the speaker who suggested it. Some results suggest that the distinction here mentioned as ‘nearby’ versus ‘far’ has more to do with conceptual accessibility than with actual or perceived distance: a referent in a position such that it would be “uncomfortable” for the speaker to reach will take the ‘far’ distal lá, while a more “comfortably reachable” position would probably lead to the choice of the ‘nearby’ ali. Consider, for instance, the following example (on which both authors’ native intuitions agree): a researcher inadvertently left an important book in her office. Later at night, already at home, she remembers the book and considers whether or not to drive back to her office to get it. If she thinks it is too much work, she might say to herself something like eu não vou lá no escritório só para pegar um livro ‘I’m not going there [= far, not comfortably reachable] to the office just to get a book’. If, however, she decides that it is better to go, she might say to her husband, concerned about her leaving home so late at night: não se preocupe, eu só vou ali no escritório pegar um livro e já volto ‘don’t worry, I’m just going there [= far, but still comfortably reachable] to the office to get a book and then I’ll be right back’. Note that what changed in the context of the second utterance was not the distance, but the speaker’s desire to reassure her husband by presenting the distance as “comfortable”, “not dangerous”.

Brazilian Portuguese Demonstratives

131

Table 5.5 Proposed analysis of the demonstrative system (singular forms only) LOCAL

Variable masc. fem. Non-var.

NON– LOCAL

Unmarked

S-centered

A-centered

Unmarked

Proximal

Distal

esse essa isso

esse aqui essa aqui isso aqui

esse aí essa aí isso aí

aquele aquela aquilo

aquele ali aquela ali aquilo ali

aquele lá aquela lá aquilo lá

This analysis is summarized in Table 5.5. The term local is proposed for the forms that refer to the speaker-and-addressee (esse) common space, and the term non-local for the forms that refer to the complementary (aquele) space. Likewise, the combinations of demonstratives with place adverbs are termed marked, while the demonstratives by themselves are unmarked. There is therefore a correspondence between syntactic markedness (phrasal structural complexity) and semantico-pragmatic markedness. The system depicted in Table 5.5 is strikingly more complicated than the demonstrative system of standard written Portuguese (cf. Table 5.1). The frequent claims that the merger of este and esse have resulted in an impoverished system can thus be shown to be incorrect: they do not take into account the marked (compound) forms, usually dismissed as colloquial. A final observation is that, as was said before, the marked forms are clearly preferred in exophoric contexts. Yet this does not mean that the unmarked forms are restricted to endophoric uses: esse and aquele, without place adverbs, are indeed not infrequent in exophoric contexts and did occur spontaneously as reactions to many of the scenes in the demonstrative questionnaire. Thus far, however, the available data do not allow a principled description of the conditions for their felicitous use (as opposed to the use of marked forms) in exophoric contexts.11 11

The formal differences, however, do suggest one possibility, which must for the time being be seen as speculative but is perhaps worthy of mention as a suggestion for further research. Place adverbs occur after the noun modified by the demonstrative (esse livro aqui ‘this book here’), the same position where other kinds of modifiers would normally occur (esse livro verde ‘this green book’, aquele livro em cima da mesa ‘that book on top of the table’). Presumably, they were initially simple modifiers, used when the speaker thought that the addressee needed more information; later on, however, they increased in frequency, thus beginning to grammaticalize as overt markers of exophoric reference, especially, at first, in cases in which an old distinction (‘near speaker’ versus ‘near addressee’) was in danger of being lost, given the ongoing merger of este and esse: this distinction was therefore transferred from the demonstratives to the place adverbs aqui and aí. They may eventually fully grammaticalize as such, becoming obligatory in exophoric contexts (like French -ci and -là in ce livre-ci ‘this book (here)’, ce livre-là ‘that book (there)’, presumably the result of a similar grammaticalization process). At the present time,

132

Sérgio Meira and Raquel Guirardello-Damian

References Batoréo, H. J. (2000). Expressão do espaço no português europeu: Contributo psicolinguístico para o estudo da linguagem e cognição. Coimbra: Fundação Calouste Gulbekian. Bechara, E. (2001). Moderna gramática portuguesa. 37th edn. Rio de Janeiro: Lucerna. Câmara, Jr., J. M. (1971). Uma evolução em marcha: A relação entre êste e êsse. In E. Coseriu & W.-D. Stempel, eds., Sprache und Geschichte: Festschrift für Harri Meier. Munich: Wilhelm Fink Verlag. Carvalho de Herculado, J. D. (1976). Systems of deictics in Portuguese. In J. SchmidtRadefeld, ed., Readings in Portuguese linguistics. Amsterdam: John Benjamins, pp. 245–266. Castilho, A. T., ed. (1991). Gramática do Português Falado. Vol. I: A Ordem. Campinas: Editora da Unicamp. Castilho, A. T. (1993). Os mostrativos no Português falado. In Gramática do Português Falado. Vol. III: As Abordagens. Campinas: Editora da Unicamp, pp. 119–147. ed. (1993). Gramática do Português Falado. Vol. III: As Abordagens. Campinas: Editora da Unicamp. Cruz-Ferreira, M. (1998). European Portuguese. In D. Hirst & A. di Cristo, eds., Intonation systems: A survey of twenty languages. Cambridge: Cambridge University Press, pp. 167–178. Cunha, C. & Cintra, L. F. L. (1984). Nova Gramática do português contemporâneo. Lisbon: Edições João Sá da Costa. De Carvalho, J. B. (1989). Phonological conditions on Portuguese clitic placement and syntactic evidence for stress and rhythmical patterns. Linguistics, 27(3), 405–436. Hutchinson, A. P. & Lloyd, J. (2003 [1996]). Portuguese: An essential grammar. 2nd edn. London: Routledge. Labov, W. (1991). The boundaries of grammar: Inter-dialectal reactions to positive anymore. In P. Trudgill & J. Chambers, eds., Dialects of English: Studies in grammatical variation. London: Longman, pp. 273–288. Mira Mateus, M. H. et al. (1989). Gramática da língua portuguesa. 2nd edn. Lisbon: Caminho. Miyaoka, O. (1996). Sketch of Central Alaskan Yupik, an Eskimoan language. In W. C. Sturtevant, ed., Handbook of North American Indians, Vol. 17. Washington, DC: Smithsonian Institution, pp. 325–363. Moraes, J. A. (1998). Brazilian Portuguese. In D. Hirst & A. Di Cristo, eds., Intonation systems: A survey of twenty languages. Cambridge: Cambridge University Press, pp. 179–194. Murray, T. E. (1993). Positive anymore in the Midwest. In T. C. Fraser. ed., Heartland English: Variation and transition in the American Midwest. Tuscaloosa: University of Alabama Press, pp. 173–186.

however, this is still not the case: the Brazilian Portuguese unmarked forms are still possible in exophoric contexts, unlike parallel French forms like ce livre ‘this/that book’. This suggests that Brazilian Portuguese marked demonstratives are at an intermediate stage of grammaticalization, with the place adverbs not quite obligatory yet, but also no longer as pragmatically marked as other modifiers would be.

Brazilian Portuguese Demonstratives

133

Naro, A. N. & Scherre, M. M. P. (2000). Variable concord in Portuguese: The situation in Brazil and Portugal. In J. McWhorter, ed., Language change and language contact in Pidgins and Creoles. Amsterdam: John Benjamins, pp. 235–255. (2007). Origens do português brasileiro. São Paulo: Parábola Editorial. Pavani, S. (1987). Os demonstrativos este, esse, aquele no português culto falado em São Paulo. Master’s thesis, Campinas, Unicamp. Perini, M. A. (2002). Modern Portuguese: A reference grammar. New Haven, CT: Yale University Press. Rodrigues, A. D. (1978). Os demonstrativos em português: Descrição morfológica sincrônica e superficial. Estudos Linguísticos, 1, 64–66. Shields, K. (1997). Positive anymore in Southeastern Pennsylvania. American Speech, 72, 217–220. Whitlam, J. (2011). Modern Brazilian Portuguese grammar. London: Routledge. Wilkins, D. P. (1999; this volume). The 1999 demonstrative questionnaire: ‘This’ and ‘that’ in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24.

6

“See This Sitting One”: Demonstratives and Deictic Classifiers in Goemai Birgit Hellwig1

1

Introduction

This chapter describes the Goemai demonstratives as they are used in noncontrastive exophoric contexts. For comparative purposes, the analysis is based on data that was collected with the help of a demonstrative questionnaire (Wilkins, 1999), supplemented with data from natural discourse. Goemai is a Chadic language (West Chadic A, Angas-Goemai group, Southern branch) that is spoken by approximately 150,000 speakers in Central Nigeria. This chapter arose out of my own research on Goemai, expanding on an earlier analysis of its demonstrative system (Hellwig, 2003: 239–284; see also Hellwig, 2011: 150–159 for a summary of the grammatical properties of demonstratives; and Hellwig, 2010 for a discussion of methodological issues that arose during fieldwork on demonstratives). Information on earlier stages of the language is taken from unpublished manuscripts written by the missionary Eugene Sirlinger (1937; 1942; 1946); and comparative evidence is taken from the closely related language Mupun, described by Zygmunt Frajzyngier (1993). The Goemai demonstratives are interesting from a number of different perspectives: • The demonstratives are used for exophoric reference and do not serve discourse anaphoric functions (with one exception). They can even cooccur with a locative anaphor, a definite article or a specific-indefinite article, suggesting that all four serve different functions. From the perspec-

1

I thank the many Goemai speakers who participated in this research, and especially the following people with whom I had long discussions about Goemai demonstratives: Louis Longpuan, Andreas Shakum, Thomas Longpuan, and Shalyen Mbai Nwang. I am very grateful to the Max Planck Institute for Psycholinguistics and to the Endangered Languages Documentation Programme for funding my research on Goemai. Many thanks to Sérgio Meira and Sarah Cutfield for comments on earlier versions of this chapter. All errors are, of course, my own.

134

Goemai Demonstratives

135

tive of Goemai, the focus of this volume on exophoric uses is thus entirely justified. • The language makes use of a two-term demonstrative system (‘proximal’ versus ‘distal’), whereby the deictic center contains both the speaker and the addressee: any object that is close to either one is considered to be within the proximal zone. This characterization holds true even for situations in which an object is located close to the addressee but far from the speaker. Such systems have not been described in the literature so far; and the only other known example is Brazilian Portuguese (Meira and Guirardello-Damian, this volume). • The demonstratives are morphologically complex forms that also code information about a referent’s posture. This postural information plays an important role for identifying a referent and is arguably more important than the location information itself. The Goemai demonstratives thus present an interesting example of the kind of information that can be coded within a demonstrative word, and of the way different semantic categories interact for the purposes of referring. • And lastly, the origins and further developments of the Goemai demonstratives are transparent and can be traced back to a presentative construction. This diachronic origin explains many aspects of their present-day usage. The chapter is structured as follows: section 2 introduces the formal properties of the demonstratives and their diachronic origins; section 3 illustrates their non-contrastive exophoric uses; section 4 presents additional naturalistic data in order to capture the attention-directing function of the demonstratives; and section 5 concludes the discussion.

2

Demonstratives and Their Diachronic Origins

Goemai demonstratives constitute a form class that is defined by its morphological properties and by its position within the noun phrase. Demonstratives are morphologically complex forms that code information about number, posture and spatial location. The structure of the demonstrative word is shown in Table 6.1. For ease of reference, this chapter uses the forms with the existential classifier (ńd’éńnòe ‘proximal’ and ńd’énáng ‘distal’) as citation forms.2 2

This chapter uses an adapted version of the practical orthography developed by Sirlinger (1937). The following symbols may not be self-explanatory: p’, t’, k’, f’, s’, sh’ = non-aspirated obstruents; b’, d’ = implosives; oe = [ə]; u = [ʉ]; o = [ɔ]. The following abbreviations are used: ADVZ = adverbializer; CL = classifier; DEF = definite article; DIM = diminutive; DEM. DIST = distal demonstrative; DEM.PROX = proximal demonstrative; FOC = focus; GEN = genitive; I = independent pronoun; INTERJ = interjection; INTERR = interrogative; LOC = locative; LOC.ANAPH = locative anaphor; LOG.SP = speaker logophoric; M = masculine; NEG

Birgit Hellwig

136

Table 6.1 The demonstrative word NMLZgòe- ‘SG’ mòe- ‘PL’

ń-

ADVZ-

Deictic classifier-

Deictic root

láng- (SG),

léng- (PL) ‘hang/move’

ńnòe ‘DEM.PROX’

t’óng- (SG), d’yém- (SG), t’ó- (SG), d’é- ‘exist’

t’wót- (PL) ‘sit’ d’yám- (PL) ‘stand’ t’óerép- (PL) ‘lie’

náng ‘DEM.DIST’

These forms function as demonstrative modifiers and as demonstrative pronouns.3 In the modifier, the last three elements (adverbializer, classifier, deictic root) are all obligatory, while the first element (nominalizer) is optional: example (1a) illustrates a modifier without the nominalizer, and (1b) a modifier with the nominalizer. In the pronoun, the nominalizer is obligatory (as in 1c). (1) a. Hòòs ń-d’é-ńnòe lá. tooth ADVZ-CL:exist-DEM.PROX pain(SG) ‘This existing tooth hurts.’ (Q00ADEM) b. Hòòs gòe-ń-d’é-ńnòe lá. tooth NMLZ(SG)-ADVZ-CL:exist-DEM.PROX pain(SG) ‘This existing tooth hurts.’ (Q01JDEM) c. Gòe-ń-d’é-ńnòe lá. NMLZ(SG)-ADVZ-CL:exist-DEM.PROX pain(SG) ‘This existing one hurts.’ (Q01JDEM)

The demonstrative modifier fills a unique slot within the noun phrase. Table 6.2 illustrates the structure of the noun phrase: it has 12 distinct slots, and elements of different slots can co-occur in the order depicted. Of special interest is the fact that demonstratives and other definiteness morphemes fill distinct slots within the noun phrase and can thus co-occur, i.e., a demonstrative in slot 10 can be followed by the locative anaphor ńnòe in slot 11 or the definite article =hok in slot 12; or preceded by the specificindefinite article ńdòe= in slot 4. It is therefore likely that they serve

3

= negation; NMLZ = nominalizer; PL = plural; POSS = possessive; PRES = presentative; S = subject pronoun; SG = singular. Additionally, Goemai has two locative adverbs: b’ák (proximal) and puánáng (distal). The proximal adverb is unrelated to the demonstratives, but the distal adverb resembles the distal deictic root. They probably both developed from the same distal adverb *náng. Note that the adverbs do not code the same spatial distinctions as the demonstratives: adverbs are entirely speaker-centered, while the demonstratives are speaker-and-addressee-centered. Moreover, the adverbs can be used in reference to non-accessible (e.g., invisible) referents, while the demonstratives cannot be used in this case.

Goemai Demonstratives

137

Table 6.2 The structure of the noun phrase Pre-head

Head Post-head

1 2 3 4 5 6 7 8 9 10 11 12

quantifier associative plural diminutive specific-indefinite article noun nominalized verb phrase free possessive pronoun modifying construction nominalized clause demonstrative locative anaphor definite article

different functions. The definite and specific-indefinite articles do not have spatial uses and are thus not discussed further (see Hellwig, 2011: 145–150, 162–165 for their functions and co-occurrences with demonstratives). More relevant for the present purpose is the locative anaphor ńnòe: it is used with all those referents that were previously introduced into discourse by means of demonstratives (see 12b for an example). In example (2), the speaker first introduces a referent with the distal demonstrative ńd’énáng ‘distal’ but then uses a wrong postural verb (t’óng ‘sit’ instead of t’ó ‘lie’). In the second sentence, he corrects this mistake by means of a demonstrative with the postural classifier t’ó- ‘lie’. And since the referent and its location are already known from the first sentence, he adds the locative anaphor ńnòe. ń-d’é-náng zák-yìt t’óng thing ADVZ-CL:exist-DEM.DIST also/however-again sit(SG) k’à ńdè. bì ń-t’ó-náng ńnòe HEAD(SG):GEN one/other thing ADVZ-CL:lie(SG)-DEM.DIST LOC.ANAPH t’óng–, t’ó k’à ńdè. sit(SG) lie(SG) HEAD(SG):GEN one/other ‘That existing thing also sits on another one. That lying thing sits–, lies on another one.’ (Q01JDEM)

(2) Bì

The exophoric and endophoric (including anaphoric, discourse deictic and recognitional) functions in Goemai are thus separated into distinct form classes: the exophoric forms ńd’éńnòe ‘proximal’ and ńd’énáng ‘distal’, and the endophoric form ńnòe ‘locative anaphor’. This formal division is very regular, and there are only two contexts that show any overlap: the anaphoric form can be used to indicate a contrast between spatial referents (see examples 8 and 11 for an illustration); and the demonstrative forms with the existential

138

Birgit Hellwig

classifier are in the process of developing some limited anaphoric uses (see Hellwig, 2011: 150–161 for details). Despite the fact that the demonstratives constitute a unique form class, they bear some resemblance to other forms in the language: the forms of the locative anaphor and of the proximal deictic root are identical; and the demonstrative word as a whole bears a striking resemblance to the presentative construction. These resemblances are not accidental but result from a grammaticalization process. Over the last century, the Goemai demonstrative system has undergone a major reorganization. Evidence for the grammaticalization process, its motivation and the various stages of reanalysis is presented elsewhere (Hellwig, 2003: 273–283). For the present purpose, it suffices to briefly outline the presumed changes, and to point out the resulting consequences. Three stages can be reconstructed: 1. Goemai had a two-term demonstrative system, which was marked tonally on the demonstrative: *nòe (proximal) and *nóe (distal). Traces of this stage are still attested in the speech of some older speakers; and a similar tonal distinction is found in the closely related language Mupun (Frajzyngier, 1993: 83–84). The exact meaning of the two forms can no longer be reconstructed. 2. The tonal distinction was neutralized through the addition of a tone-bearing nominalizing prefix ń-: ńnòe. But the semantic proximal/distal opposition was maintained through recruiting the nominalized presentative construction to take over the function of the distal demonstrative: *(gòe-)ń-d’é náng ‘(the one) which exists there’. Compare the structure of the present-day presentative clause in example (3) to the structure of the demonstrative in example (1a): the presentative particle ná is optional in the presentative construction and not part of the demonstrative word; the presentative prefix ń- was reanalyzed as the adverbializing prefix, the locative verb as the deictic classifier and the locative adverb as the deictic root. (3) Hòòs (ná) ń-d’é b’ák. tooth (PRES) PRES-exist here ‘See the tooth exists here.’ (Q00ADEM)

Note that the verb in the presentative construction has to be a locative verb (Hellwig, 2011: 380–382) – and these locative verbs are identical in form and semantics to the classifiers within the demonstrative word. The existence of this stage can be inferred from Sirlinger’s (1937; 1942; 1946) manuscripts. But again, how these forms were used can no longer be reconstructed, or whether the form ńnòe did have both spatial and anaphoric functions. 3. The proximal demonstrative ńd’éńnòe developed in analogy to the distal demonstrative ńd’énáng. The old (spatial) form ńnòe was retained for anaphoric functions. This is the present stage of the demonstrative system.

Goemai Demonstratives

139

This recent development from a presentative construction has consequences for the use of the demonstratives. It explains why the demonstratives have spatial uses, but not anaphoric or discourse-deictic uses: the presentative construction itself is only used with physical referents in physical space. More specifically, it can be shown that the demonstratives have retained some of the properties of the presentative construction: the attention-directing function, the use with accessible (e.g., visible) referents only and the importance of postural information. 3

Results from the Questionnaire

The data presented in this section were collected with the help of the Wilkins Demonstrative Questionnaire (1999). Each scene was re-enacted with real objects that were placed in real space so as to match the target situation depicted in the questionnaire. The elicitation was carried out with five male speakers of the language (three speakers above the age of 50, two speakers around 20 years of age). The results of the elicitation are shown in Table 6.3. The table is arranged in such a way that scenes that elicited similar responses are grouped together in one column. The arrows under the table indicate which scenes elicited the proximal term ńd’éńnòe, and which elicited the distal term ńd’énáng. The solid line covers those scenes where the demonstrative in question was considered the only or the best response, and the dotted line those where the demonstrative was acceptable under certain conditions. The responses to the scenes in column 3 are variable, which is reflected in the use of both the solid and the dotted lines. Goemai has a two-term demonstrative system: proximal ńd’éńnòe and distal ńd’énáng. Uncontroversial cases are those where the referent is physically close to the speaker and/or the addressee (proximal, in column 1), or where it is clearly beyond their reach (distal, in column 5). Note that it is irrelevant whether the referent is closer to the speaker or closer to the addressee: scenes 7 (proximal to speaker) and 9 (proximal to addressee) both elicit the proximal demonstrative; the distal demonstrative could not be used for scene 9. Goemai speakers thus treat speaker and addressee as a unit, and any referent that is located close to that unit is referred to with the proximal demonstrative. Columns 2, 3 and 4 depict scenes where both proximal and distal demonstratives can be used. They fall into the following categories: (i) Column 2 includes those scenes where the referent is located within reach of the addressee but far from the speaker. The proximal demonstrative still is the preferred term (since the referent is close to the unit of speaker and addressee), but the distal demonstrative is an alternative option (thereby reflecting the considerable distance from the speaker).

S

(7)

(6)

(19)

S

S

(3)

(1)

column 1

A

A

S

S

A

S

A

S

(23)

A

(20)

S

(11)

(8)

A

A

n´d'én´nòe (proximal)

S

S

S

S

S

(9)

(5)

(4)

(2)

A

A

A

A

S

A

(18)

A

S

(16)

S

(10)

column 2

A

A

(22)

S

(17)

S

column 3

A

S

A

n´d'énáng (distal)

(21)

S

(15)

SA

(14)

(13)

SA

column 5

SA

A

(12)

column 4

Table 6.3 Results of the Demonstrative Questionnaire (five consulted speakers)

A S

A S

(25)

(24)

Goemai Demonstratives

141

These three cases are especially interesting because it is here that Goemai differs most strikingly from other two-term systems described in the literature. Instead, it shows similarities to Brazilian Portuguese (Meira and Guirardello-Damian, this volume), whose speakers also distinguish between speaker-and-addressee-anchored proximal and distal demonstratives. In Goemai, the proximal demonstrative is clearly preferred for column 2, but the distal demonstrative is possible, too. In Brazilian Portuguese, by contrast, only the proximal demonstrative is possible – but it has to be augmented with the addressee-anchored adverb. That is, Brazilian Portuguese combines a two-way distinction in the demonstratives (close to speaker and addressee, far from both) with a three-way distinction in the locative adverbs (close to speaker, close to addressee, far from both). Goemai does not have such an additional adverbial resource, and thus cannot mark proximity to the addressee separately from proximity to the speaker. In fact, Goemai speakers have problems reconciling two conflicting parameters: proximal to one participant (addressee) but distant from the other (speaker). This conflict then seems to result in the acceptability of both the proximal and the distal form. Speakers also resort to a second strategy in this case: to avoid the demonstrative altogether (as in 4a), or to augment the distal demonstrative with a descriptive locative phrase that places the referent near the addressee (as in 4b). These possibilities are illustrated with responses to scene 16: the referent (a ball) is near the addressee, and both are far from the speaker. (4) a. Bóól góe-d’è sék b’ét góe ( . . . ) ball NMLZ-exist BODY:GEN BELLY 2SGM.POSS ‘The ball which exists near you ( . . . ).’ (Q00DDEM) b. Bóól ń-d’é-náng bàǹtyèm góe ńnòe ( . . . ). ball ADVZ-CL:exist-DEM.DIST FRONT 2SGM.POSS LOC.ANAPH ‘That existing ball at this your front ( . . . ).’ (Q00ADEM)

(ii) Speakers tend to avoid demonstratives in another type of situation as well: whenever the referent is not visible either to the speaker or to both the speaker and the addressee (scenes 10, 15, 18, 25 and especially scene 11; note that these scenes are not marked separately in Table 6.3). If a demonstrative is used at all, it tends to be used in reference to the visible object behind which the referent is hidden. In example (5), the speaker uses this strategy for scene 25, where the referent (a house) is located behind mountains. góe-d’è ǹk’òng p’áng ń-d’é-náng=hòe ( . . . ). settlement NMLZ-exist BACK:GEN stone ADVZ-CL:exist-DEM.DIST=exactly ‘The house which exists behind that existing mountain ( . . . ).’ (Q00ADEM)

(5) Lú

142

Birgit Hellwig

The results from the demonstrative questionnaire suggest that the relevant parameter is visibility: if a referent is not visible, speakers tend to shift away from the demonstratives. However, there are examples from naturalistic data where demonstratives are used with invisible referents that are accessible to other senses (see section 4). It is thus more likely that the relevant parameter is identifiability or accessibility, not visibility. (iii) Column 3 shows scenes where a referent is located between a speaker and an addressee, but neither one of them can reach it. Speakers accept both the proximal (better in scene 22) and the distal (better in scene 17). But again, they prefer to use a non-demonstrative strategy. Interestingly, speakers of Brazilian Portuguese have similar problems with scene 17. Meira and Guirardello-Damian (this volume) analyze it as a case where the demonstrative system breaks down because the referent is far away from both participants (i.e., it should be distal) but still clearly within the speaker-and-addressee common zone (i.e., it should be proximal). (iv) Column 4 includes scenes where the referent is located out of reach of both the speaker and the addressee. The distal term is the preferred option, but some speakers allow for the proximal term as an alternative. There seems to be some variability as to where to draw the boundary between the speakerand-addressee proximal zone and the corresponding distal zone. The presence of a third point of reference (a third person located further away than the referent) in scene 14 has probably helped to construe the referent as being within the proximal zone. The use of the proximal term with such referents shows that the system is not based on absolute distance. (v) Finally, even though speakers do use the distal demonstrative for the scenes depicted in column 5, they prefer to use the distal adverb instead. This is especially true for objects located in large-scale geographical space, as illustrated in (6), where the referent (a house) is visibly located on top of some mountains (scene 24). That is, the adverb takes over some of the demonstrative function. Note that this is only true for the distal adverb – the proximal adverb is not used in a corresponding way. (6) Lú d’è puánáng. settlement exist there/yonder ‘The house is over there.’ (Q00ADEM)

4

Evidence from Naturalistic Data

In this section, naturalistic data on the use of demonstratives are presented.4 These data confirm the questionnaire results (i.e., the demonstratives code 4

The data were partly collected with the help of stimuli that investigated the attention-directing function of demonstratives (Enfield and Bohnemeyer, 2001; Levinson et al., 2001). Other

Goemai Demonstratives

143

a speaker-and-addressee-anchored proximal versus distal distinction) and add information about aspects that were not investigated with the questionnaire (the use of deictic classifiers and the attention-directing function of the demonstrative word). All points are illustrated below. The speaker-and-addressee anchoring of the proximal and distal demonstratives is also found in the naturalistic data: there are a number of instances where a speaker uses a proximal term to refer to an object located close to the addressee but far from the speaker. For example, the speaker in (7) directs the addressee to open a cover. At first mention, both speaker and addressee are around 3 meters away from the object, and the speaker refers to it with a distal form. The addressee then moves toward it, stands next to it, but does not show any other sign of having identified it as the intended referent. The speaker directs him again, but now he refers to it with a proximal form. (7) Èèp góe-ń-t’óng-náng puánáng=hòe ( . . . ). open(SG) NMLZ(SG)-ADVZ-CL:sit(SG)-DEM.DIST there/yonder=exactly Á’à, gòe-ń-t’óng-ńnòe. no NMLZ(SG)-ADVZ-CL:sit(SG)-DEM.PROX ‘Open that sitting one over there ( . . . ). No, this sitting one.’ (M01ANCOLOR)

In addition, the naturalistic data gives information about the main function of the Goemai demonstratives: an attention-directing and identifying function. As illustrated in example (8), speakers use demonstratives to direct an addressee’s attention to a new spatial referent. Such a referent is invariably introduced into discourse by means of a demonstrative, i.e., by providing spatial information that enables the addressee to identify the intended referent. The participants will keep using demonstratives until this referent is identified. (8) N: Gòe-ń-t’ó-ńnòe

zák-yìt. demonstrative NMLZ(SG)-ADVZ-CL:lie(SG)-DEM.PROX also/however-again ‘Now this lying one.’ [= referent 5] A: ( . . . ) Gòe-ń-d’é-ńnòe=hòe=à? demonstrative NMLZ(SG)-ADVZ-CL:exist-DEM.PROX=exactly=INTERR ‘This existing one?’ [= referent 16] N: Á’à. no ‘No.’ A: Gòe-ńnòe=à? locative anaphor5 NMLZ(SG)-LOC.ANAPH=INTERR ‘This one?’ [= referent 15]

5

sources include video-recorded conversations about surrounding landmarks and flora, and participant observation. The locative anaphor can be used when contrasting spatial referents, e.g., speaker A. uses the anaphor in example (8) to contrast referents 16 and 15.

144

Birgit Hellwig

N: Kwài, gòe-ǹ-t’ó d’ì ǹtyèm góe no NMLZ-ADVZ-lie(SG) LOC.ANAPH FRONT 2SGM.POSS ń-d’é-ńnòe=hòe. ADVZ-CL:exist-DEM.PROX=exactly ‘No, this existing one that lies there in front of you.’ [= referent 5] A: Oh, hâi, à yìm gòe-b’áng. INTERJ INTERJ FOC leaf:GEN NMLZ(SG)-become.red ‘Oh, okay, (it is) a red paper.’ (M01ANCOLOR)

demonstrative

Speakers assume that addressees will be able to unambiguously identify referents on the basis of the spatial information provided. They would therefore use demonstratives only if they judge the referents to be identifiable. Typically, such referents are close to the speaker and the addressee, and they are visible to both. If this is not the case, speakers tend to shift to a non-demonstrative expression (as in examples 4 to 6 above). However, salient referents far away can also be judged to be easily identifiable, and hence trigger the use of a demonstrative form (as in 9a). Similarly, non-visible referents can be identifiable through, for example, auditory information (as in 9b). (9) a. Nà t’éng ń-d’yém-náng puánáng. see tree ADVZ-CL:stand(SG)-DEM.DIST there/yonder ‘See that standing tree over there.’ (Q00ADEM) b. Gùrùm ń-láng-ńnòe á I. person ADVZ-CL:hang/move(SG)-DEM.PROX FOC I. ‘This moving person is I.’ (I. was heard approaching, but was not visible) (D-23/01/99)

Interestingly, the distal demonstrative occurs only very rarely in natural discourse. Hanks (1992) observes that a presentative or directive function often correlates with the use of a proximal demonstrative. This also seems to be the case in Goemai: when speakers direct attention to a referent, this referent is apparently moved closer into the speaker-and-addressee common zone, thus allowing for the use of the proximal demonstrative. Speakers can use this demonstrative even with distant referents: the speaker in example (10) uses the proximal demonstrative together with the distal adverb to refer to a distant object. Speakers agree that it would be possible to replace the proximal demonstrative with the distal demonstrative in the context of (10). However, it would not be possible to replace the distal adverb with the proximal adverb: the adverb is not used to direct attention (unlike the demonstrative) but simply refers to a place and conveys its distance to the speaker. puánáng, bàǹtyèm góe ń-d’é-ńnòe=hòe. search there/yonder FRONT 2SGM.POSS ADVZ-CL:exist-DEM.PROX=exactly ‘Look over there, at this your front.’ (M01ANCOLOR)

(10) Tàng

Goemai Demonstratives

145

Table 6.4 Classifiers and their semantics Classifier

Semantics

‘hang/move’ ‘sit’ ‘stand’

referent has the potential to move within a location referent projects from location and supports itself in a stable way referent projects from location and either supports itself in a precarious way or is supported externally referent is wholly supported externally and does not project from the supporting entity no information about the support relation

‘lie’ ‘exist’

Note: These kinds of classifiers are often called “deictic classifiers” in the literature. They occur with deictic elements, have scope over the noun phrase and characterize their referent in terms of shape or position in space (see Aikhenvald, 2000: 176–183 for an overview). The Goemai deictic classifiers grammaticalized from locative verbs that are used (with the same semantics) in the presentative construction.

While information on the location of referents thus tends to get neutralized in the attention-directing function, information about the orientation of referents in space continues to play an important role. Such information is coded in the five classifiers, as summarized in Table 6.4 (for more information on their semantics, see Hellwig, 2007a; 2007b). The first four (postural) classifiers give information about the kind of support relation that holds between the referent and its location, whereas the last (existential) classifier does not code such information. Speakers have to choose between an appropriate postural classifier and the existential classifier, and this choice is largely determined by pragmatics. Two common scenarios are illustrated below. First, the posturals are used whenever information about the support relation is considered to be necessary for the addressee to identify the referent correctly. In example (11), speaker N. introduces the referent with a demonstrative containing the existential classifier plus a descriptive locative phrase. But there are several potential referents in different orientations, and speaker A. has problems identifying the intended referent. Speaker N. then uses a demonstrative containing the classifier t’ó- ‘lie’ to disambiguate between the referents. ǹtyèm góe gòe-kúl, m̀ m̀ . see NMLZ(SG)-ADVZ-CL:exist-DEM.PROX FRONT2SGM.POSS NMLZ(SG)-left yes ‘See this existing one in front of you, at the left, yes.’ [= referent 4] A: Hèn=zèm á yí gòe-b’áng. 1SG.S=like FOC leaf NMLZ(SG)-become.red

(11) N: Nà góe-ń-d’é-ńnòe

146

Birgit Hellwig Gòe-ń-d’é-ńnòe=à? NMLZ(SG)-ADVZ-CL:exist-DEM.PROX=INTERR ‘I want (to find) a red paper. This existing one?’ [= referent 3] N: Á’à. Ní bá no 3SG.I NEG ‘No. It’s not (it)’ A: Gòe-ńnòe=à? NMLZ(SG)-LOC.ANAPH=INTERR ‘This one?’ [= referent 5] N: Góe-ń-t’ó-ńnòe. NMLZ(SG)-ADVZ-CL:lie(SG)-DEM.PROX ‘This lying one.’ [= referent 4] (M01ANCOLOR)

Second, the posturals tend to be used at the first mention of a referent (as in 12a): the first reference to the object uses a demonstrative with the postural classifier, but the second reference, a demonstrative with the existential classifier. Note that this differs from contexts where the locative anaphor is used: although both are used at the second mention of a referent, the anaphoric form is only used after the referent is identified (as in 12b). As long as its identity and location are being disputed (as in 12a), the demonstratives continue to be used. (12) a. N: B’èp

gòe=éép gòe-ń-t’óng-ńnòe. do.again 2SGM.S=open(SG) NMLZ(SG)-ADVZ-CL:sit(SG)-DEM.PROX ‘Open this sitting one again.’ A: m̀ m̀ , gòe-ń-d’é-ńnòe=hòe=à? yes NMLZ(SG)-ADVZ-CL:exist-DEM.PROX=exactly=INTERR ‘Yes, this existing one?’ N: m̀ m̀ . yes ‘Yes.’ A: Okay, à gòe-pyá. okay FOC NMLZ(SG)-become.white ‘Okay, (it is) a white one.’ (M01ANCOLOR)

ńdè ń-d’yém-náng. okay one/other ADVZ-CL:stand(SG)-DEM.DIST ‘Okay, that other standing one.’ A: m̀ m̀ , gòe-ńnòe à bóró bì dwén. yes NMLZ(SG)-LOC.ANAPH FOC Fulani thing PL.LOG.SP.POSS ‘Yes, this one is (a picture of) Fulani in their own way.’ (C01ANHAND)

b. N: Tó,

Much of the burden for identifying a referent is thus carried by the classifiers, not by the deictic roots. This is especially clear in cases where there are two potential referents, one a bit closer to the participants than the other. In such cases, the proximal/distal opposition is not usually employed. Instead, if such referents are in different orientations, the postural classifier is used (as in 11

Goemai Demonstratives

147

above). But if they are not in different orientations, speakers have considerable difficulties identifying the referent, as illustrated by the exchange in (13). In this example, both potential referents (green toy cars) are in the same orientation (they “stand”). And although they are located at slightly different distances from the participants, speaker A. does not employ the proximal/distal opposition. In the end, the problem is only solved by A. leaning forward and touching the intended referent. d’è ń-gòedè là=mótò, paper NMLZ(SG)-become.white=DEF exist LOC-bottom:GEN DIM(SG)=car ǹ-t’óór góe ń-d’é-ńnòe. LOC-flank 2SGM.POSS ADVZ-CL:exist-DEM.PROX ‘The white paper is under the small car at this your existing side.’ [= referent 9] N: B’ák góe-t’óór ń-d’é-ńnòe=hòe=à? here PLACE-flank ADVZ-CL:exist-DEM.PROX=exactly=INTERR Mótò góenàng dái? Góe-d’è góe-t’óór ń-d’é-ńnòe=à? car which(SG) indeed NMLZPLACEADVZ-CL:exist-DEM.PROX exist flank =INTERR ‘(Is it) here at this existing side? Which car now? (Is it) the one being at this existing side?’ [= referent 8 or 9] (M01ANCOLOR)

(13) A: Tákàrdá gòe-pyá=hók

5

Conclusion

This chapter has shown that Goemai demonstratives are used for exophoric spatial reference. In endophoric contexts, by contrast, a distinct locative anaphor is used. Frajzyngier (1996) presents evidence that such a differentiation of functions is common in Chadic languages, and Goemai thus seems to conform to a more general Chadic pattern. In its exophoric demonstratives, Goemai distinguishes between a proximal and a distal form. Such two-term systems are common in the world’s languages (Anderson and Keenan, 1985; Diessel, 1999), but the extensions of the Goemai terms are unusual: Goemai treats the speaker and addressee as a unit, and referents are proximal or distal with respect to the whole unit. The analysis is summarized in Table 6.5. It has been suggested for some languages with a two-term system that one of the terms is, in fact, a neutral term (Dunn, this volume; Enfield, this volume). In Goemai, the best candidate for such a neutral term would be the proximal demonstrative: it has a much wider distribution than the distal term; and it is often opposed not to the distal demonstrative, but to a non-demonstrative form (the locative adverb). I do not adopt such an analysis for Goemai because there are clear instances where only the distal demonstrative is applicable (see section 3). The higher frequency of the proximal demonstrative is probably

148

Birgit Hellwig

Table 6.5 The anaphor and the demonstratives Anaphor: Demonstrative: Proximal to speaker/addressee: Distal to speaker/addressee:

ńnòe ‘locative anaphor’ ńd’éńnòe ‘proximal’ ńd’énáng ‘distal’

linked to the attention-directing and identifying function of the two demonstratives: identifiable referents tend to be proximal referents; and there is even evidence that the proximal form can be used to direct attention to some more distant referents (see section 4). It is interesting to note that the location information seems to be secondary: the demonstratives also code postural information, and section 4 has shown how speakers and addressees rely on this postural information to identify referents. There is some evidence for a language change in progress. Younger speakers are much more willing than older speakers to exploit the proximal/distal opposition in the demonstratives to disambiguate between referents, and they are less likely to shift away from using a demonstrative when the referent is non-identifiable. Considering the recent grammaticalization of the demonstratives from the presentative construction (within the life span of the older generation), it would not be surprising if the older speakers retained more of its original semantics and uses. References Aikhenvald, A. Y. (2000). Classifiers: A typology of noun categorization devices. Oxford: Oxford University Press. Anderson, S. R. & Keenan, E. L. (1985). Deixis. In T. Shopen, ed., Language typology and syntactic description. Vol. III. Cambridge: Cambridge University Press, pp. 251–308. Diessel, H. (1999). Demonstratives: Form, function, and grammaticalization. Amsterdam/Philadelphia: John Benjamins. Enfield, N. J. & Bohnemeyer, J. (2001). Hidden colour-chips task: Demonstratives, attention, and interaction. In S. C. Levinson & N. J. Enfield, eds., Manual for the field season 2001. Nijmegen: Max-Planck Institute for Psycholinguistics, pp. 21–28. Frajzyngier, Z. (1993). A grammar of Mupun. Berlin: Dietrich Reimer. (1996). On sources of demonstratives and anaphors. In B. A. Fox, ed., Studies in anaphora. Amsterdam/Philadelphia: John Benjamins, pp. 169–203. Hanks, W. F. (1992). The indexical ground of deictic reference. In A. Duranti & C. Goodwin, eds., Rethinking context: Language as an interactive phenomenon. Cambridge: Cambridge University Press, pp. 43–76.

Goemai Demonstratives

149

Hellwig, B. (2003). The grammatical coding of postural semantics in Goemai (a West Chadic language of Nigeria). PhD thesis, Radboud University, Nijmegen. (2007a). To sit face down: Location and position in Goemai. Linguistics (special issue), 45(5/6), 893–916. (2007b). Postural categories and the classification of nominal concepts: A case study of Goemai. In A. C. Schalley & D. Zäfferer, eds., Ontolinguistics: How ontological status shapes the linguistic coding of concepts. Berlin/New York: Mouton de Gruyter, pp. 279–297. (2010). Different types of data: A case study of Goemai demonstratives. Journal of West African Languages, 37(1), 7–22. (2011). A grammar of Goemai. Berlin/New York: Mouton de Gruyter. Levinson, S. C., Kita, S. & Özyürek, A. (2001). Demonstratives in context: Comparative handicrafts. In S. C. Levinson & N. J. Enfield, eds., Manual for the field season 2001. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 52–54. Sirlinger, E. (1937). Dictionary of the Goemai language. Unpublished ms. (1942). A grammar of the Goemai language. Unpublished ms. (1946). An English-Goemai dictionary. Unpublished ms. Wilkins, D. P. (1999; this volume). The 1999 demonstrative questionnaire: “This” and “that” in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24.

7

Tzeltal: The Demonstrative System Penelope Brown and Stephen C. Levinson

1

Introduction

Demonstratives have played a central role in philosophical and linguistic thought about meaning and reference (Wittgenstein, 1953; Searle, 1969; Kaplan, 1978). Pointing and the expressions that go with it have been thought to be an essential early step in both language acquisition and language evolution (Tomasello, 2014), with the expectation that demonstratives are strongly universal (Greenberg, 1966; Diessel, 1999a, b). In addition, demonstratives are the prototypical deictic words, and deixis has been a central focus in work on the pragmatics of language (Fillmore, 1982, 1997; Levinson, 1983; Hanks, 1992; 1996; 2005). Language variation in this domain is thus of special interest (Dixon, 2003). This chapter describes the demonstrative system for a local dialect of Tzeltal as used within its normal, rural context in southern Mexico. Tzeltal is a Mayan language which has VOS basic word order, ergative alignment, and complex derivational morphology among other typologically interesting features. It is closely related to the much better described Tzotzil (Haviland, 1981; Laughlin, 1988; Aissen, 1987), which, however, lacks a detailed investigation of demonstratives. The demonstrative system of Tzeltal is not straightforward to analyze due to the fact that the forms are many and combine with each other, yielding about 20 basic forms, and many further complex compounds are possible. In addition, the basic meanings (or referential constraints) are obscured by a number of usage principles (e.g., the fact that pointing licenses proximal forms even at far distances). The overall system is thus rather opaque. In this chapter we show how the use of the Wilkins Demonstrative Questionnaire (1999; this volume) greatly clarifies the underlying system. 2

The Language and Its Speakers

Tzeltal is a Mayan language spoken in the eastern highlands of Chiapas, southern Mexico, close to the Guatemalan border, by some half a million 150

Tzeltal Demonstratives

151

speakers (INEGI census, 2010). An ongoing demographic explosion has seen Tzeltal speakers in recent years spilling over into the eastern lowlands, returning to where they were probably driven out by the Spanish in the sixteenth century. Tzeltal is thus one of the few Amerindian languages with a rapidly increasing number of fluent speakers. The Tzeltal region is contiguous with those of the other Mayan languages of this area: Tzotzil just to the west and south, Chol to the north, and Tojolabal to the southeast. Altogether there are several million speakers of Mayan languages, for only one of these languages (Yucatec) are there detailed descriptions of the demonstrative system (Hanks, 1990; Bohnemeyer, 2012; this volume). Tzeltal speakers are still predominantly swidden agriculturists, who nowadays cash crop according to the ecological affordances of the local area (coffee in hot country, fruit and vegetables in cold uplands). Those who have moved into urban centers may have children monolingual in Spanish, but most families are still fully operative in Tzeltal. Minimal bilingualism in Spanish in earlier generations is now giving way to widespread bilingualism among younger generations, but literacy in either Tzeltal or Spanish is more restricted. There are radio services and primary school books produced in Tzeltal, and there is a strong sense of ethnic and political identity focused on the various municipios, the local territories that function as administrative units. This chapter describes the demonstrative system of Tenejapan Tzeltal. Tenejapa is one of the traditional administrative and political units reinforced by the Spanish conquistadors for colonial rule. Ethnic identity is reflected in distinctive traditional attire, and in syncretistic religious and civil rituals focused especially on the ceremonial centre, Lum. The inhabitants of Tenejapa number perhaps 50,000 and have a distinctive dialect. They are raised monolingually (unless they are in the border zone with Tzotzil, a closely related language) and are increasingly exposed to Spanish only through school, travel, and (especially for men) by marketing and political activities in the local ladino1 town and beyond. It is no longer exceptional for young Tenejapan men to have made the dangerous illegal trip to the USA. Men under age 40 are generally functionally bilingual in Spanish, but with varying degrees of proficiency. Interestingly, the different dialects of Tzeltal show considerable variation in their demonstrative systems. There are three main dialects, north, central, and south (Kaufman, 1972; Campbell, 1987), and the Tenejapan variety described

1

Ladino is the local term for Spanish-speaking national-culture-identified persons; many of these are of indigenous extraction.

152

Penelope Brown and Stephen C. Levinson

here belongs, like the Oxchuc variety, to the central dialect. But there are considerable differences between the Oxchuc variety as described by Polian (2013) and the Tenejapan variety described here: there are different forms for most of the demonstratives, and apparently a partially different underlying set of oppositions. Nevertheless, the formal structure of Mayan demonstratives is well conserved across the language family, with a set of proclitics and enclitics which together make up the forms in question. It should be noted here that the Yucatec system, which is well described in Hanks (1990) and Bohnemeyer (2012, this volume), has a systematicity which seems at least partially lost in Tenejapan Tzeltal. For example, in Yucatec it is at least plausible to analyze the pronouns as part of the demonstrative array, but in Tzeltal this would require analyzing focal predicator ja’ as part of the system, which seems undermotivated except for the special affinity for ja’ to occur with demonstratives (as in ja’ ini ‘it’s this one’). 3

The Forms and the Basic Oppositions

Demonstrative pronouns and adverbs and the definite article belong to a unique form class identifiable by the property of circumclisis, having paired morphemes that flank the phrase they modify, e.g., i winik=i ‘this man’, which we will represent with dots for the intervening material (e.g., i . . . i) as the general form. There is no formal relation to the set of independent personal pronouns (singular forms jo’on, ja’at’, ja’), and no relation to third person marking on the verb. The basic oppositions in the Tenejapan variety of Tzeltal are given in Table 7.1. The semantics of these terms will be clarified below, but notice that the distal/far distal distinction is restricted to the adverbial forms. That is, the underlying system opposes proximal and distal in the nominal demonstratives (adjectives and pronouns), and makes a three-way opposition in the adverbs between proximal, distal, and far distal. Mostly, the anaphoric forms are based on the distal ones. The presentational form in plays a special role, intervening across the paradigm. This much is clear from a simple inspection, but as always with deictic elements the devil is in the detail. The interpretation of “proximal”, for example, will prove quite complex and may have little to do with pure distance in space. A prior description of the essentials of the system is given in Brown (2006), and, apart from Polian’s (2013) description of the Oxchuc dialect, there is little else available. As can be seen from Table 7.1, many of these forms involve circumclisis, and to understand this we need to digress on the grammar of the noun phrase in Tzeltal and related languages.

‘here (I give you this)’

‘like this’

PRESENTATION

DEMONSTRATION

‘EXIST/be over there’

that . . . there’ in ila jich

li’ . . . ini

‘this . . . here/

this . . . over there’

i . . . ini li’i li’ . . . =i

‘here/there/over there’ ‘this . . . here/this . . . there/

ADVERB

ini i . . . =i in . . . =i

‘this/that’ ‘this NP/that NP’

PRONOUN ADJECTIVE

PROX

Table 7.1 Basic oppositions in Tenejapan Tzeltal demonstratives

tey li’ . . . =e

mene me . . . =e men . . . =e in . . . =e

DISTAL

lum (+EXIST or positional)

tey (to)

lum (to) . . . =e lumine

FAR DISTAL

tey . . . =a

tey, tey=a,

mene

ANAPHORIC

154

Penelope Brown and Stephen C. Levinson

4

Demonstratives and the Grammar of Noun and Adverbial Phrases

Tzeltal is a mildly polysynthetic head-marking language. It is an ergative language with obligatory cross-referencing markers indexing both subject and object on the verb; the set of ergative markers also marks possession on nouns. The basic noun phrase takes the form indicated in (1): (1) (DEIC) (DET) (NUM PHRASE) (ADJ) (POSS-) N (=DEIC)

A bare noun is a potential full NP in Tzeltal (e.g., la jman chitam ta lum ‘I bought (a) pig in town’). Note that the expanded frame in (1) shows two slots for deictic elements: initially, and terminally as clitics. Unusually, the language permits both an initial demonstrative element and a determiner, as in (2) (in examples, demonstrative forms and their glosses are in boldface; pointing gestures are indicated on the line above the transcription):2 (2) [2013March23_AntGO&AO: 19:01] |POINT |POINT tey ain-on jo’tik ‘w-a’y ba ay-Ø ya’tik in te there stay-1A 1PLe 2E-know where exist-3A now ID.PROX.AJ ART Marabia=a=e PLACE=TD.ANAPH=TD.ART ‘There we stayed y’know where now this the Marabia [place] is.’

The in acts in a quasi-presentational fashion (‘this here I’m showing you’), the te indicates definiteness (the known-to-both-of-us place called Marabia). 2

Tzeltal transcription conventions are based on a practical orthography; symbols correspond roughly to their English equivalents except that j = h, x = sh, and ‘ indicates a glottal stop or glottalization of the preceding consonant. Abbreviations for glosses are as follows: 1,2,3 E –1st, 2nd, 3rd person ergative prefixes (which mark both subjects of transitive verbs and noun possession), 1,2,3 A – the corresponding absolutive suffixes, 1PLe – 1st person plural exclusive, 1PLi – 1st person plural inclusive, PL – 2nd or 3rd person plural, ACS – achieved change of state clitic, ADV – adverb, AFF – affect verb suffix, AJ – adjective, ANAPH – anaphoric, ART – definite article, ASP – neutral aspect, AUX – auxiliary verb, CMP – completive aspect prefix, ICP – incompletive aspect prefix, CJ – conjunction, CL – clitic, ID – initial deictic, TD – terminal deictic, PRON – pronoun, PROX – proximal, DIST – distal, DIM – diminutive, DIR – directional, DIT –ditransitive, EXCL – exclamative, FOC – focus predicator, HESIT – hesitation marker, HON – honorific, IMP – imperative, INCH – inchoative, NAME – personal name, PLACE – place name, NC – numeral classifier, NEG – negative particle, NOM – nominalizing suffix, PASS – passive, PPrt – passive participle, PERF – perfect aspect, PREP – preposition, PT – particle, Q – question particle, QUOT – quotative particle, REFL – reflexive, RELN – relational noun, SUBJ – subjunctive. Numbers in brackets (e.g. 1.1) indicate pause lengths in speech. A text identification preceding the examples indicates their source in naturally occurring events or interactional “space games”; if unmarked they are from field notes. Gestures, where indicated, are marked on the line above the Tzeltal text, with a vertical line locating the stroke and dots indicating duration.

Tzeltal Demonstratives

4.1

155

Demonstrative Pronouns

The demonstrative pronouns and adjectives ini, ine, and mene are themselves bimorphemic: in=i contrasts with in=e, and while men=e has no such contrast, in both cases material can intervene (as in in winik=i ‘this man’, men winik=e ‘that man’). The far distal adverbial form lumine can be analyzed as lum-in=e, far-PRESENTATIONAL=DISTAL. However, when they occur affixed to each other we treat these forms as one unit (ini, ine, mene, lumine). Note that these forms cannot occur alone, for example in response to a question, “which one?”; they must be preceded by the predicator ja’ (ja’ ini ‘it’s this one’, ja’ mene ‘it’s that one’) or by another predicate. As shown in Table 7.1, these compound forms can be disassembled on either side of an NP, thereby becoming demonstrative adjectives. 4.2

Demonstrative Adjectives

The demonstrative adjectives are always expressed at least partly as clitics, nearly always with a proclitic (e.g., i or me) or initial demonstrative word (in or me(n)). (The clitic status of the enclitics is clear from the phonology of stress, the status of the reduced initial elements is less clear.) Following Hanks (1990), we refer to these elements as ID (Initial Deictic) and TD (Terminal Deictic). Examples include: (3) a. i winik=i ID.PROX.AJ -man=TD.PROX ‘this man’ b. me winike=e ID.DIST.AJ-man=TD.DIST ‘that man’

These clitics flank the noun phrase, regardless of complexity: (4) a. i j-ch’ich’el-tik=i ID.PROX.AJ 1E-blood-1PLi=TD.PROX ‘this my blood’ b. ja’ i j-muk’ul cha’am ini FOC ID.PROX.AJ 1E-big molar TD.PROX ‘It’s this my big molar (that’s hurting).’ c. s-lumal-ik tz’in me jmamal jbankil-al-tik 3E-countryman-PL PT ID.DIST.AJ HON elder.brother-NOM-1PLi ine TD.DIST ‘They are countrymen of that sir “elder brother” of ours.’ [referring to a man standing about 15 feet away]

156

Penelope Brown and Stephen C. Levinson d. i jun li’ ay- Ø ta j-pat=i ID. PROX.AJ book ID.PROX.ADV exist-3A PREP 3E-back=TD.PROX ‘this book (which) is here behind me’ e. [1995_v21sopa: 23:04] [person identification] |POINT ya to y-al me jtatik lum ta elawal ICP PT 3E-tell ID.DIST.AJ HON ID.DIST PREP face.of.hillside jtatik poxil=e HON medicine=TD.DIST ‘He still told (about it), that sir over.there on the mountain-face, sir curer.’

In this way the demonstrative adjectives are formally parallel to the definite article, which is also a discontinuous construction te . . . =e flanking the noun phrase (this is not surprising on the theory, expressed by, e.g., Lyons (1977), that definites are just demonstratives unmarked for distance): (5) a. te winik=e ‘the man’ b. te k’alal s-jul-el tal ch’ulel=e ART when 3E-arrive_NOM DIRcome soul=TD.ART ‘(at) the time when his soul arrived (i.e., he became competent)’

However, the article te . . . =e requires the =e terminal, whereas for at least the adjectival in and men and the adverbial lum it is optional. Our data shows that both (with or without a terminal) are frequent; e.g., for men: (6) [2008_v9_XntOs2: 51:51] [Identifying a referent] ja’ tz’i men y-inam jkarlos=e FOC PT ID.DIST.AJ 3E-wife NAME=TD.DIST ‘It’s that wife of Carlos.’ (7) [1991_v5 C-trees: 1:09:17] [several men working together to prune trees; Xun has the rope] yoch-a-ik tal men laso, Xun be.loose-IMP-PL DIRcome ID.DIST.AJ rope NAME ‘Loosen that lasso this way, Xun.’

One complexity introduced by circumclisis is that, in a complex NP, more than one initial deictic element may appear, calling for potentially different terminal deictics – in which case there are priority rules about which final clitic actually appears. This is the case in the following example where the final =i is motivated by the proximal adverb, which overrules the final =e expected by the definite article: (8) te jun li’ ay-Ø ta j-pat=i ART book ID.PROX.ADV exist-3A PREP 3E-back=TD.PROX ‘The book (that is) here at my back.’

Tzeltal Demonstratives

157

In the following exceptional example, two contrasting TDs occur, the anaphoric referring to addressee’s immediate preceding behavior and the proximal perhaps licensed by a point (see below): (9)

4.3

ma x-a’ pas a =i NEG ASP-2E-do ANAPH=TD.PROX ‘Don’t do this (that you just did)’ [mother to child, after he hit his brother and was hit in return]

Demonstrative Adverbs

As noted above, the demonstrative adverbs make a three-way distinction between ‘here’, ‘there’, ‘over there’. The adverbial forms may also have clitics that flank the item they modify, in this case a prepositional phrase or a verb phrase, as, for example, in the following proximal forms: (10) ay- Ø ijk’-ub-en-Ø li’ ta j-cha’am=i exist-3A black-INCH-PERF-3A ID.PROX.ADV PREP 1E-molar=TD.PROX ‘It has blackened (a tooth) here at my molar.’ (11) li’ makal-Ø ta pat witz’=i ID.PROX.ADV hidden-3A PREP back mountain=TD.PROX ‘It’s here hidden [out of sight] at the back of the mountain.’

and in the comparable distal forms: (12) lum ay-Ø ta y-elaw alal=e ID.FARDIST.ADV exist-3A PREP 3E-face child=TD.DIST ‘It’s over there at the child’s front.’ [20 feet away] (13) lum xik’il-Ø ta y-anil taj te ID.FARDIST.ADV leaning-3A PREP 3E-underneath pine.tree ART machite=e machete=TD.DIST ‘The machete is leaning over there underneath the pine tree.’

In the following example, the final distal clitic =e is triply motivated by the distal adverb lum, the first presentational in, and the final -in- (embedded in lum-in=e) which frequently collocates with lum. (14) [2012July25 MO3 15:52] (talking about potential spaces for making a house) |. . . . . . . . . . . . . . . . . . . . . |. . . . . .(other hand) ja’ nax bel lum ta tim=e sok in ta FOC PT DIRgo ID.FARDIST PREP tree=TD.DIST CJ ID.PROX PREP ba ay- Ø xan lumine where exist-3A another FARDIST ‘It’s just over there at the tree, and here is where there is another one over there.’

158

Penelope Brown and Stephen C. Levinson

As with the demonstrative adjectives, the phrase that is flanked by the deictics can be of any length: (15) [2012Jul25_XmOs1&Slu: 4:32] [talking about where she’ll sow a new crop; no gesture] lum ta banti tz’ajk y-u’un k-ijtz’in ba ID.FARDIST.ADV PREP where border 3E-RELN 1E-YoBr where ay- Ø potrero=e exist-3A pasture=TD.DIST ‘(It’s) over there where the border (is) of my younger-brother where the pasture is.’

4.4

Presentational Demonstrative

The presentational form in has a canonical use when handing something to someone, when the hander says in, holding out the object, and on taking it the receiver says ixtal, acknowledging receipt. In this usage of handing over something physical, in always occurs alone, never with terminal deictics.3 Outside such contexts of physical presentation, in recurs in compound uses mentioned above (e.g., ini, in . . . =i) where some metaphorical or non-physical presentation (e.g., drawing attention to or focalizing some object, situation, or event) is involved, as in (16): (16) [1995_v21sopa: 10:51] ja’ jich ya y-ayanta-ik in te j-tat=e FOC thus ICP 3E-narrate-PL PRES ART 1E-father=TD.ART ‘That’s how my father told it (the story).’ [father is in focus] (17) [1995_v33xanttzuj: 41:11] |POINT te k-ajwaltik in ta ba ya x-til-0 ya ART 1E-lord-1inclPL PRES PREP where ICP ASP-glow ICP ‘w-il ini 2E-see TD.PROX ‘Our lord (the sun), here’s [pointing to sun] where he’s burning.’

It is noteworthy that the combination of men (distal) plus in in this presentative sense is acceptable in cases where their functions are distinct: (18) [1995_v21sopa: 15:58] [talking about a locust plague long ago] melel tz’in te k’ulub=e, k’unk’un a tal-0 k’alal ta juju’ truly PT ART locust=TD.ART slowly CMP come-3A all.the.way PREP every

3

An alternative, pragmatically equivalent presentational form is ila, but this latter form plays no other part in the demonstrative system.

Tzeltal Demonstratives

159

y-ajk’ol jobel-etik a’ tal-0 men in k’ulub=e 3E-uphill PLACE-PL CMP come-3A ID.DIST ID.PROX locust=TD.DIST ‘Truly the locusts, they gradually came all the way to every area uphillwards of San Cristóbal, those locusts.’

The men here is spatial, the in is presentative/focal. 4.5

Demonstration Demonstrative

Jich ‘like this’ is the form used when demonstrating gesturally how something is (its size, shape, position or location) or how an event occurs. It is used, for example, when showing the size of something, as in: (19) [1995_v21_sopa: 25:46] |POINT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .SWEEPING GESTURE jich s-muk’-ul ala jil-em-0 ini ala s-tenlej thus 3E-big-NOM DIM remain.behind-PERF-3A PROX.PRON DIM 3E-level.area ‘Thus is its size (that) this flat-area (of land) has remained’ (with a size-indicating gesture).

or demonstrating the way something should be localized when pruning trees back off of a path: (20) [1991_v5 C-trees: 46:41] [man in tree asking for advice about cutting branch] bi y-ilel ya ’w-il-ik ek ini, jich bal ya x-tal-0 what 3E-appearance ICP 2E-see-PL also this.one, thus Q ICP ASP-come-3A y-ilel ini? 3E-appearance TD.PROX ‘How do you see this one (branch) as well, should it come like this, this one?’

4.6

Combinatorial Possibilities of IDs and TDs

Hanks (1990) presents a complex analysis of demonstratives in Yucatec, another Mayan language, based on separating initial deictic base (ID) and terminal deictics (TD) and analyzing the intersection as semantically compositional in complex ways. Following this lead, Table 7.2 shows the matrix of intersecting possible complex deictic expressions in Tzeltal. Roughly, the terminal deictics have the following values: =i (proximal), =e (distal), =a (distal or anaphoric),4 while in=i and in=e are themselves complex but introduce some kind of immediacy, retaining something of the presentational meaning of solo in. It follows that initial i, which is proximal, cannot take the distal enclitic =e, and initial me (distal) cannot occur with 4

The enclitic a actually isn’t strictly anaphoric, since it can refer to something that previously just happened and was in joint attention but was not actually mentioned. For example, ma’yukix a ‘There’s none of it’.

160

Penelope Brown and Stephen C. Levinson

Table 7.2 Initial and terminal deictic pairings in Tenejapan Tzeltal Terminals Initials i in me(n) li’ tey lum te jich

=i i . . . =i in . . . =i li’ . . . =i

=e

in=e me(n) . . . =e li’ . . . =e tey . . . =e lum . . . =e te . . . =e

=a

ini i . . . ini in . . . ini

tey . . . a jich . . . a

li’ . . . ini

jich . . . ini

ine

mene

in . . . ine me(n) . . . ine tey . . . ine lum . . . ine

tey a mene

jich . . . ine

jich (a) mene*

* jich mene ‘like that’ always has an anaphoric meaning.

terminal =i (proximal). But note that initial deictic in (extended from the presentational particle), and li’ (adverbial) can take both proximal and distal enclitics, while the distal initial forms can occur with both distal and anaphoric enclitics. The definite determiner te takes only the distal TD form =e, suggesting that that the distal may be historically the unmarked form, even though in terms of generality of use the proximal forms would now have best claim to unmarked status. When the terminal deictics occur alone, they seem to have lost their demonstrative functions, so that, e.g., final -i alone seems to have anaphoric functions (cf. final -i in the Oxchuc dialect, which contrasts with demonstrative final to; Polian, 2013: 713). We will not have the space here to explore the subtle meanings of all of these forms and will concentrate on the demonstrative pronouns, adjectives, and adverbials. In addition to these combinations, other compound deictics can be formed. For example, in to ta ba ay (plus obligatory pointing) is a frequently occurring compound incorporating presentational in meaning roughly ‘here (is) where it is’. The particle to specifies temporal or spatial extent (note that this is quite different from the deictic to in the Oxchuc dialect; Polian, 2013). Some idea of how these combinations work in practice can be gleaned from the following recorded examples: (21) [2013Mar23AGO: 14:56] [presentational use, no pointing gesture; talking about his sister’s family living far away in colonia Nacional] yale jk’axel, ja’ me in to ta ba ay-ik EXCL EXCL FOC PT ID.PROX.AJ PT PREP where exist-3PL ek=i too=TD.PROX ‘Goodness me, it’s that, see here all the way where they are too.’

Tzeltal Demonstratives

161

(22) [2013Pasarotk: 20:24:] [identifying where people moved to] | HAND GESTURE OVER SHOULDER |HAND GEST. REPEAT ay- Ø me ma’yuk y-ajwal lum balamilal lum to ine exist-3A PT there’s.none 3E-owner land world ID.DIST.ADV PT TD.DIST ‘There are no owners of the land/fields/countryside way over there.’

(23) [2013Mar23AGO:28:13] [talking about roads to Nacional; no gestures] ba to ya a’-tzak-ik karo yael, pero li’ to nix ta Comitan where PT ICP 2E-get-PL car PT but ID.PROX.ADV PT PT PREP PLACE ya a’-tzak-ik karo ine ICP 2E-get-PL car TD.DIST ‘where you catch a car it seems, but here just at Comitan you get a car there.’

(24) [2013Mar31XmO: 12:49:] [talking about addressee’s murdered mother; no gesture] sok k’ux-Ø obol s-ba a’w-a’y tey to metzel-Ø la and hurt-3A pity 3E-REFL 2E-know there PT lying.on.side-3A CMP s-ta-ik a k-a’y=e 3E-encounter-PL CMP 1E-hear=TD.DIST ‘And poor her, you know, they found her there lying on her side I heard.’

(25) [2013Mar31_XmO:24:04:] [talking about weather in hot country (Zintalapa); no gesture] aj, x-bujt-et- Ø k’inal Ah, ASP-clouded-AFF-3A land ‘Ah it gets cloudy there.’

tey=a there ID.DIST=TD.ANAPH

Finally, note that the proximal form li’i is always picked up in the next speaker’s response with the distal/anaphoric form tey, regardless of proximal/distal semantic relevance in the initial mention. For example: (26) [2012Jul25MOs3: 17:24:] [talking about addressee’s children, where they live; no gesture] A: ju’uk li’ to ala tzopajtik ya‘wa’y ‘No they are grouped together right here, you know.’ B: tey tz’in ‘There.’ A: tey to ya ‘wa’y ‘There, you know.’

5

The Results of the Demonstrative Questionnaire Elicitation

We used the elicitation device designed by Wilkins (1999; this volume)5 to probe exophoric demonstrative use. This elicitation tool proved especially useful for exploring the basis and bounds of notions like “proximal” versus “distal” which would be hard to extract from recorded uses, and it substantially

5

We also used earlier forms of demonstrative elicitation tools (Pederson & Wilkins, 1996; Levinson, 1999).

162

Penelope Brown and Stephen C. Levinson

changed our understanding of the system. Note that in Tzeltal the distinction exophoric vs. anaphoric forms is often hard to distinguish, since the distal forms mene and a are used anaphorically, and can be both exophoric and anaphoric simultaneously. We went through the questionnaire with three different consultants (two male, one female), in and around their homes. Elicitation sessions were always conducted in Tzeltal. They were audio-recorded, and responses entered into an Excel spreadsheet and coded for proximal/distal semantics. We also probed extensively for alternative uses and searched a large corpus of conversational Tzeltal for naturally occurring uses of demonstratives. Constraints of the field location (with houses surrounded by cornfields and banana and coffee trees, in precipitously mountainous terrain) made some minor adaptation of the questionnaire scenes necessary; in particular, it proved impossible to find an open area of 100 meters (“the size of a football field”) where participants could talk to one another from far apart (the most we achieved was about 10 meters). For the most part responses of the three consultants to the elicitation scenes were consistent. Owing to the restricted nature of the tool, only a subset of the demonstrative forms were used in responses. Among them were the demonstrative pronouns ini and mene alone, but much more frequently flanking clitics occur. Ubiquitous is some form of in, as it can be part of the stand-alone pronoun forms ini/ine or of the initial or final element to the NP (i(n), -i). In most cases in seems to have an attention-drawing function, as mentioned under its presentational uses. The adverbial li’ ‘here’ (in, e.g., li’ . . . ini) plays an important role in emphasizing the proximal forms, as in English this here. Also frequent is the use of the focal predicator ja’, which seems to have a special affinity for combining with the deictics (Polian, 2013 analyzes it as a deictic predicate in the Oxchuc dialect). The results of elicitation with the questionnaire, as they clarify the semantics of the “proximal” and “distal” categories, are summarized in Table 7.3. The table mentions just the forms in=i, men=e, and lumine which stand respectively as examples for proximal, distal, and far distal uses of all forms. The results are best explained by distinguishing the different factors involved in motivating specific forms. Three factors in particular override the more purely spatial determinants: whether or not a referent is being pointed at, whether the demonstrative is calling first attention to a referent, and whether or not it is visible to the speaker. It is best to first set these aside in order to grasp the spatial factors involved: Non-spatial Factors (i) Pointing. Elicitation makes clear that the proximal deictics based on in . . . i can be used in nearly all scenes (with one systematic exception) so long as

(7)

(6)

(3)

A

A

A

S

S

S

A

(12)

(8)

1C

A

A

Shared close space

S

S

in + =i + pointing

S's periphery

(19)

S

(22)

S

(20)

A

A

1B

ini (proximal, minus pointing)

(11) S's space

S

S

S

(1)

1A

(17)

S

(21)

S

A

Shared mid-distance

A

(14)

S A

2

Table 7.3 Results of the exophoric elicitation task for Tzeltal

(5)

(4)

(2)

3

A

A

A

A's space

S

S

S

A

S A

(18)

A

S

(16)

S

(10)

(9)

4

A

A

(23) A's periphery

S

S

5A

A S

A S

(25)

(24)

5B

lumine (ADV)

mene (distal)

Shared Far Distance

(15)

S A

(13)

S A

164

Penelope Brown and Stephen C. Levinson

they occur with a point: pointing somehow makes the referent potentially (and thus optionally) proximal. One analysis of this is that the in- of the demonstrative series is (at least in origin) the same in as the presentational in used when giving something – one is presenting the referent to view, even if at a distance, just as in English a guide can say from a kilometer’s distance ‘And lo! Here is the Taj Majal’. The proximal form does not require a point, but a point licenses the proximal form regardless of spatial distance. The systematic exception to this licensing of proximal deictics is when the referent is clearly in the addressee’s personal or peripersonal space, when the proximal forms may not be used. (ii) In Prior Attention vs. Calling Attention to a Referent. There are multiple ways to call attention to a referent, many of which are not part of the demonstrative system (e.g., ila ‘Look’). Most of the demonstrative forms can be used to draw attention to a referent except for those that are anaphoric, as the anaphoric forms presume prior mention and thus prior attention. In addition, as in English, there is a clear association between distal deixis and anaphora: The distal ine or mene forms are both typically used if the referent is already mutually in attention. Moreover, the distal deictic ine cannot be used to draw attention to a referent. (iii) Visibility. Although the system as a whole doesn’t seem to make distinctions between visible and invisible referents, the presentational in and its derivatives ini and ine clearly prefer visible referents. Thus me . . . ine can’t be used if the referent is invisible, while me . . . e can. Spatial Factors If one takes these factors into account (e.g., by holding them constant), one can get a clearer interpretation of the essential distinction between proximal and distal forms. The basic constraints are as follows (bearing in mind that ostensive pointing can license proximal forms even to far distant objects): 1. If the referent is much nearer to the addressee than the speaker, then distal forms would normally be used. Thus, the addressee’s personal and peripersonal space requires distal forms (regardless, as mentioned above, of whether the speaker points or not). 2. If the referent is far away from both speaker and addressee, distal or far distal forms would normally be used, regardless of the actual locations of speaker and addressee, e.g., whether they are sitting together or not. (Compare scenes 14 and 17 where distal mene is used both when the referent is far from speaker and addressee sitting together and when speaker and addressee are separated and the referent is still distant from both.)

Tzeltal Demonstratives

165

3. If the speaker is much closer to the referent than the addressee, proximal forms should be used. The responses make clear that ‘proximal’ and ‘distal’ are not to be understood on a purely spatial distance metric. Rather the system is person-centric. “Proximity to speaker” and “Proximity to addressee” always take priority over metric distance; when the referent is proximate to neither speaker nor addressee, then issues of relative closeness or distance to our shared or individual space arise. Where speaker and addressee are far apart and the referent in between, the referent is treated as distal. The major patterns derivable from Table 7.3 can be summarized as follows: 1. Proximal ini without pointing is specialized for areas within the speaker’s orbit regardless of visibility of referent (1A), or ecological boundaries (1B). By the same token, it can be used when both speaker and addressee are equidistant from a close referent (1C). Beyond close space it is marginal. 2. When accompanied by a point, ini and its separated form in . . . =i has an extended range, including the far distance. However, there is a gap in its distribution, since it cannot be used when referring to things in the addressee’s close or peripheral space. 3. Distal mene covers referents close or peripheral to addressee, as well as items far removed from both partipants. 4. The adverbial lumine is specialized for the far distance. Far distance, as shown by the two columns in Table 7.3, can include, say, the far end of a field (5A) or locations (visible or invisible) which would require real travel (5B). The Wilkins questionnaire can be usefully supplemented with a contrastive elicitation task, in which, for example, three referents are lined up away from speaker, say three different colored cards as in Figure 7.1 (Enfield and Bohnemeyer, 2001). Here, if one points one can use ja’ ini to refer to any of the three referents, especially if there’s a description: ini

ini

ini

Figure 7.1 Demonstratives used for three objects on a table, all within speaker’s reach

166

Penelope Brown and Stephen C. Levinson

(27) a. (closest) ja’ ini ja’ tzaj ini FOC this FOC red here ‘It’s this one this red one.’ b. (middle) ja’ yax-elel-tik ini FOC grue-DER-AJ this. ‘It’s this bluish one.’ c. (farthest) ja’ batzíl yax ini FOC true grue this ‘It’s this true green/blue one.’

If one does not point, then the demonstrative form needs supplementing to individuate a referent, since the same proximal demonstrative category is generally preferred for all referents in this close space: (28) a. (closest) li’ ay-Ø tal ta y-ox kajal ID.PROX.AJ exist-3A DIRcome PREP 3E-three layer tal=i ta ajk’ol tal moel=i. DIRcome=TD.PROX PREP uphill DIRcome DIRascend=TD.PROX ‘Here it is towards us at the 3rd layer/stack towards us towards the uphill direction here.’ b. (middle) ay ta olil=e exist-3A PREP middle=TD.DIST ‘It is at the middle there.’ c. (farthest) li’ ay-Ø ta alan=i. ID.PROX.AJ exist-3A PREP downhill=TD.PROX ‘Here it is towards the downhill direction here.’

If the items are out of reach, the distal form mene can be used, but lumine (far distal) is not possible in this tabletop context. 6

Gesture: Data from Natural Language Usage

In this section, to enrich the elicited data, we provide some naturally occurring examples to show how gesture and physical contact influence the system we have described. If the object is touched, proximal deictics are required, and both proximal and distal demonstratives are very often accompanied by an index-finger point, co-occuring with the deictic form (pointing is indicated on the line above): (29) [2013March27PaseroKun2: 24:23:] [identifying a person mentioned] |POINT. . . . . . . . . . . . . . . . . . . . . |POINT. . . . . . Pedro Lopez Moro ta tey ta ba j-tze’el jo’tik, (.5) NAME NAME NAME PREP there PREP where 1E-side 1PLe,

Tzeltal Demonstratives

167

|POINT s-tat men i Akustin, (1.2) Lopez Jiron=e 3E-father ID.DIST.AJ ID.PROX.AJ NAME NAME NAME=TD.DIST ‘Pedro Lopez Moro at, there at where our side is, the father of that August Lopiz Jiron.’ (30) [2013March27PaseroKun2: 9:12:] [identifying a location] |GESTURE ‘towards here’ . . . . . . . . . . . . . . . li’ x-tal-on jo’tik li’ ta ba ay-Ø ch’ay-el here ASP-come-1A 1PLexcl here PREP where exist-3A fall-NOM ......... |POINT tal men in in in s-pat na tal=e DIRcome ID.DIST.AJ ID.PROX/HESIT 3E-back house DIRcome=TD.DIST ‘Here we come here where there’s that sloping area towards us this- this back of the house towards us.’ (31) [2013March27PaseroKun2: 30:09:] [talking about forcibly taking men for religious cargoes long ago] |POINT ya x-tal-Ø ini li’ ta kokiltik=i ICP ASP-come-3A this ID.PROX.ADV PREP PLACE=TD.PROX ‘He’ll come this here to Kokiltik.’ [a place an hour’s walk away] (32) [2012Jul24_mammanos&AO: 27:48:] [indicating where his net bag cord is, 10 feet away; holds point till object is found] |POINT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ala chij, (1.1) tey ay-Ø, (.7) tey ay-Ø ine. (.4) sDIM cord ID.DIST.ADV exist-3A ID.DIST.ADV TD.DIST HESIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .| tey,(.6) tey. ja’ mene. ID.DIST.ADV ID.DIST.ADV FOC ID.DIST ‘The little (roll of) cord, it’s there, it’s over there. There, there. That’s it.’

Distinct points pick out distinct referents (e.g., two different sisters-in-law, who live in different directions) and can appear even in reported speech: (33) [2012Jul24_yixlelxnt&lus: 5:52:] [reporting talk about a chicken that had wandered off] |POINT, sweeping across gesture ma’yuk la k-il-Ø a jelaw-Ø nax, li’ a not.at.all CMP 1E-see-3A CMP cross-3A PT ID.PROX.ADV CMP jelaw-Ø ta s-pat na ini, ba jojk’o-be-0 a’w-a’y cross-3A PREP 3E-back house TD.PROX IMPgo ask-DIT-3A 2E-hear

168

Penelope Brown and Stephen C. Levinson |POINT to side |POINT to in front. . . . . . . . . a’-jawan y-inam Juliano, xi ek i j-jawan 2E-sister.in.law 3E-wife NAME said too ID.PROX.ADJ 1E-sister.in.law ine. TD.DIST ‘I didn’t see it cross over just here, it crossed at the back of the house this one, go ask your sister-in-law the wife of Juliano, said my sister-in-law this one.’

Pointing gestures also often accompany the demonstration adverb jich ‘like this’; in (34) the jich refers to the size of a cave and the point picks out the inside of the house that is the same: (34) [2012Jul24_mamMO&AO: 49:49:] [identifying size of cave by reference to house] |POINT to house 10 feet away ja’ jich y-util me bit’il k-ala na tey ine FOC thus 3E-inside ID.DIST.AJ how 1E-DIM house there TD.DIST ‘It’s inside (a cave) was like that one there (in) my little house there.’

Finally, contrastive combinations of the different nominal and adverbial demonstratives can highlight their semantic compatibilities, with li’ combining with ini or ine, ine combining with li’, tey or lum, men with tey or lum. The difference between ini and ine is subtle but is essentially a proximal/distal distinction and follows the generalizations in section 5, where closeness to addressee licenses distal. Compare (35) and (36), where the physical distance of the referent from speaker and addressee is greater in (35) yet proximal is used, whereas the referent is being handled by the addressee in (36); both are accompanied by a point: (35) [1980_chanit: 13:27:] [pointing out his elder brother’s house just next door] |POINT ja’ j-bankil ja’ in s-na i li’ FOC 1E-ElBr FOC ID.PROX.AJ 3E-house TD.PROX ID.PROX.ADV nax ini PT TD.PROX ‘It’s my elder brother, this is his house just right here.’ (36) [1991_v5 C-trees: 41:06:] [giving advice about handling tree branch that the addressee is holding, about 10 feet away] |POINT with machete jich xan me jich ya a’-xij koel li’ ine thus again PT thus ICP 1E-insert DIRdown ID.PROX.ADV TD.DIST ‘Like this again, thus you insert it downwards here that one [near addressee].’

Tzeltal Demonstratives

7

169

The Wider Deictic Field in Tzeltal

The demonstratives and anaphorics, naturally, form only a small part of the wider deictic field in Tzeltal (see, e.g., Brown 2006). Some parts of the wider field interact closely with the demonstratives. In addition, the spatial demonstratives are used in the temporal domain. In this section we briefly describe these interactions. 7.1

Motion and Direction in Demonstrative Usage

Tzeltal has a system of directionals, of which the most prominent are tal, ‘hither, coming, towards us’ and bel ‘thither, going, away from us’. These occur naturally with demonstratives inside noun phrases or adverbial phrases, as the following examples make clear. If there’s some kind of virtual line up of referents away from the speaker, then the ‘hither’ form tal reinforces the proximal deictics, while the ‘thither’ form bel reinforces the distal reading of the distal deictics with a kind of fictive motion: (37) baso li’ tal ini cup here DIRcome this ‘this cup here coming’ (i.e., nearest me’). (38) ay-Ø ch’ulna ta Colonia li’-Ø ay-Ø tal ini, ay-Ø exist-3A church PREP PLACE here exist-3A DIRcome this, exist-3A xan yan lum bel ine again another ID.FARDIST.ADV DIRgo TD.DIST ‘There’s a church in the Colonia here it is coming (i.e., towards us) [about 1 km away], there’s another over there awaywards [about 15 km].’ (39) [2012July25XmOsil&lus: 14:12:] [talking about lack of mud in addressee’s land in Cintalapa [100 km away], in contrast to here; no point] li’ ta a’w-ala lumal ek=e, a’w-ala k’inal ID.PROX.ADV PREP your-DIM land too=TD.DIST, your-DIM land bel ek=e DIRgo too=TD.DIST ‘Here in your land too, your land away there.’

The demonstratives also interact with the absolute (cardinal direction) reference system, which is based on a notional “uphill” (S), “downhill” (N), “across” (E & W) set of oppositions (again, see Brown, 2006 for the full system). These have as unmarked or default interpretations that the deictic centre is the origo of the angle, so, for example, ta ajk’ol (PREP ‘uphill’) by default will be read as ‘uphillwards (S) from here’. When combined with demonstratives, these terms have the useful attribute that they specify an angle or arc in which to search for the referent (the choice of proximal vs. distal demonstratives will rest on the principles given above):

170

Penelope Brown and Stephen C. Levinson

(40) li’ ta ajk’ol ini ‘Here uphillwards (from here).’ (41) tey ta ajk’ol ine. ‘There uphillwards (from here).’ (42) ja’ me tey ta ajk’ol ine, ja’ me tey ta ajk’ol=e, ja’ me ta alan=e ‘It’s that one there uphillwards (from here), it’s that one there uphill, it’s that one downhill.’

7.2

Demonstratives, Directionals, and Time Reference

Finally, there has been considerable theoretical interest in the mapping from space to time. The details in the Tzeltal case can be found in Brown (2012). Here we just illustrate some of the ways that the demonstratives, motion verbs, and directionals (tal ‘come’, bel ‘go’, k’ax ‘go past’) are utilized in time expressions.6 First, the proximal deictics specify the present moment, as in (43) below, or the temporal span in the future which includes the present moment, as in (44): (43) a. b. c. d. e.

ya’tik ta ora ini ‘exactly now [lit.: “now at this moment”]’ ban ta ora ini ‘go right now’ la’ ta ora ini ‘come right now’ in k’aal ini ‘this day, today’ ayotikix ta yach’il ja’wil, sba lijkibal k’aal ya’tik ini ‘We are now in the new year, its first day is exactly today [lit.: “this day”].’

(44) a. ya bal stzob sba komite ta xemana ini? ‘Will the committee members gather in this week? (the current week)’ b. ay kuxebal li’ ta cheb u’ ‘Easter is in two months (from now).’ (45) [2012July25XmalOsil&lus: 7:24:] [talking about farming work in Zintalapa, how addressee (L) can no longer keep up with it; M is saying but you’ll finish it and come back here in 2-3 weeks] M: ja’ pero ju’uk, ya x-walkoj-Ø tal koel FOC but no, ICP ASP-turn.around-3A DIRcome DIRdescend a’w-u’un li’ ta che’-oxeb semana=i 3E-RELN here PREP 2-3 week=TD.PROX ‘But no, it will come down around for you here in two-three weeks [from now].’

6

Of the four deictic motion verbs/directionals in Tzeltal, only two (tal ‘coming’, bel ‘going’) are used in time expressions, along with the non-deictic k’ax (‘pass by/through’). The other two (jul ‘arrive here’ and k’ot ‘arrive there’) are not possible in reference to time, e.g., one cannot say *yax jul kuxibal ‘Easter arrives here’.

Tzeltal Demonstratives

171

L: ya x-koj-Ø tal li’ ta jo’lajuneb k’al ICP ASP-descend-3A DIRcome here PREP 15 day bel =i DIRgo=TD.PROX ‘It will come down here in 15 days awaywards [from now].’

One interesting detail is that tal, the “hither” particle, now generally refers to a past temporal unit, while the “thither” directional bel refers to future temporal referents (although it may be possible to shift the deictic centre into the past in narrative). The directionals thus serve to indicate the temporal direction from temporal “here”. Temporal distance in the future can be indicated by proximal or far distal deictics; distal me(n) appears to be unacceptable. (46) ya xsujton tal li’ to ta chaneb xemona . . . . . . . . . . . . . . . ja’ to. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . lum to. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . lum to bel. . . . . . . . . . . . . . . . . . (47) a. b. c. d. e.

‘I’ll return 4 weeks from now’ ‘not until, not before 4 weeks’ ‘far in 4 weeks’ ‘far away in 4 weeks’

li’ ta ja’wil i/ yan jaw’il i ‘Here in next year (2015).’ li’ bel ta yantik ja’wil ‘Here awaywards in another-PL year (2016).’ lum bel ta 2020 ‘Far away in 2020.’ lum bel lajuneb xan ja’wil ‘Far away in 10 more years (2024).’ ya xtalon li’ ta jun/yan ja’wil, ta junio (*me junio *ine) ‘I’ll come here in one/next year, in June.’ [temporal, not spatial ‘here’]

The proximal deictics do not, however, apply readily to time periods entirely included in the past (not overlapping the present); instead, a “time passing towards now” metaphor is employed, along with a “downhill/below is earlier” metaphor: (48) k’ax tal xemana Kuxibal ‘Easter was last week [lit.: “passed coming a week”]’ k’ax tal Kuxibal ‘Easter has gone past’ k’ax tal jun u ‘one month has passed’ (49) k’ax-Ø tal jukeb xemana ta y-anil kuxibal ma la pass-3A DIRcome seven week PREP 3E-underneath Easter NEG CMP j-ti’-tik ti’bal 1E-eat.meat-1PLexcl meat ‘There were seven weeks before [lit.: “below”] Easter that we didn’t eat meat.’ (50) [2013PasaroKun 19:10:] [talking about when K was a boy, 60 years ago] K: yu’ ja’ ya a’-jojk’oj yael te lum to because FOC ICP 2E-ask EV DET IDFARDIST PT tal yael ine DIRcome EV TD.DIST ‘What do you say, then, is it that you are asking about way back then coming (towards now)?’

172

Penelope Brown and Stephen C. Levinson

A: ta k’alal to ta kerem-at to tal=e bit’il a, bi’- mach’a PREP when PT PREP boy-2A PT DIRcome=CL, what what- who la y-ak-at och-at ta kunerol-il? CMP 3E-give-2A enter-2A PREP municipal.president-Vl ‘When you were still a boy coming (i.e., growing up), how was it, how, who got you to be the municipal president?’ (51) la bal stzob sbaik komite k’ax tal xemana i / ja’wil ini? ‘Did the committee members meet in this past week/year?’ la bal stzob sbaik spisil kirsanoetik ta kuxibal k’ax tal ja’wil=e ‘last year’? la bal stzob sbaik spisil kirasanoetik ta kuxibal k’ax tal li’ ja’wil=i ‘this last year’

For events far removed in time, either future or past, the distal demonstratives can be used: (52) lum to bel ta Decembre sk’in Jalame’tik, lum bel ta enero sk’in Kajkanantik [bel = FUT] lum toix tal ine – very long (many years) ago namej toix – long ago (ancestral time) and bel can be used into the future on any scale: (53) pajel, chawe’ bel – in 1–2 days from now ja’wil bel ine – next year lum ta ja’wil bel – far away next year

These constraints are diagrammed in Figure 7.2, with the slope angled to reflect the conceptual mapping onto the “uphill/downhill” distinctions (downhill is earlier in time). FU

TU

RE

be

l

lum

PR

ES

EN ini

T(

no

w)

ta

l

li’i,

lum PA

ST

Figure 7.2 Demonstrative and directionals in time reference

Tzeltal Demonstratives

8

173

Conclusions

We have shown here that, as in other Mayan languages, Tzeltal demonstratives are mostly compounded from an initial and final deictic and may include yet other deictic elements. The formal complexity is matched by a set of semantic oppositions that are obscured by various pragmatic factors. The demonstratives questionnaire (Wilkins, 1999) helped to reveal the underlying system, which is essentially based on zones of proximity to speaker, or to addressee, or zones of distance from both speaker and addressee. The encoding of a notion like “peripersonal space” of a speech act participant was predicted by Kemmerer (1999) on the basis of neurological findings about the treatment of space in neurocognition. He was, however, disappointed not to find it clearly reflected in his own review of demonstrative systems. Tzeltal – as well as the descriptions of Yélî Dnye (Levinson, this volume) and Dalabon (Cutfield, this volume) – offer interesting support for his theory. A special feature of Tzeltal demonstratives is that a co-extensive gesture licenses a proximal deictic in all cases except references to objects in the addressee’s peri-personal space. This is because the proximal forms are associated with ostensive or presentative displays, which of course are also attention directing. Finally, the demonstratives interact with the system of absolute (cardinal) directions, allowing directional vectors to be encoded linguistically with the demonstratives (as they can also be gesturally, of course). In the temporal uses, the collapse of the three-dimensional spatial distinctions onto the onedimensional time line still leaves direction (past or future) unspecified – the directionals (“hither/thither”) can then then be used to indicate past or future. Tzeltal speakers thus speak as if the future is “thither (away from now), uphill” and the past “hither (coming towards now, from downhill)”, in yet another manifestation of this interesting deictic system. References Aissen, J. (1987). Tzotzil clause structure. Dordrecht: Reidel Publishing Co. Bohnemeyer, J. (2012). Yucatec demonstratives in interaction: Spontaneous vs. elicited data. In A. C. Shalley, ed., Practical theories and empirical practice: A linguistic perspective. Amsterdam: John Benjamins, pp. 99–128. Brown, P. (2006). A sketch of a grammar of space in Tzeltal. In S. C. Levinson & D. Wilkins, eds., Grammars of space: Explorations in cognitive diversity. Cambridge: Cambridge University Press, pp. 230–272. (2012). Time and space in Tzeltal: Is the future uphill? Frontiers in Psychology, 3, 212. Campbell, L. (1987). Tzeltal dialects: New and old. Anthropological Linguistics, 29, 549–570.

174

Penelope Brown and Stephen C. Levinson

Diessel, H. (1999a). The morphosyntax of demonstratives in synchrony and diachrony. Linguistic Typology, 3(1), 1–49. (1999b). Demonstratives: Form, function and grammaticalization. Amsterdam: John Benjamins. Dixon, R. M. W. (2003). Demonstratives: A cross-linguistic typology. Studies in Language, 27(1), 61–112. Enfield, N. J. & Bohnemeyer, J. (2001). Hidden colour-chips task. In S. C. Levinson & N. J. Enfield, eds., Manual for the field season 2001. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 21–28. Fillmore, C. (1982). Towards a descriptive framework for spatial deixis. In R. J. Jarvella & Wolfgang Klein, eds., Speech, place and action: Studies in deixis and related topics. Chichester: Wiley, pp. 31–59. (1997). Lectures on deixis. Stanford, CA: CSLI Publications. Greenberg, J. H. (1966). Universals of language. Cambridge, MA/London: MIT Press, pp. 43–76. Hanks, W. F. (1990). Referential practice: Language and lived space among the Maya. Chicago: University of Chicago Press. (1992). The indexical ground of deictic reference. In A. Duranti & C. Goodwin, eds., Rethinking context: Language as an interactive phenomenon. Cambridge: Cambridge University Press, pp. 1–42. (1996). Language form and communicative practices. In J. J. Gumperz & S. C. Levinson, eds., Rethinking linguistic relativity. Cambridge: Cambridge University Press, pp. 232–270. (2005). Explorations in the deictic field. Current Anthropology, 46(2), 191–220. Haviland, J. B. (1981). Sk’op Sotz’leb; El Tzotzil de San Lorenzo Zinacantán. Mexico City: Centro de Estudios Mayas, Instituto de Investigaciones Filológicas, Universidad Nacional Autónoma de México. Instituto Nacional de Estadistica y Geografia (INEGI) (2010). Census. (Available online at www3.inegi.org.mx/.) Kaplan, D. (1978). On the logic of demonstratives. Journal of Philosophical Logic, 8, 81–98. Kaufman, T. (1972). El proto-tzeltal-tzotzil: Fonología comparada y diccionario reconstruido. Centro de Estudios Mayas, Cuaderno 5. Mexico City: UNAM. Kemmerer, D. (1999). ‘Near’ and ‘far’ in language and perception. Cognition, 73, 35–63. Laughlin, R. M. & Haviland, J. B. (1988). The Great Tzotzil dictionary of Santo Domingo Zinacantán, with grammatical analysis and historical commentary. Smithsonian Contributions to Anthropology, No. 31. Washington, DC: Smithsonian Institution Press. Levinson, S. C. (1983). Pragmatics. Cambridge: Cambridge University Press. (1999). Deixis and demonstratives. In D. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 29–40. (Available online at http://fieldmanuals.mpi.nl/volumes/1999/deixis-demonstratives/.) Lyons, J. (1977). Semantics. Vol. II. Cambridge: Cambridge University Press. Pederson, E. C Wilkins, D. P. (1996). A cross- linguistic questionnaire on ‘demonstratives’. In S. C. Levinson, ed., Manual for the 1996 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp.1–14.

Tzeltal Demonstratives

175

Polian, G. (2013). Grammatica del Tseltal de Oxchuc. Mexico City: Centro de Investigaciones y Estudios Superiores en Antropología Social. Searle, J. (1969). Speech acts: An essay in the philosophy of language. Cambridge: Cambridge University Press. Tomasello, M. (2014). A natural history of human thinking. Cambridge, MA: Harvard University Press. Wilkins, D. P. (1999; this volume). The 1999 demonstrative questionnaire: ‘this’ and ‘that’ in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24. Wittgenstein, L. (1953). Philosophical investigations. G. E. M. Anscombe & R. Rhees (eds.), G. E. M. Anscombe (trans.). Oxford: Blackwell.

8

Yucatec Demonstratives in Interaction: Spontaneous versus Elicited Data Jürgen Bohnemeyer

This chapter compares Hanks’ (1990; 2005) “practice” approach to the demonstratives of Yucatec Maya based on the recording of spontaneously occurring interactions to the results obtained by the author with the elicitation questionnaire developed by Wilkins (1999). The study of the meaning and use of demonstratives represents particular challenges to linguistic data gathering because of their context dependency and the role of interactional factors such as attention sharing. The questionnaire study disconfirmed any direct impact of the location of the addressee on the choice of demonstrative and showed a systematic contrast between simple forms used for joint attention and augmented ones used for attention-direction. It is argued that observation of spontaneous interactions and elicitation should be pursued in tandem. 1

Introduction

This chapter discusses the demonstrative system of Yucatec, a Mayan language spoken on the Yucatán Peninsula in Mexico and Belize. The emphasis here is on the uses of these demonstrative forms for exophoric reference, i.e. for reference to real or imagined entities or events present in space at the moment of utterance. Other typical uses of demonstratives, e.g. for anaphoric reference tracking, “textual deixis” (Lyons, 1977), and “recognitional” uses (Himmelmann, 1996), are only considered to the extent that they shed light on the question of just how much exophoric reference is actually a semantic property of demonstrative forms rather than the result of pragmatic inferences. The vantage point from which the discussion proceeds is a comparison of different approaches to the analysis of demonstrative systems. A widely known study of the Yucatec system by Hanks (1990; 2005) is compared with results obtained by the author in fieldwork applying the Demonstrative Questionnaire developed by David Wilkins at the Max Planck Institute for Psycholinguistics (see section 4). The methodological backdrop to this comparison is the lack of 176

Yucatec Demonstratives

177

standard techniques for analyzing the semantics of indexical expressions. In the case of spatial deixis, this problem becomes particularly obvious, because the distinctions that are made are more complex than in other deictic domains. Numerous debates over the analysis of particular demonstrative systems bear witness to this; one example is the debate that has gone on for decades over the Turkish demonstrative šu. Some (e.g. Kornfilt, 1997) have claimed it to be used to refer to objects at mid-distance from the speaker, while others (e.g. Lyons, 1977) have considered it to be used for reference to objects close to the addressee. However, there never was a methodology in place to settle the question. Standard techniques of semantic analysis seek to determine an invariant of reference, something that is being referred to across all the contexts in which the particular expression is used, and to extract the meaning of the expression by eliminating all context dependencies. But the meanings of “indexicals” are particular kinds of context dependency.1 What is invariant across the contexts in which indexicals are used is not what is referred to but how it is referred to (Kaplan, 1989). So what is needed in order to study the use of demonstratives for exophoric spatial reference is a methodology that allows one to keep track of the interactional parameters of the speech context in which these forms are used.2 This includes the participants, their locations in real and in social space, and the location of the reference object (or “denotatum”) in these co-ordinate systems; e.g. the attention sharing among the speech act participants and the information status of the referent in discourse, but also possession of the object referred to by one of the participants. For example, Özyürek (1998) presents evidence suggesting that the use of šu does not depend on the location of the reference object relative to speaker or addressee, but rather on whether or not a joint focus of attention has been established among the two that includes the referent (see also Küntay and Özyürek, 2002). The two approaches discussed here attempt to provide a methodology for studying the interactional parameters in the meaning and use of spatial deixis. Hanks’ “practice” approach is based on recordings of spontaneous interactions in culturally typical settings such as the household, corn field, religious ceremonies, etc. Hanks developed coding schemas for the participants when deictically referring to places and objects in these settings. These coding schemas showed the spatial layout of the settings, and Hanks assigned numbers to 1

2

I use the term “indexical” as a cover term for any kind of expression which triggers retrieval of a referent through a context dependency that is part of the lexical meaning of the expression. This includes deictic expressions, anaphoric expressions, and definite descriptions. The study of interactional aspects of linguistic meaning is one of the leitmotifs of the volume edited by Schalley (2012), where this chapter first appeared; the book is dedicated to Dietmar Zaefferer’s work on interaction, a theme to which the present chapter hopes to make a small contribution.

178

Jürgen Bohnemeyer

certain prominent locations and objects to code the interactions he recorded in the settings. Based on these data, Hanks carried out a detailed analysis of the use of the demonstrative forms in the interactions he had observed. However, Hanks did not attempt to go beyond usage and venture into an analysis of the underlying semantics of the forms. In contrast, the Demonstrative Questionnaire (Wilkins, 1999; Chapter 2, this volume) gives instructions for the enactment of 25 scenarios, specifying for each scenario the relative locations of the participants and the object to be referred to, but also the referent’s status in discourse and the object’s status with respect to the interlocutors’ focus of attention (see section 4). This is an instance of the method of controlled elicitation with non-verbal stimuli, which plays a crucial role in semantic typology; see Senft (2012) for further examples. Enacting the 25 scenes, usage preferences and judgments of acceptability of demonstrative forms for exophoric reference by adult native speakers of Yucatec were elicited. The results thus obtained do permit a (partial) assessment of the semantics of the demonstrative forms, since they involve systematic contrasts in the elicitation scenarios and negative evidence concerning the use of forms in particular scenarios, i.e. evidence of what native speakers do not consider acceptable in a particular context. Some of the outcomes of the questionnaire study are surprising from the point of view of Hanks’ analysis, in the sense that they would not have been directly predicted from it. However, conversely, many of Hanks’ findings concerning the use of the demonstrative forms in spontaneous interactions could not possibly be predicted from the results of the questionnaire study. An analysis of elicited data can only make generalizations over scenarios of the kind that are tested during the elicitation and over the interactional parameters that are controlled for during the elicitation. More generally, meaning is only a partial predictor of use, just as use is only a partial predictor of meaning. Hence, neither the study of spontaneous interactions nor controlled elicitation can provide a complete picture of the meaning and use of demonstratives by themselves. Rather, the combined use of both methods is advocated here. The discussion in section 5 generalizes this maxim beyond the study of indexicals to all forms of the collection of data about languages from their speakers. Researchers are bound to miss important insights if they restrict themselves to either approach. Section 2 introduces the relevant structural details of the Yucatec expressions used in exophoric spatial reference. The system is a fairly intricate one. Section 3 summarizes Hanks’ (1990; 2005) analysis and discusses the evidence for and against a possible addressee-based analysis of some of the forms at the outset of the questionnaire study. In section 4, the Yucatec responses collected with the Demonstrative Questionnaire are discussed. The study has produced three main findings: (a) the location of the addressee does not seem to have a direct impact on the use of Yucatec demonstrative forms; (b) the distal (in Hanks’ terms,

Yucatec Demonstratives

179

“non-immediate”) forms are semantically distance-neutral and arguably not even specified for exophoric reference; and (c) the proximal-distal (in Hanks’ terms, “immediate” versus “non-immediate”) opposition intersects with a contrast between simple forms used with a pre-established focus of attention and augmented forms used for attention-direction. Performance of the two approaches on the Yucatec data is compared in section 5. 2

A Sketch of the Expression of Spatial Deixis in Yucatec

Yucatec is spoken in the northeast of the Mayan area, all across the Yucatán Peninsula, by approximately 759,000 people in Mexico3 and 6,000 in Belize (Lewis, 2009). Dialect differentiation is low; all contemporary varieties are readily mutually intelligible. On historic and sociolinguistic grounds, a western variety spoken in the outskirts of the cities of Mérida and Campeche and the region between these cities may be tentatively distinguished from an eastern variety spoken everywhere else (Edmonson, 1986; Pfeiler, 1995). On this classification, Hanks’ field site in Oxkutzkab in the Mexican state of Yucatán would likely fall in the western dialect region, whereas the author’s field site in Yaxley in the Mexican state of Quintana Roo is situated in the eastern region. The two dialects are mutually intelligible without restriction, and the differences between them are quite subtle; they concern only a few lexical items, some morphophonemic processes, and certain grammatical operators, such as aspectual and modal markers. The basic grammatical system, including word order, phrase structure, and inflectional and derivational morphology, is the same across the two dialects. The dialect described here lacks the clause-final particle -be’ for referents that are audible but not visible (see Table 8.1). With this exception, all examples in Hanks (1990; 2005) seem inconspicuous to me, judging on the basis of 20 years of field experience in the eastern dialect region. To get a first impression of the expression of spatial deixis in Yucatec, consider Table 8.1, based on Hanks (1990: 18–19).4 Where Indo-European languages use a single form to mark indexical reference, say demonstratives like this and that or place adverbs like here and there, Yucatec uses combinations of two morphemes that occur in different positions in the clause. One part of these combinations occurs in the positions where 3 4

2005 census data shows a decline by more than 40,000 speakers age 5 or older since 2000 (cf. PHLI, 2009 and PerfilMayaweb, 2005). Hanks’ synopsis of Yucatecan indexicals also includes personal pronouns, temporal adverbs, and an indexical manner adverb that translates ‘like this/that’. The representation in Table 8.1 is restricted to indexical forms used for reference to places and objects in space, and it is couched in the terminology that is used throughout this chapter; so the labels deviate from Hanks’. Two other studies of deixis in Yucatec should be mentioned here as well: Hanks (1984) studies the interactions of deictic forms with factors of evidentiality, and Vapnarsky (1999: ch. 3) examines the deictic temporal adverbials of the language.

180

Jürgen Bohnemeyer

Table 8.1 Synopsis of Yucatecan spatial indexicals (based on Hanks, 1990: 18–19) Clause-final indexical particle Non-final indexical stem =a’ Presentative he’l

=o’

-be’

=i’

=e’

‘Here it is’

he’la’ / he’ . . . =a’

‘There it is’

he’lo’ / he’ . . . =o’ he’l . . . -be’ Adverbial te’l

‘There it comes (audible)’ ‘Right there/here’

te’la’ / te’ . . . =a’

‘There’

te’lo’ / te’ . . . =o’ ti’ . . . =i ’

ti’ way tol Determiner le

Gloss

tolo’ / to . . . =o’

‘There (anaphoric)’ way . . . = e’ ‘(In) here’ ‘(Out) there’ ‘This’

lela’ / le . . . =a’

‘That’

lelo’ / le . . . =o’ le . . . =e’

‘As for that one’

English speakers expect them: as place adverbs, presentative adverbs like French voilà, or determiners. The other component is a clitic particle that always appears in clause-final position. Whenever one of the determiners or adverbs in the leftmost column of Table 8.1 is used, it co-occurs with one of the terminal particles. So the determiners and adverbs are triggers of the terminal particles. Each triggering adverb or determiner co-occurs only with a subset of the terminal particles. Exophoric spatial reference is largely restricted to the particles =a’ and =o’. Both particles occur with the presentative adverb he’l, the locative adverb te’l, and the determiner le. With the last two, =a’ and =o’ may be said in first approximation to distinguish proximal and distal reference. In the same sense, te’la’ may be glossed as ‘here’ and te’lo’ as ‘there’, and similarly lela’ as ‘this’ and lelo’ as ‘that’. However, the discussion in the following sections shows that the proximal-distal characterization is inadequate. I therefore adopt Hanks’ (1990; 2005) labels immediate for =a’ and nonimmediate for =o’.

Yucatec Demonstratives

181

In combination with the presentative adverb he’, the contrast is slightly different. In addition, the particle =e’ needs to be considered, but only in combination with the locative adverb way ‘here’. Consider a few examples, starting with the presentative adverb he’l. In (1), he’l occurs as the main predicate of the clause:5 (1)

He’l hun-p’íit ts’àak = a’! PRSV one-bit cure\ATP = D1 ‘Here’s some medicine!’

With these words, the speaker would typically hand over the medicine to the addressee. Note the immediate particle =a’, which obligatorily accompanies he’l in this function. He’l also occurs as a noun-phrase-internal modifier, as in example (2): (2)

K-u=bin Xokempich le=bèeh he’l=a’? IMPF-A3=go Xokempich DET=way PRSV=D1 ‘Does this way here go to Xokempich?’

In this case, he’l can occur with either =a’ or =o’, and its function is no longer presentative; instead, it is used to call the addressee’s attention to the referent. This function is further discussed below. It is also possible to use he’ nominalized in the same function, as in example (3): (3)

5

Ba’x le=he’l=o’? Ba’x what DET=PRSV=D2 what(B3SG) u=k’àaba’? A3=name(B3SG) ‘What’s this? What’s its name?’

Examples (1)–(3), (7)–(8), (11), and (18) are based on Blair and Vermont-Salas, 1965–1967. Examples (4) and (6) are from the unpublished text Bix u meta’l hump’éel k’axbil nah by Esteban Pool Kaaw recorded and transcribed by Christian Lehmann with the aid of Ramón May Cupul. In some cases, the examples have been simplified for expository purposes. Examples (5), (8)–(9), (14)–(17), and (20)–(21) were elicited with the Demonstrative Questionnaire (Wilkins, 1999); cf. section 4. The remaining examples have been collected in other contexts. The orthographic representation in this chapter is morphemic rather than morphophonemic. The orthography applied is based on Lehmann (1998). In the interlinear morpheme glosses, the following conventions are used: ‘-’ for affixes; ‘=’ for clitics; ‘+’ for compounding; ‘\’ for subsegmental realization or infixation. Abbreviations in the glosses include the following: 2 – 2nd person; 3 – 3rd person; A – set-A (‘ergative’/possessor) clitics; ATP – antipassive derivation; B – set-B (‘absolutive’) suffixes; CAUSE – causal preposition; D1 – immediate clause-final indexical particle; D2 – non-immediate clause-final indexical particle; D3 – text-deictic clause-final particle; DET – determiner stem; EXIST – existential/locative/possessive predicate; F – feminine prefix; IMPF – imperfective aspect; INC – incompletive aspect; INCH – inchoative derivation; INSTR – instrument nominalization; IRR – irrealis modality; NEG – negation; PREP – generic preposition; PROG – progressive aspect; PROSP – prospective aspect; PRSV – presentative stem; PRV – perfective aspect; SG – singular; SR – subordinator.

182

Jürgen Bohnemeyer

Moving on to the locative adverb te’l, example (4) shows it forming an adverbial expanded by a prepositional phrase: (4)

U = hòol+nahken u = bin A3 = hole+house SR.IRR

A3 = go

te’l t-u = mòoy = a’. there PREP-A3 = apse = D1 ‘The door is what will end up here in the apse’

Like he’l, the stem te’l does not by itself have an indexical meaning. The gloss ‘there’ is thus somewhat misleading. Te’l translates as ‘there’ in combination with the non-immediate particle =o’, but as ‘here’ in combination with the immediate particle =a’. This adverb does not modify a noun phrase by itself. However, a relative clause headed by the locative/existential predicator yàan can be constructed around it, as in (5): (5)

le=lìibro yàan te’l=o’ DET=book [EXIST(B3SG) there=D2]S ‘the book that’s there’ (distal or anaphoric!)

And of course it is again possible to nominalize te’l and use it as a noun phrase head itself, as in (6): (6)

Le=te’l=a’, es que kul-ub DET=there=D1 is.which sit-INSTR(B3SG) ‘This one here, it’s a pillar (lit. thing for sitting).’

Finally, the determiner le acts as a proximal demonstrative when combined with the immediate particle =a’: (7)

A = ti’a’l A2 = property(B3SG) ‘Is this house yours?’

le = nah = a’? DET = house = D1

There is an alternative form lel- that constitutes a noun phrase head itself, as in (8): (8)

A=ti’a’l lel=a’? A2=property(B3SG) DET=D1 ‘Is this yours?’

Now when le is combined with the non-immediate particle =o’, it can be used for both distal deictic reference, as in (9), and the marking of definiteness, as in (10):

Yucatec Demonstratives

(9) A = ti’a’l A2 = property(B3SG) ‘Is that book yours?’

183

le = lìibro = o’? DET = book = D2

(10) Káa = h-òok káa = PRV-enter(B3SG)

le = x-ch’úup chak u = nòok’ = o’, (...) DET = F-female red(B3SG) 3 = garment = D2 ‘(And then) the woman dressed in red entered, (...)’

In fact, any definite description whose lexical head is a common noun has to be accompanied by a clause-final particle, and in case the referent has been mentioned before or is assumed by the speaker to be uniquely identifiable to the addressee, the non-immediate particle =o’ (and some form of the determiner le(l)) is used. And similarly, the counterpart lelo’, which constitutes a noun phrase (or “determiner phrase”) by itself, can be used anaphorically, as in (11): (11) Ba’x k’ìin k-uy=úuch-ul lel=o’? what sun IMPF-A3=happen-INC DET=D2 ‘What day does that usually happen?’

So effectively, the presence of a clause-final particle indicates that the clause contains an expression other than a pronoun which is used indexically. It has been shown that the forms that trigger the non-immediate particle =o’ occur with both deictic (9) and anaphoric (11) reference or as definite markers (10). This is also true of the locative adverb combination te’lo’ and the manner adverb combination bèeyo’. This gives rise to the hypothesis that it is really only the combinations with the immediate particle =a’ that are semantically specified for exophoric reference, whereas the non-immediate =o’-forms merely have a more general indexical meaning which does not exclude exophoric use but does not entail it either. This hypothesis is pursued further in section 4. Since the determiners and adverbials obligatorily co-occur with the terminal particles, it is a non-trivial task to distinguish the contribution of the determiner/adverbial to the meaning of the combination from that of the terminal particle. At this point, I will confine myself to preliminary observations. The adverbs indicate that the referent is a place, the determiner that it is an entity (a person, animal, or object). The different syntactic categories are of course also associated with different ranges of syntactic functions as arguments versus

184

Jürgen Bohnemeyer

adjuncts, etc. The adverb or determiner also indicates that the referent is given indexically. However, in the case of the place adverb te’l and the determiner le it is only the terminal particle that distinguishes between deictic and anaphoric reference. 3

Demonstratives in Spontaneous Interactions: Hanks (1990; 2005)

In this section, I summarize and discuss Hanks’ (1990; 2005) analysis of how Yucatec speakers use the demonstrative system for spatial reference in everyday interactions. Of central concern in the following is the question as to the role of the addressee in this system and whether it is semantically encoded. As mentioned above, Hanks avoids the terms “proximal” and “distal”, arguing that these are “obscured in standard approaches to deixis which take as their touchstone ‘real’ space rather than social interaction” (1990: 488). Instead, Hanks uses the labels “immediate” and “non-immediate”, which I adopt here. Among the determiners, the immediate versus non-immediate opposition is the only opposition there is. But in the adverbial system, the immediate–nonimmediate opposition between te’la’ ‘here’ and te’lo’ ‘there’ semantically intersects with an “inclusive-exclusive” opposition between waye’ ‘here’ and tolo’ ‘there’. In other words, there are two heres and two theres in Yucatec. Hanks calls the “inclusive–exclusive” opposition between waye’ and tolo’ “egocentric”. This distinction presupposes some kind of perimeter around the speaker, such that waye’ refers to the inside of that perimeter and tolo’ to its outside. The perimeter can be defined by the boundaries of for example the house, the field, the village, or the state where the conversation takes place. The addressee is normally inside the perimeter as well. Tolo’ is used in indiscriminate reference to things that are “out there” in the relevant respect. Table 8.2 summarizes Hanks’ analysis of the space-deictic determiners and adverbs of Yucatec. Waye’ ‘here’ and tolo’ ‘(out) there’ cannot normally be contrasted in reference to places that speaker and addressee have visual access to, as such places would be within the perimeter and hence entirely inside the domain of waye’. Similarly, if there are multiple possible referents for waye’, they are concentric and thus cannot easily be distinguished gesturally. Therefore, both terms can be used without accompanying gestures, and the only gestures that do accompany them are gestures that do not point to specific places.6

6

The “egocentric” terms play only a marginal role in responses to the Demonstrative Questionnaire; therefore, they are not discussed further. Way . . . =e’ ‘here’ did not occur at all, and tol . . . =o’ ‘out there’ occurred only once in a consultant’s first response. Interestingly, the two scenes one would predict to be most likely to trigger tol . . . =o’ based on Hanks’ “perimeter” analysis, 20 and 21, failed to elicit tol . . . =o’.

Yucatec Demonstratives

185

Table 8.2 The semantics of the space-deictic determiners and adverbs of Yucatec according to Hanks (1990) Inclusive

Adverbs Determiners

Immediate

Non-immediate

way . . . =e’ ‘here’ te’l . . . =a’ ‘there’ lel=a’ ‘this one’ le . . . =a’ ‘this’

te’l . . . =o’ ‘there’ lel=o’ ‘that one’ le . . . =o’ ‘that’

Exclusive tol . . . =o’ ‘there’

In contrast to the “egocentric” “inclusive–exclusive” distinction, the “immediate–non-immediate” opposition between te’l . . . =a’ ‘here’ and te’l . . . =o’ ‘there’ and the determiners lela’/le . . . =a’ ‘this’ and lelo’/le . . . =o’ constitutes what Hanks calls a “sociocentric” system. He observes that these forms are used contrastively with respect to speaker and addressee, respectively: immediate forms are used for reference to objects or places closer to the speaker than to the addressee, while non-immediate forms are used in reference to objects or places closer to the addressee. Hanks notes that the usage patterns his analysis ascribes to the immediate and non-immediate forms differ “in two details: (i) the relative remoteness of the ( . . . ) possible referents, and (ii) the foregrounding of the addressee rather than the speaker. The second feature is motivated by the fairly consistent association between the ‘there’ of te’lo’ and the addressee’s location” (Hanks 1990: 437). Consider some of the examples that Hanks quotes in support of this analysis. These are examples in which speaker and addressee are in relatively close proximity, such as (12)–(13) in which a child is chided by an adult while both are in the same room and in the second case even less than 2 meters apart. Yet the speaker picks the non-immediate form to refer to the child’s location: (12) Mak

a=chi’ te’l=o’, páal! Close(B3SG) A2=mouth there=D2 child ‘Shut up over there, kid!’ (Hanks, 1990: 438)

(13) Ts’a’

le=ba’l te’l=o’! Give/put(B3SG) DET=thing there=D2 ‘Put that thing down there!’ (Hanks, 1990: 438)

The selection of the non-immediate forms – the determiner lelo’ / le . . . =o’ and the place adverb combination te’lo’ / te’l . . . =o’ – in these examples raises the question as to whether these forms are semantically addressee-based; i.e., whether they encode proximity to the addressee, rather than distance from the speaker, or whether the association with the addressee’s location Hanks

186

Jürgen Bohnemeyer

observed is merely a property of how these forms are used in situations such as those in (12)–(13). Hanks (1990) does not address this question. An example of a language with an addressee-based demonstrative is Japanese. Japanese has a demonstrative ko for referents close to the speaker, a demonstrative so for referents close to the addressee, and a demonstrative a for referents that are in the proximity of neither the speaker nor the addressee.7 Addressee-based terms like Japanese so are found somewhat regularly in three-term demonstrative systems; they compete with other types of three-term systems that distinguish three degrees of distance from the speaker or two degrees plus one distanceneutral term, as in the case of Turkish. Applying an addressee-based analysis to the immediate–non-immediate contrast in Yucatec straight away runs into the problem that the latter is a binary contrast. So one would have lela’ / le . . . =a’ or the adverb combination te’la’ / te’l . . . =a’ for entities and places close to the speaker, lelo’ / le . . . =o’ or the adverb combination te’lo’ / te’l . . . =o’ for entities and places close to the addressee, and then the question arises as to what to use for entities and places that are neither in the speaker’s nor in the addressee’s zone of proximity. For this reason, a two-term demonstrative or deictic adverb system is not very likely to include an addressee-based term; and indeed, the typological surveys of Anderson and Keenan (1985) and Diessel (1999) do not include a single example of such a system – only three-or-more-term systems may include addressee-based terms. However, a two-term system with one speaker-based and one addressee-based term is by no means impossible. One conceivable realization of such a system might be found in a language in which demonstratives or deictic adverbs are simply not used in reference to objects that are close neither to the speaker nor to the addressee. The hypothetical language would employ other means to this end, such as explicit locative descriptions. But this is very clearly not the case in the dialect of Yucatec discussed here.8 Hanks (1990: 490) in fact observes that the =o’ forms are used in reference to entities and places in both the addressee’s zone and the “common field”. It is not completely clear to me how this “common field” is to be construed (see Enfield, 2003 for a possibly similar analysis). At any rate, the “common field”

7 8

This is a simplified account based on unpublished research by Sotaro Kita. See Kita and Walsh Dickey (1998: 66) and Senft and Smits (2000: 69) for summaries. As mentioned above, the “egocentric” adverb tol . . . =o’ ‘out there’ is according to Hanks used for vague reference to places outside some perimeter around the deictic center. In a hypothetical two-term system with forms for the speaker’s and the addressee’s zones, this would indeed be a solution to the problem of referring to objects and places that are in neither zone. However, it would be a solution only for those special circumstances in which tol . . . =o’ is used (i.e., there is a salient perimeter around the deictic center, and the reference object/place is situated outside it.) As likewise mentioned above, tol . . . =o’ plays only a marginal role in the responses to the Demonstrative Questionnaire.

Yucatec Demonstratives

187

would presumably cover a significant part of the space outside both speaker’s and addressee’s “zones” (i.e. areas of proximity). Leaving aside the issue of how entities outside the “common field” would be referred to, the main question that arises is how to reconcile the “foregrounding” of the addressee by the non-immediate forms with the fact that they are also used for reference to objects and places in the common field outside the addressee’s zone. Hanks (2005) suggests that this foregrounding is a pragmatic rather than semantic effect:9 The rule of thumb is therefore simply, in pragmatically contrastive contexts such as greetings and scoldings, to treat [the speaker’s] field as a’ and [the addressee’s] field as o’. When I state this association as a rule of thumb I mean to underscore that it is not part of the semantics of Yucatec deixis, since it is easy to find examples in which the association is canceled. It is, however, part of the routine handling of types of exchange that happen throughout any ordinary day. (Hanks, 2005: 206; emphasis mine)

Hanks’ use of the term “cancellation” suggests that his “rule of thumb” is a Gricean stereotype implicature, i.e. that the addressee’s zone of proximity is in many instances the stereotypical search domain of the non-immediate forms. This, however, implies that those “easy to find” situations in which the nonimmediate forms are used in exophoric reference to entities or places outside the addressee’s zone of proximity are somewhat less typical. It is one of the strengths of elicitation approaches such as the one presented in the following section that they permit the realization and testing of reference in such atypical situations. This puts the researcher in a position to distinguish between semantic and pragmatic meaning components. In order to determine the role of the addressee’s location in the use of the non-immediate forms, their use needs to be examined in contexts in which the relative locations of speaker, addressee, and reference object are systematically varied. Controlling these variables is one of the main goals of the Demonstrative Questionnaire discussed in the following section. However, the realization of the questionnaire scenes with Yucatec speakers failed to produce evidence of prototype effects associated with the addressee’s zone. If the prototypical referent of the non-immediate forms is in the addressee’s zone, while entities and places in the common field outside the addressee’s zone are more peripheral instances of the extensions of the non-immediate forms, this entails that native speakers use the non-immediate forms more readily and more consistently in reference to objects and places in the addressee’s zone than in reference to objects and places in the common ground outside the addressee’s zone. This is not consistent with the findings from the Demonstrative Questionnaire study presented in the following section – these 9

It is unclear whether this clarification was in fact prompted by earlier versions of the present chapter.

188

Jürgen Bohnemeyer

data indicate that the addressee’s zone plays no direct role whatsoever in the use of the spatial deictics. Note that this by no means precludes “indirect” addressee-based effects as discussed by Meira (2003; cf. also Enfield, 2003). That is, the addressee may well have an impact on what counts as immediate or proximal for the speaker. For example, it is argued below that physical accessibility is one of the parameters that determine whether a place is judged as proximal. And the presence of the addressee may of course influence the accessibility of the reference object or location to the speaker. Consider again examples (12)–(13). While the reference entity/place is close to the speaker in both cases, it is not immediately physically accessible to the speaker. This may well be explained in part with reference to the fact that the speaker is referring to entities and places controlled by the addressee. Control – in a sense of the term “control” that still remains to be specified – would then be one possible cause of indirect addressee-based effects. 4

Demonstratives in Elicited Productions: The Questionnaire Study

The Demonstrative Questionnaire (Wilkins, 1999; this volume) was developed by D. P. Wilkins for the Space Project at the Max Planck Institute for Psycholinguistics. Since 1999, it has been applied to the study of spatial deixis in numerous languages spoken around the world. The questionnaire was designed to study the form of utterances that make non-contrastive exophoric reference to single objects present in space at varying degrees of distance from speaker and addressee at the moment of utterance. It describes 25 scenes which the researcher is meant to enact together with native-speaker consultants. The variables controlled in the Demonstrative Questionnaire were of course determined on the basis of prior research both within and outside the Space Project. Hanks’ (1990) influential study of demonstrative use in spontaneous interactions in Yucatec was among the sources that were considered in the design of the questionnaire. A major goal of Hanks (1990; 2005) is to show that the meaning and use of demonstratives are primarily governed by interactional variables rather than by purely spatial properties such as in particular measurable distance. In the design of the Demonstrative Questionnaire, both spatial and interactional variables are controlled for. The descriptions specify for each scene a setting (e.g. inside a walled-off space; on a ballgame field); a spatial configuration of speaker, addressee, and reference object, and optionally a bystander, within that setting; the kind of reference object at issue (one of the speaker’s teeth, a bug, a radio, book, or ball); and a number of additional properties such as whether joint attention between speaker and addressee is on the referent at the moment of utterance or is rather directed to it by the speaker in the course of the utterance, whether the object has been mentioned before in the course of the conversation, and whether the object is owned by one of the

Yucatec Demonstratives

189

interlocutors.10 The spatial configurations vary the distances between speaker, addressee, reference object, and bystander and the visibility and accessibility of the object from the vantage point of speaker and hearer. Distance from speaker and/or addressee is varied in terms of a seven-point scale, according to which the object is a body part vs. in contact with the body vs. within arm’s reach vs. within easy access a few steps away vs. tens of meters away vs. more than a hundred meters away vs. several kilometers away.11 The descriptions are realized as verbal instructions to the researcher supported by diagrams; Figures 8.1–8.4 show examples of these diagrams; the full set is reproduced in the chapter explaining the task. During the enactment, a native-speaker consultant is meant to assume the role of speaker and another or the researcher that of the addressee. The researcher describes the scene for the speaker, records the utterance the speaker considers most appropriate in each scenario and/or the range of utterances the speaker considers acceptable, and optionally asks follow-up questions to clarify properties of the elicited utterances and/or test the influence of additional variables. The Yucatec questionnaire data were collected in August 1999 with five adult native speakers, four men and one woman, aged between 25 and 52. All speak Yucatec as their first and dominant language but have some command of Spanish as well. The 25 questionnaire scenes were enacted with the consultants as speakers and the author as addressee. The enactments were conducted at the appropriate scale except for the far-distant scenes 13–18 and 24–25, which were enacted at a reduced scale. In order to judge the significance of the data (given the small number of consultants), it will be worth pointing out that the five consultants generally showed a high degree of convergence in their responses. For example, in their first choices between an immediate and a non-immediate form (regardless of whether they also considered a form of the complementary set applicable, and whether they volunteered that other form or merely agreed to its applicability), all five consultants agreed with respect to 15 of the 25 scenes; and only three scenes elicited a two-to-three split in this regard. Moreover, in two of the three scenes that elicited the largest amount of variation, scenes 2 and 4, in fact all consultants agreed that both immediate and non-immediate forms would be applicable, depending on the proximity between the speaker’s pointing gesture and the reference object. This suggests that the data do in fact permit viable generalizations about the knowledge of Yucatec native speakers regarding the use of demonstrative forms in exophoric spatial reference.

10 11

The questionnaire does not specify the relevant concept of ownership. Scenes in which the object is equidistant from speaker and addressee vary distance according to the last five of these seven points.

190

Jürgen Bohnemeyer

The Yucatec questionnaire study has produced three major findings. First of all, there is no evidence suggesting that the relative location of the addressee with respect to the speaker or the reference object has any direct impact on the selection of forms for exophoric reference (notwithstanding indirect effects such as discussed at the end of section 3). It is not even the case that nonimmediate forms are applied more readily and/or consistently in reference to objects and places close to the addressee than in reference to objects or places distant from both speaker and addressee. This finding is somewhat surprising, given Hanks’ observation, quoted in the previous section, of a “fairly consistent association between the ‘there’ of te’lo’ and the addressee’s location” (Hanks, 1990: 437). In particular, the results of the questionnaire study do not support the hypothesis that the addressee’s zone of proximity constitutes a focal area within the extension of the non-immediate forms. Secondly, use of the immediate forms is much more restricted than use of the non-immediate forms. In general, immediate forms may be replaced by nonimmediate forms, while the opposite does not necessarily hold. However, the non-immediate forms are not used within very close proximity of the speaker, in particular in reference to his/her body parts, to objects that are attached to his/her body, or to objects (s)he is pointing to at close range. And finally, there are in fact two overlapping systems for spatial deixis, a simpler one used only under joined focus of attention and a more complex one used for attention-direction. Both systems operate on binary distance distinctions, but the cut-off points on the two distance scales are different. These three findings are now addressed in turn. 4.1

The Impact of the Addressee’s Location on Demonstrative Choice

To determine the impact of the addressee’s location on the choice of deictic forms, responses to scenes that only differ in the addressee’s location need to be compared, such as scenes 13 and 16, depicted in Figure 8.1. In both scenes, the speaker and the reference object are on opposite ends of a football field, but the addressee is very close to the speaker and far away from the object in one case and very close to the object and far away from the speaker in the other case. All five consultants unanimously use non-immediate forms under both conditions, regardless of the location of the addressee. A typical response is (14): (14) Le=ràadyo=o’ (yàan te’l=o’), hach ma’+lóob. DET=radio=D2 EXIST(B3SG) there=D2 really NEG+bad(B3SG) ‘That radio (that is over there) is really nice.’

A non-immediate form is optionally augmented by the deictic locative adverb te’l. The choice of whether or not the more complex form is used depends on the attention parameter; this is discussed below. The consultants just as readily

Yucatec Demonstratives S

191 S

A

A

(13)

(16)

Figure 8.1 Demonstrative scenes 13 and 16

S

A

S

(9)

A

(12)

Figure 8.2 Demonstrative scenes 9 and 12

used the non-immediate forms in reference to an object distant from both speaker and addressee in scene 13 as they did in reference to an object close to the addressee in 16; there is thus no evidence suggesting that places and objects in the addressee’s proximity play any special role in the reference of the non-immediate forms. A similar point can be made with respect to scenes 9 and 12, depicted in Figure 8.2. In 9, the reference object is close to the addressee and out of the speaker’s reach. In 12, the object is equidistant from speaker and addressee and out of either’s reach. Given Hanks’ observations of a privileged association between the nonimmediate forms and the addressee’s zone, it may have been expected that the non-immediate forms are more readily applied in scene 9 than in 12. And on the hypothesis that reference to the addressee’s zone is the semantic prototype of

192

Jürgen Bohnemeyer

the non-immediate forms, this is in fact clearly predicted. But if anything, the opposite is the case: all five consultants prefer non-immediate forms in scene 12, but only four out of five do so in 9. Again, both simple non-immediate forms and augmented constructions were used, depending on whether the addressee’s attention was assumed to be on the object prior to the utterance. Notice, though, that the augmented form in this case is formed with the presentative adverb he’l, not with the locative adverb te’l. A typical example is (15). (15) A=ti’a’l le=lìibro (he’l)=o’? A2=property(B3SG) DET=book PRSV=D2 ‘Is that book (there) yours?’

4.2

The Semantics of the Non-immediate Forms

Having failed to find a direct addressee bias in the use of the non-immediate forms, the hypothesis that the non-immediate forms are semantically specified for exophoric reference to places and objects outside the speaker’s proximity needs to be considered. A glance at the overall distribution of immediate versus non-immediate choices across the 25 demonstrative scenes, as presented in the appendix to this chapter, shows that this cannot be correct. A single scene in which the use of non-immediate forms is excluded – reference to the speaker’s own body in scene 1 – contrasts with no less than 12 scenes that exempt the use of immediate forms. A response to scene 1 is reproduced in (16). All five speakers rejected the use of non-immediate forms in this context (see Appendix): (16) Tèen=e’ mu’n bèey-tal in=meyah, tuméen túun me=D3 NEG.PROSP:A3 thus-INCH.INC A1SG=work CAUSE PROG:A3 ki’nam in=koh he’l=a’ hurt A1SG=tooth PRSV=D1 ‘Me, I can’t work, because this tooth here of mine is hurting.’

The 12 scenes that elicited exclusively non-immediate responses are found in the right-most column of the table in the Appendix. An example is (17), a response to scene 10: (17) A=ti’a’l le=lìibro yàan te’l=o’? A2=property(B3SG) DET=book EXIST(B3SG) there=D2 ‘Is that book over there yours?’

Non-immediate forms infringe on immediate territory all the way up to that first scene, while the immediate forms are clearly confined to the speaker’s zone of proximity. This distribution suggests a privative rather than an equipollent opposition, with the immediate forms as the marked terms. Only the immediate forms are semantically specified for exophoric reference to a particular region,

Yucatec Demonstratives

193

namely, the speaker’s zone of proximity. But what, then, is the common denominator in the uses of the non-immediate forms? Perhaps the non-immediate forms express “neutral deixis”, i.e. exophoric reference without restriction to a particular region of space. However, the nonimmediate forms are also used for anaphoric reference and definiteness marking, as illustrated in (18) and (19), respectively (repeated from section 2): (18) Ba’x k’ìin k-uy=úuch-ul

lel=o’? what sun IMPF-A3=happen-INC DET=D2 ‘What day does that usually happen?’ le=x-ch’úup chak u=nòok’=o’, ( . . . ) káa=PRV-enter(B3SG) DET=F-female red(B3SG) 3=garment=D2 ‘(And then) the woman dressed in red entered, ( . . . )’

(19) Káa=h-òok

The non-immediate particle is obligatory in the contexts illustrated in (18)–(19), although other anaphoric and definite expressions – in particular, third person pronouns and proper nouns – do not trigger it. In light of these endophoric uses of the non-immediate forms, the neutral-deixis hypothesis can only be maintained under an additional assumption of polysemy. Therefore, in the absence of further evidence, Occam’s razor appears to favor an analysis of the non-immediate forms as generic indexicals which do not semantically distinguish between exophoric and endophoric reference. Given that the non-immediate forms are semantically neutral regarding the immediate–non-immediate contrast, why are they dispreferred for reference to objects/places in the speaker’s proximity? The semantic analysis just outlined cannot explain this, so the answer has to be sought in the pragmatics of the system. A traditional Gricean analysis would most likely argue for “pre-emption” of the non-immediate forms from the immediate domain, i.e. a generalized conversational implicature based on Grice’s (1975) first maxim of Quantity (“Make your contribution as informative as is required”) or Levinson’s (2000) equivalent “Q-heuristic” (“What isn’t said, isn’t”). This mechanism yields an inference to the non-applicability of the marked term wherever the marked term is not chosen. In the case of the spatial deictics of Yucatec, preemption generates a default interpretation of the non-immediate forms according to which they do not refer to objects/places in the speaker’s proximity, based on the reasoning that if the speaker were in fact referring to his or her region of proximity, why would (s)he not use an immediate form, given that the immediate forms are positively specified for this reference? This mechanism is invoked in the analyses of demonstrative systems proposed by Enfield (2003) and Levinson (2006).12 12

Fillmore (1997) suggests a similar analysis to explain why Tuesday is inferred to not mean ‘today’ in case it is uttered on a Tuesday.

194

Jürgen Bohnemeyer

Gricean mechanisms generate the implicatures of an utterance from its entailments, or “what is said” by the utterance, in Grice’s parlance, in relation to what else could have been said in the same context, loosely speaking. In the case at hand, the implicature of distance attributed to the non-immediate forms is generated on the basis of a putative entailment of proximity by the immediate forms. A preemption analysis presupposes the existence of an “entailment scale” or “Horn scale” (after Horn, 1972; see Levinson, 2000 for discussion): the stronger (marked) term shares the entailments of the weaker (unmarked) term but has additional entailments not shared by the latter. This could be argued in the case at hand along the following lines: the immediate forms entail exophoric reference to the speaker’s proximity; the non-immediate forms entail indexical reference; exophoric reference is a special case of indexical reference. But can demonstratives really be said to entail proximity or distance of the referent? Suppose I say, pointing to a mountain peak on the horizon, I’ve climbed this mountain. Certainly my choice of demonstrative would be pragmatically odd; but would it render my statement false? What are the truth conditions of demonstratives? Linguists and philosophers have grappled with this question for a long time. One analysis that has been fairly influential in contemporary discussions, summarized, e.g., in Kaplan (1989), stresses that the location of the referent (or even the exophoric–anaphoric distinction) is not normally part of the contribution demonstratives make to the truth conditions of an utterance. Now, it is not too difficult to see how a Gricean preemption analysis could be extended to cases of scales not constituted by entailments, but by mere inclusion of one term’s meaning in that of another. Under such an analysis, the immediate forms preempt the non-immediate forms from their domain of use, not because their use in a given utterance affects the truth conditions of that utterance in such a way as to entail the applicability of the non-immediate forms, but merely because they have more narrowly defined meanings and hence are informationally richer. Levinson (2000: 86–104) discusses a variety of applications of such implicatures. However, it remains to be seen how an implicature analysis can be rigorously validated under such circumstances. Implicatures are identified by their defeasibility, i.e. essentially negatively, namely in contrast to entailments. Such a contrast does not appear to apply in the case at hand. A more specific problem with the preemption analysis of the Yucatec demonstratives is that it fails to account for the exemption of the nonimmediate forms from reference to the speaker’s own body, as in scene 1. This exemption in fact extends to objects pointed at or touched by the

Yucatec Demonstratives S

A

(6) immediate forms

S

A

(7)

195 S

S

A

(8)

A

S

(9)

A

(10)

non-immediate forms S

A

(12)

Figure 8.3 Accessibility in the demonstrative scenes

speaker at close range, as in scenes 2 and 4.13 Consultants consider the use of non-immediate forms in these contexts decidedly odd. However, it seems conceivable that similar phenomena may be encountered in bona fide cases of preemption as well. Perhaps the flip side of Grice’s “Make your contribution as informative as is required” is that speakers who use a truth-conditionally weaker term where evidence for the validity of the additional conditions of the stronger term is clearly available are felt to not make their contribution informative enough. For example, to say Steve ate some of the cookies famously implicates but does not entail that Steve did not eat all of the cookies. But to say this holding the empty cookie jar (Steve still chewing) might well be interpreted, although not strictly false, as inaccurate in some contexts. In sum, then, a preemption analysis must allow for the stated refinements in order to successfully account for the data. Before the matter of the immediate–non-immediate opposition in Yucatec can be left, the question of the cutoff point between the immediate and nonimmediate domains must be addressed. As Hanks (1990: 488) points out, the knowledge Yucatec speakers have of this cutoff point cannot be represented in terms of measurable spatial distance. One crucial parameter appears to be the accessibility of the reference object from the point of view of the speaker. Consider the sequence of scenes in Figure 8.3: speaker 13

Intriguingly, Yucatec immediate and non-immediate forms cannot be used contrastively within the speaker’s zone of proximity. Thus, a Yucatec speaker pointing to two objects in table-top space in front of him/her would not say the equivalent of thís object versus thát object, but would use proximal forms for reference to both objects.

196

Jürgen Bohnemeyer A

S

(11)

Figure 8.4 Demonstrative scene 11

and addressee are sitting next to each other, facing in the same direction, close enough so that both can reach and grab the object in the scene depicted in the central picture (scene 8). In this setting, accessibility is determined by whether the speaker can grab the object without having to get up. That is the case in scenes 6–8, but not in 9–10. Immediate forms are used when the speaker can grab the object, and non-immediate forms when (s)he cannot. And in scene 12, depicted in the lower half of Figure 8.3, the object is equidistant from speaker and addressee, as in 8, but this time it is several steps away from both, so that the speaker cannot reach it without getting up. Unanimously, consultants prefer non-immediate forms in this case. There is evidence suggesting that the critical measure of what counts as accessible depends on the setting. For example, Hanks (1990: 432–433) observes that the goal of a motion event which the speaker is en route to is often referred to using immediate forms, irrespective of how far away it is. Also, it is virtually impossible to use the immediate forms in any of the questionnaire scenes without accompanying pointing gestures. Consider also scene 11 (depicted in Figure 8.4), where the addressee is facing the reference object, but the latter is within the speaker’s easy reach, although the speaker cannot look at the object without turning around, which (s)he is not supposed to do according to the instructions. All consultants strongly prefer immediate forms here. So manual access apparently overrides visual access. 4.3

The Role of Attention-Direction

Another interesting finding that has come out of the questionnaire study is the consistent use of augmented forms for attention calling. Why does

Yucatec Demonstratives

197

attention-direction play such a prominent role in spatial deixis? The facts seem quite simple: spatial deictics do not provide descriptions of their referents,14 but merely information about where to look for them. In order to determine the speaker’s intended referent, the addressee must attend to what the speaker is attending to, following his or her gaze and point (cf. Diessel, 1999). In short, exophoric spatial reference requires a joint focus of attention. Under these conditions, the speaker may use attention-calling forms to alert the addressee to the effect that the addressee’s attention may not be on whatever the speaker has shifted attention to (see Küntay and Özyürek, 2002; Levinson, 2004). Based on these considerations, two distinct functions may be isolated in spatial deictic reference acts: one may be dubbed deictic anchoring, the other one attention calling. Deictic anchoring alone is done in Yucatec using the simple immediate and non-immediate forms, at least as far as reference to objects rather than to places is concerned. These forms operate on the accessibility scale, where the cutoff point is between entities/places that are readily accessible to the speaker and entities/places that are not so easily accessible. For attention calling, the augmented forms are used. These are the forms expanded by the presentative adverb he’l, where the distinction simply projects down from the accessibility scale (see Figure 8.4), and the form expanded by the deictic locative adverb te’l, which requires a relative clause to modify noun phrases (see section 2). The cutoff point between the forms expanded by the presentative adverb and the forms expanded by the locative adverb seems to be the difference between referents that are easily identifiable in the visual field so that attention is shifted to them easily and referents that are not so easily identifiable. It appears that in the latter case reference is reinforced by the place adverb te’l because it is easier in such cases to refer to the location of the referent than to the referent itself. One instantiation of the cutoff point of the attention-calling system emerges from a comparison of the scenes depicted in Figure 8.1 (scenes 13 and 16) with the scenes 14 and 17 (see Appendix). The only difference across these two pairs of scenes is that the reference object is on the far side of the ball park from the speaker’s point of view in 13 and 16, but in the center of the ball park in 14 and 17. In the former case, the augmented forms with te’l are used for attention calling (see example 14), while in the latter case, the augmented forms with he’l are used to this end, as in (20): 14

Thanks to G. Senft for reminding me that this does not hold for classificatory demonstratives, or, as they are more commonly, though perhaps misleadingly, known, “demonstrative classifiers” (see Klein, 1979; Barron and Serzisko, 1982, and, for recent descriptions, Hellwig, 2003; Senft, 2004, and O’Meara, 2010). However, in such expressions, the classificatory meaning is independent of the deictic one, and the two semantic components can often be traced to distinct diachronic sources. The deictic component is, as far as I can see, always non-descriptive.

198

Jürgen Bohnemeyer le(l) ...–a’

le(l) ...–o’

accessible to speaker

not accessible to speaker

le(l) ... he’l–a’

le(l) ... he’l–o’

Deictic anchoring

le(l) ... yàan te’l–o’

Attention calling reference object easily identifiable in visual field

reference object not easily identifiable in visual field

Figure 8.5 Anchoring and attention calling in Yucatec spatial deictics (20) A=ti’a’l le=ràadyo (he’l)=o’? A2=property(B3SG) DET=radio PRSV=D2 ‘Is that radio (there) yours?’

However, the object does not have to be far away to be difficult to attend to. Another possibility is that the object is close by but occluded from vision, as in scene 10, where the addressee’s body covers the object from the speaker’s sight (this is the rightmost scene depicted in Figure 8.3). Note that the speaker has no problem directing the addressee’s attention to the object in this context as long as the speaker knows where the object is, because the object is not blocked from the addressee’s vision. Four out of five consultants demand in response to this scene the form used for attention-direction to objects not easily identifiable in the visual field: (21) A=ti’a’l le= bùulto le= yàan A2=property(B3SG) DET=bag DET=EXIST(B3SG) te’l=o’? there=D2 ‘Is the bag that is over there yours?’

Thus, the visibility condition, just like the accessibility condition, refers primarily to the speaker rather than to the addressee. This effectively introduces a kind of evidential aspect to the use of the demonstrative forms: if the speaker has some form of knowledge of the referent being there, but cannot see it, they are required to mark this fact, thereby effectively flagging their knowledge of the existence and/or location of the object as not based on the

Yucatec Demonstratives

199

strongest possible evidence. This connection between exophoric reference and evidentiality is expanded in the dialect studied by Hanks by the existence of an additional terminal particle used for referents that are audible, but not visible (see Hanks, 1984; 1990; and cf., e.g., Behrens, 2012 and Gärtner, 2012 on evidentiality). 5

Discussion

Hanks’ (1990; 2005) “practice” approach is based on the observation of deictic usage in spontaneously occurring interactions. As has been demonstrated here, this approach falls somewhat short of clearly establishing category boundaries or cutoff points. In the case at hand, Hanks observes a frequent association between the use of non-immediate forms and the addressee’s location. Hanks (2005) clarifies that this association is not a semantic property of the non-immediate forms. The questionnaire study has clearly confirmed this: the non-immediate forms are readily applied to places and entities outside the addressee’s proximity and in fact even in the speaker’s proximity, right up to the boundaries of the speaker’s body. However, Hanks’ analysis still suggests that the addressee’s zone of proximity plays a privileged role in the use of the non-immediate forms. Perhaps the addressee’s zone of proximity is a focal area in the extension of the non-immediate forms and their interpretation is therefore biased toward this focal point by stereotype implicatures. But the questionnaire study has also failed to produce any evidence in support of such typicality effects: the non-immediate forms were just as readily and consistently used in reference to entities and places outside the addressee’s zone of proximity as they were used with referents that were near the addressee. This suggests that the apparent addressee bias Hanks observed may in fact be nothing more than a statistical correlation: the non-immediate forms were used most frequently with referents close to the addressee simply because Hanks’ database of observed interactions is biased toward referents that are close to speaker, addressee, or both. This conclusion should be understood as tentative and preliminary: it should be checked, on the one hand, against a quantitative analysis of Hanks’ database and, on the other, against psycholinguistic studies of the production and comprehension of Yucatec demonstrative forms. If there is an addressee bias, it should manifest itself, for example, in word association tests and in faster processing of non-immediate forms when used in reference to the addressee’s region of proximity compared to when used in reference to other places. The Demonstrative Questionnaire approach is based on controlled elicitation of usage under artificial conditions. This method offers the following closely related principal advantages over the observation of spontaneous data:

200

Jürgen Bohnemeyer

1. Elicitation can generate evidence of how speakers and hearers behave in situations that occur less commonly or even marginally in spontaneous interactions. Such evidence can provide important clues about the underlying categories, representations, and procedural knowledge speakers and hearers rely on. To put it in more general and abstract terms, any scientific analysis seeks to describe and explain the dependencies between the variables that affect the phenomena at issue. Such an analysis will be incomplete unless it covers all possible variable–value combinations, including combinations that are difficult to study except under artificial conditions. Reference to entities and places that are neither in the speaker’s nor in the addressee’s zone of proximity in the present study is arguably a case in point. 2. Realization of all possible variable-value combinations in a grid-like design may help avoid misinterpretations of statistical correlations as causal relations. This is exemplified in the present study by the association between the non-immediate forms and the addressee’s location in Hanks’ analysis, which the questionnaire study has identified as a possible result of a statistical bias in Hanks’ database. 3. A grid design may also help uncover systematic relations between variables that otherwise elude the researcher due to their complex nature. A case in point is the systematic correlation between attention-direction and the use of augmented forms the questionnaire study has shown. 4. Finally, the elicitation of native-speaker judgments may produce negative evidence of particular variable–value combinations not only not occurring spontaneously, but being excluded from occurrence due to ungrammaticality, semantic anomaly, or pragmatic infelicity, or a combination thereof. It may also show certain combinations to be acceptable but nonidiomatic. Conversely, showing that certain infrequent and/or atypical combinations are nevertheless possible may help identify pragmatic implicatures. In the questionnaire study, this latter principle was applied twice. It failed to support a stereotype implicature analysis of the apparent addressee bias in the use of the non-immediate forms in Hanks’ data. But it did produce evidence of a scalar implicature or preemption effect pragmatically excluding the non-immediate forms from the speaker’s zone of proximity. On the downside, one drawback of elicitation is that it is by itself blind, as it were. A questionnaire design such as the one used above needs to be informed by specific research questions and will only provide answers to the questions implemented. As mentioned above, Hanks’ meticulous study of demonstrative use in spontaneous interactions was in fact an important source in the design of the Demonstrative Questionnaire. Beyond this, Hanks (1990; 2005) makes a

Yucatec Demonstratives

201

number of intriguing observations that the questionnaire study failed to replicate, simply because the relevant variable was not implemented in the questionnaire. For instance, many of Hanks’ examples refer to motion events rather than stative locations. Hanks (1990: 432–433) notes that speakers consistently use immediate forms in reference to motion goals they are en route to and nonimmediate forms in reference to motion sources that they have already left. The Demonstrative Questionnaire has no way of detecting this phenomenon, because motion is not coded in the demonstrative scenes. The second principled drawback of elicitation is that it determines only what native speakers do under simulated conditions. This immediately raises questions of validity. One aspect where the Demonstrative Questionnaire proves artificial in a way that may well limit the validity of any study conducted with it has to do with the role of joint attention and attentiondirection in demonstrative usage. The questionnaire study has produced evidence suggesting that attention-direction is grammaticalized in the Yucatec systems of spatial deixis. While the simple immediate and nonimmediate forms are used when a joint focus of attention on the reference object or place has been established prior to the reference act, complex forms augmented with the presentative adverb he’l or the place adverb te’l are used to direct the addressee’s attention to the reference object or place. The choice between the te’l forms and the he’l forms depends not on the physical accessibility of the reference object or place, as with the choice between immediate and non-immediate forms, but on identifiability of the reference object/place in the visual field. However, attention is coded in the questionnaire in instructions to ask the consultant to imagine that the researcher as the addressee is, say, not aware of the reference object, that (s)he may not have noticed it, etc. In essence, this means asking the consultant to imagine that (s)he is not thinking about something! Obviously, this is methodologically unsatisfactory. But controlling the focus of attention under experimental conditions is an extremely difficult task. In the absence of a technique for doing this (see Enfield and Bohnemeyer, 2001 for a possible solution), the observation of natural interactions may be our best bet in the study of the role of attention-direction in spatial deixis. The upshot of this comparison of the two approaches to the study of spatial demonstratives, observation of spontaneous interactions and controlled elicitation, seems clear enough: to ensure optimal results, the two are best pursued in tandem. Moreover, the observation of spontaneously occurring speech must naturally take the lead role in this combination. Elicitation is essentially a clean-up job that helps to sort out and make sense of the results of spontaneous observation.

(1)

S

Immediate forms only

S

S

(6)

(2)

(11)

S

A

A

A

A

S

(22)

S

(7)

(3) A

S

Immediate forms preferred; non-immediate forms possible

S

A

(8)

(20)

S

(19)

S

A

A

S

S

S

(9)

(5)

(4)

A

A

A

Immediate and non-immediate Non-immediate forms preferred; forms equally good immediate forms possible

S

S

(23)

A

(16)

S

AS

(24)

(18)

A

S

(14)

A

AS

(25)

(21)

S

(17)

S

(15)

(13) SA

SA

(12)

A

SA

A

S

(10)

A

Non-immediate forms only

Appendix: Distribution of Yucatec immediate and non-immediate preferences across the 25 Demonstrative Scenes

A

Yucatec Demonstratives

203

References Anderson, S. & Keenan, E. (1985). Deixis. In T. Shopen, ed., Language typology and syntactic description. Vol. III. Cambridge: Cambridge University Press, pp. 259–308. Barron, R. & Serzisko, F. (1982). Noun classifiers in Siouan languages. In H. Seiler & F. J. Stachowiak, eds., Apprehension: Das sprachliche Erfassen von Gegenständen. Teil II: Die Techniken und ihr Zusammenhang in Einzelsprachen [Apprehension: The linguistic representation of objects. Part II: The techniques and their relation in individual languages]. Tübingen: Gunter Narr, pp. 85–105. Blair, R. W. & Vermont–Salas, R. (1965–1967). Spoken (Yucatec) Maya. Chicago: University of Chicago, Department of Anthropology. Diessel, H. (1999). Demonstratives: Form, function, and grammaticalization. Amsterdam: John Benjamins. Edmonson, M. S. (1986). Heaven born Merida and its destiny: The book of Chilam Balam of Chumayel. The Texas Pan American Series. Austin: University of Texas Press. Enfield, N. J. (2003). Demonstratives in space and interaction: Data from Lao speakers and implications for semantic analysis. Language, 79(1): 82–117. Enfield, N. J. & Bohnemeyer, J. (2001). Hidden colour-chips task. In S. C. Levinson & N. J. Enfield, eds., Manual for the field season 2001. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 21–28. Fillmore, C. J. (1997). Lectures on deixis. Stanford, CA: CSLI Publications. Grice, H. P. (1975). Logic and conversation. In P. Cole & J. L. Morgan, eds., Speech acts. New York: Academic Press, pp. 41–58. Hanks, W. F. (1984). The evidential core of deixis in Yucatec Maya. In J. Drogo, ed., Papers from the twentieth regional meeting of the Chicago Linguistic Society. Chicago: Chicago Linguistic Society, pp. 154–172. (1990). Referential practice: Language and lived space among the Maya. Chicago: University of Chicago Press. (2005). Explorations in the deictic field. Current Anthropology, 46(2), 191–220. Hellwig, B. (2003). The grammatical coding of postural semantics in Goemai (a West Chadic language of Nigeria). Doctoral dissertation, Radboud University Nijmegen. Himmelmann, N. P. (1996). Demonstratives in narrative discourse: A taxonomy of universal uses. In B. Fox, ed., Studies in anaphora. Amsterdam: John Benjamins, pp. 205–254. Horn, L. R. (1972). On the semantic properties of logical operators in English. Mimeo. Bloomington: Indiana University Linguistics Club. Kaplan, D. (1989). Demonstratives: An essay on the semantics, logic, metaphysics, and epistemology of demonstratives and other indexicals. In J. Almog, J. Perry & H. K. Wettstein, eds., Themes from Kaplan. New York: Oxford University Press, pp. 481–563. Kita, S. & Walsh Dickey, L., eds. (1998). Annual report 1998. Nijmegen: Max Planck Institute for Psycholinguistics. Klein, H. (1979). Noun classifiers in Toba. In M. Mathiot, ed., Ethnology: Boas, Sapir, and Whorf revisited. The Hague: Mouton, pp. 85–95.

204

Jürgen Bohnemeyer

Kornfilt, J. (1997). Turkish. London: Routledge. Küntay, A. C. & Özyürek, A. (2002). Joint attention and the development of the use of demonstrative pronouns in Turkish. In B. Skarabela, S. Fish & A. H. J. Do, eds., Proceedings of the 26th annual Boston University conference on language development. Somerville, MA: Cascadilla Press, pp. 336–347. Lehmann, C. (1998). Possession in Yucatec. Munich: Lincom Europa. Levinson, S. C. (2000). Presumptive meanings: The theory of generalised conversational implicatures. Cambridge, MA: MIT Press. (2004). Deixis. In L. Horn, ed., The handbook of pragmatics. Oxford: Blackwell, pp. 97–121. (2006). The language of space in Yélî Dnye. In S. C. Levinson & D. P. Wilkins, eds., Grammars of space. Cambridge: Cambridge University Press, pp. 157–205. Lewis, M. P., ed. (2009). Ethnologue: Languages of the world, sixteenth edition. Dallas, TX: SIL International. (Available online at www.ethnologue.com/). Lyons, J. (1977). Semantics. Vol. II. Cambridge: Cambridge University Press. Meira, S. (2003). “Addressee effects” in demonstrative systems: The cases of Tiriyó and Brazilian Portuguese. In F. Lenz, ed., Deictic conceptualisation of space, time, and person. Amsterdam: John Benjamins, pp. 3–12. O’Meara, C. (2010). Seri landscape classification and spatial reference. Doctoral dissertation, University at Buffalo – SUNY. Özyürek, A. (1998). An analysis of the basic meaning of Turkish demonstratives in faceto-face conversational interaction. In S. Santi, I. Guaitella, C. Cave & G. Konopczynski, eds., Oralité et gestualité: Communication multimodale, interaction. Paris: L’Harmattan, pp. 609–614. PerfilMayaweb (2005). Perfil sociodemográfico de la población hablante de maya [Socio-demographic profile of the Yucatec-speaking population]. Aguascalientes: Instituto Nacional de Estadística, Geografía e Informática. (Available online at www.inegi.gob.mx/prod_serv/contenidos/espanol/bvinegi/productos/censos/pobl acion/poblacion_in digena/PerfilMayaweb.pdf.) Pfeiler, B. (1995). Variación fonológica en el maya yucateco [Phonological variation in Yucatec Maya]. In R. Arzápalo & Y. Lastra, eds., Vitalidad e influencia de las lenguas indígenas en Latinoamérica. Mexico City: Universidad Autónoma de México, Instituto de Investigaciones Antropológicas, pp. 488–497. PHLI (2009). Perfil sociodemográfico de la población que habla lengua indígena [Sociodemographic profile of the speakers of indigenous languages]. Aguascalientes: Instituto Nacional de Estadística y Geografía. (Available online at www.inegi.org .mx/prod_serv/contenidos/espanol/bvinegi/productos/censos/poblacion/poblacio n_in digena/leng_indi/PHLI.pdf.) Schalley, A. (2012). Many languages, one knowledge base: Introducing a collaborative ontolinguistic research tool. In A. Schalley, ed., Practical theories and experimental practice. Amsterdam: John Benjamins, pp. 129–155. Senft, G. (2012). Referring to colour and taste in Kilivila: Stability and change in two lexical domains of sensual perception. In A. Schalley, ed., Practical theories and experimental practice. Amsterdam: John Benjamins, pp. 71–98. Senft, G. & Smits, R., eds. (2000). Annual report 2000. Nijmegen: Max Planck Institute for Psycholinguistics.

Yucatec Demonstratives

205

Vapnarsky, V. (1999). Expressions et conceptions de la temporalité chez les mayas yucateques (mexique) [Expressions and concepts of temporality among the Yucatec Mayans (of Mexico)]. Doctoral dissertation, Université de Paris X, Nanterre. Wilkins, D. P. (1999; this volume). The 1999 demonstrative questionnaire: “This” and “that” in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24.

9

Lavukaleve: Exophoric Usage of Demonstratives Angela Terrill1

1

Introduction

This chapter explores the exophoric usage of demonstratives in Lavukaleve, a Papuan language of the Solomon Islands. Demonstratives are a complex and interesting aspect of Lavukaleve grammar. There are four separate paradigms of demonstrative modifiers, of which one is not relevant to the task in hand, as it does not have an exophoric function. The rest include one paradigm for proximal use, i.e. close to speaker – one paradigm for distal use, that is far from the speaker, and an indeterminate term which is used for when neither of the others is appropriate. While all three forms are speaker anchored, the middle term is not a true medial distance term, but rather is distance-neutral. The structure of the chapter is as follows. Section 2 describes the language and its speakers. Section 3 discusses the morphological field of deixis in Lavukaleve and describes the different form classes relevant to exophoric demonstration. Section 4 discusses the results of the Wilkins (1999) Demonstrative Questionnaire on exophoric demonstrative usage. Section 5 elaborates on the questionnaire results with further evidence on exophoric demonstratives from naturalistic observation and from contrastive usage. Section 6 presents the conclusions. 2

The Language and Its Speakers

2.1

Lavukaleve in the Solomon Islands

The Solomon islands boasts a huge amount of linguistic diversity, with around 60 Oceanic languages and 4 Papuan languages (Tryon and Hackman, 1983), 1

I would like to thank the Lavukal speakers who helped work with me on this task; I hope I have done justice to their insights. I would also like to thank Sérgio Meira and Sarah Cutfield for useful comments on an earlier draft. Also I am deeply grateful to David Wilkins for many insightful discussions on this topic.

206

Lavukaleve Demonstratives

207

among a population of 609,883 people.2 Lavukaleve is a Papuan language spoken by about 1,500 people in the Russell Islands, a small island group in the Central Province of the Solomon Islands. Lavukaleve has not been conclusively shown to be related to any other of the handful of Papuan languages in the Solomon Islands, or indeed to any other language. Perhaps the most substantial attempt is Ross (2001) who tentatively suggests a genetic relationship between these languages based on pronoun paradigms, although leaving conclusive demonstration of the relationship to future research. 2.2

Lavukal People

The ethnonym for Lavukaleve speakers is Lavukal (from an unanalysable root lavu plus plural suffix -kal). Most Lavukal children are monolingual Lavukaleve speakers, but most people learn to speak the lingua franca of the Solomon Islands, Solomon Island Pijin (a creole, despite its name) as well, usually during early adolescence. Today all but the oldest of adults and youngest of children can speak Solomon Islands Pijin. However, in the West Russells at least, Solomon Island Pijin is used only for talking to non-Lavukals. Such situations arise when non-Lavukals visit the villages (e.g. when the district priest or malaria workers come), or when Lavukals go to the provincial centre, or leave the Russells for school or work. However, in the East Russells, Pijin is taking over from Lavukaleve as the medium of daily conversation, thus rendering Lavukaleve highly endangered. 2.3

Typological Overview of Lavukaleve

Lavukaleve has a strict SV/AOV constituent order. There are nine basic participant marking systems based on the type of predicate and the focus status of the arguments. There is a gender system with three genders, marked on all nominal modifiers, and nouns are inflected for dual and plural number in a complex system of declensions which also interacts with the gender system. There is a system of focus marking, with focus markers agreeing in person, gender and number with their referent. Agreement of focus markers shows the scope of focus. Thus, argument focus markers follow and index the focused argument. Predicate focus markers follow and index the predicate, and sentence focus marks focus on the entire sentence, showing default third person singular neuter marking. There are two nominal cases, Locative and Perlative.

2

CIA World Factbook. (Available online at www.cia.gov/library/publications/the-world-factbook /geos/bp.html, accessed 6 October 2014.)

208

Angela Terrill

Table 9.1 1st/2nd person personal pronouns

1 EXCL 1 INCL 2

SG

DU

PL

ngai

el mel imil

e me imi

inu

Table 9.2 3rd person demonstrative pronoun ‘he/she/it’

SG

DU

PL

MASC FEM NEUT MASC FEM NEUT

PROXIMAL

NEUTRAL

DISTAL

UNSPECIFIED

fona fo foga fonala fol fogala fova

foina foia foiga foinala foiaol foigala foiva

feana fehea feaga feanala feheaol feagala feava

foana fohoa foaga foanala fohoaol foagala foava

Predicates can be joined by various means, including serial verb constructions, and compounds, verb-adjunct constructions, and an auxiliary construction expressing habitual aspect. Clauses can be combined using subordinate, coordinate, or cosubordinate constructions. Adverbial subordinate clauses employ a split-ergative participant marking system in which first and second person subjects follow a nominative/accusative participant marking system, whereas third person subjects follow an ergative/absolutive marking system. See Terrill (1999; 2003) for more detail on Lavukaleve grammar. 3

Deictics and Demonstrative Systems in Lavukaleve

3.1

Pronoun and Demonstrative Forms

In Lavukaleve, deictics are a highly elaborated area of the morphology, consisting of pronoun/demonstratives, adverbial locative deictics and a very rare set of demonstrative identifiers. The pronoun/demonstratives consist of first and second person personal pronouns (Table 9.1), two paradigms of third person demonstrative pronouns (Tables 9.2 and 9.3) and a paradigm of demonstrative modifiers (Table 9.4). The exophoric demonstrative questionnaire elicited demonstrative modifiers, as ideally suited to expressing spatial-based modification of a referential head. While third person demonstrative pronouns

Lavukaleve Demonstratives

209

Table 9.3 3rd person demonstrative pronoun ‘the other one’. The paradigm is defective.

SG

DU

MASC FEM NEUT MASC FEM NEUT

PL

PROXIMAL

NEUTRAL*

oga -

oina oia oiga oinala oial oigala oiva

* The terms ‘proximal’ and ‘neutral’ are not so immediately relevant for the oi- demonstrative but are retained because of morphological parallels with the foi- and hoi- demonstratives.

Table 9.4 Demonstrative modifier ‘this’

SG

DU PL

MASC FEM NEUT MASC FEM NEUT

PROXIMAL

NEUTRAL

DISTAL

UNSPECIFIED

hona ho hoga honala hol hogala hova

hoina hoia hoiga hoinala hoiaol hoigala hoiva

heana hea heaga heanala heaol heagala heava

hoana hoa hoaga hoanala hoaol hoagala hoava

also express spatial modification, they are the heads of NPs rather than being modifiers to a head as in the demonstrative modifiers, thus being less useful in these contexts. The stems in Tables 9.2 and 9.4 clearly suggest a morphological analysis whereby the forms consist of an initial element expressing the type of form (for h-), a medial element expressing distance (broadly construed) and a gender/ number suffix. Table 9.3 shows a further demonstrative paradigm, a rather reduced one, used only for reference-tracking purposes; it has the function of reactivating deactivated referents (Terrill, 2001: 183–190). The paradigm in Table 9.3 shows that the root of all the demonstratives in Tables 9.2–9.4 is not the first element, as the oi- pronouns, in that sense, have no first element. Rather, if there is a root it must be the middle element, which receives prefixation (f, h or ø) to show the function of the form, and a gender/number suffix agreeing with the

210

Angela Terrill

referent. Note that the same gender/number suffixes are widespread in other areas of the language as well (Terrill, 2003: chapter 10). 3.2

Demonstrative Morphology

The demonstratives mark gender, number and distance of the referent (to be elaborated in this chapter), plus making another distinction in a fourth form between specified and non-specified distance. Both the pronoun/demonstratives and locative deictics can appear with certain suffixes: Group suffixes -sa (ungendered) and -ha (feminine), the Extended -la and the Presentative and Predicative suffixes -ri and -o/om/v respectively. Only certain of the deictics can appear with each of these suffixes. Of these, the Presentative suffix is important to our discussion, as its use with the demonstrative modifiers is associated with pointing; this in turn affects the choice of demonstrative. Normally when people point they use a presentative form of the demonstrative, and when they use a presentative demonstrative, they point. For example, when a group of speakers was looking at pictures of fish on a poster, in order to tell me the Lavukaleve name of each one, people pointed to each fish and asked each other, as in example (1):3 (1)

H-o-na-ri? MOD-PROX-sgm-PSNV ‘This one?’

The demonstrative modifiers consist of an initial element h-, a stem -o-/-oi-/-ea-/ -oa- and a gender/number suffix as shown above. They mean ‘this [N], that [N], that[N] far’ (examples 2–4): (2)

foto-n Karumulu-n me-a h-o o-na middle-LOC Karumulu-LOC SPEC-sgf MOD-PROX.sgf 3sgfO-in . . ., ‘in the middle of this village, Karumulun here.’

(3)

Gaikoko na a-na f-oi-va h-oi-va holou canoe(m) sgmArt 3sgmO-in PN-NTRL-pl MOD-NTRL-pl sink foa-re lei-v. go.down-NF exist-pl ‘In the canoe, all of them sink down.’

3

Abbreviations: ACTion particle, ANTerior, Art article, DISTal, EFOC focus marker from heo paradigm, f feminine, FOCus, FUT Future tense, HABitual, LOCative, m masculine, MODifier, n neuter, NEGative, NeuTRaL, NonFinite, Object, pl plural, POSSessive, ProNoun, PRESent, PROXimal, PreSeNtatiVe, QuestionFOCus, SBD = Subordinate verb, SPEC Specifier (adjective), Subject, sg singular, UNSPecified.

Lavukaleve Demonstratives (4)

211

Aka Suvala h-ea-na fin a-ea-re-m then Suvala(m) MOD-DIST-sgm 3sgmFOC 1sgS-talk.about-FUT-sgm hin. 3sgmEFOC ‘I will talk about that Suvala (island) over there.’

The fourth term, -oa- (glossed as UNSPecified), is used for non-specific referents which are far away at some unspecified place. Its most frequent use is for referring to entities which do not actually exist, or which are used merely as generic instances of a type. While rare in general discourse, these forms occur frequently in elicitation. For example, if in elicitation one asks something like the following example (5), the unspecified demonstrative will commonly be used: (5)

Ali h-oa-na songi-nu. man(m) MOD-UNSP-sgm swim-PRES.sg ‘That man is swimming.’

The man, while grammatically definite, is semantically a generic instance of a type – any man. Therefore the unspecified form is enabled. This fourth term hoa- is not strictly speaking a distance form – its semantics can imply a referent far away, but more importantly that referent is in some unspecified place, or is a generic referent, or is in some faraway place which the speaker does not know. It says nothing about position of an object with respect to speaker or some other location, and if there is a semantically definite referent, it cannot occur at all, so it is incompatible with all of the demonstrative questionnaire scenes. The three remaining demonstrative modifiers, ho-, hoi- and hea-, do, however, operate to express distance contrasts. The semantics of these contrasts are discussed in section 4. All three forms, ho-, hoi- and hea-, can also be used anaphorically to express time: hoi- is by far the most frequently used of the three. Example (6) shows a proximal temporal use: (6)

ta h-o-na hola na a-na uruala fo’sal time(m) MOD-PROX-sgm tree.sp(m) sgmArt 3sgmO-in properly fish(m) vo-kuru-re hi-la ma-lei. 3plO-hit-NF do/say-NEG 3plS-exist ‘ . . . now [lit: at this time] they don’t [know how to] kill with fish poison [hola] properly.’

There is also a set of locative deictics, built on the same morphological material as the demonstrative modifiers, which function as adverbs. The forms hoka ‘here’, hoika ‘there, neutral’, heaka ‘there, distal’, hoaka ‘there’, unspecified

212

Angela Terrill

make the same distinctions, in their forms at least, as the demonstrative modifiers. They too consist of the stems -o-, -oi-, -ea- and -oa-, with a fixed prefix h- and a fixed element -ka as a suffix. The fourth term, hoaka, as for the demonstrative modifier hoa-, marks unspecified distance. Example (7) shows a reference to a college miles away on a different island. Hoaka expresses a degree of distance far away yet unspecified and unknown: (7)

Selwyn hoaka nun kini vau-m Selwyn there.UNSP from ACT go.seawards-sgm ‘It went seawards from way over there at Selwyn [College].’

It is not uncommon for a locational adverb and a demonstrative modifier to cooccur in one utterance, as in example (8). When this happens, the same stem is almost always used. (8)

3.3

E-me-ge fi hoka namu vo-le 3sgnO.SBD-continue-ANT 3sgnFOC here.PROX places 3plO-see ngo-me-v h-o-va. 2sg-HAB-pl MOD-PROX-pl ‘It went on, and now you can see these places here.’

A Three-term System

A first obvious hypothesis of the three demonstrative forms ho-, hoi- and hea- suggests a three-way distance distinction: proximal, medial and distal. Each of the demonstratives is speaker anchored, and a first suggestion would be that each form expresses a proto-typical distance from the speaker: ho-, proximal, is used within or close to touching distance; hoi-, medial, used within the vicinity of the speaker; and hea-, distal, used for referents which are very far away. Schematically, this analysis would be as set out in Table 9.5. That is, the claim (as in fact suggested in Terrill, 1999) would be that each term had a specific location or area associated with it, based on ever-enlarged degrees of distance from the speaker. The results of the questionnaire show that this is not the case.

Table 9.5 Possible analysis of exophoric demonstratives

hohoihea-

Close to S

Small distance from S

Far from S

+ − −

− + −

− − +

Lavukaleve Demonstratives

3.4

213

The Exophoric Demonstrative Questionnaire (Wilkins, 1999)

The exophoric demonstrative elicitation was carried out monolingually in Lavukaleve. For practical reasons I divided the elicitation into two halves and did each half with two groups of people. Two groups of two people each did the first half of the elicitation, and two groups, one of one person and one of three people respectively, did the second half. Speakers were males and females, aged between their early twenties and late fifties. The pictures were not shown to consultants. Rather, I set out the scenes as shown in the pictures using real objects. For scenes with the object not on the body but within immediate reach and conversational space, I located the consultants in a room of a house and used a book, a small radio or ball as the object whose location was questioned. For scenes in which the object was on the body, the pictures suggest using a caterpillar. For logistical reasons I did not use a real caterpillar, but rather speakers were asked to imagine a spider there. In Lavukal culture, spiders are not feared or loathed, and thus although it is not ideal that the object was imaginary, at least it did not have an obviously different emotional significance than the radio, book or ball of the other scenes. For scenes such as 13–18, in larger space, we walked to an open part of the village, and a soccer ball was used as the referent. For scenes 24 and 25, we stood at the edge of the sea and looked at the hill on the island about 2 kilometres across the lagoon. An example of the types of interactions used in the task is shown in (9): (Scene 8) (9)

Legis ho-ga ngo-legis mi? book(n) MOD.PROX-3sgn 2sgPOSS-book(n) QFOC.3sgn ‘Is this book your book?’ Tamu. Nga-legis tave fi. NEG 1sgPOSS-book(n) be.not.3sgn 3sgnFOC ‘No, it is not my book.’

4

Experimental Results

The results of the Demonstrative Questionnaire (Chapter 2, this volume), presented in Table 9.6, show that the analysis of a three-degree distance distinction is clearly problematic. The difficulty lies with the role of the middle term, hoi-. The use of the other two terms seems relatively clear. Ho- is used for objects close to the speaker; it seems clearly to be a proximal term. It is used for scenes 1, 2 and 3 with objects on the speaker’s body. It is used for 7 and 11, in which the object is notably close to the speaker, and notably closer to the speaker than the addressee. Note that in scene 11 it is clear that it does not

S

S

(8)

(7)

(3)

(1)

A

S

A

S

HOGA

S

A

(23)

S

(11)

S

(2)

A

A

S

S

(22)

A

(4)

A

HOIGA

S

S

(14)

SA

(13)

SA

(10)

A

(9)

A

A

A

HEAGA

(18)

A

S

(17)

S

(16)

S

(15)

SA

AS

AS

S

S

(25)

(24)

(12)

(5) A

A

S

S

(21)

A

(20)

A

(19)

S

(6)

S

A

A

HOGA or HOIGA or HEAGA

heaga best, hoga fine, one speaker disallowed hoiga, one allowed it

hoiga best, hoga / heaga possible

hoga best, hoiga / heaga possible

hoga best, hoiga / heaga possible

Table 9.6 Summary of results of the exophoric demonstratives questionnaire for Lavukaleve (cited in neuter sg forms)

Lavukaleve Demonstratives

215

matter if the speaker cannot see the object. In scenes 8 and 23, the object is equidistant from speaker and addressee, but it is still saliently close to the speaker, in reach, and within a few steps, respectively. Hea- is used for objects not saliently close to the speaker. This notion of not being saliently close to speaker can arise from various factors. For instance, with objects that are genuinely a long way away, such as hills or islands in the distance (e.g. scenes 24, 25), hea- is used exclusively. Similarly in scenes 13–18, in which the object is a significant distance out of normal conversation space, hea- is used. Further, hea- is used if an object is much closer to the addressee than the speaker, as in scene 5, in which the object is on the addressee’s shoulder, and scenes 9 and 10 in which two people are sitting side by side and the object is much closer to the addressee than the speaker. Such scenes suggest that, even though there is no special close-to-addressee term, the distal-to-speaker term can be chosen by virtue of the fact that the object is saliently much closer to someone else than to the speaker. The form hoi- has a much wider application of use than either proximal hoor distal hea-. It can be used for most scenes, which suggests that it is spatially neutral. Those scenes in which it cannot be used include scenes in which the object is actually on the body of the speaker (for which only proximal ho- is appropriate) and scenes in which the object is far away in the distance (for which only distal hea- is appropriate). In contrast with the previous analysis of hoi- as a distance-medial term, the two scenes for which only hoi- was used are not scenes in which the object is an intermediate distance from speaker. In scene 4, the speaker points to an object on the addressee’s shoulder. In this scene, the object is not noticeably at an intermediate distance from the speaker. However, the object is closer to the addressee than the speaker, so proximal ho- is not appropriate, and yet it is not far from the speaker either, so distal hea- is not appropriate. Rather, the object is saliently close to the addressee, for which there is no term in Lavukaleve (see below for discussion of this point). The proximal form ho- was rejected here by speakers. Note that in scenes 9 and 10 the conversational space is greater, enabling the use of distal hea. In scene 4 the conversational space is smaller, disallowing hea. In scene 22, the speaker is in a house, the addressee five paces away outside the house, and the object is outside the house, equidistant between them. Again, the object is not particularly close to the speaker, so proximal ho- is not appropriate. Neither is the object particularly far from the speaker, being only a few steps away, so distal hea- is not appropriate. It is for this reason that hoi- is used. A picture is emerging of ho as proximal, hea as distal and hoi as occupying a no-man’s-land, used when neither proximal ho- nor distal heaare appropriate.

216

Angela Terrill

It is interesting to compare scene 22 with 12, as only distal hea- was used for the latter. In both scenes, the object is equidistant from speaker and addressee; the two scenes differ in that 22 has an established social boundary between object and speaker; whereas 12 does not. One might predict that distal heawould be better with scene 22 and hoi- with scene 12, but this was not borne out. I suspect that hea- was used in scene 12 because in this configuration the addressee was not taken into account; the use of hea- was calculated solely on the relationship between the object and the speaker; and for the speaker the object was far enough away for hea- to be appropriate. Scene 22 is far more complex. Not only is there an intervening strong social boundary between speaker and object (making the object perceptually saliently closer to the addressee than speaker), but also the object intervenes between speaker and addressee. Further, the speaker and addressee are about 10 paces apart, which confuses calculations of conversational distance. The use of hoi- may be accounted for by the fact that so many confounding factors made distal headifficult to apply (and in any case proximal ho- would not be appropriate). Therefore, hoi- was the appropriate term. Comparison between scenes 22 and 23 is also interesting. Scene 23 differs from 22 only in that the speaker is outside the house and the addressee is inside. For this scene, either ho- or hoi- was used. These scenes were intended to test whether significant established boundaries in social space make an impact on demonstrative use. In Lavukaleve, the boundary between house, speaker and object is critical. Where the speaker and object are both on one side of the inside/outside the house boundary, the proximal term can be used. Where the speaker and object are divided by the house boundary, the proximal term is not used. Thus, the social boundary of inside/outside a house overrides distance in determining the proximal/non-proximal use. Comparison between scenes 4 and 5 is also fruitful. In both of these scenes there is an object on the addressee’s shoulder. The scenes differ only in that in scene 4 the speaker points to the object, whereas in 5 there is no pointing. For scene 4, with the point, hoi- was used. For 5, without the point, distal hea- was used. This suggests that pointing has a role in bringing an object into closer proximity to the speaker, thus making the distal term less appropriate. Note that even with the point, the proximal term was still not appropriate. This is possibly because, as explained above, the proximal demonstrative is less appropriate for an object saliently much closer to the addressee than the speaker. This again supports an analysis of hoi- occupying a kind of no-man’s-land, used where neither of the other forms is appropriate. This analysis is further supported by the contrastive use of these demonstratives (discussed below) – the proximal and distal terms are used in contrastive usage but the hoi- term is not used in this function – as it has no natural spatial domain of reference. This echoes a semantically general demonstrative in Laos, nii4, which ‘lacks any

Lavukaleve Demonstratives

217

specification of the spatial location of the referent, but often picks up an inferred meaning “here” by paradigmatic opposition to the available more semantically specific term nan4, which encodes “not here”’ (Enfield, 2003: 107). In scenes 6, 19, 20 and 21 (shown at the far right of Table 9.6), all three demonstratives were possible. Scenes 19, 20 and 21 all have the confounding factor of the boundary between inside and outside the house. It seems that the correct choice of demonstrative when speaker and addressee are separated by a social boundary of this nature depends on whether the speaker takes pure distance into account or whether the social boundary is felt to override this. Scene 6 is problematic for the analysis. In scene 6 the speaker and addressee are sitting side by side, with the object next to the speaker away from the addressee. Speakers agreed that proximal ho- was best for this scene, but either neutral hoi- or distal hea- were also possible. The analysis would predict that ho- would be used; but it is difficult to see why hea- or even hoi- should be appropriate. One possibility is that the speaker may have some freedom to take into account the relationship between addressee and object in determining the applicability of the proximal term. Another possibility is that visibility may be a factor: the speaker may be unsure whether the addressee can see the object, thus enabling a distal or no-man’s-land demonstrative. 5

Further Evidence

5.1

Natural Observational Data

The impact of established social boundaries on demonstrative choice was mentioned above. A further instance supporting this observation came during a conversation between three people: me, Christa and Gloria. We happened to be arrayed as the points of an equilateral triangle, each about five steps away from the other. Christa and Gloria were each standing, and holding a baby on their left shoulder (baby Angela and Holden respectively). I was sitting on the steps of my house. Gloria said to Holden (the baby on her shoulder) “Angela tula ho, Angela bakel hoia” (Angela small this, Angela big that), pointing at each one as she said it. So even though baby Angela and I were the same distance away from Gloria when she spoke, she used the ho- term for baby Angela and the hoi- term for me. In contrastive use, often the same demonstrative is used for both objects of contrast (Meira and Terrill, 2005), so Gloria did not need to use different demonstratives to refer to the two objects. Further, order of mention is not likely to have been important (Meira and Terrill, 2005). Rather, the main difference appears to be that Christa and Gloria were both on the ground outside the house, whereas I was on the steps, at the entrance to the house.

218

Angela Terrill

This seems to have been enough to have enabled the use of the proximal term for baby Angela but made it less appropriate for reference to me. Other social factors may also have been present: the fact that Gloria, Christa, Holden and baby Angela are all closely related, whereas I am a foreigner. Social factors of this kind may subtly influence the appropriateness of the proximal term. Another possible analysis of this situation is that Gloria and Christa were on a level, whereas I was sightly raised (Lavukal houses are raised on stilts, and I was sitting on the top step, so I was slightly higher than Gloria’s head). It is possible that items on different vertical planes may be less proximal than those on the same vertical plane. However, other observational data does not bear out this possibility. For example, in another situation a group of people were looking in the tree-tops for a possum which they had seen a moment before. There were about five people looking for it, in a cluster of three or four trees. One boy spotted it above his head in the tree closest to him, and called “Honari!”, using the proximal term with the Presentative suffix (see above). Thus, even though the possum was some distance above his head, the proximal term was chosen. There is a further common use of the proximal term which did not come out in the experiment. Hoga ‘this’ is the polite way to refer to someone in the third person. For instance, when a neighbour was bringing a plate of food to my house one time, her relative asked her who the food was for. My neighbour lippointed to me, and said “Ho-ham fi” (MOD.f.sg.PROX-Purposive 3sg.neutral. Focus) ‘It’s for this one’). It is rude to finger-point to people, but lip-pointing is acceptable, and the correct way of referring to me in this instance, particularly in this case when I could hear and see the conversation, was with the proximal demonstrative. 5.2

Contrastive Demonstrative Usage

In contrastive usage we see intriguing similarities with non-contrastive exophoric usage. For contrastive reference to three objects in table-top space, all could be referred to using proximal ho-, or all using distal hea-, or the closest one (or two) with ho- and the farthest one or two using hea-, but in no circumstances were three objects aligned in a row referred to using ho- proximal, hoi- medial and hea- distal. The hoi- form is not a distance-based term and has no distance-based semantics apart from what it picks up from implicature (as described above). In contrastive use as well as non-contrastive exophoric use, it is a neutral form, perhaps best seen as occupying a no-man’s-land spatially, not having its own locational semantics, and used when for various circumstances other forms are not applicable (Meira and Terrill, 2005).

Lavukaleve Demonstratives

6

219

Conclusions

The results of the Wilkins exophoric demonstrative questionnaire confirm the hypothesis that ho- is a proximal term and hea- a distal term. The main issue raised by the experiment is the role of hoi-, the middle term, in the system. The experiment shows that hoi- seems not to be a medial-distance term. Crucial to the understanding of the role of hoi- is the observation that if three objects are lined up in space, in whatever scale (e.g. from table-top space, to conversation space or large-scale geographic space), speakers never use the three terms neatly, one for each object. This suggests that the three terms do not divide space into three sections based on distance from one deictic centre. Rather, speakers almost invariably use ho- for objects close to the speaker, and hea- for the other(s). If the area involved is table-top space, i.e. the objects are all within the speaker’s reach, they can all be referred to with ho-. Alternatively, they can all be referred to using hea-, if they are construed as far from the speaker. In any case, at whatever scale, ho-hoi-hea- are never used to refer to three spaced-out objects (Meira and Terrill, 2005). This casts serious doubt on a three-term-distance-system account. Further, while there is an ideal space associated with use of the proximal term, space which can be roughly defined as within reach of the speaker, and while there is an ideal space associated with the distal term, which can be roughly defined as large-scale geographical space, there is no space ideally associated with the middle term. A possible account is that hoi- is an addressee-anchored term, referring to objects close to the addressee. Scene 4 is suggestive of this. However, scenes like 2, 5, 9 and 22 refute this analysis. In scene 2, proximal ho- is just as good as hoi-, which would not be predicted if hoi- was a close-to-addressee term. In scene 5, hoi- is not used at all; distal hea- is used (if the speaker does not point). In scene 9, the object is very close to the addressee, yet distal hea- is used, not hoi-. In scene 22, hoi- is preferred, even though the object isn’t closer to the addressee than the speaker. So these are contexts in which the object is saliently close to the addressee, yet hoi- is not used; or scenes where the object is not saliently closer to the addressee than the speaker, yet hoi- is used. The analysis that seems to fit best is that ho- is used if the object is saliently close to the speaker (e.g. scenes 1, 3, 7, 8), hea- is used if it’s saliently far from the speaker (e.g. scenes 24, 25), and hoi- is used if the object is not particularly saliently close to or far from the speaker. As a result of scalar implicature, hoi- can appear to function as a medial demonstrative. However, it is not semantically specified for location and is used when either of the other two is for some reason proscribed. There is a selectional hierarchy in the paradigm, an asymmetry in usage whereby if the proximal or distal forms can be used, they will be, and otherwise the

220

Angela Terrill

Table 9.7 Functions of exophoric demonstratives

hoga hoiga heaga

Close to S

Far from S

+ − −

− − +

neutral form will be used. This suggests a schematic view of Lavukaleve’s exophoric demonstratives as in Table 9.7. It was stated above that there is an ideal space associated with the proximal term and an ideal space associated with the distal term. Thus, if speakers are asked what ho- means, they will typically suggest it be used for an object within reach; and if they are asked what hea- means, speakers will typically suggest distances involving large-scale geographic space, for instance the next island, or beyond the lagoon. However, the system is rather compressible. For example, if two groups of objects are placed on a table within reach of the speaker, one closer to the speaker and one further away, hea- is acceptable for reference to the further of the groups. Indeed in this context, both groups of objects can be referred to by hea-, as the speaker can choose not to contrast their location in terms of distance. The summary above (Table 9.7) shows the middle term, hoi-, as neutral with respect to distance meanings. It is worth also noting that hoi- is the most frequently used term in anaphoric usage, although both the proximal and the distal term can be used anaphorically; and further, that it is the neutral term in the demonstrative pronoun system which has been grammaticalized to mean something like ‘okay; that’s it then’ (Terrill, 2003: 175). These appearances of the neutral term in other non-exophoric demonstrative usages suggest that this term is not just distance-neutral but that it is semantically relatively neutral in other respects as well. References Enfield, N. J. (2003). Demonstratives in space and interaction: Data from Lao speakers and implications for semantic analysis. Language, 79, 82–117. Meira, S. & Terrill, A. (2005). Contrasting contrastive demonstratives in Tiriyó and Lavukaleve. Linguistics, 43, 1131–1152. Ross, M. (2001). Is there an East Papuan phylum? Evidence from pronouns. In A. Pawley, M. Ross & D. Tryon, eds., The Boy from Bundaberg: Studies in Melanesian linguistics in honour of Tom Dutton. Canberra: Pacific Linguistics, pp. 301–321.

Lavukaleve Demonstratives

221

Terrill, A. (1999). Lavukaleve: A Papuan language of the Solomon Islands. Doctoral thesis, Australian National University. (2001). Activation in Lavukaleve pronouns: Oia versus foia. Linguistic Typology, 5, 67–90. (2003). A grammar of Lavukaleve Berlin: Mouton de Gruyter. Tryon, D. T. & Hackman, B. D. (1983). Solomon Island languages: An internal classification. Canberra: Pacific Linguistics C–72. Wilkins, D. P. (1999; this volume). The 1999 demonstrative questionnaire: ‘This’ and ‘that’ in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24.

10

Tiriyó: Non-contrastive Exophoric Uses of Demonstratives Sérgio Meira

1

Introduction

This chapter discusses the non-contrastive exophoric use of Tiriyó demonstratives, as reflected in the answers to a demonstrative questionnaire (Wilkins, 1999; this volume) developed as an elicitation tool at the Max Planck Institute for Psycholinguistics. Tiriyó demonstratives are first described in Meira (1999a) as forming a speaker-based system with three degrees of distance (proximal, medial, distal), plus a few non-spatial (anaphoric, invisible referent) extra terms. The data from the questionnaire has mostly confirmed this initial analysis, with a few extra details, namely: pointing and touching affect demonstrative choice; the “invisible” forms are really not simply about visibility but about ‘non-visible sensorial saliency’; and the two proximal inanimate forms (se(nï) and sere; cf. Table 10.1) are not spatially distinct but, rather, seem to instantiate a distinction between “new” (and in need of attention) versus “old” (already attended to). In the subsequent sections, a sociocultural and linguistic introduction to Tiriyó is offered (section 2), followed by a sketch of the initial analysis of the demonstrative system (section 3). The results of the demonstrative questionnaire are then presented (section 4) and discussed (section 5). Independent observational data and the results of other relevant investigations of the meanings of Tiriyó demonstratives are presented and compared with the questionnaire results (section 6), leading to a (slightly) revised analysis of the Tiriyó system and some considerations of its place in a typology of demonstratives (section 7). 2

Tiriyó and Its Speakers

Tiriyó is a language of the Taranoan (Tiriyó) sub-branch of the Cariban language family (Kaufman, 1994; Gildea, 1998; 2012; Meira, 2000b; 2005; Campbell, 2012). It has approximately 2,000 speakers living in several 222

Tiriyó Demonstratives

223

villages scattered on both sides of the Brazil–Surinam border in northern Amazonia. There are two main (mutually intelligible) dialects, called h-Tiriyó and k-Tiriyó in Meira (1999a); the data in the present work was provided by h-Tiriyó speakers from the Missão Tiriós village in Brazil. There are two in-depth descriptions of Tiriyó grammar to date: Meira (1999a) and Carlin (2004); earlier studies, dealing mostly with some aspects of the morphology, include an unpublished manuscript by Claude Leavitt, a missionary (1971), and De Goeje’s grammar sketch (1909). Other articles and studies have appeared more recently; cf., specifically on the topic of demonstratives, Meira (2003a; b), Meira and Terrill (2005), and on other topics, Meira (1998; 1999b; 2000a; 2001; 2004; 2005; 2006; 2007; 2011), Carlin (1999; 2003; 2006; 2011). The Tiriyó call themselves tarëno [taɽəːno] (etymologically, ‘someone from here, a local person’); this autodenomination, however, can also be used to refer to members of neighboring groups (when opposed to Westerners, Mestizos, or Blacks) or to Amerindians in general (e.g., other native South and North Americans seen in television documentaries). The word tïrïjo [tɨɽɨːjɔ] is used by people from neighboring groups, and in its Westernized forms Tiriyó (in Brazil) and Trio (in Surinam) by non-Amerindians. The Tiriyó recognize this word and use it to refer to themselves when dealing with outsiders, but they consider it a “foreign word”. In all Tiriyó villages, the Tiriyó language is the normal means of everyday communication. Almost all the Tiriyó are monolingual. Only a few people know any of the neighboring Amerindian languages (usually those who had long-term contact with the group whose language s/he learned, for example by marrying one of its members or having lived in one of its villages). The surrounding national languages, Portuguese (in Brazil) and Dutch (in Surinam), are spoken fluently only by a few individuals, despite being taught in the village schools. Even Sranantongo, the main lingua franca of Surinam, does not enjoy wide currency. The few outsiders who have intensive contacts with the Tiriyó (missionaries, village teachers, nurses) have found it necessary to learn at least some of the language. Of course, this situation is changing: as more and more contacts are established with the surrounding societies, the level of exposure of the Tiriyó to other languages and cultures (and the consequent bilingualism) are increasing. The Tiriyó economy is based on hunting and swidden agriculture (basically cassava). Traditionally, they lived in small villages with up to 100 inhabitants. A group of nearby villages formed a larger agglomeration in which economic and social exchanges were frequent; different agglomerations lived more or less in isolation. Contact with Westerners was infrequent (mostly explorers in the nineteenth and early twentieth centuries), and the traditional way of life was maintained.

224

Sérgio Meira

Table 10.1 Tiriyó third-person pronouns (after Meira, 1999a: 154) Inanimate Non-collective Anaphoric Demonstrative Visible Proximal Medial Distal Invisible

irë se(nï) sere mërë ooni më(nï)

Animate

Collective

Non-collective

Collective

irëto(mo)

nërë

namo

sento(mo) serëto(mo) mërëto(mo) oonito(mo) mënto(mo)

mëe

mëesa(mo)

mëërë ohkï më(kï)

mëëja(mo) ohkïja(mo) mëkïja(mo)

However, the arrival of Western missionaries (Franciscan Catholics in Brazil, American Protestants in Surinam) disrupted the traditional settlement pattern, leading to the concentration of the Tiriyó in larger agglomerations (Frikel, 1961; Rivière, 1969). Cultural influence has also led to the adoption of Western clothing, medicine, technology (firearms, vehicles, radios), and, to some extent, food (in Brazil, the Tiriyó now raise cattle). 3

The Previous Analyses of Tiriyó Demonstratives (Third-person Pronouns)

The two previous works which attempt to analyze the use of Tiriyó third-person pronouns are Meira (1999a) (with further details described in Meira, 2003a, b; Meira and Terrill, 2005) and Carlin (2004), which offer almost identical conclusions (with small differences that will be mentioned when relevant), summarized in Table 10.1. For other Cariban languages, similar terminology (except for the “visible” and “invisible” labels), and usually cognate forms, are also attested; see Derbyshire (1999: 53–55) for a cross-Cariban overview. As can be seen in Table 10.1,1 the Tiriyó demonstrative system is sensitive to animacy and number. The animacy category is semantically quite transparent: animate forms are used for persons and animals, including insects, and inanimate forms are used for everything else (including trees, plants, and moving

1

The transcription system used here is taken from Meira 1999a. Possibly ambiguous symbols: ï = [ɨ] ~ [ɯ]; ë = [ə] ~ [ʌ]; o = [ɔ] ~ [o]; e = [ɛ] ~ [e]; j = IPA [j], s = [s]̠ ~ [ʃ]; r = [ɽ] ~ [ɾ]̠ ; n = [n], but [ŋ] if word-final or followed by k, [m] if followed by [p], and [ɲ] if followed by j. Syllables in parentheses change form depending on the following material. For instance, (mo) is mo [mɔ] if, within the same phonological word (including affixes, clitics, etc.), it is followed by a non-CV syllable, but n [ŋ] otherwise. Thus, irëto(mo) occurs by itself as irëton [iɽəːtɔŋ], but, with the clitic nkërë ‘still’, it becomes irëtomo=nkërë [iɽəːtɔmɔŋkəɽə]. See Meira (1999a: 77–94; 1999b).

Tiriyó Demonstratives

225

objects like planes or cars).2 The number category is traditionally analyzed as opposing ‘less than all’ (non-collective) to ‘all’ (collective) rather than ‘one’ (singular) to ‘more than one’ (plural). This seems to be by and large true for Tiriyó, though some sporadic inconsistencies suggest that more research on the semantics of number is necessary. Meira (1999a) presents the first major distinction in the third-person pronoun class as that between anaphoric (used only to refer to what is known from previous discourse, i.e. endophorically) and demonstrative (used also to refer to entities in space, i.e. exophorically) pronouns.3 Within the demonstrative subclass, the following distinction was based on visibility: visible forms typically describe physical entities which the speaker and/or addressee can perceive visually, while invisible forms are used to refer to physical entities that cannot be perceived visually but somehow affect the speaker and/or addressee (e.g., a person singing in an adjacent room). The visible subclass is further subdivided according to the distance of the referent: proximal, medial, and distal terms are used when talking about entities that are “near”, “in the middle range”, or “far away” from the speaker and/or addressee.4 The obvious main target of the Demonstrative Questionnaire (Chapter 2, this volume) were the visible demonstratives. The objective was to examine in more detail the spatial meaning of these forms, so as to have a first idea about what the borders between “near”, “in the middle range”, or “far away” could be. In descriptive grammars, distance-related terms like proximal or distal are often used as more or less automatic labels for demonstratives, without much attention being paid to actual cross-linguistic differences between the terms that are given the same label (cf. the striking differences in use between the “proximals” or between the “distals” in the two-level systems represented by English this and that and by Russian etot and tot [cf. Wilkins’ and Dunn’s contribution to the Max Planck Annual Report; Van Geenhoven and Warner, 1999: 55–58]). Answers to a standard questionnaire, being more comparable crosslinguistically, can be used as a first step toward a better understanding of the language-specific features of these terms from an extensional perspective, to be later refined with data from more natural sources (texts, dialogues, in situ observations, etc.). 2

3

4

Stars are the only known exception: they are treated as animate. Note that other celestial bodies (sun, moon, clouds) are treated as inanimate. Among younger speakers, even stars are often treated as inanimate; older speakers, however, quite consistently prefer animate forms. Note that demonstrative pronouns can have endophoric uses, which are not further discussed here. Anaphoric pronouns, on the other hand, can only be used endophorically: they have no exophoric uses. Although anaphoric pronouns are never used with demonstrative (deictic) value, the converse is attested: demonstrative (visible) pronouns are sometimes used with anaphoric value. In the present work, as its title indicates, only the spatial demonstrative (exophoric) use of the demonstrative pronouns is examined; their anaphoric uses are left for future studies.

226

Sérgio Meira

4

Applying the Demonstrative Questionnaire

The Demonstrative Questionnaire (Wilkins, 1999; this volume) was used as an elicitation tool with four different speakers from the Missão Tiriós area in Northern Brazil. The speakers were all male, with ages ranging from 18 to 40–45. Two of them were monolingual, a third one could speak a little Portuguese, while the fourth one was fluent in Dutch. Each of the scenes proposed in the questionnaire was presented to the speakers with the help of visual props (i.e. real objects were placed so as to match the target situation); the speakers were told to interact with the investigator in a realistic manner.5 Their reactions, judgments, and comments were recorded (on video, for two speakers; in a notebook, for the other two). Since inanimate objects were used, all answers contained inanimate demonstratives (se(nï)/serë, mërë, ooni). To check if the animate forms would differ significantly from the inanimate forms, animate props (an ant and, for more distant scenes, local animals and people) were used in one further application of the questionnaire, with one of the four selected speakers; his answers showed the same behavior for the animate forms (mëe, mëërë, ohkï) with respect to distance as for their inanimate counterparts (a fact the speaker in question also mentioned spontaneously on several occasions during the work; e.g., when discussing scene 13, he remarked that “mëërë is just like mërë, but when we’re talking about animals, people”). After that, the answers containing inanimate demonstratives were compared and classified. The final results are summarized in Table 10.2.6 5

Discussion of the Results

Table 10.2 obeys the following conventions: (i) The first section contains pictures and numbers representing the proposed situations from the Demonstrative Questionnaire. Pictures are placed in

5

6

To give one example, in order to act out scene 3, I first described what was going to happen (“there is a bug on your shoulder, and you want me to look at it”). When I was sure the speaker had understood everything, I placed a prop (a pebble) on the speaker’s shoulder and pretended to be doing something else. The speaker would then call my attention to the bug on his shoulder in whatever way seemed most natural to him. (The speaker had been asked to respond as spontaneously as possible before we started working with the questionnaire.) All explanations and interactions were carried out in Tiriyó. There are, of course, certain problems with this kind of approach. First, although the speakers did understand the situations and tried to produce natural reactions, there is always the danger, inherent to elicitation, that their answers did not entirely reflect their actual usage. Second, since every individual has his or her own personal way of expressing levels of doubt and certainty, it was necessary to interpret their reactions to each situation, which introduced an inevitable amount of subjectivity. Despite these shortcomings, the results seem consistent enough to be significant.

A

A

SE(NÏ) / SERË

Usually with pointing and/or touching. The speaker’s body parts are always se(nï) / serë.

(7)

(6)

(3)

S

Demonstratives are not good. Only one speaker used se(nï) / serë thought other strategies (‘the one behind me’) equally good.

S

S

(1)

S

Se(nï) / serë (4 sp.)

(11)

A

Other strategies (3 sp.); se(nï) / serë (1 sp.)

S

(19)

S

(4)

(2) A

A

A

With touching or close pointing, se(nï) / serë becomes the only possible choice.

Se(nï) / serë much better than mërë

S

S

S

A

A

(23)

A

(22)

S

(20)

S

(8)

A

MËRË

With touching or close pointing, se(nï) / serë becomes the only possible choice. Mërë is used if ‘it’s a bit further away’; se(nï) / serë becomes ‘better’ if there is (not necessarily close) pointing. The one speaker who accepted ooni still preferred mërë.

Se(nï) / serë better than mërë (2 sp.) mërë better than se(nï) / serë (2 sp.) and than ooni (1 sp.)

S

A

(21)

S

(12)

(17)

S

(16)

S

(14)

S A

(13)

S A

A

A

(18)

A

S

(15)

S A

S

A S

A

(25)

(24)

OONI

Distance was considered the important factor: using ooni ‘means that it is more distant’ than using mërë would.

Visibility was never mentioned as an important factor.

Distance (‘if it’s further away, ooni is better’) sometimes mentioned. Visibility was never considered an important factor.

A

A

Se(nï) / serë is not good, but becomes better if there is pointing. Ooni suggests that it is ‘more distant’.

(10)

A

(9)

(5)

S

Mërë better than Ooni much better ooni (3 sp.) ooni than mërë (2 sp.) slightly better ooni only (2 sp.) than mërë (1 sp.)

S

A

Sërë better than se(nï) / serë (4 sp.) Mërë better than ooni and than ooni (1 sp.) (4 sp.)

S

S

Table 10.2 Results of the demonstrative questionnaire (total of four consulted speakers)

228

Sérgio Meira

the same column if the corresponding situations received approximately the same treatment from the speakers in terms of demonstrative choice. (ii) The second section contains the actually occurring demonstratives, with the number of speakers who preferred a given term in parenthesis. “Other strategies” refers to non-demonstrative solutions (i.e., answers which, instead of a demonstrative, contained some other indication, such as “the one behind me”, “the one which we can’t see”, etc.). (iii) The third section contains comments based on observation of the speakers’ reactions to the situations, as well as comments and observations by the speakers themselves. (iv) The arrows under the table graphically indicate the observed extension of the demonstrative terms. Solid lines cover areas for which the term in question was considered a perfectly good, usually the best, solution. Broken lines cover areas for which the term in question was accepted, but only conditionally, or as a “less good” choice. A first inspection of the data in Table 10.2 suggests that Tiriyó demonstratives can be analyzed as a speaker-centered distance-based system: se(nï)/serë is ‘proximal’, mërë is ‘medial’ and ooni is ‘distal’ from the speaker. Before elaborating on this analysis, however, let us first discuss some specific consequences from the data in Table 10.2. (1) For certain situations, demonstratives are not optimal. Scene 11, in which the speaker refers to an object situated behind his back, was problematic for most speakers. The spontaneous answers had descriptive phrases (e.g., “the thing behind my back”) rather than demonstratives. Only one speaker used a demonstrative (se(nï)/serë), accompanied by face pointing (i.e., he looked over his shoulder in the general direction of the referent; see (4) below), and he also accepted descriptive expressions as equally good. For the other three, a simple demonstrative was clearly insufficient: it would not be of much help for the addressee to locate the referent. (2) Pointing and/or touching was relevant for the choice of the right form. For both situations in which the speaker touched the referent (1, 2), se(nï)/serë was obligatory; this was also true for any variation of the other situations that involved touching the referent. Close pointing, even without touching, was already enough to make se(nï)/serë strongly preferred (although mërë was occasionally accepted). Pointing from a certain distance was perfectly compatible with mërë (e.g., 5, 9, 10, 12, 21) or even ooni (e.g., 24, 25), but it frequently made se(nï)/serë at least possible. As is described in detail in the next section, spontaneous uses of se(nï)/serë in real-life situations to refer to objects not close to the speaker were invariably accompanied by pointing. It is as though pointing would bring referents closer to the speaker and addressee.

Tiriyó Demonstratives

229

(3) None of the terms refers specifically to the addressee. Unlike, for instance, spoken Brazilian Portuguese (see Meira and Guirardello-Damian, this volume), there are no specific addressee-anchored terms. For instance, scenes 9, 10, and 12 are treated alike (the proximal terms se(nï)/serë were less adequate than the medial term mërë) regardless of distance from addressee, while scene 8, in which it was possible to see the referent as close to the speaker, was treated differently (se(nï)/serë better than mërë). The same is also true for scenes 13, 14, 16, and 17, as well as for scenes 15 and 18: the position of the addressee could be changed at will without rendering one of the terms impossible for that situation. This, however, does not mean that the position of the addressee is irrelevant: as was mentioned in Meira (2003a), in scenes like 16 or 18, the first spontaneous answer was usually the medial term mërë, even though the distal ooni was considered “also OK”. In other words, the presence of the addressee next to a speaker-distant referent makes it more likely (though not necessary) that the medial term mërë will be used rather than the distal term ooni, perhaps because the speaker and addressee feel “conceptually closer” to each other, by virtue of their shared engagement area (see Enfield, 2003 for this concept). (4) There was no observable difference between the proximal terms se(nï) and serë. For all situations in the questionnaire, all speakers agreed that if se(nï) was appropriate, then so was serë. The semantic distinction between these two terms is apparently not related to distance. In fact, Meira (2003b) points to the hypothesis that the main difference between them is whether or not the referent is considered by the speaker to be “new”: serë indicates “novelty”, while se(nï) is better suited for objects that are not “new”.7 (5) The so-called “invisible” term never occurred. Më(nï), listed in Table 10.1 as “inanimate invisible”, was never spontaneously used for any of the situations proposed in the questionnaire, although some of them (e.g., scenes 11, 15, 25) involve referring to objects that the speaker, or the addressee, or both, could not see. This casts some doubt on whether “invisible” is the best label for the semantics of më(nï) and its animate counterpart më(kï).

7

Carlin (2004: 152) claims that there is a syntactic difference between se(nï) and serë: the latter can be used to modify a noun, whereas the former cannot. In my own field data, however, exceptions to this claim (i.e., cases of serë followed by a noun) are not infrequent. The difference between these terms is thus not syntactic, but semantic (see section 7 below). Carlin also mentions a distinction between the locative phrases sen po and serë po ‘here’ (both with the locative postposition po ‘at’): sen po would be more specific (‘this very place’) and serë po more general (‘this general area’). There does seem to be a correlation, but still, in this case, there are also significant exceptions (see Meira, 2003b for a discussion). This specific topic needs further research.

230

Sérgio Meira

A closer look at the available data reveals that “invisible” demonstratives are used when the speaker wishes to refer to physical entities that can be heard, but not seen, a pattern first mentioned by Carlin (2004: 148). Thus, akï mëkï? ‘who’s that?’ can refer to a person singing in an adjacent room, or to someone who is talking aloud outside, or to an animal making noise on the roof, and atï mën? ‘what’s that?’ to a far-away machine making noise (e.g., an airplane, or a diesel electricity generator, or fireworks). One might wonder if being away from the place occupied by the speaker and the addressee is an important part of the semantics of these forms; however, this is not so. During the elicitation of the questionnaire, one speaker, after refusing to use më(nï) for a certain scene, added that it could be used if the object was making noise inside of a box. This was tested with the other three speakers, who all spontaneously asked atï mën? ‘what’s that?’ about an alarm clock audibly tick-tocking inside of a cardboard box. It thus seems that the typical referent of an “invisible” form is something that is audibly, but not visually, perceptible to the speaker and the addressee.8 Interestingly, objects that were not directly visible but did not produce any audible noise were also not referred to with the “invisible” demonstratives; rather, if there was a perceptible feature (e.g., a bulge under one’s shirt), “visible” demonstratives were used, and if no such perceptible feature was present, then non-demonstrative solutions (“the pen that was here before”, “the pen I saw earlier”, etc.) were used. Such cases suggest that the most important feature may be whether or not the speaker can physically situate the object (it is either visible or has a visible effect in a certain location), in which case “visible” demonstratives are used, as opposed to perceiving its noise but not being able to determine its location, in which case “invisible” demonstratives are more appropriate. (6) There was no situation for which the medial term mërë was used exclusively. Surprisingly, though there were situations in which all speakers agreed that only se(nï)/serë could be used and situations in which at least two speakers agreed that only ooni could be used, there were no situations in which only mërë was possible: all speakers agreed that, even though some situations are best described with mërë, other demonstratives were also acceptable, albeit less felicitously. This lack of a “mërë-only scene”, together with the fact that mërë was applied to a larger number of scenes than either ooni or se(nï)/serë, suggests 8

Note that, as one might expect, entities which are not only audibly but also visually perceptible were referred to with the “visible” demonstratives. Objects perceptible with other senses caused inconsistent answers. When asked how to say “what is that?” when something tasty is being cooked and its smell is perceptible, four consulted speakers used various different nondemonstrative strategies (e.g., “what is producing smell?” or “what is the thing I smell?”).

Tiriyó Demonstratives

231

a possible analysis whereby mërë is the “vaguest” exophoric demonstrative, while ooni and se(nï)/serë are its more specific counterparts, indicating explicitly that the referent is close or far away. In other words, mërë appears to be a neutral term which does not really encode distance but is used when speakers are not sure that the situation requires the more specific forms ooni or se(nï)/serë (i.e., by default, “medial” situations). Although appealing, this idea is contradicted by other results, as is discussed below, in section 7.

6

Observational Data

Ideally, the exophoric use of demonstratives as revealed by the questionnaire should be compared with non-questionnaire data for an independent evaluation. However, most of the data usually collected in field trips (texts, conversations, elicitation) is typically not satisfactory for comparison, even when is do contain demonstratives used exophorically, since crucial facts like the actual distance of the referent or whether there was pointing and/or touching by the speaker are not recorded. In view of that, a certain amount of observational data, consisting of real-life scenes observed by the researcher in which demonstratives were used exophorically, was recorded during the last field trip (typically by taking notes in a field notebook immediately after the event). A representative sample of these real-life scenes is given in the following subsections. 6.1

Distal Scene

In this scene, the collective animate distal pronoun ohkïja(mo) was used. Observation (A). During the execution of a task that consisted of aligning toy animals on a table so as to reproduce a scene that had been observed on another table a few moments before (the “Animals-in-a-Row” task; cf. Levinson and Schmitt, 2003; Pederson et al., 1998), one speaker (a 25year-old male monolingual), after being told what to do, asked for confirmation by saying ohkïjan apo? ‘like those ones there?’ to me (i.e., ‘should I align these animals exactly like the ones on the other table we saw before?’). The other table was in a different room, 20–25 meters away from where we were and invisible to both of us. Two facts seem noteworthy. First, although the toy animals were made of plastic, he used the animate form ohkïja(mo). Second, they were not visible anymore; yet, he did not use the so-called “invisible” forms mënto(mo) (inanimate) or mëkïja(mo) (animate).

232

Sérgio Meira

6.2

Medial Scenes

In these scenes, the non-collective medial pronouns mëërë (animate) and mërë (inanimate) were used. Observation (B). I was walking down a path with two young (18- to 20-yearold) Tiriyó friends. At one point, we saw two people on the same path, coming from the opposite end, probably 150–200 meters away from us. One of them was clearly closer to us, and his face was already visible; it was possible to recognize him. The other one was behind him, too far for recognition (as far as I could tell). One of the young Tiriyó turned to his friend and said: akï mëërë? (‘who is that one?’). His friend replied: aano? (‘which one?’), to which he answered: wapono (‘the first one’). Two facts seem noteworthy. First, although the two men were very far away, the distal term ohkï was not used. Second, the use of the medial term mëërë was apparently insufficient for the addressee to identify which of the two men the speaker had in mind, despite the clear difference in distance between the two men. Observation (C). I was next to the river, looking for a certain man, but I did not know exactly where his house was. At that moment, I saw an old woman washing dishes by the river. I squatted next to her and asked where the man’s house was. After answering me, she asked me to hand her the soap, which was lying next to her, between us (within arm’s length of both of us). Since she spoke rather fast, I failed to understand the word for ‘soap’ (soopu, an obvious borrowing from one of the English-based pidgins of Surinam) and said to her: atï se manan? ‘what do you want?’ Her answer was mërë! (‘that one!’), accompanied by lip-pointing. This scene is virtually identical to scene 8, one of those for which there was variation among speakers as to whether the best solution was one of the proximals (serë/se(nï)) or the medial (mërë). It is interesting to note that the woman could not point, since she was washing dishes and her hands were busy (presumably, pointing would have made serë/se(nï) more likely). 6.3

Proximal Scenes

In these scenes, the non-collective inanimate proximal pronouns se(nï) and serë, and their animate counterpart mëe, were used. Observation (D). A male, 50-year-old monolingual speaker once pointed at a crucifix that was hanging on the wall (inside the Mission’s house) approximately 2–3 meters behind my back and said Jesu ikuhtu serë (‘This is

Tiriyó Demonstratives

233

a representation of Jesus’). Immediately after that, he repeated the utterance, this time with lip-pointing. Observation (E). Two young men were talking about an airplane that had been announced by radio about two hours ago and was probably about to become visible. I had already finished talking to them, and was about to leave, when suddenly the noise of the airplane became audible. We soon saw the airplane (a small single-engine plane), but it disappeared behind the leaves of the nearby trees. After a few seconds, one of the men saw the airplane again, but the other one did not, because he had been looking in the wrong direction. The man who saw the airplane pointed to it and called his friend’s attention to it by saying: serë (‘this one’). The airplane was preparing to land, but it was still flying probably several hundred meters above us. Observation (F). While video-recording an interview with a middle-aged monolingual male speaker, I asked at one point: anpo nai Taratarahpë ‘where is (the village of) Taratarahpë?’. His answer (accompanied by fingerpointing) was: amohtëe, serë tae ‘upstream, along this one’. The proximal serë was used to refer to the nearby river, approximately 300–400 meters away from us, and not visible. In (D–F) above, the proximal serë was used to direct the addressee’s attention to something that was probably “new” to the addressee (even in D, in which both speakers had already been talking about the airplane, one of them had lost sight of it). The demonstrative was accompanied by pointing in all three cases. It is noteworthy that the proximal forms were used for non-nearby referents (the airplane in (E) was obviously very far away, and the river in (F) was also far from the place where the interaction happened), or for referents close to the addressee (D). Observation (G). I asked an older woman (a monolingual speaker), who was grating manioc outside, what animal the word soni (‘vulture, buzzard’) referred to. Since there were some buzzards at a distance, eating the carcass of a dead dog, she just pointed at one of them and said mëe, mëe (‘this one, this one’). They were approximately 20 meters away from us. Here, the animate form mëe, like the inanimate serë, is used to refer to something “new” to the hearer. Presumably, mëe would also correspond to se (nï). It is somewhat surprising that two proximal forms exist only for inanimates. Observation (H). As I was sitting outside next to the school classroom (in which there is a television set), I saw a middle-aged male monolingual speaker tell a woman (who was sitting on a small wooden stool with a baby in her arms about 10–15 meters away from the classroom door) that he was going home

234

Sérgio Meira

first to talk to his daughter and then would return to watch television. Asking for confirmation, she then said: sen tao? ‘in this one? in this room?’ (i.e., “are you going to watch television here in the classroom, or somewhere else?”). They were facing each other, and the classroom door was on the woman’s left. Her utterance was accompanied by face and lip-pointing (she could not use her hands because of the baby). In this case, se(nï) was used to refer to something conceivably not “new” (the topic of conversation was watching television, which should immediately suggest the classroom, in which there is a television set; besides, the woman was asking for confirmation, not calling the man’s attention to something that he had failed to notice). 7

Towards a Better Understanding of Demonstratives in Tiriyó and in General

As was said in the introduction, the aim of this chapter is to consider how data on the non-contrastive exophoric usage of Tiriyó third-person pronouns, as elicited with the Demonstrative Questionnaire (Wilkins, 1999; this volume), can contribute to a better analysis of the system. Of course, such data cannot be complete, not only because the focus is on one aspect of usage (exophoricity), but also because staged scenes are inescapably less natural than ordinary everyday situations and thus will, in principle, be only imperfect reflections of normal usage. The spontaneous situations described in section 6 show this fact: in them, distance (near versus far) was often not sufficient to determine demonstrative choice, which was clearly guided by other considerations (e.g., directing the speaker’s attention to a new object), whereas in the questionnaire results distance was often enough for certain choices to be excluded. Despite all that, there were sufficiently many interesting and consistent results from the questionnaire, clearly relevant to any attempt at explaining the semantics of Tiriyó demonstratives. These results, already listed in section 5 above, were as follows: (1) (2) (3) (4) (5)

certain situations are not felicitously described with demonstratives; pointing and touching affect demonstrative choice; there are no addressee-anchored terms; the proximal terms se(nï) and serë are not differentiated by distance; the so-called “invisible” forms are not used for referents that are simply “invisible” (see case A in section 6), and the so-called “visible” forms can also be used for referents that are not visible (see case F in section 6); (6) there was no situation for which only the medial term was adequate.

Tiriyó Demonstratives

235

The first result is probably not simply a Tiriyó phenomenon (it was also noted for spoken Brazilian Portuguese; see Meira and Guirardello-Damian, this volume), but a more general feature of demonstrative semantics: there are situations, possibly not all language-dependent, which make demonstratives a bad choice to help the addressee identify the referent meant by the speaker. It is as if demonstratives occupied “areas” or “fields” around the speaker and/or addressee that did not cover the entire area where their interaction takes place. Judging by the Tiriyó case (scene 11, where the referent is behind the speaker but visible to the addressee), demonstratives are felicitously used in a space defined by the direct perception field (especially visual or auditory) of the speaker and addressee. The second result is likewise not simply a Tiriyó phenomenon, as can be seen in other contributions to this volume: close pointing or touching generally has the effect of making proximal demonstratives more appropriate, indeed often obligatory. All in all, it seems that the “areas” or “locations” that demonstratives refer to (“here” versus “there” versus “over there”, “reachable” versus “non-reachable”, etc.) are the result of constant negotiation between speaker and addressee, and can change form or even be abandoned as their interaction evolves (cf. the concept of “engagement area” as the region around the speaker and/or addressee defined by their interaction, their joint and/or individual actions and activities, as discussed in Enfield, 2003) via acts often as simple as pointing or touching. Despite traditional views, “absolute” distance or location is not really the factor guiding demonstrative choice, but a dynamic understanding by the speaker of the current state of his/her interaction with the addressee and how this affects their conceptualization and use of the space where they are interacting, to say nothing of other possibly relevant factors or domains (e.g., shared cultural knowledge, social hierarchies, logical presuppositions, shared individual history, etc.). The spontaneous scenes discussed in section 6 certainly give further support to this view by showing that “proximals” can felicitously refer to far-away objects in certain situations. The import of (physical) space or distance must therefore be that of an external set of parameters, not under the speaker’s (or addressee’s) control, with a limiting effect on the possible construals of the various areas in the interactional space of speaker and addressee. The third result is certainly specific to Tiriyó, since there are languages (like spoken Brazilian Portuguese, in this volume) which have strongly addresseeanchored terms in their demonstrative systems. Note, however, the “addressee effects” alluded to above (and discussed in Meira, 2003a), which make the selection of medial terms more likely (without excluding distals) if the referent is close to the addressee. The actual physical location of the addressee is clearly an important factor influencing the construal of the various interactional space

236

Sérgio Meira

areas by speaker and addressee, even in languages which, like Tiriyó, lack addressee-anchored terms. The fourth result concerns the proximal inanimate terms se(nï) and serë. As was observed above, (physical) spatial distance is not the crucial factor in demonstrative choice (it acts more like a set of limitations or weights added to possible construals); nevertheless, it is noteworthy that a questionnaire which produced results that tended to overemphasize the role of spatial distance (in that terms were deemed inappropriate reportedly on account of spatial distance) failed to highlight any consistent influence of distance on the choice between se(nï) and serë. This fact is compatible with the observations in section 6 (and in Meira, 2003b) about se(nï) and serë differing not in spatial cues, but in whether or not the referent in question is “new” (= as yet unattended to by the addressee) or “old” (= already present in the addressee’s awareness). The fifth result concerns the actual usage of the so-called “invisible” terms, më(nï) (inanimate) and më(kï) (animate): there are situations in which non-visible referents are not referred to with these terms, but with their “visible” counterparts. As was seen above, a better characterization of their actual usage is that they rely on audibility, i.e., they refer to entities that can be heard but not seen (other sensorial perception modes, e.g., smell, produced inconsistent results). Here, the area of demonstrative usage (deixis) comes close to that of evidentiality (source of knowledge), a connection already mentioned elsewhere (see Hanks, 2012; see also the introduction to this volume). The sixth and final result concerns the medial term mërë (and presumably its animate counterpart mëërë). Its proximal and distal counterparts were considered the only possible choices by at least some speakers for certain scenes (1, 3, 6, and 7 for the proximals se(nï)/serë, 24 and 25 for the distal ooni); mërë, however, was never considered the only choice, though it often was the best choice. One possible explanation would rely on the pragmatic difference between encoded and implied/inferred meaning (see Levinson, 1983 and especially 2000): mërë might be a “spatially neutral” term which did not encode any distance or any “construed region” in the speaker-addressee interactional space, but which, when contrasted with other choices (semantically marked proximal and distal terms that do encode distance distinctions), ends up being used mostly in reference to “middle-range” locations. In Grice’s (1975) terms, a failure to use a proximal or a distal if the conditions for their use are met would be a violation of the maxim of Quantity; only in the other (middlerange) contexts would the use of mërë respect this maxim. Similar neutral or unmarked “medials” exist in other languages (see Terrill on Lavukaleve, this volume). However, additional evidence involving contrastive situations, in

Tiriyó Demonstratives

237

which speaker and addressee must deal with more than one possible referent, suggest that mërë is not really semantically unmarked, and that (construed) distance is indeed encoded as part of its meaning (for a comparison between Tiriyó and Lavukaleve dials” in contrastive contexts, see Meira and Terrill, 2005). Taking into account the above results from the Demonstrative Questionnaire, as well as those from contrastive elicitation (Meira and Terrill, 2005) and from spontaneous situations such as those described in section 6, the following analysis for the Tiriyó system of third-person pronouns is proposed: 1. The first semantic opposition is between forms that refer only to discourse (“anaphorics”) and forms that can refer to non-discourse (“demonstratives”). The anaphoric pronouns (irë, nërë in Table 10.1) can never be used for exophoric reference, which is the exclusive province of the demonstrative pronouns (se(nï)/serë, mërë, ooni, më(nï) and their animate counterparts mëe, mëërë, ohkï, më(kï)).9 2. Within the set of demonstratives, the first opposition is between më(nï)/ më(kï) ‘audibly, but not visually perceptible’ and the others. Clearly, më(nï) (inanimate) and më(kï) (animate) are semantically more specific: they can be used only to refer to entities which exist in the speaker-addressee interactional space but whose existence is ascertained via a secondary mode of perception (hearing, rather than seeing). These forms will be here termed audible demonstratives, while all the others, semantically unmarked for this feature, will simply be called general demonstratives. 3. The set of general demonstratives has preferred uses: it pragmatically implicates a “normal” situation for the referent. In other words, once “audibility” is not at stake, the felicitous exophoric use of a Tiriyó general demonstrative suggests a certain “default” extension and structure of the interactional space between speaker and addressee that excludes unusual positions, such as behind the speaker’s back, or surprising situations, such as when a book (= referent) is hanging on a rope, as well as situations in which the number of possible referents is too large (say, several books on a table, only one of which the speaker wants the addressee to attend to). In these cases, speakers will prefer non-demonstrative identifying expressions (“the one behind my back”, “the one on the rope”, “the largest one”, “the one on top”, etc.). The fact that demonstratives are still marginally acceptable in at

9

Interestingly, at least some demonstrative pronouns do have endophoric uses: for example, the proximal “new referent” pronoun serë can be used cataphorically, to refer to something that the speaker is about to say (“and the problem was this: he came too late and I was already asleep”). An examination of such uses is beyond the scope of this paper and will be left aside for future publications.

238

Sérgio Meira

least some of these cases points to a pragmatic analysis: one could propose as a working hypothesis that the choice of a general demonstrative (rather than a descriptive phrase) pragmatically implicates that the referent occupies a “normal”, “not surprising” position (typically within the field of vision of the speaker and the addressee) and is itself sufficiently salient (e.g., there are not too many other potential referents around). 4. The further structure within the set of accessible demonstratives is not entirely clear. The contrastive evidence suggests that the medial mërë does encode distance as part of its meaning, but the absence of a “prototypical” scene that excludes other demonstratives suggests a certain vagueness. Noting, in Table 10.2, that the “prototypical” distal scenes 24 and 25 were not unanimously considered to exclude mërë (two speakers considered mërë possible, though ooni was better, while two speakers considered mërë impossible and ooni obligatory), one may suggest a tentative analysis whereby mërë is an “unmarked distal” and ooni a “marked distal” (reminiscent, though not equivalent, to the difference between the place adverbs ali and lá in Brazilian Portuguese; see Meira and Guirardello-Damian, this volume). In other words, as (construed) distance increases, the use of ooni would be favored over mërë, while the latter would, in principle, still be possible (presumably, e.g., in cases in which it is not clear whether the referent is really very far away or not). This would imply a first-order distinction between proximals and distals, with a further opposition between an unmarked distal and a far-away distal. Figure 10. 1 provides a schema of the different oppositions proposed above, and Table 10.3 presents a consequent revision of Table 10.1. discourse

anaphoric pronouns (irë)

Referent in: audible but not visible

më(nï)

elsewhere (space) all other cases

distal

proximal

marked

ooni

unmarked

mërë

‘new’

se(nï)

‘old’

serë

Figure 10.1 Oppositions in Tiriyó third-person pronouns (inanimate forms only for convenience)

Tiriyó Demonstratives

239

Table 10.3 The Tiriyó system of third-person pronouns: a revised version Inanimate

Anaphoric Demonstrative General

Distal Proximal

Audible

Unmarked Marked “Old” “New”

Animate

Noncollective

Collective

Noncollective

Collective

irë

irëto(mo)

nërë

namo

mërë ooni se(nï) serë më(nï)

mërëto(mo) oonito(mo) sento(mo) serëto(mo) mënto(mo)

mëërë ohkï mëe

mëëja(mo) ohkïja(mo) mëesa(mo)

më(kï)

mëkïja(mo)

References Carlin, E. (1999). WYSIWYG in Trio: The grammaticalized expression of truth and knowledge. Journal of the Anthropological Society of Oxford (JASO), 30(3), 233–245. (2003). The Experiencer role in the expression of mental states and activities in Trio (Cariban). Amérindia, 23, 161–182. (2004). A grammar of Trio: A Cariban language of Suriname. Duisburger Arbeiten zur Sprach- und Kulturwissenschaft, vol. 55. Frankfurt am Main: Peter Lang (Europäischer Verlag der Wissenschaften). (2006). The verbalizers in Trio (Cariban): A semantic description. In G. J. Rowicka & E. B. Carlin, eds., What’s in a verb? Studies in the verbal morphology of the languages of the Americas. LOT Occasional Series. Utrecht: LOT, pp. 143–157. (2011). Theticity in Trio (Cariban). International Journal of American Linguistics, 77, 1–31. Campbell, L. (2012). Classification of the indigenous languages of South America. In L. Campbell & V. Grondona, eds., The indigenous languages of South America: A comprehensive guide. The World of Linguistics series, vol. 2. Berlin/Boston: De Gruyter Mouton, pp. 59–166. De Goeje, C. H. (1909). Grammaire et vocabulaire de la langue trio. In Études linguistiques caraïbes. Verhandelingen der Koninklijke Akademie van Wetenschappen te Amsterdam, Afdeeling Letterkunde, Nieuwe Reeks, Deel X, no. 3. Amsterdam: Johannes Müller, pp. 180–243. Derbyshire, D. C. (1999). Carib. In R. M. W. Dixon & Alexandra Y. Aikhenvald, eds., The Amazonian languages. Cambridge Language Surveys. Cambridge: Cambridge University Press, pp. 23–64. Enfield, N. J. (2003). Demonstratives in space and interactional: Data from Lao speakers and implications for semantic analysis. Language, 79(1), 82–117. Frikel, P. (1961). Fases culturais e aculturação intertribal no Tumucumaque. New Series of the Boletim do Museu Paraense Emilio Goeldi (Antropologia), vol. 16. Belèm: Museu Paraense Emilio Goeldi.

240

Sérgio Meira

Gildea, S. (1998). On reconstructing grammar: Comparative Cariban morphosyntax. Oxford Studies in Anthropological Linguistics, vol. 18. New York/Oxford: Oxford University Press. (2012). Linguistic studies in the Cariban family. In L. Campbell & V. Grondona, eds., The indigenous languages of South America: A comprehensive guide. The World of Linguistics series, vol. 2. Berlin/Boston: Mouton De Gruyter, pp. 441–494. Grice, H. P. (1975). Logic and conversation. In P. Cole & J. Morgan, eds., Syntax and semantics. Vol. III: Speech acts. New York: Academic Press, pp. 41–58. Hanks, W. F. (2012). Evidentiality in social interaction. In J. Nuckolls & L. Michael, eds., Evidentiality in interaction. Special edition of Pragmatics and Society, 3(2), 169–180. Kaufman, T. (1994). The native languages of South America. In C. Moseley & R. E. Asher, eds., Atlas of the world’s languages. New York, London: Routledge, pp. 46–76. Leavitt, C. (1971). Tiriyó grammar. Unpublished manuscript. Levinson, S. C. (1983). Pragmatics. Cambridge: Cambridge University Press. (2000). Presumptive meanings: The theory of generalized conversational implicatures. Cambridge, MA: MIT Press. Levinson, S. C. & Schmitt, B. (1993). Animals in a row. In S. C. Levinson, ed., Cognition and space kit (version 1.0): July 1993. Nijmegen: Cognitive Anthropology Research Group, Max Planck Institute for Psycholinguistics, pp. 65–69. Levinson, S. C. (2003). Space in language and cognition: Explorations in cognitive diversity. Cambridge, Cambridge University Press Meira, S. (1998). Rhythmic stress in Tiriyó (Cariban). International Journal of American Linguistics, 64(4), 352–378. (1999a). A grammar of Tiriyó. Doctoral thesis, Rice University, Houston. (1999b). Syllable reduction and ghost syllables in Tiriyó. In S. J. J. Hwang & A. R. Lommel, eds., XXV LACUS Forum. Fullerton, CA: The Linguistic Association of Canada and the United States (LACUS), pp. 125–131. (2000a). Reduplication in Tiriyó (Cariban). Languages of the World, no. 17. Munich: LINCOM Europa. (2000b). A reconstruction of Proto-Taranoan: Phonology and morphology. Munich: LINCOM Europa. (2001). Linguistic theory and linguistic description: The case of Tiriyó [h]. International Journal of American Linguistics, 67, 119–135. (2003a). “Addressee effects” in demonstrative systems: The cases of Tiriyó and Brazilian Portuguese. In F. Lenz, ed., Deictic conceptualization of space, time and person. Pragmatics & Beyond New Series, vol. 112. Amsterdam/Philadelphia: John Benjamins, pp. 3–12. (2003b). Les démonstratifs proximaux non-animés de la langue tiriyo (caribe): Une étude de corpus. Amerindia, 28, 183–200. (2004). Mental state postpositions in Tiriyó and other Cariban languages. Linguistic Typology, 8, 213–241. (2005). Cariban languages. In K. L. Brown, ed., Encyclopedia of language and linguistics. 2nd edn. Oxford: Elsevier, pp. 199–204. (2006). Tiriyó body part terms. Language Sciences, 28, 262–279.

Tiriyó Demonstratives

241

(2007). Orações relativas em línguas Karíb. Boletim do Museu Paraense Emílio Goeldi, Ciências Humanas, 2(1), 105–121. (2011). Tiriyo. In E. Bonvini, J. Busuttil & A. Peyraube, eds., Dictionnaire des langues. Paris: Presses Universitaires de France, pp. 1495–1501. Meira, S. & Terrill, A. (2005). Contrasting contrastive demonstratives in Tiriyó and Lavukaleve. Linguistics, 43(6), 1131–1152. Pederson, E., Danziger, E., Wilkins, D. P., Levinson, S. C., Kita, S. & Senft, G. (1998). Semantic typology and spatial conceptualization. Language, 74(3), 557–589. Rivière, P. (1969). Marriage among the Trio: A principle of social organization. Oxford: Clarendon Press. Van Geenhoven, V. & Warner, N., eds. (1999). Annual report 1999. Nijmegen: Max Planck Institute for Psycholinguistics. Wilkins, D. P. (1999, this volume). The 1999 demonstrative questionnaire: ‘this’ and ‘that’ in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24.

11

Trumai: Non-contrastive Exophoric Uses of Demonstratives Raquel Guirardello-Damian

1

Introduction

1.1

The Present Study

The focus of this study is the Trumai demonstrative system and how it works in non-contrastive exophoric uses (i.e., when one is referring to something in the extralinguistic context). The system has been described in previous studies (Guirardello, 1992; 1999) but some gaps in the data and description did not lead to a complete view of how demonstratives work in this language, which pointed to the need of further analysis. A new investigation was conducted using the Wilkins Demonstrative Questionnaire (1999; this volume), a fieldwork elicitation tool especially designed for the study of demonstratives. The present chapter reports the results of this investigation, which lead to new findings and the clarification of several questions. They are discussed in detail in sections 2 and 5, but for an overview, it could be mentioned: • Additional demonstrative forms were recognized (inde, ka’in), filling gaps in the previous studies. • The uses of some demonstrative forms (ni and in) were better understood. • The forms beginning in in were also better understood. • The organization of the Trumai system is now clearer: it is not a two-term, but rather a three-term system. • There is also a relationship between demonstrative locational semantics and adverbial semantics (ni’de and nina, ka’ne and kaina, etc.). • When the demonstratives are used pronominally, there is a distinction between human, animate, and inanimate referents. • Whether the addressee is estimated to be able to reach the referent influences demonstrative choice in Trumai (inde versus ka’ne). • In some scenes of the Wilkins questionnaire, speakers reported that it was inappropriate to use a demonstrative or that it was preferred to use 242

Trumai of Demonstratives

243

a descriptive NP. This is part of the evidence of the semantics of the demonstratives and the nature of oppositions in the paradigm. The chapter is organized as follows: section 1.2 provides a brief introduction to the Trumai people and their language. Section 1.3 describes previous studies about the demonstratives and the issues that needed further investigation. Section 2 explains how the elicitation with the Wilkins questionnaire was conducted and its main results (presented in Table 11.4). Section 3 adds further information about the use of the demonstratives, taken not from elicitation but rather from naturalistic data.1 In section 4, we have some remarks about contrastive uses of the demonstratives (i.e., when two entities are pointed and contrasted). Next (section 5), we have a summary of the main findings and the proposal of a new analysis for the Trumai demonstrative system. Section 6 presents a general conclusion about the facts observed in the study, with additional comments. 1.2

The People and the Language

The Trumai people live in the central region of Brazil, in a multiethnic indigenous reserve. The group does not live in one single site, but rather is divided among three different villages and other locations in the area. The language, which is genetically isolated (Rodrigues, 1906; Derbyshire & Pullum, 1986), has a reduced number of speakers: although the population consists of more than 130 individuals, less than 50 speak it. In the older generation, almost all adults are bilingual in Trumai and Brazilian Portuguese (the official language of the country), with different degrees of proficiency in the latter. For the younger generations, Portuguese is the language employed in daily life, although some young women still speak Trumai. There are various historical factors that lead to the current scenario (see Guirardello, 1999: preface). There are four open classes in Trumai: nouns, verbs, adjectives, and adverbs. Personal and demonstrative pronouns can be considered a subclass of nouns, since they receive nominal morphology (e.g., case markers) and exhibit almost the same syntactic behavior – they can also be the head of the NP (see Guirardello, 1999: section 2.2.2). 1.3

Previous Studies on Trumai Demonstratives

The demonstrative system was previously described in Guirardello (1992: 72–76) and Guirardello (1999: 28–34) as a two-term system, based on distance 1

The data is represented using the orthography of the language (used in their local schools). The majority of the letters correspond to the IPA symbols, except for these cases: t (IPA: t̪ ); ţ (IPA: t); ‘ (IPA: ʔ); ch (IPA: ʃ); tl (IPA: tɬ); r (IPA: ɾ); y (IPA: j); ï (IPA: ɨ).

244

Raquel Guirardello-Damian

from the speaker: proximal or distal. It was observed that the forms were used pronominally, and besides distance, they would also distinguish gender and number. It was further mentioned that the masculine forms (ni’de, ka’ne) could also occur, modifying the head noun of the NP (1–2). (1)

ha pudits’ ni’de pike yi-ki 1 like this house YI-Dat ‘I like this house.’

(2)

ha pudits’ ni’de yi-ki 1 like this.one YI-Dat ‘I like this one.’

Tables 11.1 and 11.2 below present the demonstrative forms attested in the previous works. Apart from them, other forms were attested: ine, inatl, inak a/inak ana, inak wan. However, these forms were not analyzed as demonstrative, but rather as personal pronouns, meaning respectively: ‘he’, ‘she’, ‘they (dual)’, ‘they (plural)’. They were considered to be personal pronouns rather than demonstrative ones for two reasons:

Table 11.1 Demonstrative pronouns (previous studies)

Masc SG Fem SG Masc/Fem DU Masc/Fem PL

Proximal (close to speaker)

Distal (far from speaker)

ni’de ni’datl ni’dak a ni’dak ana (non-Abs) ni’dak wan

ka’ne2 ka’natl ka’nak a ka’nak ana (non-Abs) ka’nak wan

Table 11.2 Demonstrative modifiers (previous studies)

2

Proximal (close to speaker)

Distal (far from speaker)

ni’de

ka’ne

According to one consultant (an older man), nowadays people say ka’ne, ka’natl, ka’nak wan, but in old times the demonstrative pronouns were actually kande, kandati, kandak wan. He still prefers to use the older forms.

Trumai of Demonstratives

245

Table 11.3 Place adverbs and demonstratives (previous studies)

Place adverbs Demonstratives

Proximal

Medial

Distal

nina ni’de

ina

kaina ka’ne

(i) in the corpus at that time, there was no natural data in which a form beginning in in would be used as a demonstrative modifying a noun; (ii) when some consultants were asked if ine could be used to modify a noun, they did not accept it, or at least there was some controversy if it could be used. Nevertheless, in Guirardello (1999b) some questions were also raised. It was observed that the language has three place adverbs, apparently based on distance from the speaker: proximal, medial, distal (see Table 11.3). Compared to the place adverbs, the demonstrative set would have a gap. Thus, would it be the case that Trumai had a “medial” demonstrative too? Some possibilities were pointed out (Guirardello, 1999b: 33–34): We could speculate that the “missing” form in the chart would be the personal pronoun ine, which originally could have been a demonstrative pronoun rather than a personal one . . . However, if ine was originally a demonstrative pronoun, now it seems to be only a personal one, because its use as modifier of a head noun in an NP seems not to be possible, or at least it is controversial . . . The question is to know if it was originally the medial demonstrative pronoun (with the full demonstrative set being: ni’de – ine – ka’ne ‘this’ – ‘that’ – ‘yon’) or if Trumai has another form for the medial demonstrative pronoun.

Questions about the place adverb ina were also raised: although this form was analyzed as indicating medial distance from the speaker, there was one example that pointed toward another possible analysis: “in a letter to a friend, a Trumai used nina (here) when referring to her own village and ina (there) when referring to the friend’s village” (Guirardello, 1999b: 33): (3)

hamuna in hi wan chomta-tke: ina, nina ha wan hilakan-n? where Foc 2 PL play-Des there here 1 PL village-Loc ‘Where do you want to play soccer: there (friend’s village), or here in our village (speaker’s village)?’

The example suggested that ina might indicate proximity to addressee, but this was not totally transparent because the friend’s village had actually an ambiguous characteristic: it was the place where the addressee was, but it was also a place at a moderate distance from the village of the speaker. The example did not represent a clear situation, and therefore it could not help us to understand

246

Raquel Guirardello-Damian

the status of ina. There were other examples in the corpus that were equally ambiguous. Finally, two further forms were mentioned, which seemed to be demonstrative pronouns as well: ni and in. It was speculated that they could be neuter forms (as opposed to the masculine and feminine ones) and that they could be aligned with them, but again there would be a gap in the pattern:

Masc

Fem

Neuter

ni’de ine ka’ne

ni’datl inatl ka’natl

ni in

Thus, given this problem and other unanswered questions, a further investigation on the topic was necessary. 2

Elicitation with the Wilkins Questionnaire

A new elicitation was performed using the Wilkins (1999) questionnaire, and the scenes presented in the questionnaire were the basis for setting up real situations. For the elicitation, a concrete object was used as the element being referred to.3 The Trumai consultant was the speaker, and the researcher was the addressee. Their positions were changed according to the configuration of each scene. The elicitation was conducted in two occasions: in 1999, when the first results were obtained, and in 2000, when questions or possible hypotheses were verified. Four consultants participated in 1999: two of them were middleaged (a man and a woman), and two were in their twenties (two young women). In 2000, the data were rechecked with three of the consultants of the 1999 elicitation (the middle-aged man and the two young women), and further data were obtained from a new consultant (a middle-aged woman). All consultants were native speakers of Trumai. The consultants provided data with demonstrative forms as modifiers (Dem N) and as pronouns. Through the application of the questionnaire, new facts came up: A. It was observed that ni and in can occur not only as pronouns, but also as the modifier of a noun in an NP when the referent is an object: 3

With all consultants, the element being referred to was a comb. With one consultant, further elicitation was conducted, using a person (a man) and a dog as referents. The Trumai demonstratives were also investigated through further questions or observation of natural uses.

Trumai of Demonstratives (4)

ha pudits’ ni kuaw yi-ki 1 like this(close.spk) comb YI-Dat ‘I like this comb.’

(5)

ha pudits’ in kuaw yi-ki 1 like that(close.addr) comb YI-Dat ‘I like this comb.’

247

B. New demonstrative terms were identified: inde (indicating closeness to addressee) and ka’in (indicating distance from both speaker and addressee). (6)

inde tartaï yi-k hi ki hat’ke de that(close.addr) ant YI-Erg 2 sting in.future already ‘That ant will bite you.’ [the ant is in the addressee’s shoulder]

(7)

ha pudits’ ka’in kuaw yi-ki 1 like that(far.spk.addr) comb YI-Dat ‘I like that comb.’

These terms fill the gaps observed in the previous studies. Trumai does have a three-term demonstrative system. The full sets of demonstrative forms that occur as modifiers are: ni’de inde ka’ne ni in ka’in

C. Apparently ni’de and ni (and similarly, inde and in, ka’ne and ka’in) can freely substitute each other when referring to objects (i.e., inanimate entities). This occurs when they are used both as a modifier and as a pronoun (8–9)4. Both forms can be used with pointing gestures. (8)

a. ni’de kuaw ‘this comb’ b. ni kuaw ‘this comb’

(9)

a. ha pudits’ ni’de yi-ki 1 like this.one YI-Dat ‘I like this one.’ b. ha pudits’ ni yi-ki 1 like this.one YI-Dat ‘I like this one.’

4

According to an older speaker, if the referent is an object (i.e., an inanimate entity), the demonstrative form to be selected should be ni – in – ka’in. However, during the elicitation of the questionnaire, the speakers would freely alternate the forms (sometimes using ni – in – ka’in, sometimes ni’de – inde – ka’ne), saying that both uses would be acceptable.

248

Raquel Guirardello-Damian

D. The status of the demonstratives in, inde and the adverb ina were clarified. As explained in section 1.3, the main question was whether these forms refer to objects close to the addressee or rather to objects at a medial distance from the speaker (with no reference to the addressee). Scenes 16 and 18 of the Wilkins questionnaire were crucial in clarifying this issue, showing that what anchors the forms beginning in in is proximity to the addressee (Table 11.4 below). E. When is an object considered to be close to the addressee? Apparently, speakers may have different opinions about that, with older speakers being slightly stricter than the younger ones. The ideal situation to say that an object is close is when: (i) it is in contact with the body of the addressee, or (ii) it is within his/her arm’s reach, or (iii) it is in an area that belongs to the addressee – like scene 23 of the questionnaire, where the object is not in arm’s reach of the addressee but is in the entrance of his house. Otherwise, the distance becomes more susceptible to personal interpretations, which will then influence the speaker’s choice between inde (close to addressee) and ka’ne (far from speaker and addressee).5 Table 11.4 on the next page summarizes the results of the elicitation of Trumai demonstratives using the Wilkins questionnaire. The scenes are organized in groups, depending on the demonstrative form used for them. Under the table, we have lines that indicate the extensions of each form. Solid lines indicate that the use of the demonstrative in question was considered good for those particular scenes. Broken lines indicate that, for those scenes, the use of the demonstrative in question was considered (i) fine under certain circumstances, or (ii) possible, but not preferred. 3

Naturalistic/Observational Data

The results of the questionnaire were compared to naturalistic data, i.e., examples extracted from texts or from observation of daily life situations (the Trumai corpus is now bigger than it was at the time of the previous studies). In the naturalist data, the following patterns are observed: A. The form ni’de (and others whose initial syllable is ni) is consistently used to refer to entities close to the speaker. Ni’de can occur pronominally or it can be found modifying a noun in an NP. It can also be found functioning as a presentative (10). Finally, ni’de can be used to contrast situations (11): 5

For example, for scenes 16 and 18 of the questionnaire, if the addressee is standing close to the object being referred to, younger speakers may use inde, but older speakers will prefer ka’ne instead, because in their interpretation, the object is not clearly close to the addressee, since it is not in his/her arm’s reach.

(7)

(3)

A

S

ni'de / ni

S

(1)

S

S

(20)

A

(19)

S

A

S

A

(11)

S

A

ni'de

The form used is ni'de / ni Additional but pointing or description additional may also be description is used also necessary

(22)

A

(8)

(6)

A

ni'de / ni

S

S

(4)

(2) A

A

Otherwise: inde / in

ni'de / ni if speaker touches the object.

S

S

S

A

(10)

A

(9)

(5)

A

inde / in

S

S

A

A

inde

(23)

S

(18)

A

S

(16)

S

(12)

A

ka'ne / ka'in was the most frequent one

All possible answers occurred.

S

A

Possible (if distance is not so big): ni'de / ni inde / in (Y Sp)

Preferred: ka'ne / ka'in

(17)

S

(14)

S A

Table 11.4 Results of the elicitation conducted with the Wilkins (1999) Demonstrative Questionnaire

(24)

ka'ne

ka'ne / ka'in

A S

S

(21)

A

(13)

S A

(25)

but additional description is necessary

ka'ne / ka'in

A S

(15)

S A

250

Raquel Guirardello-Damian

(10) iyi pumu-n henikik dat-ea-tl, iyi pita-n hen: ni’de IYI enter-3Abs firstly home-3Poss-Dat IYI exit-3Abs then presentat. ‘She entered into her house, then she came back: “Here it is!” (she said).’ (11) fa maxke-s de aloke ha a hat’ke kill/hit Aux-Cond already fast 1 Dual in.future ni’de, ofa tak de ha a hat’ke, in.contrast kill/hit Neg already 1 Dual in.future aleitak de ha a hat’ke long already 1 Dual in.future ‘If we kill a lot (of fish), we will be back soon. In contrast, if we do not kill (anything), we will stay longer.’

B. The form ka’ne (and others whose initial syllable is ka) is consistently used to refer to entities far from the speaker and far from the addressee. Like ni’de, ka’ne can occur pronominally or can be found behaving as a modifier (Dem N). So far, there are no attested examples with ka’ne as a presentative, but this use is probably possible. C. With regard to inde, there are also some natural-speech examples with this form, occurring as a pronoun or a modifier (12). Inde is employed to refer to entities close to the addressee. It can also occur as a presentative (13). (12) inde kasoro yi ka_in pïx_tak that(close.addr) dog YI Foc/Tens small ‘That dog is small.’ (13) inde, Tata dacha-n that(close.addr) Tata back-Loc ‘There it is, behind Tata.’ [Tata was sitting on the floor near the addressee (i.e., in an area that belongs to the addressee). The speaker was far from Tata]

D. As for ine, the situation is slightly different: in the data obtained with the Wilkins questionnaire, ine was employed by one consultant (an older man) as a modifier when the referent was close to the addressee (i.e., instead of inde, this consultant systematically used ine). In natural speech, it is possible to find ine employed as a modifier, but it is not because the referent is close to the addressee. It is rather because ine is referring to an entity previously mentioned in discourse (i.e., it is an anaphoric element). For example: (14) esak-(k)i hen wan elka-n ale hammock-Dat then PL buy-3Abs hearsay ine esak-(k)i wan tsula-n ale men Dem(anaph) hammock-Dat PL lie.down-3Abs hearsay Frust ‘They bought a hammock. In this hammock, they tried (in vain) to lay down.’ (15) ine chahnini-s, faxdola pila-n Dem(anaph) night-Temp hear Aux-3Abs ‘In this night (a night previously mentioned), he heard (things).’

Trumai of Demonstratives

251

With regard to pronominal uses of ine in natural speech, we observe it in two situations: (i) when the speaker wants to point to the referred entity, instead of merely mentioning it (16b); (ii) as an anaphoric pronoun, referring to an entity mentioned earlier in the discourse. When this is the case, Trumai speakers often translate ine into Portuguese as ‘he’ (17). (16)

a. pu’dits-e hi-tl like-3Abs 2-Dat ‘He likes you.’

[just mentioning]

b. ine pu’dits hi-tl Dem(close.addr) like 2-Dat ‘That one (close to you) likes you.’ [pointing to the person] (17) chahnini-s ka_in laketsi ke oke_yar yi, night-Dat Foc/Tens promenade KE sorcerer YI hu’tsa tak hak yaw chï ine-tl see Neg purp human.being Cop Dem(anaph)-Dat ‘The sorcerer walks in the night, (so that) nobody sees him.’

The same occurs with the other pronominal forms in in (inatl, inak a/inak ana, inak wan), which are often translated into Portuguese as ‘she’, ‘they two’, ‘they (plural)’, respectively. For instance: (18) herohen ka_in ke anïr_kik yi beautiful Foc/Tens KE woman.from.woods YI ha ï ţ í inatl lots’. 1 fear Dem(anaph) Ablat yaw apane inatl yi-k Human.being chase Dem(anaph) YI-Erg ‘The woman from the woods (a kind of ghost) is beautiful. I am afraid of her. She chases people.’

Thus, comparing the various kinds of data, we can say that the results of the Wilkins questionnaire overall match the naturalistic/observational data with regard to the demonstratives ni’de – inde – ka’ne. With regard to ine, there is a slight mismatch. While the elicitation with the Wilkins questionnaire showed that ine can be employed as a modifier indicating closeness to the addressee, the naturalistic data showed that ine is actually used as an anaphoric modifier. However, it should also be pointed out that only one consultant used ine during the elicitation (an older speaker); all the other consultants preferred to use inde. Thus, it might be that ine and inde – which were possibly just variants of the same demonstrative – are now taking different roles: ine is becoming more specialized as an anaphoric modifier, while inde is becoming specialized as a modifier that indicates closeness to the addressee.6 6

So far, there are no examples of ni’de and ka’ne used as anaphoric modifiers in a text.

252

Raquel Guirardello-Damian

As for ine as a pronoun (and similarly, the other pronominal forms begining in in), the elicitation with the Wilkins questionnaire shows only part of what happens in the language. The questionnaire indicates that ine can be used to point to an entity that is close to the addressee but does not show that ine can also be used as an anaphoric element. The anaphoric use only becomes transparent when we analyze data from natural speech, especially narratives. 4

Contrastive Uses

The results of Wilkins’ Demonstrative Questionnaire also match what is observed in contrastive uses of the demonstratives. When Trumai speakers are asked to contrast two or three objects, the selection of the demonstrative form depends on the frame of reference, i.e., if the frame includes the speaker only or if it includes the addressee as well: (1) When only the speaker is included in the frame of reference, the demonstratives beginning in ni and ka are selected, but not the ones in in. For example, three sandals were on the floor (represented below by triangles). They were all located in relation to the speaker, and the addressee was not included in the scenario. When the Trumai speaker was asked “Which sandal do you want?”, the answers were: kaina chïk (lit.the one there)

ka'ne

ni Speaker

(2) When the addressee is also included in the scenario, the demonstratives in in are employed. For example, there were two sandals on the floor. One of them was close to the addressee (in arm’s reach). When the Trumai speaker was asked “Which sandal do you want?”, the answers were:

Trumai of Demonstratives

253

inde

Addr

ni'de Speaker

5

Trumai Demonstratives: A New Analysis

The new investigation revealed that the previous analysis needed to be revised, indicating that the Trumai demonstrative system is not a two-term, but rather a three-term system, and that the forms beginning in in (and the adverb ina as well) are addressee anchored. It also showed that ine, inatl, inak a/inak ana, inak wan are not personal but rather demonstrative pronouns, since they do have demonstrative roles. This analysis is reinforced by the fact that there are strong parallelisms between the in forms and the ni and ka forms with regard to gender and number distinction (see Table 11.5 below). Additional investigation revealed that when the demonstrative forms occur pronominally, there is a distinction between human, animate, and inanimate entities. It is interesting to note that such a distinction is also observed in other parts of the language: • the use of Dative markers, which clearly draws a distinction among nouns referring to humans, animates, and inanimates (see Guirardello, 1999: chapter 7); • the use of the Pluralizers a ‘Dual’ and wan ‘Plural’, which distinguishes between nouns that refer to animate entities (with a human and non-human subdivision) and inanimate entities (see Guirardello, 1999: chapter 2); Table 11.5 Demonstrative pronouns Close Sp Human Beings

Masc SG ni’de Fem SG ni’datl Masc/Fem DU ni’dak a ni’dak ana (non-Abs) Masc/Fem PL ni’dak wan Animals ni’de Objects (inanimate entities) ni ni’de

Close Addr

Distant Sp, Distant Addr

ine inatl inak a inak ana (non-Abs) inak wan ine in inde ~ ine

ka’ne ka’natl ka’nak a ka’nak ana (non-Abs) ka’nak wan ka’ne ka’in ka’ne

254

Raquel Guirardello-Damian

Table 11.6 Demonstrative modifiers

Human beings Animals Objects (inanimate entities)

Close Sp

Close Addr

Distant Sp, Distant Addr

ni’de

inde ~ ine

ka’ne

ni ni’de

in inde ~ ine

ka’in ka’ne

Table 11.7 Place adverbs Close Sp

Close Addr

Distant Sp, Distant Addr

nina7

ina

kaina

• the use of the Imperative particles wa versus waki, which also exhibit a contrast between animate and inanimate entities (see Guirardello, 1999: chapter 3); • the use of the positional verbs and auxiliaries, which again reflect a distinction between human, animates, and inanimates (see GuirardelloDamian, 2007: section 5.2). The revised set of Trumai demonstratives is presented in Table 11.5; these are the forms used pronominally. The spatial distinctions encoded are distance (close–far) and anchoring (speaker versus addressee). If the referent is a human being, there are different forms for masculine and feminine, singular and nonsingular. With regard to preemption, as we can see in scenes 2 and 4 of Table 11.4 above, the selection of ni’de and ni is forced under conditions of touching; i.e., when the speaker touches the referent, the use of ni’de is favored, even if the referent is very close to the addressee. The forms that can occur as the modifier of a noun are presented in Table 11.6. The use of demonstratives as modifiers is favored when the referent is an object, but it would be disfavored when the referent is a human being or animal; in this case, the pronominal use would be preferred – e.g., speakers prefer to say ine ‘that one (close addr)’ instead of ine axos ‘that boy (close addr)’. Table 11.7 presents the revised analysis for place adverbs.

7

Sometimes ni can be used instead of nina to indicate place (‘here’). This use is still not well understood. For example: (1) ni-a hi chï? 2-Quest 2 Cop ‘Are you here?’

Trumai of Demonstratives

255

As a final point, it should be mentioned that there are situations in which the use of demonstratives is not considered appropriate: it is when the addressee does not have his/her attention on the object being referred to. In this case, the consultants consider that it is better to have a descriptive NP, instead of using a demonstrative term, because this avoids misunderstandings. For example: (19) a. ha pudits’ ni’de kuaw yi-ki 1 like this comb YI-Dat ‘I like this comb.’ b. ha pudits’ kuaw yi-ki, ha heni-n chï-k 1 like comb YI-Dat 1 side-Loc Cop-Nzr ‘I like the comb, the one by my side.’ [this would be clearer to the addressee]

Such use also occurs in natural data. In texts, we can observe the use of descriptive NPs to refer to people or things: (20) Kamayula, ma tak wan-e: pïţík, mot’e, puk Kamayurá eat Neg PL-3Abs monkey jacu curasaw nina chï-k, a’dï wan ma-n kodetl-es here Cop-Nzr many PL Eat-3Abs animal-Dat ‘The Kamayura people, they do not eat: monkey, jacú (a kind of bird), curasaw (a bird). The ones here (i.e. the Trumai people), they eat many (kinds of) animals.’

6

Conclusion

The investigation performed with the Wilkins questionnaire helped to elucidate several issues and provided a better picture about the language. In terms of typology, demonstrative systems can be classified as distance-based or personbased (Anderson & Keenan, 1985; Diessel, 1999). After this new study, we can say that Trumai is not a representative of the first, but rather of the second type. The demonstratives of this language are sensitive to the position of not only the speaker but also the addressee. In some ways, Trumai aligns with spoken Brazilian Portuguese, which has a person-oriented system as well (especially if we take into account the combination of demonstratives + place adverbs: esse aqui ‘close to speaker’, esse aí, ‘close to addressee’; see Meira and Guirardello-Damian, this volume). However, Trumai seems to be more selective about what counts as being close to the addressee. If the referred object is on the floor in front of the addressee, a Brazilian Portuguese speaker could classify it as proximal, even if the addressee is standing and cannot quite touch the object. For a Trumai speaker, this would represent a different situation – contact with the addressee’s body and reachability of the referent are important factors to be taken into

256

Raquel Guirardello-Damian

consideration. Older speakers follow this principle more strictly, while younger speakers seem to be more flexible about it (see footnote 5). This might be an influence from Portuguese, given that the younger individuals are fully bilingual. The study also allowed us to have a more complete idea about the set of demonstrative pronouns, with the distinction between humans, animals, and inanimates. The inanimate set was probably just ni – in – ka’in originally, having become more diverse with the spread of ni’de – inde – ka’ne to it. It might be that this spreading may eventually cause the inanimate versus animal distinction to disappear, with only the human versus non-human distinction remaining. Again, one may wonder if Portuguese may have had an influence on this, given that Portuguese demonstratives present a distinction of gender and number, but not of animacy (for example, the Portuguese demonstrative aquele ‘that(masc)’ can be used for either a man, a dog, or a comb). This is a topic for future exploration. References Anderson, S. & Keenan, E. (1985). Deixis. In T. Shopen, ed., Language typology and syntactic description. Vol. III. Cambridge: Cambridge University Press, pp. 259–308. Derbyshire, D. C. & Pullum, G. K., eds. (1986). Handbook of Amazonian languages. Vol. I. New York: Mouton de Gruyter. Diessel, H. (1999). Demonstratives: Form, function, and grammaticalization. Typological Studies in Language, vol. 42. Amsterdam: John Benjamins. Guirardello, R. (1992). Aspectos da morfossintaxe da língua Trumai e de seu sistema de marcação de caso. Master’s thesis, State University of Campinas (Unicamp), Brazil. (1999). A reference grammar of Trumai. Doctoral thesis, Linguistics Department, Rice University, Houston, TX. (2007). Locative construction and positionals in Trumai. Linguistics, 45(5/6), 917–954. Rodrigues, A. D. (1986). Línguas Brasileiras: Para o conhesimento das línguas indígenas brasileiras. São Paulo: Edições Loyola. Senft, G. & Smits, R., eds., (2001). Annual report 2000. Nijmegen: Max Planck Institute for Psycholinguistics. Wilkins, D. P. (1999; this volume). The 1999 demonstrative questionnaire: “This” and “that” in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24.

12

Saliba-Logea: Exophoric Demonstratives Anna Margetts1

1

Introduction

Saliba-Logea is a Western Oceanic language spoken in Milne Bay Province, Papua New Guinea. It belongs to the Suauic family of the Nuclear Papuan Tip Cluster (Cooper, 1975; Ross, 1988: 190ff.). The language is spoken by approximately 2,500 people on Saliba Island and Logea Island, adjacent islands and parts of the mainland. The differences between the dialects lie mostly in the lexicon (for a discussion of emerging grammatical differences in number marking, see Margetts, 2010). English is the modern lingua franca of Milne Bay Province – rather than Pidgin as in most parts of Papua New Guinea. Its influence on the local languages is strong and is reflected in extensive borrowing. Saliba-Logea is still acquired as a first language on both islands, but education from Grade 1 onwards is entirely in English. Young people tend to have around six years of schooling and are mostly fluent in Milne Bay English. Many older people read Suau, a local mission language which has a bible translation. People also readily read Saliba-Logea if presented with it, though the language has no written tradition. The language shows a three-way deictic distinction between ‘close to speaker’, ‘close to addressee’ and ‘distal’ across several form classes of demonstratives and place adverbs. The speaker-based forms are used to denote objects close to the speaker, prototypically within arm’s reach. The addresseebased forms are used to denote objects which are closer to addressee than to speaker, while the distal forms denote referents which are close to neither the speaker nor the addressee. Table 12.1 summarizes the deictic distinctions and shows the associated forms. The data presented here draws on elicitations with the Wilkins Demonstrative Questionnaire (1999a; 1999b; this volume) as well as on 1

As always my sincere thanks go to the Saliba and Logea communities who have supported me in learning about their language. The fieldwork on which this study is based was made possible through generous funding by the Max Planck Institute for Psycholinguistics and the DoBeS Program of the Volkswagen Foundation.

257

258

Anna Margetts

Table 12.1 Three-way deictic contrast across classes of spatially deictic terms NEAR SPEAKER

NEAR ADDRESSEE

DISTAL

objects close to SPKR typically within arm’s reach ina, teina, =te, te, (te-)inai

objects close to ADDR and closer to ADDR than to SPKR meta, temeta, =me, me, (te-)metai

objects close to neither SPKR nor ADDR nem, tenem, =ne, ne, (te-)nemai/(te-)menai

observations from natural interactions and on a text corpus collected between 1995 and 2008. Section 2 of this chapter provides a discussion of previous research on spatial deictics in Saliba-Logea and closely related languages, and section 3 an overview of the relevant form classes. In sections 4 and 5 I then discuss the methodology of working with the Wilkins questionnaire (1999a) as a research tool, the findings from the questionnaire, as well as some observational data. Section 6 provides a discussion and conclusion. 2

Previous Research

The main studies on Saliba-Logea spatial deictics to date are Margetts (2004b) and Dawuda (2009). The Saliba sketch grammar by Mosel (1994) makes only cursory reference to demonstratives and place adverbs. The grammatical sketch in Margetts (1999: 29) discusses some of the relevant forms but does not consistently identify the semantic distinctions expressed.2 Cleary-Kemp (2006) provides an analysis of the clitics =ne and =wa which form a paradigm with the adnominal demonstrative clitics. Section 3.3 reviews some of her findings. Margetts (2004b) presents a first analysis of the Saliba-Logea demonstratives. The study is based on the Saliba dialect and includes some data elicited with Wilkins’ two demonstrative questionnaires (1999a; 1999b; this volume). The focus of this earlier article lies on the different form classes rather than the semantic distinctions discussed here. It is based on elicitations with a smaller number of speakers and does not draw on extensive text data. Dawuda (2009) presents a detailed study of discourse functions of demonstratives and place adverbs with exophoric reference in the Logea dialect. The functions discussed include providing additional information (e.g. illustrating and explaining uses of demonstratives) but also the maintenance of discourse topics, i.e. introducing, tacking and dismissing topical referents. 2

The sketch lacks any discussion of the clause-final demonstrative; it discusses the adnominal clitics but fails to provide accurate glosses for some of the forms. The description of the free demonstratives suffers from similar problems; teina is correctly labelled as proximal but temeta and tenem are both preliminarily glossed as ‘distal’.

Saliba-Logea Demonstratives

259

Table 12.2 Buhutu spatial deictics (adapted from Cooper, 1992: 96) ite buka te buka mate luma ni luma au

‘this book’ (close to speaker) ‘that book’ (close to hearer) ‘that house’ (away from speaker and hearer) ‘that house’ (far away)

A major finding of this study is that exophoric demonstratives can perform such tracking functions which are typically discussed only for endophoric devices. There is some work on languages closely related to Saliba-Logea which is worth mentioning here. Cooper (1992) discusses discourse functions of the addressee-based demonstrative in Buhutu, another language of the Suauic dialect chain (Ross, 1988). He presents a paradigm of demonstrative forms (given in Table 12.2), which shows both similarities to and differences from the Saliba-Logea paradigms. For Buhutu, Cooper describes only the category of ‘close to speaker’ as involving two elements which bracket the noun. When functioning as adnominal modifiers, one speaker-based demonstrative precedes the noun while another follows. For the addressee-based and the distal categories, only a post-nominal demonstrative is listed. Of interest is also the Buhutu distinction between two distal categories ‘away from speaker and hearer’ versus ‘far away’, as described by Cooper. Saliba-Logea has a corresponding distinction in only one of the paradigms of spatial deictics (see section 3.3). It remains unclear from Cooper’s discussion whether Buhutu has only a single demonstrative form class or whether other paradigms which might be described as demonstratives are simply not listed, since Cooper is focusing on the discourse function of a particular form. Finally, a study by Peteliyaki and Lithgow (not dated) discusses the use of the determiners -ne and -wa in Auhelawa and a number of other Papuan Tip languages (besides Auhelawa the study presents data from Suau, Tubetube, Bunama and Iduna). These two forms are cognate to the Saliba-Logea adnominal clitics =ne and =wa (see section 3.3 below). The functions of these suffixes, as outlined by Peteliaki and Lithgow, are very similar to those of the Saliba-Logea forms: -wa indicates something previously mentioned, an object no longer in use, . . . -ne refers to a specific one, a particular one, . . . (Peteliaki and Lithgow, not dated)

The authors make no reference to other forms in a paradigmatic relation with -wa and -ne for any of the languages they discuss, while the corresponding

260

Anna Margetts

Table 12.3 Three-way contrast across classes of spatial deictics

clause-final DEM free DEM adnominal DEM interjections full place adverbs

NEAR SPEAKER

NEAR ADDRESSEE

DISTAL

ina te-ina =te* te (te-)ina-i

meta te-meta =me me (te-)meta-i

nem te-nem =ne ne (te-)mena-i (te-)nema-i**

* For some speakers, the proximal clitic =te is in free variation with the form =ta which may be a loan from the mission language Suau. ** The Saliba form of the place adverb shows metathesis, while the form in the Logea dialect does not.

Saliba-Logea forms =wa and =ne are in a paradigmatic relation with the adnominal demonstrative clitics. 3

Description of Form Classes

There are five classes of spatial deictics in Saliba-Logea. Two of them function as place adverbs, the other three function as demonstratives pronouns and adnominal demonstratives. Each of these five form classes makes a three-way deictic distinction between ‘near speaker’, ‘near addressee’ and ‘distal’. Table 12.3 presents an overview. The functional differences between these form classes are discussed in Margetts (2004b) and are presented here in a summarized and updated form. 3.1

Clause-final Demonstratives

The clause-final demonstratives consist of the base form of spatial deictics: ina ‘near speaker’, meta ‘near addressee’ and nem ‘distal’. Both the free demonstratives and the full place adverbs are derived from these base forms. Clausefinal demonstratives may not stand alone as a one-word utterance (e.g. as the answer to questions such as ‘which one?’ or ‘where?’). They occur as the last element of a clause and generally seem to function as demonstrative pro-forms, as in (1) to (5). (1)

Teina ya-kasiyebwa ina NEAR.SPKR 1SG-sick NEAR.SPKR ‘I’m being sick here (I can’t go and fetch water).’ (Gesila01_BC_0226)

Saliba-Logea Demonstratives

261

(2)

Kaiteya yona tobwa ina? who 3SG.POSS bag NEAR.SPKR ‘Whose bag is this (close to me)?’ (DemQuest)

(3)

Sinebada yona tobwa meta old.lady 3SG.POSS bag NEAR.ADDR ‘That’s the old lady’s bag (close to you).’ (DemQuest)

(4)

Kaiteya natu-na nem? who child-3SG.POSS DIST ‘Whose child is that?’ (Bagi_02EO_0107)

(5)

Ka-duwui-kasayai nem 1EXCL-dive.for-in.vain DIST ‘We dived for him there to no avail (he had drowned).’ (Conversation_01AN_0057)

The difference in use between the clause-final and the free demonstratives (section 3.2) still deserves further study. Both can function as demonstrative pro-forms, but they are not always interchangeable. While the free demonstratives generally refer to nominal entities, the clause-final forms – at least at times – appear to make reference to whole situations, activities or events. Consider the minimal pair in (6). Both utterances were set in a context where the addressee is carrying a palm leaf back to the village which is needed for the preparation of a feast. (6) a. Nabada meta! enough NEAR.ADDR ‘That’s enough!’ (You’ve cut enough leaves, you can stop. We’ll use that one but we don’t need any more leaves than that) b. Temeta nabada NEAR.ADDR enough ‘That’s enough!’ (Never mind that leaf. We won’t bother with that one.)

The addressee-based form of this paradigm, meta, has developed into a topic markers, as discussed by Dawuda (2009: 52–55). (7)

Hewali=wa meta hasala=wa i=tawasola-ya-ko young.man=ANA TOPIC young.woman=ANA 3SG-marry-3SG-already ‘The young man had already married the girl.’ (Adapted from Mosel, 1994: 19)

The same extension of the addressee-based demonstrative to topic marker has also been observed for a cognate morpheme in the related language Buhutu (Cooper, 1992). See also François (2005: 142–143) on similar extension in Mwotlap, an Oceanic language of Vanuatu.

262

Anna Margetts

3.2

Free Demonstratives

The free demonstratives are composed of the base forms ina ‘near speaker’, meta ‘near addressee’ and nem ‘distal’ plus a prefix te-. They can function as demonstrative pronouns or adnominal modifiers. As pronouns, functioning as the head of a noun phrase, they may optionally appear with the adnominal clitics, generally with the matching spatial distinction (but see section 3.3 for the discussion of the distal form =ne). Consider the examples in (8): (8)

a. teina=te b. temeta=me c. tenem=ne NEAR.SPKR=NEAR.SPKR NEAR.SPKR=NEAR.SPKR DIST=DIST ‘this one (here)’ ‘that one (near you)’ ‘that one (there)’

As modifiers, the free demonstratives precede the head noun, and they differ in this respect from other nominal modifiers which all follow the head (lexical modifiers, numerals and quantifiers, as well as the adnominal demonstrative clitics). When they function as adnominal modifiers, the free demonstratives typically combine with the adnominal clitics, forming a bracket around the noun, as in (9) to (14). (9)

Siya teina bosa=ta nige kabi-na se-kata 3PL NEAR.SPKR basket=NEAR.SPKR NEG nature-3SG.POSS 3PL-know ‘They don’t know this type of basket.’ (Gulewa_01AH_0206)

(10) Teina

isimulita=ta nige daam ka-gina-ginauli NEAR.SPKR new.generation= NEAR.SPKR NEG mat.type 1EXCL-RED-make ‘(We,) this new generation here, we don’t make daam mats.’ (Daam_01AH_0291)

(11) Yau

bena se-hai-gau ya-saema 1SG OBLI/WANT 3PL-get-1SG 1SG-come.up temeta soi=me unai NEAR.ADDR feast=NEAR.ADDR PP.SG ‘They wanted to get me to join one of those feasts.’ (Giyahi_01AA_0499)

(12) Temeta

woyawa=me kabo ye-henaku-go NEAR.ADDR feast=NEAR.ADDR TAM 3SG-chase-2SG ‘That scary thing might chase you.’ (Bagodu_01AH_0108)

(13) Tenem

taubada=ne iya ena aroli labui hinage DIST man=DIST 3SG 3SG.POSS story two also ‘That man, there are also two stories about him.’ (Boneyawa_28DZ_0068)

(14) Meta

tenem kaiwa=ne unai meta kumakala ye-pesama TOPIC DIST tree-DIST PP.SG TOPIC lizard 3SG-come.down ‘From that tree, a lizard came down.’ (TwoHeads_01AR_0103-0104)

There are some rare examples where the free demonstratives occur as adnominal modifiers without a following demonstrative clitic on the noun, as in (15) and (16).

Saliba-Logea Demonstratives

263

(15) Teina tautau hesau meta wawayao hayona tautau-di NEAR.SPKR picture other TOPIC children three picture-3P ‘This other film is about three children.’ (Abs-Rel2_02DO_0001-2) (16) Teina sinebada yona taubada, Samarai unai ye-laoma,. . . NEAR.SPKR old.woman 3SG.POSS old.man Place.Nam PP.SG 3SG-come ‘This woman’s husband, when he comes back from Samarai . . . ’ (Conversation)

Such cases also occur with the particle doha ‘like’ as in (17) and (18). (17) Doha teina weku like NEAR.SPKR stone ‘Like this stone here.’ (BudoiNualele_01CY_0402) (18) Doha tenem kabikabi-na like DIST style-3SG.POSS ‘Like that style.’ (Beyabeyana_02CZ_0066)

There is no separate class of manner demonstratives (see Diessel, 1999; Dixon, 2003; König, 2013), but doha ‘like’ in combination with a free demonstrative may be used to express this function, as in (19) to (25). (19) se-bodi-gogoi doha teina 3PL-sew-together like NEAR-SPKR ‘They sew it together like this.’ (Kemuluwa_01AS_0099) (20) Hekana yona gadosisi doha temeta some 3SG.POSS love like NEAR.ADDR ‘Some people’s wants are like that (i.e. what you just said).’ (Giyahi_01AA_0788)

The particle doha ‘like’ more commonly precedes the demonstrative but may also follow it, as in (21) to (23). (21) Tabu teina doha u-gina-ginauli don’t NEAR-SPKR like 2SG-RED-do ‘Don’t behave like this.’ (Bagi_02EO_0324) (22) Temeta doha NEAR.ADDR like ‘Like that.’ (Beyabeyana_01AH_0136) (23) lya ede tenem doha 3SG PRSUP DIST like ‘That’s it, like that.’ (Tawasola_02CL_0050)

Sometimes the demonstrative is preceded by doha ‘like’ and followed by a second occurrence of doha or by dohagi ‘how/like what’ as in (24) and (25): (24) Kita eda laulau doha tem doha 1INCL 3INCL.POSS way/style like DIST like ‘That’s our way.’ (Tawasola_02CL_0039)

264

Anna Margetts

(25) Edi sunuma doha tem dohagi 3PL.POSS faith like DIST how ‘Their belief was like that.’ (CanoeBuilding_01BC_0048)

3.3

Adnominal Demonstrative Clitics

The adnominal clitics function as demonstrative modifiers. They are unstressed and follow the head noun. As mentioned, they can combine with the free demonstratives as in (9) to (14) above, but they commonly occur without the free forms, as in (26) to (30): (26) Laugagayo=ta doha begdi doha kabi-di ka-kata law=NEAR.SPKR like little like nature-3PL.POS 1EXCL-know ‘We know the customs a little.’ (HairCutting_03BC_0010) (27) Kawa-gu=te ye-kamkamna tooth-1SG.POSS=NEAR.SPKR 3SG-hurt ‘My tooth here hurts.’ (DemQuest) (28) Taba baela=me hesau ku-hai-ya-ma if banana=NEAR.ADDR other 2SG-get-3SG-hither ‘If you could give me one of those bananas.’ (Boneyawa_23DZ_0211_) (29) Tobwa=me kaiteya ye-halusi? bag=NEAR.ADDR who 3SG-weave ‘Who wove that bag (near you)?’ (DemQuest) (30) Ne! kuduli=ne hesau DIST hill=DIST other ‘There! The hill over there!’ (Taukulupokapoka_02AC_0005)

What I have termed the distal clitic =ne does not have the same deictic restrictions as the speaker-based and the addressee-based forms. It is often translated into English as a definite article, and while the proximal forms =te ‘near speaker’ and =me ‘near addressee’ seem to only co-occur with their spatially appropriate counterparts of the free demonstratives, =ne is not restricted in this way and may occur with all three of the free forms. (31) Meta teina beya-na=ne se-ulapwasi TOPIC NEAR.SPKR ear-3SG.POSS=DIST 3PL-burst ‘So (these) his ears burst.’ [Speaker touching her ears] (BudoiNualele_02CY_111) bosa=ne ye-miya? (32) Temeta NEAR.ADDR basket=DIST 3SG-stay ‘Does a basket such as this (the one the addressee is making) last long?’ (BasketWeaving_03CW_0015)

Saliba-Logea Demonstratives

265

Furthermore, =ne may occur with nouns referring to the speaker’s or the addressee’s body parts. In these cases the referents are clearly within the spatial domain associated with the speaker- and addressee-based proximal forms. (33) Malatomtom ka

ya-hedede kulu-gu=ne se-yali morning TAM 1SG-tell head-1SG.POSS=DIST 3PL-shave ‘In the morming I’ll tell them to shave my head.’ (Conversation_06BO_0100)

(34) Nima-m=ne

ku-hai-gabae! hand-2SG=DIST/DEF 2SG-take-away ‘Take your hand away.’ (FirstBoat_01CK_0042)

There are two possible explanations for this behaviour, one is that =ne is an inherently distal form which has lost some of its deictic selection restrictions and is grammaticalizing towards a definite article, which can be used independently of the location of the referent. The second explanation is that =ne is in fact a distance-neutral demonstrative which acquires a distal interpretation through contrast with the specialized speaker-based and addressee-based proximal terms which, as a default, leave only the non-speaker, non-addressee and therefore distal space as a felicitous interpretation of its use. These possibilities require further study in order to ascertain whether the distal reading is an inherent semantic specification or one acquired pragmatically through contrast. Besides the three forms =te, =me and =ne, the paradigm of adnominal modifiers has the clitic =wa as a fourth member. This form is not spatially deictic but marks a noun as given information, as in (35) to (37). (35) Taubada=wa

haedi? old.man=ANA where ‘Where is the old man (you know who I mean)?’

(36) Waila=wa

kalo-na=wa unai se-talu water=ANA inside-3SG.POSS=ANA PP.SG 3PL-land ‘They landed in the water.’ (FrogStory_02AZ_0113-0115)

(37) Nige

baela=wa ku-tole NEG banana=ANA 2SG-put ‘You didn’t put the bananas down.’ (Boneyawa_19DL_0034)

There are no forms in the other paradigms of spatial deictics corresponding to =wa, i.e. there is no corresponding free demonstrative or place adverb. Also, unlike the other adnominal clitics, =wa does not combine with the free demonstratives. The combinations in (38) and (39) are not attested. (38) ?Teina=wa NEAR.SPKR=ANA ‘This one (given).’

266

Anna Margetts

Table 12.4 Distribution of =ne and =wa according to discourse mode =ne Narrative Non-narrative TOTAL

10 86 96

=wa 10% 90% 100%

258 53 311

83% 17% 100%

(39) ?Teina baela=wa NEAR.SPKR banana=ANA ‘This banana (given).’

The clitic =wa functions as an anaphoric marker (as defined by Fox, 1984) which traces the continuity of identity of discourse participants. Anaphoric =wa is in complementary distribution with the distal clitic =ne. Both mark identifiable referents and are commonly translated as definite articles. In a study comparing the functions of these two forms, Cleary-Kemp (2006: 70–72) found that =wa typically occurs in narrative clauses tracking identifiable referents. In contrast, =ne typically occurs in non-narrative clauses (e.g. in procedural texts or conversations, and in direct speech within narratives). Table 12.4 shows the distribution of the two forms by clause type (adapted from Cleary-Kemp, 2006: 48). The majority of referents marked by =wa across Cleary-Kemp’s corpus of 14 texts are explicitly evoked or inferable from the preceding discourse, whereas referents marked by =ne are typically identifiable via the immediate or wider extra-linguistic situation (2006: 72). Accordingly, following her analysis =wa marks identifiability, while =ne indicates definiteness. Within narratives, =ne mostly occurs in direct speech, as in (40) and (41), or with entities which are considered identifiable even at their first mention, as in (42) and (43). (40) “Kabo sinebada=ne ta-unui na then woman=DEF 1INCL-kill CONJ kabo boga-na=ne ta-nigwa-i na ” then stomach-3SG.P=DEF 1INCL-cut-TR CONJ ‘“We will kill that woman and cut open her stomach”.’ (WekuSinibu_01AC_0188) (41) Beala=ne ku-hai-ya-ma to natu-gu=ne ye-kai. banana=DIST 2SG-get-3SG-hither CONJ child-1SG.POSS=DIST 3SG-eat ‘Give me the banana for my child to eat.’ (Boneyawa_17AH_0015-16) (42) Papua Niu Gini=ne PNG=DEF ‘Papua New Guinea [first mention] . . . ’ (Bagodu_01AH_0001)

Saliba-Logea Demonstratives

267

(43) Yaumai=ne ye-lotalu. wind=DEF 3SG-blow ‘The wind [first mention] was blowing.’ (Bagodu_01AH_0065)

The clitics =ne and =wa are both attested to mark entire clauses as given. Such constructions are often translated as subordinate clauses into English. There is no formal indication of a hierarchical relationship, but the clause followed by the clitic is backgrounded by being marked as presupposed and given information. Similar constructions have also been described in Mwotlap by François (2005: 143). (44) Sada ye-mwalae=wa nige gonowana betelnut 3SG.SUBJ-climb=ANA NEG able ‘He tried to climb the betelnut palm but couldn’t do it.’ ‘Having tried to climb the betelnut palm he couldn’t do it.’ (Boneyawa_10BD_0009) (45) Se-tolo=wa pologi=wa ye-wasabu-ko 3PL.SUBJ-get.up=ANA frog-ANA 3SG.SUBJ-go.away-already When they woke up, the frog had already jumped out and disappeared.’ (FrogStory_01AW_0017) (46) Air ye-gehe=ne meta kabo . . . tem dohagidi air 3SG-finished=DIST TOPIC TAM DIST like/how ‘When the air is finished . . . we hold it (the hand) like that.’ (Diving_01DP_0136 -139)

ta-tole 1INCL-put

Both clitics are also attested to form brackets around a clause which follows a noun. The reading can be that of a relative clause modifying the preceding noun, as in (47) and (48). (47) Na ede yama=ne se-kai-gwali=ne ku=dabu-dabu? CONJ PRSUP fish=DIST 3PL-spear=DIST 2SG-RED-abstain ‘So you don’t eat the fish that they catch? . . . ’ (Tautolowaiya_01AG) (48) Nigwa=wa unai boxi ya-soke=wa haedi? knife=ANA PP.SG box 1SG-open=ANA where ‘Where is the knife I opened the box with?’

3.4

Interjections

The interjections te ‘near speaker’, me ‘near addressee’ and ne ‘distal’ are homophonous with the adnominal clitics but differ in that they bear stress and can occur as one-word utterances. They function as place adverbs and may occur as single-word answers to questions and as attention-drawing exclamations, as in (49) to (54).

268

Anna Margetts

(49) Q: Leah ede? A: Te! Name PRSUP NEAR.SPKR ‘Where is Leah?’ ‘Here, she’s with me’ (Conversation) (50) Yau ede te 1SG PRSUP NEAR.SPKR ‘I’m here.’ (Conversation) (51) Kami gulewa ede me 1EXCL.POSS clay.pot PRSUP NEAR.ADDR ‘Your clay pot is there.’ (Tautolowaiya_01AG_0093) (52) Me! Mwata=me! NEAR.ADDR snake=NEAR.ADDR ‘There! The snake (near you)!’ (Bagi_02EO_0418) (53) lya ede ne! 3SG PRSUP DIST ‘There it is!’ (KidsAndRecorder_04BO_0021) (54) Ne! Wale sinebada=ne DIST Wari woman=DIST ‘There! The lady from Wari Island!’ (Gulewa_01AH_0026)

As place adverbs they generally occur with non-verbal predicates. Following verbal predicates, as in (55) and (56), they are only rarely attested. (55) Ka-dobi te Hibokote 1EXCL-go.down NEAR-SPKR Place.Name ‘We went down to Hibokote here.’ (Sipoma_01AU_0122) (56) Kagu sada=ne kesega=mo na meta ko-kai me 1SG.POSS betelnut=DIST one=only CONJ TOPIC 2SG-eat NEAR.ADDR ‘I had only one betelnut – and you ate it!’ (Boneyawa_07BC_0031-32)

A fourth member of this paradigm is the interjection uuu or auu (with a lengthened vowel and change to falsetto voice), which denotes a remote location, as in (57) to (59).3 (57) Se-sae uuu Kulugwagwana 3PL-go.up REMOTE Place.Name ‘They went up there, to Kulugwagwana.’ (Kawakipakipa_01AT_0034-35) (58) Na ye-sikwa-hepito ye-sae=wa auu kewa-na=wa unai CONJ 3SG-poke-bounce 3SG-go.up=ANA REMOTE top-3SG=ANA PP.SG ‘And having poked it, making it bounce, it landed there on the top.’ (Mouse6_05 BQ_0028) 3

It is possible that there is a corresponding remote form =uuu/ =auu in the paradigm of adnominal demonstrative clitics discussed in section 3.3 (parallel to au in Buhutu, as discussed by Cooper, 1992). I believe examples such as numa=auu ‘the house yonder’ may occur; however, the database does not include any clear instances.

Saliba-Logea Demonstratives

269

(59) lya ede uuu waga bagu-bagu-na=ne 3SG PRSUP REMOTE boat RED-first-3SG.POSS=DIST ‘She is there yonder, in the first boat.’ (Legusisala_01AS_0534)

Remote spatial deictics with long vowels (and typically with falsetto voice) are also attested in other Milne Bay languages, including au in Buhutu as described by Cooper (1992) (cf. Table 12.2 in section 2 above) and the remote distal forms beyuuu and -neee from Kilivila, as described by Senft (2004: 64). 3.5

Full Place Adverbs

The full place adverbs are composed of the base forms ina ‘near speaker’, meta ‘near addressee’ or nem ‘distal’ plus the locative suffix -(a)i. (In the Saliba dialect, the distal form undergoes metathesis and occurs as men-ai.) Consider the examples in (60) to (63): (60) Inai ye-miya here 3SG-stay ‘He stayed here.’ (FamilyOrigin_07 CM_0120) (61) Tabu metai ku-lao-laoma! don’t there.NEAR.ADDR 2SG-RED-come ‘Don’t come that way.’ (Boneyawa_21DL_0029) (62) Nemai ye-miya there.DIST 3SG-stay ‘He stayed there.’ (Ulawa_01DG_0249) (63) Menai se-keno there.DIST 3SG-sleep ‘They slept there.’ (Torres_01AC_0051)

For emphasis the place adverbs may combine with the same prefix te- which also occur as part of the free demonstratives (section 3.2). (64) Teinai pasa ye-tole-di here flowers 3SG-put-3PL ‘He put flowers here.’ (Bagodu_01AH_0072) (65) Temetai kwa-hekaiyawasi there.NEAR.ADDR 2PL-stop ‘Stop it there!’ (Giyahi_01AA_0482) (66) Temenai ya-katuni=ne there.DIST 1SG-catch=DIST ‘There’s where I caught it.’ (Conversation_01AN_0133)

270

Anna Margetts

4

Applying the Demonstrative Questionnaire

4.1

Elicitation Procedure

I conducted elicitations based on the Wilkins questionnaire during two field trips in 1999 and 2000 with seven speakers, six of them women in an age range from their late teens to early forties and one male speaker in his late teens. With three speakers the elicitation was thorough and I not only recorded spontaneous responses to the scenes but also tested the acceptability of other demonstratives for the same scene.4 With the four remaining speakers, I recorded first responses and mainly engaged in discussions when these deviated from the previous results. All seven speakers were fluent in Milne Bay English, but the elicitations were conducted basically monolingually in Saliba.5 The sessions were both videorecorded and transcribed in a notebook on the spot. The scenes in the questionnaire were recreated with real objects, positioning them in the place and at the relative distances shown in the questionnaire’s schematic images.6 In most cases it was possible to recreate scenes very closely to how they are depicted in the questionnaire, including scenes involving confined spaces (e.g. conversations through a window or door) and scenes where the referent object is located in some hills across a body of water. In all cases the Saliba speaker was assigned the role of speaker while I took the role of addressee.7 Several methodological issues arose in the course of the first elicitations. First, assumed possession or control over the referent object tends to trigger the use of possessive classifiers rather than demonstratives, and location of the referent near a participant was generally sufficient as a trigger (see the discussion in section 4.2.4). So utterances with meanings like ‘Where did you get that bag?’ or ‘That bag is nice’ would literally be rendered as ‘Where did you get your bag?’ and ‘Your bag is nice’. To prompt responses with demonstratives and avoid possessive constructions, I generally elicited (among others) the 4

5

6

7

At the time of working with the demonstrative questionnaire, I was not aware that the remote interjection auu is in a paradigmatic relation to the demonstrative interjections te, me and ne and therefore I did not test for the acceptability of the remote interjection in comparison to the distal form ne. My impression is that a subset of the scenes which allow for the use of the distal form would allow the remote form as an alternative. In explaining the purpose of the task, I generally made reference to the English demonstratives this and that and showed the problems which arise when translating the English terms into Saliba and vice versa, i.e. translating from a two-term system to a three-term system. This helped to ground the task and to explain my interest in spatial deictics. The referent object was generally a bag or bucket and for the extremely distal scenes a house or garden. For scenes which involve a small object located on the speaker or addressee, the referent was an imaginary gecko. I also tried the reverse assignment of roles, which proved not to be a good set-up since it necessarily involves a shift in perspective (the Saliba speaker has to tell me – being in the role of SPEAKER – what she would say in my place).

Saliba-Logea Demonstratives

271

sentences ‘Whose X is this/that?’ where ownership is questioned and therefore not presupposed. As a second point, it proved to be important to actually make speakers utter the words they would say rather than restricting the elicitation with the questionnaire scenes to a metalinguistic discussion of what speakers WOULD say if coming across a depicted scene in real life. It could be observed several times that when speakers actually said an utterance out loud in a specified scenario, this triggered the use of a different demonstrative from the one they thought and had stated they would use. (This is especially relevant for the same scenes with and without pointing, see section 4.2.2.) It was therefore helpful to treat the task as a game with the set rule that the speaker would actually say the utterance with the demonstrative, even though this involved a lot of repetition. 4.2

Experimental Results

The results of the elicitations with the questionnaire are summarized in Table 12.5. In this table, results are separated in terms of the three deictic distinctions ‘near speaker’, ‘near addressee’, and ‘distal’. The five form classes of spatial deictics (clause-final, free demonstratives, adnominal clitics, interjections and place adverbs) are not separated here since in terms of deictic distinction they behave exactly parallel. That is to say that, if a free form may be used for a particular scene, the clause-final form or adominal clitic of the same deictic distinction will also be allowed and vice versa.8 I have chosen the free forms to represent the deictic distinctions in the table. This means the use of the free form teina in Table 12.5 indicates that a ‘near speaker’ form – from any of the three demonstrative paradigms – may be used in the respective context. For columns where only a single demonstrative appears, all speakers preferred this term. For the columns in which two demonstratives are listed, there is variation either across speakers or within the responses by the same speaker (e.g. depending on the exact distance of the referent or the presence of pointing). The criteria which influence or determine the choice of demonstrative so that more than one term may be acceptable for a given scene are discussed below. With the exception of scenes 17 and 19 (discussed in section 4.2.1 below), each picture is presented only once in the table. The arrows underneath the table illustrate the domain of use of each of the demonstratives (or rather of the three deictic distinctions which are represented in the table by the free demonstratives). The solid lines show the core domain of use of a term; the dotted lines represent areas of overlap between terms. Note that there is overlap between all three terms. This overlap cannot easily be 8

This holds with the exception of the distal form of the adnominal clitics which, as discussed in section 3.3, has a wider distribution than the distal forms of the other paradigms.

S

A

S

(8)

(7)

(11)

S

S

A

A

A

TEINA (near SPKR)

(6)

S

(3)

(1)

S

TEINA

S

S

(19)

S

(17)

S

(4)

(2)

A

A

A

A

TEINA / TEMETA

S

S

S

A

A

(23)

S

(18)

A

S

(16)

S

A

TEMETA (near ADDR)

(10)

A

(9)

(5)

A

TEMETA

(19)

S

(17)

S

A

A

TEMETA / TENEM

Table 12.5 Results of the demonstrative questionnaire (seven speakers)

S

A S

A

S

(24)

(21)

TENEM (distal)

(14)

S A

(13)

S A

(12)

A

TENEM

A S

(25)

(15)

S A

other strategies (& TEMEM)

S

A

A

TEINA (near SPKR)

(22)

S

(20)

A

A

(19)

S

(17)

S

TENEM / TEINA

Saliba-Logea Demonstratives

(20), (22) teina, tenem (17), (19)

(1), (3), (6) teina (7), (8), (11)

(12), (13), (14) tenem (21), (24)

(17), (19) teina, temeta, tenem

(2), (4) teina, temeta (17), (19)

273

(17), (19) tenem, temeta

(5), (9), (10) temeta (16), (18), (23)

Figure 12.1 Circle schema of results presented in Table 12.5

represented graphically within the table. Therefore scenes with overlap, i.e. 17 and 19, are listed more than once and the table should be read as a circular diagram (i.e. both ends meet as indicated by the second teina-arrow on the right). The format of Table 12.5 is adopted here to allow easy comparison across this volume. A better graphic presentation of the Saliba-Logea data would be in a circle diagram, as shown in Figure 12.1. In the following sections, I further discuss the results presented in Table 12.5.

4.2.1

Deictic Distinctions

The results from the questionnaire show that the primary spatial distinctions encoded in the Saliba-Logea demonstrative system are distance (near versus far) and anchoring (speaker-based versus addressee-based). Some observations help specify the parameters involved. First, prototypically the speaker-based term is used when referring to objects which are within arm’s reach of the speaker. For example, with three of the speakers, scene 7 (object located right in front of the sitting speaker) was tested in two ways: (a) with the speaker sitting cross-legged, and (b) with out-stretched legs. In both cases the object was positioned just in front of the speaker’s legs. When sitting cross-legged, the speaker could reach the object and the speaker-based form was used. But when sitting with stretched legs, the speaker could not easily reach the object and the distal term was typically preferred. This shows that reachability is more important than mere proximity to the speaker’s body. The second observation is that the criterion for the use of the addresseebased terms is not just proximity to the addressee but more specifically that the

274

Anna Margetts

referent is closer to the addressee than to the speaker. This can be seen in scenes where the object is proximate and equidistant from speaker and addressee, i.e. scene 8 in comparison with scenes such as 9 where the referent is at a comparable distance to the addressee but not in reach of the speaker. (See also scene 2 in comparison to 4, as discussed in section 4.2.2.) The distal demonstrative is used when the referent object is close to neither speaker nor addressee. But even in scenes in which the referent object is out of reach for both speaker and addressee, the distal forms at times compete with the speakeror addressee-based proximal forms, as, for example, in scenes 17 and 19. It seems that the distance between speaker and addressee influences the use of the proximal forms. If speaker and addressee are located at a relatively long distance from each other, the space between them may nevertheless be construed as ‘proximal space’ in the sense that it is the interactional space within which the speech act takes place (see Wilkins, 1999a: 50; this volume). For scene 17 the referent object is located between the speaker and the addressee, but at quite a distance from both. On the grounds of spatial distance alone, this could count as the domain of use of the distal demonstratives, as seen in scene 12. In spite of this, several speakers used the addressee-based demonstratives rather than the distal forms to refer to the object in scene 17. It is interesting that for this scene, where the object is equidistant from both participants but out of the reach of either, the distal form competes with the addressee-based form rather than the speaker-based one. In proximal space, when speaker and addressee are located closer to each other, the speaker-based form always has preference over the addressee-based form.9

4.2.2

Pointing and Touching

Pointing (e.g. with fingers, gaze or head) is natural with all three deictic categories but seems to be required with the speaker-based forms. A finger point accompanying an utterance can affect the choice of demonstrative term. For scene 19, some speakers used the distal term when merely looking at the referent but the speaker-based term when pointing to it. Similarly for scene 22, though speakers preferred the speaker-based forms, some allowed the distal forms when there was no finger point accompanying it. Pointing also plays a role in the choice between the speaker-based and the addressee-based terms. Without a finger point, scenes 2 and 4, where the referent object is in contact with the addressee’s body, require the addressee-based form, and even together with a finger point the addressee-based form is normally preferred. But when accompanied by a very close finger point or by touching of the object, the 9

There could be a visual explanation for the use of the addressee-based form in scene 17 along the lines that an object relatively far away but equidistant from speaker and addressee may, for the speaker, appear closer to the addressee due to perspective.

Saliba-Logea Demonstratives

275

speaker-based term must be used in both of these scenes. Compare examples (67) and (68), which both referred to the same object at about 2 metres from the speaker: (67) Ginauli=te thing=NEAR.SPKR ‘This thing (near me)’ (with finger point) (68) Ginauli=ne thing=DIST ‘That thing (distal)’ (without finger point)

4.2.3

Givenness and Visibility

In general, if the referent object has already been introduced in the discourse it is likely to be marked by the anaphoric marker =wa rather than by a demonstrative form. An invisible object is likely to trigger the use of =wa because – since it is invisible – some prior knowledge of the object is required in order to refer to it. This means it is the pragmatics that come with invisible objects rather than the visibility parameter itself which is responsible for the choice of the anaphoric marker. Similarly, as previously discussed, a referent which is identifiable even at its first mention tends to be marked by =ne. An object in a distant location and not visible to either speaker or addressee, as in scenes 15 and 25, typically triggers the use of the distal place adverb tenemai/temenai rather than a demonstrative. The anaphoric marker =wa and the distal/definite clitic =ne and the interjections ne ‘distal’ and auu ‘remote’ are also allowed. Since the referents are invisible, they are generally assumed to have already been mentioned, to be inferable or to be identifiable at first mention. Clause-final and free demonstratives were generally rejected for these scenes where the referent is distant and invisible, but some speakers considered them marginally acceptable. In contrast, within proximal space, it is possible to use demonstratives for invisible objects near the speaker or near the addressee, as in scenes 6, 10 and 11. For scene 11, where the referent is visible to the addressee but not to the speaker, demonstratives were used but had to be accompanied by a finger point and typically also by a head nod towards the object.

4.2.4

Ownership and Control

As mentioned, demonstrative modifiers tend to compete with possessive markers in Saliba-Logea. As discussed in Margetts (2004a), possessive constructions in Saliba-Logea, as in many other Oceanic languages, are commonly used for a broader range of scenarios than may be familiar from European languages, including benefaction and intended or future possession (see also Song, 1997; 1998; 2005). In Saliba-Logea, even when the speaker has no actual knowledge

276

Anna Margetts

of a possessive relation, the mere proximity of the referent object to a participant is often enough to trigger the use of a possessive construction rather than a demonstrative, as in example (69). (69) Yom tobwa kaiteya ye-halusi? 3SG.POSS bag who 3SG-weave ‘Who wove your bag?’

While the free demonstratives are in a paradigmatic relation with the possessive classifiers (both occurring in prenominal position), the adnominal demonstrative clitics may co-occur with possessive classifiers, as in (70). (70) Yom tobwa=me haedi ku-hai? 3SG.POSS bag=NEAR.ADDR where 2SG-get ‘Where did you get your bag?

4.2.5

Social Boundaries

The deictic distinction between ‘near speaker’, ‘near addressee’ and ‘distal’ is generally based on actual spatial distances rather than on social boundaries. However, scene 19, where the referent is in a room with the addressee, and the speaker looks in from the outside, also elicited a couple of ‘near addressee’ responses even though the object is spatially closer to the speaker. This suggests that location within defined domains can at times influence the choice of demonstrative.10 5

Observational Data

Observations of natural interactions in context confirm the findings of the questionnaire elicitations. Speaker-based and addressee-based forms tend to occur as substitutional pairs in conversation. This holds especially for the free demonstratives and place adverbs. Consider the exchanges between speakers A and B in (71) and (72) below: (71) A: Kaputi=wa ku-hai-ya-ma cup=ANA 2SG-get-3SG-hither ‘Give me the cup.’ B: Teina? NEAR.SPKR ‘This one?’ A: Ah, temeta INTRJ NEAR.ADDR ‘Yes, that one.’ 10

None of the speakers suggested the addressee-based form for scene 22 or the speaker-based form for 23.

Saliba-Logea Demonstratives

277

(72) A: Ku-tole metai 2SG-put there.NEAR.ADDR ‘Put it there.’ B: Haedi? Inai? where here ‘Where, here?’ A: Ah temetai INTRJ there.NEAR.ADDR ‘Yes, there.’

As discussed earlier, the use of speaker-based proximal terms can be extended from their prototypical use in personal space to interactional space (Wilkins, 1999a: 50; this volume). Observational data shows that proximal terms may also be extended beyond this to home range and geographic space if the referent object is accordingly big (e.g. ‘this village’, ‘this island’, ‘this country’). Finally, there is evidence that the spatial deictic constraints of the demonstratives can be over-ruled in contrastive use of the terms. It is possible, for example, to use the distal term within proximal space if contrasted with another object in proximal space and, vice versa, the proximal form may be used within distal space, when contrasted with a more-distally-located referent. The absolute selection restrictions of the non-contrastive use are translated into relative selection restrictions in contrastive use. If the two contrasted objects are not equidistant from the speaker, but one is closer than the other, the proximal form may refer to only the closer of the two objects and the distal form may refer to only the one which is further away. Consider (73): (73) Tenem teha=ne ku-ginauli na teina=te sola DIST side=DIST 2SG-do CONJ NEAR.SPKR=NEAR.SPKR still ‘You did that side but this one not yet.’

So, in non-contrastive use the near-speaker form refers to an object which is near the speaker but in contrastive use it refers to an object which is equidistant or nearer to the speaker than other objects with which it is in contrast. The reverse holds for the distal forms. This shows the importance of investigating contrastive and non-contrastive uses of demonstratives independently (see Wilkins, 1999b). 6

Discussion of Results and Proposed Analysis

Below I provide a summary of the findings in terms of extendability, competition with other forms and topics for future research.

278

Anna Margetts

Extendability Both the speaker-based and the addressee-based forms typically occur within personal space of the respective participant (touching, or within reach; see Wilkins, 1999a: 50; this volume). The use of these proximal forms is extendable beyond the proximal domain into interactional space (the social space of conversation). In this case they indicate relative closeness to either the speaker or the addressee within the interactional space (see 4.2.1). The speaker-based proximal terms may be extended further, into home range and geographic space (section 5). Contrastive use of demonstratives may extend beyond their typical spatial domain and indicate proximity or distance relative to a contrasting referent. An extension of the use of speaker-based proximal terms beyond personal space into interactional space can also be triggered by an accompanying finger point. An object within interactional space but out of the speaker’s reach is commonly referred to by a proximal term if the speaker is pointing at the same time. The same object may be referred to by a distal term but in that case typically without an accompanying finger point. Similarly, the use of speaker-based terms may be extended into the personal space of the addressee if accompanied by a very close finger point or by touching of the referent object (section 4.2.2). Finally, social boundaries, such as enclosed living spaces, may affect the choice of a deictic term and may extend it beyond the typical spatial domains of its use (section 4.2.5). But such effects appear to be rather marginal in Saliba-Logea and do not occur consistently. Competition with Other Forms Demonstratives compete with a number of other terms, in particular with possessive expressions, place adverbs, the definite use of =ne and the anaphoric marker =wa. The use of possessive morphology in Saliba-Logea commonly indicates control over or association with the referent object rather than an actual possessive relation. Proximity of the referent object to a person is therefore typically enough to encode the referent object as a possessed noun rather than as a noun phrase with a demonstrative (section 4.2.4). The form =ne can occur with referents in all kinds of spatial configurations, and it may combine with free demonstratives of all three deictic distinctions. As a spatial deictic it functions as a distal form in the paradigm of adnominal clitics, but it should be considered an open question as to whether it has inherently distal semantics or is assigned the distal value pragmatically through contrast with the two proximal forms (section 3.3).

Saliba-Logea Demonstratives

279

In contexts where the referent object is far away and invisible to both speaker and addressee, place adverbs, the clitic =ne or the anaphoric marker =wa are used rather than demonstratives (section 4.2.3). If an object has previously been introduced to the discourse, or it can be inferred from the extra-linguistic context, it tends to be marked by =wa rather than by a spatially deictic demonstrative form. This tendency also accounts for the use of this maker for referents which are invisible to both the speaker and the addressee. If an object is invisible but referred to as specific, it is typically given information. Referents which are visible to at least one speech act participant, either the speaker or the addressee, can generally be referred to by demonstrative terms. In conclusion, Saliba-Logea shows a three-way deictic distinction between ‘close to speaker’, ‘close to addressee’ and ‘distal’ across five different form classes of demonstratives and place adverbs. In some of the form classes, these three spatial distinctions participate in contrasts with additional forms, such as the anaphoric adnominal clitic =wa and the remote interjection auu. The speaker-based forms are used to denote objects close to the speaker, prototypically within arm’s reach. The addressee-based forms are used to denote objects which are close to the addressee and closer to addressee than to speaker. If a referent object is close and equidistant to both speaker and addressee, the speaker-based demonstrative forms generally have preference over the addressee-based forms (but see section 4.2.1). The distal forms denote objects which are close to neither the speaker nor the addressee. Topics for future research include, in particular, the functions of the clausefinal demonstratives in comparison to the free demonstratives. The addresseebased form meta of the clause-final demonstratives has taken on discourse functions and developed into a topic marker – a development which is also attested in a number of other Oceanic languages (section 3.1) and which also deserves further study. Abbreviations ADDR APP CONJ DET DIST EMPH EX IMMED INC LOC

addressee applicative conjunction determiner distal free emphatic pronoun exclusive immediateness marker inclusive locative

NEG POSS PL PP PROX PRSUP RED SG SPKR TAM

negative possessive plural postposition proximal presupposition reduplication singular speaker tense/aspect/mode

280

Anna Margetts

References Cleary-Kemp, J. (2006). Givenness, definiteness, and specificity in a language without articles: The use of NP markers in Saliba, an Oceanic language of Papua New Guinea. Honours thesis, Monash University, Melbourne. Cooper, R. (1975). Coastal Suau: A preliminary study of pan-dialectal relationships. In T. Dutton, ed., Studies in the languages of South-East Papua. Canberra: Pacific Linguistics, pp. 93–124. (1992). That’s what I’m talking about: Discourse level deixis in Buhutu. Language and Linguistics in Melanesia, 23, 95–105. Dawuda, C. (2009). Discourse functions of demonstratives and place adverbs with exophoric reference in Logea, an Oceanic language of Papua New Guinea. Melbourne: Monash University. Diessel, H. (1999). Demonstratives: Form, function, and grammaticalization. Amsterdam: John Benjamins. Dixon, R. M. W. (2003). Demonstratives: A cross-linguistic typology. Studies in Language, 27, 61–112. Fox, B. (1984). Participant tracking in Toba Batak. In P. Schachter, ed., Studies in the structure of Toba Batak. Los Angeles: University of California, pp. 59–79. François, A. (2005). A typological overview of Mwotlap, an Oceanic language of Vanuatu. Linguistic Typology, 9, 115–146. König, E. (2017). The deictic identification of similarity. In Y. Treis & M. Vanhove, eds., Similative and equative contructions: A cross-linguistic perspective. Amsterdam: John Benjamins. Margetts, A. (1999). Valence and transitivity in Saliba: An Oceanic language of Papua New Guinea. Doctoral dissertation, Radboud University Nijmegen. (2004a). From implicature to construction: Emergence of a benefactive construction in Oceanic. Oceanic Linguistics, 43, 445–468. (2004b). Spatial deictics in Saliba. In G. Senft, ed., Deixis and demonstratives in Oceanic languages. Canberra: Pacific Linguistics, pp. 37–57. (2010). Spread of the Saliba-Logea plural marker. In B. Evans, ed., Discovering history through language: Papers in honour of Malcolm Ross. Canberra: Pacific Linguistics, pp. 413–429. Mosel, U. (1994). Saliba. Munich/Newcastle: LINCOM Europa. Peteliyaki, B. & Lithgow, D. (not dated). The use of suffixes -WA and -NE in Auhelawa. Working paper. Ukarumpa: SIL. Ross, M. (1988). Proto Oceanic and the Austronesian languages of Western Melanesia. Canberra: Pacific Linguistics. Senft, G., ed. (2004). Aspects of spatial deixis in Kilivila: Deixis and demonstratives in Oceanic languages. Canberra: Pacific Linguistics. Song, J. J. (1997). The history of Micronesian possessive classifiers and benefactive marking in Oceanic languages. Oceanic Linguistics, 36, 29–64. (1998). Benefactive marking in Oceanic languages: From possessive classifiers to benefactive marking. In A. Siewierska & J. J. Song, eds., Case, typology and grammar: In honor of Barry J. Blake. Amsterdam/Philadelphia: John Benjamins, pp. 247–275. (2005). Grammaticalization and structural scope increase: Possessive-classifierbased benefactive marking in Oceanic languages. Linguistics, 43, 795–838.

Saliba-Logea Demonstratives

281

Wilkins, D. P. (1999a; this volume). The 1999 demonstrative questionnaire: ‘This’ and ‘that’ in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24. (1999b). Eliciting contrastive use of demonstratives for objects within close personal space ((all objects well within arm’s reach). In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 25–28.

13

Warao Demonstratives Stefanie Herrmann

1

Introduction

In the following I will present a discussion of the semantics of Warao demonstrative pronouns and the variables governing their use.1 Most of the relevant data was gathered during six weeks’ fieldwork in the Northern autumn of 2000. The Demonstrative Questionnaire (Wilkins, 1999a; this volume) served as the main elicitation tool.2 This data was completed with observational data and corpus data. Whereas in the literature there are descriptions of a three-term distancesensitive and possibly hearer- versus speaker-anchored system, I argue that Warao has four demonstrative terms which subdivide in two sets. One set is shown to be distance-sensitive and speaker anchored, accompanied by obligatory pointing gestures and calling attention to a referent that is in shared vision space. When used in discourse, this set serves to establish the referent in the universe of discourse. The other set is distance-neutral, has no obligatory accompanying gesture, and can also be used when the referent lies outside a shared vision space. This set is used when the referent is already in the joint attention of both addressee and speaker and serves to keep track of referents in discourse. I begin my discussion with an introduction to the Warao language and its speakers (section 1.1). After reviewing information from written sources (section 2), a morphological analysis of the encountered forms is proposed (section 3). In section 4, experimental results and the delimitation of language-specific variables are examined in detail. I begin with a discussion of the exophoric noncontrastive uses of the demonstratives and then complement this questionnaire

1 2

As for the final version of this chapter, I am especially grateful to Sarah Cutfield who read the draft and made many valuable suggestions. I am grateful to Gunter Senft who invited me to the Max Planck Institute for Psycholinguistics, Nijmegen (MPI) and brought me into contact with the the Space Project of the Language and Cognition Group.

282

Warao Demonstratives

283

data with data from other sources and briefly discuss discourse uses (section 4.5). Section 5 sums up my analysis of the Warao demonstratives. 1.1

The Warao Language and Its Speakers

The Warao, speakers of the Amerindian language Warao, have traditionally lived in the lower Orinoco Delta. For historical and ecological reasons, some Warao groups migrated to the western Delta and to adjoining areas in Venezuela and Guyana in the course of the last century. In Venezuela they form a steadily increasing population of approximately 30,000 individuals.3 The Warao are the second largest indigenous group in Venezuela; although the total indigenous population of the country comprises only about 2.7%, they make up for 25% of the population in their homeland district Delta Amacuro. The latest official census (Venezuela, 2011) rates 89.4% of the population in Delta Amacuro speaking an indigenous language4 and notes that this is the only district where the number of speakers of an indigenous language has increased slightly. Of these, 36% are rated bilingual in Spanish-Warao. The Warao communities in Guyana are reported to be bilingual or trilingual in Warao, English, and Spanish but seem to have a tendency to shift away from speaking Warao as their mother tongue (Forte, 2000). In Venezuela fluency in Warao varies according to the impact state and religious education policies had on the respective community. Communities in the lower Delta that have lived within the reach of Spanish Capuzin missionary boarding schools have long shown a tendency to shift to speaking Spanish. Nevertheless communities in the western part of the delta (river Pedernales and tributaries) are soundly Warao speaking.5 The Chavez government (2000–2013) granted better access to school and university education for the indigenous and rural population and recognized the indigenous languagues as official languages. As a result there is more motivation for Warao speakers to pursue a bilingual education.6 It is important to point out that Warao groups are not homogeneous. There is regional variation in culture (Heinen and Garcia-Castro, 2000) as well as in language. Some authors have suggested that there are no dialects at all (Romero-Figeroa, 1997) or only slight variations (Osborn, 1966a: 108). Nevertheless, I observed differences in vocabulary, syntax, and phonology

3 4 5 6

In 2000 there were approximately 1,000 individuals living in Guyana (Forte, 2000). Warao is the only indigenous language in this part of Venezuela. I visited the communities repeatedly within the last 14 years and noticed no change in attitude toward the language. For more detailed information on the language policies of state and church as well as on the perspective Warao speakers hold themselves see Herrmann (2007).

284

Stefanie Herrmann

across the Delta during my field trips and in the analysis of my field material and published texts.7 Warao as a language is considered an isolate (Weisshar, 1982). Typologically Warao is an agglutinating language and has no articles. As regards the “basic” or unmarked word order in Warao, there is no general agreement. Osborn (1966b) takes a SOV word order to be basic, whereas Romero-Figeroa (1997) argues in favor of OSV.8 2

Claims in the Literature

There is no broad or systematic treatment of demonstratives in Warao in the existing literature. Information concerning demonstratives is nonetheless found in older and newer works on the Warao language. However, much of this description is inconsistent, and most of the data is devoid of contextual information. This does not allow for the delimitation of variables that govern the use of demonstratives, nor does it allow us to determine the semantics of the terms mentioned. In the remainder of this section, I first discuss older analyses (section 2.1), followed by more recent descriptions (section 2.2). I will then identify outstanding issues in the description of Warao demonstratives in section 2.3. 2.1

The Older Sources from the Lower Delta

The older publications on Warao describe a three-term distance-sensitive system of demonstrative pronouns, some of which are also used adnominally. Vaquero, a Capuchin missionary, reports a set of two apparently distancesensitive demonstrative terms with singular and plural forms which he calls “adjectivos demonstrativos”(Vaquero, 1965: 61–64). Two of them are reported to function as personal pronoun for third person singular (tai) and plural (tatuma). Barral (who was also a Capuchin) integrated his own findings and data from his fellow brethren into an extensive lexical work (Barral, 1979). In Table 13.1 definitions are cited as they are found under the respective Warao terms in the dictionary.9,10 Neither Barral nor his fellow brethren had received extensive linguistic training, so they more or less mapped the Warao data onto the Spanish 7 8 9

10

Text analyses are part of my ongoing teaching of Warao at the University of Stuttgart. They have not yet been published. A systematic study of dialect differences remains to be done. For an overview see Herrmann (2004). I will use the official Spanish based spelling for the Warao words where “j” corresponds to [h] in the IPA alphabet. [r] and [d] are allophones of /r/. There are some dialectal variations regarding the use of these allophones: accordingly the official spelling uses both graphemes. The glosses given here stem from my analysis; justification for these will be given in the remainder of the discussion.

Warao Demonstratives

285

Table 13.1 Barral’s person-oriented system

Warao term tama-ja* tama-tuma oma-ja

ama-ja

Person and word class according to Barral (1979)

Spanish translation by Barral

1st pers. sing. dem. pron. 1st pers. pl. dem. pron. 1st pers. pl. dem. adj. 2nd person dem. pron. 2nd person dem. adj.

Este, esta, esto Estos, estas Estos, estas Ese, esa, eso. Esos, esas Ese, esa, eso. Esos, esas Aquel, aquella, aquello Aquellos, aquellas Ahí. Ese lugar. Allí está. Vélo ahí. el, ella, lo Ese, esa, eso Aquél, aquélla, aquéllo Ellos, ellas Aquellos, aquellas

dem.pron. 3rd sing. and pl.

adv. of location tai

3rd pers. sing. pron.,

tatuma

3rd and 2nd sing. dem. pron. 3rd pers. pl. pers. pron. 3rd and 2nd pers. pl. dem. pron.

My gloss DEM.SPK.RCH–EXIST DEM.SPK.RCH–GEN.PL DEM.DIST –EXIST DEM.DIST –EXIST

3SG.DEM.PRON

3PL.DEM.PRON–GEN.PL

* The morpheme -ja has multiple uses. It can work as a nominalizer or emphasizer; here it is necessary as the stem cannot stand alone. Its most abstract meaning could be glossed as ‘existence of X’.

demonstrative system which itself was inspired by the person-oriented model of classical grammar. In this analysis este is considered ‘near speaker’, ese ‘near hearer’, and aquello as ‘not near hearer or speaker’.11 2.2

More Recent Works from the Western Delta

The most recent work on Warao is a reference grammar (Romero-Figeroa, 1997). It describes a class of “demonstrative determiners” (1997: 51) with three members (ta, ama, tama) which combine with the suffix -tuma.12 The determiners are described as having two different degrees of distance in their English glosses. One term is labeled ‘proximal’, while the other two are both glossed as ‘distal’. Nominalized forms are derived with the suffix -ja (Romero-Figeroa, 1997: 60). The forms tai and tatuma, however, do not have nominalized forms. They function as both pronoun and determiner in his 11 12

This is still the current explication in Spanish grammar books of the 1990s. See, for example, Reumuth and Winkelmann (1993: 50). -tuma denotes an associative plural (Corbett, 2000: 101), as in Maria-tuma ‘Maria and her friends/family’ or najoro-tuma (food-tuma) ‘the things that have to do with food’.

286

Stefanie Herrmann

Table 13.2 Romero-Figeroa’s (1997) system Warao term

category

glosses

tama tama-tuma tama-te ta(i) ta-tuma ta-te ama ama-tuma ama-te tama-ja tama-ja-tuma ama-ja ama-ja-tuma

dem. determiner dem. determiner deictic locative dem. determiner / dem. pronoun dem. determiner / dem. pronoun deictic locative dem. determiner dem. determiner deictic locative dem. pronoun dem. pronoun dem. pronoun dem. pronoun

proximal proximal proximal distal distal distal distal distal distal proximal proximal distal distal

system, as presented in Table 13.2. The suffix -te denotes ‘location’ and the suffix -i is glossed as ‘subordinating’ by Osborn (1966b).13 While Romero-Figeroa (1997) seems to consider ama and tama free forms,14 they were not observed without suffixes in my field data or in texts reviewed thus far. Osborn (1966b: 261) also refers to them as bound forms or stems, in his structuralist overview of Warao. He identifies three demonstrative base terms:15 tama- ‘this’, ama- ‘that yonder’, ta- ‘that (near you)’. According to him they belong to an overlapping class sharing some characteristics of nouns and others of relationals. He states that the “noun-like forms” are obtained through the suffixation of -tuma (as in tama- tuma ‘these’) and all have (with the exception of ta-) alternates formed with -ja (as in tama-ja ‘this’). The nounlike demonstratives may modify another noun and then function as determiner. In contrast to Romero-Figeroa, Osborn does not claim a separate class of demonstrative determiners and does not mention the combination of the suffixes -ja and -tuma in a row (as in tama-ja-tuma). The “relational like forms”, on the other hand, are classified as demonstrative adverbs. They are a combination of the demonstrative stem and the suffixes -te or -ta as in tama-te ‘there’ or ta-ta ‘there’. 13

14

15

Osborn names the suffix -i (number 729) “subordinating” and states that it only occurs after the suffix -mejere (number 601: desire command), the relative kuta or within the word ta-i (< ta-i) (Osborn, 1966b: 254). The only example given by Romero-Figeroa himself is not in the context of the discussion of pronouns: “tama jobaji yakera” (this land is good) (Romero-Figeroa, 1997: 8). In Warao identical vowels that are only separated by a voiced [h] (written “j”) are easily assimilated in fast speech and in long-standing place names. This may have occurred in the data cited here by Romero- Figeroa. In this case it would have been “tamaja jobaji yakera”. The English translations are Osborn’s.

Warao Demonstratives

2.3

287

Open Questions

While older works from the lower delta and Osborn map the Warao data onto a traditional three-term person-oriented system, Romero-Figeroa’s more recent language description distinguishes only two degrees of distance. Osborn is the only author who, albeit indirectly, describes hearer-anchored semantics, glossing the stem ta- as ‘that (near you)’ (Osborn, 1966b: 261). We are thus left with several unanswered questions: if there is a distance-sensitive demonstrative system in Warao, how many degrees of distance are encoded? Is the system speaker- and/or hearer-anchored? And do differences between authors stem from dialectal variants? 3

Form and Class

Before answering the questions about usage, I will briefly consider the form and class of Warao demonstratives. Considering all terms found in the literature and in my field data, the inventory of morphemes can be seen in Table 13.3. The term omaja occurs in Barral’s dictionary and in texts from the lower delta (Lavandero Perez, 1991; 1992; 1994; 2000), but it did not occur in my field data nor in Osborn’s articles (1966a; 1966b; 1967) or Romero-Figeroa (1997). Also, Barral notes that amaja can be used as a synonym for omaja. I therefore suppose that the stems ama- and oma- are dialectal variants. Otamaja ‘over there’, on the other hand, was not mentioned in the literature at all. My consultants either uttered ote amaja or the contracted form otamaja. Interestingly, Barral notes a word tamajote ‘over there far away’ which he analyzes as being composed of tamaja and ote. This suggests that ote ‘far away’ may be combined with amaja in the western part of the delta, whereas it combines with tamaja in the lower delta in combination. On a morphological level this analysis shows that ta- patterns differently to the other stems insofar as it is the only stem combinable with -i and the only stem not combinable with -ja or -te. On the other hand, all stems are combinable with -tuma directly. -tuma may also be suffixed after -i or -ja.

Table 13.3 Combinability of demonstrative stems combines with demonstrative stem

-te (LOC)

-ja (EXIST)

-tuma (GEN.PL)

-i (SUBORD)

tamaama-/ omaotamata-

yes yes yes no

yes yes yes no

yes yes yes yes

no no no yes

288

Stefanie Herrmann

A demonstrative stem and -te form a demonstrative adverb. All other possible combinations can be said to be demonstrative pronouns. These can be used both pronominally and adnominally, which is not surprising given that nouns in Warao can be used to modify other nouns. I suggest there is thus no separate class of demonstrative determiners.16 4

Experimental Results

4.1

Scope and Methodology of Investigation

I undertook a 12-month fieldwork stay in the Warao village of Wakajara de La Horqueta from 1998 to 1999 before starting to investigate demonstrative pronouns in a subsequent field trip in 2000. The area lies, roughly speaking, in the western part of the Delta. During my first stay I had been integrated into a family by ways of fictive kinship,17 a position that defines certain rights and obligations toward consultants who were all members of this family. People were used to my presence, but the more controlled methodology of a psycholinguistic questionnaire was new for all of us. Following the methodology of the manual, I first focused on non-contrastive exophoric use. My main elicitation tool was the Demonstrative Questionnaire (Wilkins, 1999a; Chapter 2, this volume). Subsequently, I included data from participant observation and open interview as well as corpus data obtained from published transcribed Warao texts in order to clarify the usage of demonstratives and investigate discourse usage. All the questioning was conducted in Warao, the language I used in my everyday interaction with the population of the entire village. I enacted the questionnaire with five consultants, four of whom were male – aged 16, 17, 30, and 45 – and one of whom was female, aged 20. They were all Warao mother tongue speakers. The 25 different scenes and their variants as specified in the questionnaire were conducted in a naturalistic setting. Sessions sometimes additionally involved neighbors or family members who were present and started participating or commenting. As sessions tended to become rather complex, it was essential to audio record them for later analysis. Videotaping would surely have been desirable, especially with regard to pointing and other nonverbal behavior. To compensate at least partly for this shortcoming, I included verbal comments as to whether a consultant was touching an object. 16

17

This is in line with Diessel’s (1999: 158) observation that “some languages do not have a particular class of demonstrative determiners, instead, using a demonstrative pronoun with a coreferential noun in apposition.” But it differs from Romero-Figeroa (1997) who claims a distinguishable set of demonstrative determiners. There are terms for adopted kinship in Warao. They were not applied to me. I recieved the position of a “niece” with regard to the owner of the house, an elderly woman.

Warao Demonstratives

289

It is important to note that I did not always ask the speakers for alternative possibilities, so that my data is not always explicit on whether another form could have been chosen. An additional test for contrastive use as included in the fieldwork manual was carried out with three male consultants including the two younger men and a 33-year-old. Answers for contrastive use were written down in a notebook. 4.2

Table of the Answers Given to the Questionnaire

Table 13.4 lists the answers given to the scenes of the questionnaire. Each column groups together scenes which elicited the same choice of demonstrative terms. These terms are then listed in row 2 of the table. Row 3 shows if tai was also used in the scenes. Finally, the arrows below Table 13.4 show the respective “reach” of each demonstrative. Here the possibility to use tai is inferred from later analyses. Answers contained linguistic material other than demonstrative pronouns such as relative clauses (“the bottle that we saw before”), demonstrative adverbs or possessive constructions (“give me your bottle”). They were employed when the referent was not visible to either speaker or addressee, as in the scenes depicted in Figure 13.1. Speakers also made comments and formulated ideal definitions of the terms in question or predicted their use in a hypothetical scene. This metalinguistic commentary is discussed in section 4.4.1. 4.3

Analysis of the Questionnaire Results

First I shall analyze the character of the scenes grouped together by the usage of specific demonstrative terms before proceeding to the identification of the language-specific variables governing the use of demonstrative pronouns in Warao.

A

S

S A

(11)

(16)

A S

(25)

Figure 13.1 Scenes where no demonstrative pronoun was used

Stefanie Herrmann

290

Table 13.4 Usage of the demonstrative pronouns in the questionnaire (consulted speakers) S

S

A

S

S

A

S

A

A A

(1)

(2) S

S

(17)

(12)

(4)

S

SA

A

S

(21) S

A

A

(3) S

(13)

(6) A

S

(18)

(5) SA

S A

A

S A

(7) S

(14)

(8) A

S

(23)

(15)

A

AS

(9)

(10)

(24)

S A

A

(19)

S

(20)

A S

(22)

tamaja

amaja / otamaja

tai

tamaja

amaja / otamaja

In scenes 1, 2, 3, 6, 8, 9, 10 tai was also used.

In scenes 12 and 13, tai was also used.

otamaja

amaja

tai

Tai was the only pronoun used in these scenes.

Warao Demonstratives

4.3.1

291

Comments on the Scenes and Their Respective Demonstrative Pronouns

• Scenes where tamaja was used. In all scenes where tamaja was chosen the referent is either within direct reach of speaker (spkr) or can be easily brought within spkr’s reach (i.e. by taking a step toward the object or by bowing toward it) as in scenes 9, 10, 19, 20, and 22. In scenes 19 and 22 the referent is out of easy reach of the addressee (addr). Also the referent remains always visible to both spkr and addr except in scenes 6 and 10 where the object is difficult but not impossible to see for one of the two participants. In most of the scenes tai was also used. In scene 4 tai was chosen or demonstratives were omitted altogether. • Scenes where both amaja and otamaja/ote amaja were used. In all of the scenes where otamaja/ote amaja and amaja were used, the referent is at least some steps away from spkr and addr. Still all referents are visible to both participants. • Scenes where only otamaja/ote amaja were used. In scenes 17, 18 and 23 ote amaja or its contracted form otamaja was the only form used. In these scenes the referent is at least some steps away from spkr, and in 17 it is additionally some steps away from addr. All referents are visible to both. • Scenes where only amaja occurred. There is one scene (21) where amaja is the only from attested. In this scene addr and spkr are both some meters away from the referent and it is visible to both. • Scenes where only tai was used. In most of the scenes, tai was used as a secondary possibility (1–6, 8–13, and 15). In the scenes where tai was the only demonstrative pronoun used (4, 5, 15), scene 5 is the only scene of the questionnaire for which it was specified that the consultant should not point or grab, and in scene 15 the referent was hidden completely from the sight of both speakers and addressee (the same applied to scene 25 and although tai was not chosen for this scene by my consultants it would have been a possible choice according to informants). In scene 4 and 5, the referent is very close to both speaker and hearer, but as it is on the hearer’s body it is more close to him than to the speaker. • Scenes where no demonstrative pronouns were chosen. In some scenes demonstrative pronouns were omitted altogether. In scene 11 the referent is hidden from spkr’s sight and in scene 25 from spkr’s and addr’s sight. In these scenes the referent is either several meters away or at large geographic scale distance away from at least one of the participants. In scene 16 addr is far away from spkr but the referent is right next to the addr. In all of these scenes the object is quite inaccessible for the spkr.

292

Stefanie Herrmann

4.3.2

Identifying the Language Specific Variables for Warao

In the following discussion I will start with the easily definable variables and then proceed to the ones that were more challenging to identify. Visibility When the referent is visible, all demonstrative forms occur in the data. But when the referent was difficult to see or invisible to one participant only, consultants preferred to use relative clauses or else used tai. They applied formulations such as ‘the bottle behind your back’ or ‘the bottle in front of you’ or ‘your bottle’ or ‘the bottle we were talking about’. Most of my consultants then tried to bring the referent into shared vision space by grabbing it and holding it up. When I asked explicitly for the use of demonstrative pronouns other than tai, they would suggest walking over and having a look at the object. It is therefore apparent that “shared visual space” is required for all demonstrative pronouns other than tai. In scenes 6 and 10, which have the referent half hidden from spkr or addr, tamaja could be used. Note that these scenes are on the “edge of visibility” and that it depends on the participants whether they consider the referent visible or invisible. This is probably why tamaja was also possible in these scenes. In scene 16 there is visibility, but still no demonstrative was chosen by my informants. Spkr is far away from addr and the referent is right next to the addr. The avoidance of demonstratives could therefore also be triggered in situations in which the spkr feels that the object is not within an easily shared vision space of spkr and addr. But then scenes 18 and 13 are also of that kind. Gesture Throughout the questionnaire the term tamaja was the one most often used, and it became quickly apparent that it was always accompanied by finger pointing. As my audiotaped data was not always explicit enough on gesture, I discussed this variable with consultants and all of them stated that tamaja, amaja, and otamaja were obligatorily accompanied by finger pointing or touching. In the scene where hands could not be used (5), the demonstrative tai was used. When I proposed a hypothetical scene where spkr was holding a baby in her hands, only tai was judged possible by my consultants. Although pointing with the mouth is a common practice in Warao culture, according to my consultants it would not be allowed to substitute finger pointing with tamaja, amaja, and otamaja. In at least one instance, I recorded head pointing together with the usage of amaja in participant observation. All in all it is safe to state that there is “obligatory (finger) pointing/touching” when using the demonstrative pronouns other than tai. Only tai can be used

Warao Demonstratives

293

without gesture. This indicates that gesture influences demonstrative selection in Warao. Distance Clearly distance from speaker is a factor in Warao demonstrative use, while distance from addressee is not. In some of the scenes in which tamaja was used (19, 22), the referent is within easy reach of the spkr but out of easy reach of addr. The same holds true for the usage of ote amaja/otamaja. Looking at scenes (13) and (18), here the object lies right in front of addr but very far away from spkr, and therefore tamaja cannot be chosen. These observations are evidence against Osborn’s (1966b: 261) claim that the demonstrative ta- should be glossed as ‘that (near you)’. According to my analysis distance from addr is not part of the semantics of Warao demonstratives. This leaves speaker with the problem in scenes like (4), where the referent is in the shared vision space of hearer and speaker and reachable for the speaker: the distance-sensitive demonstratives do not signal that the referent is located inside the hearer’s personal space.18 In such a situation it may be infelicitous to use tamaja (the speaker-proximal form), as the referent is so much closer to the addressee. In fact, only one informant said that the use of tamaja would have been possible in (4). Given that the speaker is close to the referent, the use of one of the distal forms does not appear to be an appropriate choice either. In fact, speakers solved this problem by using tai or by omitting demonstratives all together.

4.3.3

Preliminary Results

To sum up the findings so far, it can be stated that ‘distance from spkr’ governs the choice of all terms but ta-i. As the stem ta- is also the only one combinable with -i and the only stem not combinable with -ja or -te, it is then possible to define two sets of demonstrative pronouns for Warao on morphological as well as on semantic grounds: a distance-sensitive set which is speaker anchored containing the terms tamaja, amaja, and otamaja and a distance-neutral set comprising tai alone. Within the distance-sensitive set, a clearly definable line becomes apparent between tamaja, on the one hand, and amaja and otamaja on the other. In scenes with tamaja, the referent was always within spkr’s reach or could be easily brought within his or her reach. In scenes other than these, amaja and otamaja were chosen. Nevertheless the difference between these two non-proximal terms is not apparent on the basis of the non-contrastive questionnaire data alone.

18

I am grateful to Sarah Cutfield for pointing this out to me.

294

Stefanie Herrmann

4.4

Additional Data Types and Variables

In order to probe the difference between the two non-proximal terms amaja and otamaja, additional data types were included. I used the speakers’ metalinguistic commentary in the form of an ideal definition of terms, together with naturalistic observational data as well as elicited contrastive use data for this purpose.

4.4.1

Consultants’ Ideal Definition of a Term

In Table 13.5 consultants’ ideal definition of the demonstratives that denote distance are listed. The first row gives the demonstrative term. Rows 1 to 4 list the answers given by four different consultants. The ideal description of tamaja confirms the observations made earlier. The object used to explain tamaja is on the speakers’ knees and the expressions used are reinforced by witu ‘really (near)’. In the descriptions of amaja in row 3, the speaker takes his or her own position as point of reference, confirming that the term is speaker anchored. As to the difference between amaja and otamaja, otamaja seems to be used ideally for objects that are some 100 meters away, while amaja seems to indicate objects which are well out of spkr’s reach but not really far. So the question arises whether amaja should be glossed as medial19 or whether it makes more sense to label the two terms as ‘distal’ and ‘marked-distal’. This motivates our interest in the contrastive use of the set of distance-sensitive terms.

4.4.2

Contrastive Use

Differences between the two non-proximal terms became more apparent when I conducted the additional questionnaire for contrastive reference included in the same MPI field manual (Wilkins, 1999b). Here six scenes were enacted using paper cards or plates/cups as referents.20 Distance ranged from table scale (on the lap of consultant) to room scale (up to several steps away from consultant). They were done with three male consultants and were specified as follows: scene

setting

distance from spkr

B1 B2 B3 B4 B5 B6

2 paper cards in a vertical row 3 paper cards in a vertical row 2 paper cards in a horizontal row (equidistant from spkr) two plates in a horizontal row (equidistant from spkr) two plates in a vertical row three plates in a vertical row

on spkr’s lap on spkr’s lap on spkr’s lap all within spkr’s reach 1 within spkr’s reach 1 within spkr’s reach

19 20

If the term medial implies that it denotes proximity to hearer, it is not appropriate for Warao, as it was shown that only distance to speaker is a factor in Warao demonstrative use. Scene B6 was added by me to the five scenes proposed in the questionnaire.

kawere witu kaOBJ.1PL‘really near us’

3

4

tamatika witu tama -te -uka DEM.SPK.RCH -LOC -PREC ‘really right here’

2

awere near

tamaja karata tama -ja DEM.SPK.RCH -EXIST (object on speakers’ knees) ‘this book here’

1

tamaja

witu really

witu really

karata book

awere ‘near’ /taitika ‘there’

utira sabuka jaine utira sabuka ja far_away quite is ‘I am quite far from it’

‘there’

allá (spanish)

-ine -1SG

amaja karpa ama -ja karpa DEM.DIST -EXIST plastic (object ~ 3 paces away) ‘this plastic foil over there’

amaja

Table 13.5 Metalinguistic commentary and ideal definitions given by consultants

motor motor

(points at an object 200 meters away)

ote amaja motor ote ama -ja over_there DEM.DIST-EXIST (obj. ~ 300 meters away) ‘that motor over there’

ote amaja/otamaja

296

Stefanie Herrmann

Both plates in B4 and the most proximate plate in B5/B6 were reachable by the consultant when bowing toward them. In order to reach the remaining plates, one would have had to get up and take at least one step. Tamaja was the demonstrative pronoun used in all scenes with the following exceptions: the more distal plate in B5 was addressed with amaja or otamaja, and in B6 amaja was used for the middle plate and otamaja for the distal one. So comparing this with the consultants’ ideal definition of a term, it becomes clear that in contrastive use, the two distal terms can be used at a distance closer to spkr than in non-contrastive use. But tamaja is still the only available choice for reachable referents. In B6 amaja is used to refer to the paper card lying in the middle. In a set of two cards (B5), both amaja and otamaja can be used to refer to the more distal. In a row of three, otamaja (ote-amaja) seems to carry the semantic notion of being ‘the farthest away’/‘further away than others’. It could therefore be called ‘marked distal’.21 That would leave us with the gloss ‘distal’ for amaja and ‘marked distal’ for otamaja. The semantics of the demonstrative ote when used alone conveys that the referent is out of sight of spkr and addr and only reachable by a displacement reaching from a short walk to a prolonged trip by boat. It is also used to correct an assumption about the location of a referent as in “no it is not near us but really far away”. These semantics would also support a ‘marked distal’ analysis.

4.4.3

Attention on Object

It proved difficult to determine on the basis of the questionnaire data alone whether the parameter “attention on object” played a role for Warao. During the successive enactment of the different scenes, the referent was practically always in focus especially as I had been talking the setting over with the informants before enacting it. Here observational data was useful. In the following example, the consultant is trying to show a louse to me she just found in one of her grandchildren’s hair, holding it on the tip of her finger: tai mi! tamaja! (DEM.3SG look! DEM.SPK.RCH – EXIST!) She first used tai, but when she realized that I was not looking right at the referent, she used the appropriate term from the distance-sensitive set (tamaja). She could have used tamaja right away, as she was touching the louse. So in order to draw attention to an object, she employed a term from the distancesensitive set. By contrast, when attention is already on the referent tai is usually used, as in scenes similar to the next example: a young girl of about 12 was sent by her grandmother to fetch one of her grandfather’s trousers for mending. 21

See also Dunn (this volume) and Meira (this volume) for discussion on the opposition between marked and unmarked distals.

Warao Demonstratives

297

The girl walked over to another part of the house about 10 meters away and held up a pair of trousers. (1)

girl: tamaja? (DEM.SPK.RCH – EXIST) grandmother: tai! ([yes] DEM.3SG!)

There was hardly any variation in scenes like this one. After the first participant had asked “tamaja?” the second always responded “tai!” when a positive answer was required. So tai was used for a referent that was already in the shared attention/vision space of both spkr and addr, whereas tamaja was used to call attention to the referent and to present it as being located in shared vision space. This is not surprising, as the semantics of tamaja include information on where the referent is, while tai does not. There were nevertheless situations resembling scene (4) of the questionnaire where tamaja was not used to call attention to the referent although this spkr proximal form would have been a logical choice. Typically these situations involved an insect sitting on the leg or arm of addr, preparing to suck blood. Then sentences like: “Jijara jotia!” (sucking your arm!) would prototypically be used. 4.5

Use of Demonstratives in Discourse

As shared vision space was identified as a prerequisite for all demonstrative pronouns of the distance-sensitive set, it is reasonable to ask how spkr draws attention to a referent in discourse. The following example is taken from a text commissioned by an oil company produced by Heinen and Warao consultants (Heinen et al., 1999).22 In this text of 834 words, tai was used a total of 44 times and tamaja occurred only once, in the second sentence: (2)

tamaja kiri’tiana, tatuma aribu ‘petróleo’, tama-ja kiritiana ta-tuma a-dibu petróleo DEM.SPK.RCH-EMPH Criollo DEM.3SG-GEN.PL POSS-word oil(Spanish) oko warao karibu isia ‘ore’ oko warao ka-dibu isia ore 1PL person OBJ.1PL-word with oil ‘This is non-Indian, their speech is petróleo (oil). We the Warao, with our speech (say) ore.’23

Here the topic, “oil”, is defined and translated and attention is drawn to it as the subject of the following text. Elsewhere tai/tatuma are exclusively used as pronoun or demonstrative determiner.

22 23

It is a description of the manufacturing process for fuel from oil, intended to inform the Warao inhabitants of the area about the activity of the oil company. Translation is by Heinen.

298

Stefanie Herrmann

In a small corpus, which I created using transcribed Warao texts available in the literature,24 tai was observed 23 times while tamaja occurred 9 times. tamaja occurred either at the beginning or at the end of a text, other demonstratives of the distance-sensitive set occurred only in direct speech. At the beginning/end, tamaja25 always states the genre of the narration and calls attention to the fact that the story is now starting or drawing to an end. The following example is the last sentence of an everyday narrative: (3)

tamaja isia akwa. Mate tamaja deje kwai tama-ja isia a-kwa mate tama-ja deje kwai DEM.SPK.RCH–EXIST with POSS-head still DEM.SPK.RCH-EXIST story high ‘Herewith it is its head (end of). This one is yet a story from the upper part (recent).’

Attention is called to the fact that this is the end of the story and the whole narration is summed up by the expression tamaja deje kwai (this is a story from the upper part) and qualified as recent or non-mythological story.26 So tamaja is used in a way that would be labeled as discourse-deictic use according to the different uses of demonstrative as typologized by Himmelmann (1996: 224). It points to a discourse entity thereby establishing it (as a mythological story) in the universe of discourse. The expression “tamaja X” is nearly always used to start a story. X will then be the genre specification such as “denobo” (story of old) or “dejiro” (recent story). As noted by Himmelmann, discourse-deictic use and situational use have in common that they establish a new referent in the realm of discourse. The difference is that the situational use points to an entity or location, while the discourse-deictic use refers to a discourse segment that will follow immediately. Within a story tai is one possibility used to keep track of constituents.27 This would correspond to the tracking use as defined by Himmelmann (1996: 226). It has to be noted here that, strictly speaking, Warao has no third person pronouns but uses the demonstratives tai/tatuma instead. This is in line with Himmelmann’s (1996: 243) observation that third person pronouns can develop from tracking use.

24

25 26

27

Myths transcribed by J. Wilbert (1964; 1969) are included with 6,384 words or 41,925 characters as well as 10 narratives of everyday life comprising approximately 70 pages of text transcribed by Lavandero Perez (1992). In the Lavandero Perez texts, I found three occurrences of omaja. The term behaved just as described for tamaja. ‘its head’ means ‘its end’ and ‘a high story’ means a ‘recent story’. In Warao world view the tree is a symbol of life: its roots are the ancient and the crown with its new branches and leaves the recent. Warao has no article, so often possessive constructions are used to specify a character in a story. So in a sentence such as ‘the older sister went out’ this is expressed as ‘her older sister went out’ in Warao.

Warao Demonstratives

5

299

Conclusion

Summing up the results from the questionnaire and the information provided by other data types, I would like to propose my analysis of Warao demonstrative pronouns. Two sets of Warao demonstratives are distinguishable (see Table 13.6). On the one hand, a distance-sensitive set formed on the stems tama-, ama-/oma-, and the combination ote- ama- (otama-) and, on the other hand, a distance-neutral set consisting of tai and tatuma built on the stem ta-. The following language-specific factors governing exophoric use of Warao demonstrative pronouns were found in my data (Table 13.7). The first set has obligatory finger pointing / touching to it, requires shared visual space, denotes distance from speaker (is speaker anchored), and is used to call attention to an object or to establish a referent in the universe of discourse. Tai and tatuma are used without an obligatory gesture, can be used for referents outside shared vision space, and do not convey distance information. They are used to keep track of a referent and are probably on the way to functioning as pronouns. Table 13.6 Warao demonstrative forms and definitions Distance-sensitive set: tamaja, amaja, otamaja

Distance-neutral set: tai, tatuma

Warao term

Definition

Warao term

Definition

tama-

within spkr’s reach /easily brought within spkr’s reach (proximal) well out of spkr’s reach but not really far away (distal) really far away (marked distal)

ta-

mentioned / established

amaotama- / ote ama-

Table 13.7 Language-specific variables for Warao demonstratives Parameter tested for

Language-specific factor found for distance-sensitive set: tamaja, amaja, otamaja

Language-specific factor found for distance-neutral set: tai, tatuma

use of gesture

obligatory finger pointing / touching shared visual space required denotes distance from speaker calls attention to an object (tamaja) establishes a new referent in the universe of discourse (discourse-deictic use)

no obligatory gesture

visibility of object distance of object attention on object mentioned

no shared visual space required does not denote distance does not call attention to object refers to entity that has been established (tracking use)

300

Stefanie Herrmann

Table 13.8 Warao demonstratives in participant and distance anchoring

Speaker Addressee

Proximal

Distal

tama- tata-

ama-/oma- otama- / ote ama- taama-/oma- otama- / ote ama- ta-

From the discussion in sections 4.3.2 and 4.4.3, it becomes clear that there is a “gap” in the spatial values of the Warao demonstratives namely ‘addresseeanchoredness’.28 Because it is infelicitous for the speakers to use either the proximal or the distal demonstratives when an object is in the personal space of the addressee, speakers resort to either using the distance-neutral form tai, or omitting the use of a demonstrative altogether. This gives rise to an implicature that an object referred to with tai is addressee anchored. Table 13.8 presents the distribution of Warao demonstratives relative to both participant and distance anchoring. It shows that only the distance-neutral demonstrative tai can be used in the cell ‘addressee-proximal’. It makes sense that a form which indicates that a referent is established may also have an association with the addressee, as in ‘that which the addressee knows about’. Burenhult (this volume: section 5) writes about an “addressee-anchored accessible” form (ton) in Jahai that is exclusively associated with referents which have the addressee’s attention. Nevertheless addressee-anchoredness is not a part of the semantics of tai – as seen in Table 13.8, where it is used in all cells. It is rather by implicature in pragmatic use that this information is conveyed. This parallels Enfield’s (2003) discussion of the demonstratives nii4 and nan4 in Lao where nii4 “implies (but never entails) that the referent is something ‘here’” (Enfield, 2003: 102) in contrast to nan4, which is “genuinely specified for ‘awayness’” (Enfield, 2003: 102). The implicature of tai as ‘near hearer’ arises from its opposition to the terms of the distance-sensitive set which have a semantic specification of ‘how far the referent is from speaker’. Tamaja isia akwa! ‘This is the end’.

28

I am grateful to Sarah Cutfield for pointing this out to me.

Warao Demonstratives

301

Appendix 1 Glosses used: 1SG 3SG.DEM.PRON DEM.DIST DEM.SPK.RCH EXIST GEN.PL LOC OBJ.1PL POSS PREC

first person singular third person singular demonstrative pronoun demonstrative distal demonstrative within speaker’s reach existence / nominalizer generic plural locative object marker first person plural possessive prefix precisely/exactly

References Barral, B., de (1979). Diccionario Warao-Castellano, Castellano-Warao. Caracas: UCAB. Corbett, G. (2000). Number. Cambridge: Cambridge University Press. Diessel, H. (1999). Demonstratives: Form, function and grammaticalization. Amsterdam: John Benjamins. Enfield, N. J. (2003). Demonstratives in space and interaction: Data from Lao speakers and implications for semantic analysis. Language, 79(1), 82–117. Forte, J. (2000). Amerindian languages of Guyana. In F. Queixalós & O. RenaultLescure, eds., As línguas amazonicas hoje. São Paulo: Instituto Socioambiental, pp. 317–331. Heinen, D. et al. (1999). Vocabulario Warao referente al petróleo / Warao petroleum vocabulary. Versión Trilingue. Unpublished manuscript. Heinen, D. & Garcia-Castro, A. (2000). The multi-ethnic network of the Lower Orinoco in early colonial times. Ethnohistory, 47, 561–579. Herrmann, S. (2004). Warao. In P. Strazny, ed., Encyclopedia of linguistics. Vol. I. London: Routledge, p. 1304. (2007). Perspektiven auf die Waraosprache und das Waraosprechen. Doctoral dissertation, Philipps University of Marburg. (Available online at http://archiv.ub.unimarburg.de/diss/z2011/0044/.) Himmelmann, N. P. (1996). Demonstratives in narrative discourse: A taxonomy of universal uses. In B. Fox, ed., Studies in anaphora. Amsterdam: John Benjamins, pp. 205–254. Lavandero Perez, J., ed. (1991). I Ajotejana, Mitos. Caracas: Hermanos Capuchinos. (1992). II Ajotejana, Relatos. Caracas: Hermanos Capuchinos. (1994). III Uaharaho: Ethos narrativo. Caracas: Hermanos Capuchinos. (2000). IV Noara y otros rituales. Caracas: Hermanos Capuchinos

302

Stefanie Herrmann

Osborn, H. (1966a). Warao I: Phonology and morphophonemics. International Journal of American Linguistics, 32, 108–123. (1966b). Warao II: Nouns, relationals, and demonstratives. International Journal of American Linguistics, 32, 253–261. (1967). Warao III: Verbs and suffixes. International Journal of American Linguistics, 33, 46–64. Reumuth, W. & Winkelmann, O. (1993). Praktische Grammatik der spanischen Sprache. Wilhelmsfeld: Gottfried Egret Verlag. Romero-Figeroa, A. (1997). A reference grammar of Warao. Munich/Newcastle: Lincom. Vaquero, A. (1965). Idioma Warao: Morfología, sintaxis, literatura. Estudios Venezolanos Indígenas. Caracas: Editorial Sucre. [Government of] Venezuela. 2011. Censo Indígena de Venezuela 2011. (Available at www.ine.gov.ve/index.php?option=com_content&view=article&id=648%3A201 4-03-11-19-01-12&catid=149%3Ademograficos&Itemid=2.) Weisshar, E. (1982). Die Stellung des Warao und Yanomama in Beziehung zu den indigenen Sprachen Südamerikas nördlich des Amazonas: Studien zur genetischen und areal-typologischen Klassifikation. Doctoral dissertation, Eberhard-KarlsUniversität, Tübingen. Wilbert, J. (1964). Warao oral litrerature: Instituto Caribe de Antropología y Sociología. Fundación La Salle de Ciencias Naturales. Monograph no 9. Caracas: Editorial Sucre. (1969). Textos folklóricos de los Indios Warao. Latin American Studies, vol. 12. Los Angeles: University of California, Latin American Center. Wilkins, D. P. (1999a; this volume). The 1999 demonstrative questionnaire: ‘this’ and ‘that’ in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24. (1999b). Eliciting contrastive use of demonstratives for objects within close personal space (all objects well within arm’s reach). In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 25–28.

14

Chukchi: Non-contrastive Spatial Demonstrative Usage Michael J. Dunn

1

The Language and Its Speakers

Chukchi is the language of one of the indigenous groups of the Russian Arctic. The speakers are traditionally nomadic reindeer herders, although into the historical period there have been increasing numbers of Chukchi speakers found also in the coastal villages of the Asiatic Eskimos. According to the 2010 census of the Russian Federation, there are about 15,908 people registered as ethnic Chukchis, of whom 4,563 speak Chukchi; there are also about 500 non-Chukchi speakers of Chukchi, mostly from other local indigenous groups.1 About 80 per cent of the Chukchis live in the province of Chukotka, and most of the remainder inhabit neighbouring regions of Sakha (Yakutia) and Kamchatka. During the Soviet era Chukotka was rapidly industrialized, and the population rapidly increased up to the economic collapse of the Soviet Union. However, most of the population growth was temporary migrants coming for work, and in the years since the end of the Soviet Union the population of the province has fallen sharply, as migrants leave for the easier conditions of central Russia. The current population is about 50,000. The number of ethnic Chukchis has been fairly stable throughout. Ethnic Chukchis born before the 1960s are usually speakers of Chukchi, although almost all are also bilingual in Russian (there were only 82 ethnic Chukchis in the 2010 census reported not to speak Russian). It is much rarer to find Chukchi speakers among ethnic Chukchis born from the 1970s on, and so despite having several thousand remaining speakers Chukchi should probably be considered highly endangered (Dunn, 1999; Vakhtin, 2001).

1

In the Russian Federation the folk notion of ‘ethnicity’ (Russian nacional’nost’) generally aligns with the administrative category recorded on each individual’s identity document.

303

304

Michael J. Dunn

2

Linguistic Background

The following is a formal description of the spatially deictic elements of Chukchi (following Dunn, 1999) as well as an outline of previous claims. Chukchi has two independent series of spatially deictic morphemes. Both these series of morphemes indicate distance along a cline, but the number and range of spatial distinctions they make are different. One set of stems is used to form the demonstrative pronouns (which also function as modifiers) and the demonstrative adverbs. The other set of morphemes is a morphologically invariant set of deictic particles. 2.1

Formal Properties of Demonstrative Pronouns

Chukchi demonstrative stems are structured into paradigms of pronouns and adverbs. The demonstrative pronouns are formed regularly by means of the suffix -qen(a-), a suffix which does not combine with any other suffixal form (Table 14.1). Demonstrative pronouns are used as heads of single-word noun phrases as well as for modification within multiple-word noun phrases (1). (1)

ŋotqen kaara-n turɣ-in mən-jaa-ɣʔa-n=ʔm2 PROX.DEM.ABS nursery.sled.ABS 2pl-POSS 1PL.INT-use-TH-3SG=EMPH ‘We’ll use this nursery sled of yours . . . ’ (cy413)

All Chukchi nominals are marked for core case functions: ergative (transitive subject), essive (copula complement) and absolutive (all other core syntactic roles). Pronouns, including demonstratives, differ from common nouns in the set of cases they may select; some of the more semantically complex cases (e.g.

Table 14.1 Demonstrative stems stem*

with demonstrative pronoun suffix -qen(a-)

ŋut- PROX ən- NEUT ŋaan- DIST ŋoon- FAR.DIST

ŋotqen(a-) ənqen(a-) ŋaanqen(a-) ŋoonqen(a-)

* The glosses used here presume the analysis which will ultimately be presented.

2

All Chukchi forms are presented in a broad phonological transcription which, following orthographic tradition, writes five surface vowels rather than the underlying three value system with vowel harmony i~e, e~a and u~o.

Chukchi Demonstratives

305

certain locative relationships and types of accompaniment) are not available to pronouns. The demonstratives take the inflections of the high animate class of nominal, which differs from the common (low animate/inanimate class) nominal mostly in that it has plural forms for non-absolutive cases. Most Chukchi NPs consist in a single phonological word; multiple-word NPs in cases other than the absolutive do not occur in colloquial usage.3 Complex NPs with multiple lexical stems can be formed in non-absolutive cases by compounding. Like other modifiers in NPs, demonstratives can be compounded; see examples (2)–(3). The proximal demonstrative, like other pronouns, has an incorporated stem which is not identical to the free form or the stem form used in other derivations (see Table 14.1). (2)

ɣəm=ʔm ŋan Jareŋŋa-jɣəm ŋutin-nute-k 1SG.ABS=EMPH PCL? personal.name-1SG.ABS PROX.DEM-land-LOC tə-nʔel-ək 1SG-become-1SG ‘I, Jareŋŋy, came to be in this land.’

(3)

Ii emelke etʔəm ŋoten-rocɣ-etə ɣ-it-linet Yes it.seems like PROX.DEM-river.bank-ALL PERF-be-3PL ‘Yes, it seems they moved over to this bank of the river.’

Some of the demonstrative pronouns can also be used as locative adjuncts in their citation (absolutive case) form, as in (4). (4)

ənqorə / thence

ŋotqen Roclow-əna ləɣe-taŋ-əməlʔo PROX. RoslovINTS-INTS-all DEM.ABS ERG .ABS ʔorawetlʔa-n/ rə-ra-nŋa-w-jəw-nen . . . person-ABS CS-house-acquire-CS-COLL-3A.3O

qynut like

ŋelwəlʔ-əkin herd-SRC .ABS ‘After that Roslov resettled absolutely all the herding people into houses here’ (=PROX.DEM.ABS) . . .

A form identical to the absolutive form of the demonstrative ənqen also occurs in a non-argument role as a semantically vacuous discourse connective; the spatial meanings of the other demonstratives are always present, and they have no, for example, grammaticalised discourse functions. In (5), the demonstrative pronoun functioning as a pronoun would have to have the plural form ənqena-t; lack of number agreement shows that the word is not being used as an argument. (5)

3

ənqen=ʔm qoranmat-ɣʔa-t so=EMPH slaughter.reindeer-TH-3PL ‘So they slaughtered reindeer.’

Counterexamples to this, occasionally cited in pedagogical grammar materials, are artificial sentences in the literary standard.

306

Michael J. Dunn

Table 14.2 Demonstrative stems and the corresponding demonstrative adverbs stem.

Locative -kə, -ku

Allative -kəri, -ri(lə)

Ablative -qo(rə)

ŋut- PROX ən- NEUT ŋaan-, ŋen- DIST ŋoon- FAR.DIST

ŋutku ənkə ŋaankə, ŋenku ŋoonkə

ŋutkəri, ŋutri(lə) waj-ənre(lə)* ŋaanre(lə), ŋenri(lə)

ŋotqo(rə) ŋaanqo(rə) ŋoonqo

* The spatial meaning of the derived form waj-ənre(lə) comes from the compounded spatial particle waj; see section 2.3.

2.2

Formal Properties of Demonstrative Adverbs

The demonstrative adverbs are slightly irregular, but the demonstrative stems are evident both in form and in meaning. Table 14.2 shows demonstrative stems with markers for (static) locative, allative and ablative functions (these are the most common; others also exist). The demonstrative adverbs may enter the same syntactic slot as any locative adjunct, as in (6). (6)

ŋutku ləɣi-telenjep ny-twa-qinet wʔej-pəlek-əlʔ-ət . . . PROX.ADV INTS-long.ago.ADV HAB-be-3PL grass-shoe-NMZR-ABS.PL ‘Long ago there were Japanese (“Grass-shoe people”) here . . . ’

The two DISTAL forms given in the table above look like they might be different stems; however, they seem to indicate the same distance, as shown in (7). (7)

ŋaanre n-in-ʔemet-qinet, ŋenku ləɣen n-ine-tril-qinet DIST.ADV.ALL HAB-TR-drag-HAB.3-PL DIST.ADV EMPH HAB-TR-put-HAB.3-PL ‘He dragged them way away thither, put them there.’

Demonstrative adverbs can combine with a nominalized locative copula walʔən (< wa-/-twa-) to form an element which is syntactically interchangeable with a demonstrative pronoun. 2.3

Formal Properties of Deictic Particles

The deictic particles waj and raj/caj 4 are used frequently in discourse. Their morpho-syntactic behaviour is clitic-like, in that they are invariant, they don’t take stress and they cannot be used as single-word utterances. Omitting one of these particles never makes a sentence ungrammatical, and never changes the truth value of the sentence. 4

The two variants reflect men’s and women’s dialect respectively (see Dunn, 2000b).

Chukchi Demonstratives

307

Deictic particles frequently co-occur with demonstrative adverbs and pronouns, as in (8). This shows that the semantic functions of the two stem classes are not contradictory. (8)

wʔej-pəlek-əlʔ-ət ... ŋutku waj nə-twa-qenat grass-shoe-NMZR-ABS.PL PROX.ADV PROX.PRES HAB-be-3PL muri kale-tko-ma 1PL.ABS write-ITER-CONC ‘The Japanese . . . were here while we were in school.’

The particles waj and raj/caj are frequently used with the demonstrative ənqen to form a further binary opposition (proximal contrasted to distal) which operates parallel to distinctions made by the demonstratives ŋotqen, ənqen, ŋaanqen and ŋoonqen. These particle forms have an attention-directing function; see section 4 and example (9). (9)

waj=ənqen ŋan nəmnəm qə-jʔo-ɣən / PROX.PRES=that.ABS DEF? village.ABS INT-reach-3SE ŋencilə qə-ŋewənju-cqik-wi / DEVUSHKA qə-piri-ɣən DIST.ADV.ALL INT-find.bride-PURP-TH girl INT-take-3SG ənke-kin PROX.DEM-REL.ABS ‘Visit that settlement, go there to find a bride, take a girl from there.’

A further invariant form, ŋoot, is an alternative form for the absolutive singular demonstrative pronoun ŋotqen when it is acting as an NP head, i.e in an identificational function to mean ‘(it’s) this one’ as an answer to a question. These are illustrated in the exchange (10a–b): (10) a. rəmaɣtə-ŋqa umkuum / ŋoot? cross.over-SIDE tree.ABS this(one) ‘ . . . the tree on the far side . . . [is it] this one?’ b. etlə / aha / ənqen. no oh DEM.ABS ‘No . . . oh[yes], that’s it.’

2.4

Earlier Analyses

Chukchi has been the subject of serious academic investigation for over a century. The first in-depth fieldwork was carried out by Waldemar Bogoras, starting in the 1890s and culminating in his monumental ethnography (Bogoras, 1904–1909) and detailed grammatical sketch (Bogoras, 1922). A two-volume grammar was published by Skorik (1961–1977). This was written as a prescriptive, pedagogical reference, which makes it somewhat misleading when used for purposes of linguistic typology. The third grammar

308

Michael J. Dunn

(Dunn, 1999) is descriptive and usage based, with a focus on spontaneous narrative rather than structured elicitation. Two dictionaries are available; the earlier one (Moll and Inenlikej, 1957) includes some key grammatical information about words, as well as indicating phonological alternates, whereas the later one (Inenlikej, 1976) mostly gives only translation equivalents in citation form. Below I will discuss the analyses of demonstratives present or implicit in Bogoras (1922), Skorik (1961–1977) and Moll and Inenlikej (1957). Further research has shown that the analysis in Dunn (1999) is also wanting in the light of the research presented in this chapter, and I will ignore it in favour of my current best analysis presented here. One major difference in analyses which should immediately be evident is the disagreement between writers over what constitutes the set of demonstrative elements. Bogoras describes a set of ten demonstratives, Skorik nine, and Moll and Inenlikej only six. These are summarised in Appendix 1. Work on Bogoras’ grammar of Chukchi was interrupted by the Russian Civil War, Revolution and the First World War, and ultimately Boas, rather than Bogoras, was responsible for the final arrangement of the materials in the published work (Bogoras, 1922: 634). There are suggestions in Soviet sources that Bogoras was dissatisfied with the final product (see Vdovin, 1954). Skorik’s list of Chukchi demonstratives and their meanings occurs in a footnote, implying that they are considered a class. His translations of the terms mostly include specification of various locations of the object with respect to the speaker, although in his analysis distance from addressee and visibility are also given as parameters in some cases. Skorik’s set of demonstratives reflects the Bogoras set quite closely, but this may be because they are based on Bogoras’ description, rather than both being accurate descriptions of the same phenomenon. The Moll and Inenlikej set are quite different. The number and identity of the terms they describe is the same as the set observed in usage in 1995–2000, although the glosses to the terms are not very transparent. The terms I do not believe are really part of the set of demonstratives are discussed below. ŋunqen. This term is only given by Bogoras, has never appeared in any of my texts and is not mentioned by Skorik or Moll and Inenlikej. It is possible this term appears in error or is dialect interference from some other Koryako-Chukotian language. ŋotənqen. This seems like a well-formed term, but I think it is the result of an analytic error. The terms waj=ənqen and raj=ənqen/ caj=ənqen are formed by means of a deictic particle (waj or raj/ caj) cliticized/fused to the neutral demonstrative stem, and so a term formed by the deictic particle ŋoot might also be

Chukchi Demonstratives

309

expected to be well-formed (assuming it had a long vowel, i.e. *ŋootənqen). However, ŋoot is not the same sort of deictic particle as waj or raj/caj: ŋoot forms the head of an (oneword, absolutive) NP, and, unlike the other deictic particles, cannot appear as an NP modifier. ɣanqen/ɣaanqen. This term does occur in my data, but it is not a demonstrative pronoun, it is a particle. Clear evidence of this is the lack of number or case forms, and its adverbial function in sentences. The term is also phonologically divergent: what is written here with the phoneme /ɣ/ is actually pronounced as a pharyngeal fricative, i.e. [χaanqen]. ŋanqen. Vowel length distinctions in Chukchi are not very well established (long vowels only occur as the result of a recent phonological change resulting in loss of intervocalic approximants; Dunn, 1999). The distinction between ŋanqen and ŋaanqen is illusory. Chukchi storytellers make frequent use of vowellengthening of the first syllable of distal demonstratives as a stylistic device to emphasise distance. The remaining terms (i.e. the complete Moll and Inenlikej set) all occur in my texts in spatial functions. As mentioned in the description of form classes, my investigations also show an auxiliary construction used in the exophoric modifier function (section 3.6). This construction does not appear in any of the dictionaries or the grammars; the former are concerned with single words, the latter are mostly focused on morphological alternations and have very little to say about syntax. 3

Results from the Questionnaire

The results reported here were collected using Wilkins’ (1999; this volume) Demonstrative Questionnaire. The questionnaire was carried out with six people aged from 30 to 60. The speakers were all first language Chukchi speakers, with near-native fluency in Russian acquired in monolingual Russian-medium boarding schools. All experimental subjects had tertiary qualifications (education or agriculture/veterinary) and were untroubled by the unnatural format of the task. Past experiences with older, near monolingual, Chukchi speakers made me unwilling to attempt this task with them: despite the obvious attraction (to a descriptively oriented linguist) of their metalinguistic naiveté and the lesser risk of Russian interference, such speakers tend to be puzzled by structured linguistic elicitation and may even be affronted that the researcher is apparently not interested in their life experience and ethnographic knowledge.

310

Michael J. Dunn

All situations were constructed in real space, as described in the questionnaire. The experimental subjects easily grasped what was required and regularly produced sentences containing modifier demonstratives as a response to most scenarios. 3.1

Proximal

The term ŋotqen is a proximal, which can be used for all scenes where the target is within the range of manipulation of the speaker. If the target is close to the speaker but is situated on or beyond the addressee, other factors must be considered. When the target is on the addressee, but the speaker is gesturing towards it or touching it, ŋotqen is still appropriate (scenes 2 and 4), but without gesture or contact (as in scenes 5 and 10), ŋotqen is not used. This suggests that ŋotqen is strongly speaker anchored. Boundaries set up different zones; a boundary between speaker and addressee on one side, and a close target on the other (scene 20) can allow use of a distal term as well. Conversely, a boundary between speaker and target on one side, and addressee on the other, brings speaker and target closer together, such that the proximal term can be used even when the target is somewhat beyond the range of manipulation. 3.2

Distal

The term ŋaanqen indicates a target distant from both speaker and addressee, although still within the range of visibility. The distance involved seems to be anywhere from a little beyond the range of manipulation right up to the horizon. Note that this term is not anchored on speaker alone. 3.3

Neutral

The term ənqen can apparently be used in all scenes at whatever range from speaker and addressee, and irrespective of visibility. It is most preferred for scenes in which the referent does not have a salient ‘proximal’ or ‘distal’ location. The more semantically specified terms are more commonly used when participant-anchoring is salient (Table 14.3). This system can be described in terms of scalar implicature. Note that the previous accounts of this term are all approaching elements of this function: Skorik describes ənqen as an addressee-anchored proximal, which is semantically incorrect, but pragmatically implied. Moll and Inelikey give the vaguest possible gloss (‘this’ or ‘that nearby’), which also concurs with my analysis of the spatial semantics of this term.

Chukchi Demonstratives

311

Table 14.3 The preferred and alternative (bracketed) demonstratives at different ranges from speaker and addressee

close ADDR far ADDR

close SPKR

far SPKR

ŋotqen (ənqen) ŋotqen (ənqen)

ənqen ŋaanqen, ŋoonqen (ənqen)

The demonstrative subsystem in Chukchi comprising the three terms ŋotqen, ənqen and ŋaanqen provides a structural intermediate between proximal/neutral/distal systems and near-speaker/near-addressee/distal systems, as well as a possible path for syntactic change between the two types. 3.4

Contrasting the Two Distals

The two distal forms ŋaanqen and ŋoonqen pose an analytic problem; ŋaanqen can be used at all ranges, whereas ŋoonqen can be used only at the most distant ranges. Meira (this volume) observes a similar opposition in Tiriyó, and his analysis seems also to apply here. The term ŋaanqen is a general distal, and the term ŋoonqen has emphatic meaning, so whereas ŋaanqen has a spatial element indicating ‘far’, ŋoonqen has a spatial element meaning ‘very far’. 3.5

The Terms waj=ənqen and raj=ənqen/caj=ənqen

These two terms do not seem to enter into semantic or pragmatic oppositions with the other demonstratives, which suggests that their meaning and function is in some way different. A couple of observations can be made on the basis of the questionnaire data and speakers’ metalinguistic reflections: The term waj=ənqen is strictly associated with a pointing gesture.5 The term raj=ənqen/caj=ənqen does not require a pointing gesture (although it still frequently occurs with one). Speakers claim that it has a sense of ‘beyond a boundary’, but I have not managed to elicit systematic responses sensitive to this. The distance component of these terms’ semantics is organised along slightly different principles: • The demonstrative waj=ənqen is used with targets near to speaker and addressee’s shared zone. Thus, waj=ənqen can be 5

The particular gesture used is usually the palm-down index finger point, which is used to indicate spatially contained targets, rather than the flat hand, vertically oriented, point used to indicate routes and paths.

312

Michael J. Dunn

contrasted to ŋoonqen, which is not used when the target is in the addressee’s zone (e.g. 5, 10, 16, 18). • The demonstrative raj=ənqen/caj=ənqen is used with targets far from speaker and far from addressee. Its range of application is very similar to that of the distal ŋaanqen but may also be used at any range when the target is beyond the addressee. The Wilkins questionnaire in itself does not generate evidence to account for how these two terms differ from the other demonstratives, but other observations of naturally occurring usage, and supplementary elicitation, suggest that these terms have an attention-directing function (see section 4). 3.6

Invisibility

Invisible targets are referred to using an auxiliary construction, which features a spatial adverb and a nominalised form of the locative copula. This construction was used in the questionnaire responses for targets which were also distal as well as invisible. There were two terms, nenku walʔən (‘which is there’; based on the adverb cognate to ŋaanqen and the substantivised form for the verb (-t)wa- ‘to be’) and ŋoonko walʔən (also ‘which is there’; based on the adverb cognate to ŋoonqen). These terms were apparently used interchangeably. Although it was not elicited as a response to any item in the Wilkins questionnaire, supplementary investigation showed that the analogous proximal construction ŋutku walʔən (lit. ‘which is here’) also exists and is also used when the target is invisible. 4

Naturalistic Observation: Beyond the Experimental Results

I carried out fieldwork investigations on Chukchi between 1995 and 2000, during which period I spent a total of some 15 months in the field. The observations reported below relate to this entire period. Chukchi is not the main language of daily communication in any of the areas where I was able to stay for extended periods; my database is biased towards narrative monologues by elderly people, supplemented by discussions of these texts with younger bilinguals. My experience includes relatively little observation of Chukchi used in spontaneous conversation, although I do have recordings of discussions rich in spatial demonstratives produced by people cooperating on picture matching tasks, such as ‘Man and Tree’ (Levinson et al., 1992). The first observation of demonstrative use from outside the instrument is that most demonstratives are only used spatially. The exception is ənqen, the spatially neutral term, which is the demonstrative used for all non-spatial functions, such as discourse deixis, the recognitional, temporal deixis, etc.

Chukchi Demonstratives

313

The existence of spatial-only demonstratives seems linguistically well motivated: if there are more spatial categories marked by demonstratives than there are narrative functions, at least some of the demonstratives must be specialized to spatial use alone. Levinson (this volume) points out a similar fact for Yélî Dnye, which has a rich demonstrative system involving many terms usable in spatial functions only. Thus, the following (11) was spoken from within one of the houses being discussed: (11) ŋotqen jara-tko-n sovqoc-ken PROX.DEM.ABS house-COLL-ABS state.farm-REL.ABS qənwettom ɣat-əmɣo-ɣʔe=ʔm Finally be.created-INCH-TH=EMPH ‘. . . this group of houses of the State Farm began to be formed.’ (he051)

When talking about a legendary person, the demonstrative ənqen is the only one used in non-spatial narrative functions, e.g. the recognitional function in (12). (12) qənwet ənqen teŋ-ənjiw rəju-lqət-ɣʔi=ʔm finally DEM.ABS good-uncle stand.herd-PURP-TH=EMPH ‘Finally the good uncle came to stand watch.’ (cy022)

Temporal deixis is also done with ənqen. Example (13) is from a historical narrative: (13) ənkʔam ənqena-jpə=ʔm qənur n-arojw-ʔaw and DEM-ABL=EMPH like HAB-strong-ADV nə-le-qin remk-ən miɣciret-ək HAB-go-3SG folk-ABS.SG work-INF ‘And thenceforth (lit. “from that”) it’s like people went stronger in their work.’ (he029)

The distance scale investigated by the questionnaire is subject to compression in suitably delimited spaces. All examples I have of this happening occur under conditions of explicit contrast between two objects. Thus, it is possible to build a contrast between ŋotqen and ŋaanqen in table-top space for a pair of objects located in closer and further zones, but it is not possible to use a distal term to pick one target object out of many. For example, ŋotqen (or its invariant identificational form ŋoot) is always used in identifying pictures on table top in the ‘man and tree’ task. The demonstratives waj=ənqen and raj=ənqen/caj=ənqen are used for attention direction (14–15). (14) waj=ənqen ətlon Rultəŋew PROX.PRES=DEM.ABS personal.name ‘There’s that Rultəŋew.’ [gesture towards the village]

314

Michael J. Dunn

(15) caj=ənqen neməqej ənp-əŋew dist.pres=DEM.ABS also old-woman ‘That woman too.’ [in the house over there behind you]

The questionnaire does not systematically control for attention, which makes it difficult to distinguish between presentative and non-presentative functions of demonstratives. The Chukchi data is enlightening here, since presentative and non-presentative functions are indicated by different forms. Example (16) shows a presentational demonstrative. (16) waj=ənqen ... meŋin / PROX.PRES=DEM.ABS who? ŋan Təle.lʔ-əwətr-əqej / ʔOmrən-en ekək DEF? pers.name-similar-DIM.ABS pers.name-GEN.ABS son-ABS ‘There’s that one . . . what’s he called, who looks like Təlelʔən, ʔOmrən’s son.’ (kr006)

Non-presentative demonstrative pronouns are not typically used with reference to invisible items (see section 3.6), unlike the presentative demonstratives waj=ənqen and raj=ənqen/caj=ənqen, which do not have any visibility restrictions. 5

Discussion and Proposed Analysis of the System

From the number of terms present, the Chukchi demonstrative system is necessarily complex (Figure 14.1). It can, however, be divided into a number of subsystems, and these subsystems do not look so exotic in comparison to the other languages which have been studied under this method. The ‘backbone’ of the system is the contrast between the pragmatically marked proximal and distal terms ŋotqen and ŋaanqen, and a pragmatically unmarked, neutral term ənqen which can refer to any distance. This contrast is similar to Lavukaleve SPATIALLY NEUTRAL nqen e

SPATIALLY SPECIFIED

PROXIMAL otqen

DISTAL aanqen oonqen } e

e

e

waj= nqen

raj= nqen/caj = nqen

Figure 14.1 The exophoric demonstratives in Chukchi along a proximal–distal cline

attentiondirection

Chukchi Demonstratives

315

(see Terrill, this volume) The distal term ŋaanqen can also be opposed to a ‘far’ or ‘emphatically’ distal ŋoonqen, which functions in the same way as the twoway distal contrast in Tiriyó (see Meira, this volume). Independently of these four terms is a two-way contrast between attention-directing demonstratives: proximal waj=ənqen and distal raj=ənqen/caj=anqer; lack of the third term here would seem to follow naturally from the semantic incompatibility of an unmarked, neutral term with the markedness of the attention-directing function. Thus, despite having a large number of different spatial terms, the Chukchi demonstrative system ultimately can be resolved to a set of semantic parameters which are very similar to those in systems with many fewer terms. Abbreviations ABSolutive (singular unless specified plural), ADVerb, ALLative, CauSitive, COLLective, CONCurrent, DEMonstrative (neutral unless specified proximal or distal), DIMinutive, DISTal, EMPHatic, ERGative, FAR DISTal, GENitive, HABitual, INCHoative, INFinitive, INTentional, ITERative, LOCative, NEUTral, NoMinaliZeR, PartiCLe, PERFect, PLural, POSSessive, PRESentative, PROXimal, PURPosive, RELative, SinGular, THematic affix, TRansitivity marker, – morpheme break, = clitic

Appendix 1 Descriptions of Chukchi spatial demonstratives from Bogoras, 1922; Moll and Inenlikej, 1957 and Skorik, 1961–1977

ŋotqen

Bogoras, 1922

Skorik, 1961–1977

Moll & Inenlikej, 1957

this

‘this’ (closer to the speaker than the addressee, and visible to the speaker or both) orig.: этот (ближе к говорящему, чем к слушающему, видимый обоим или только говорящему) ‘that’ (closer to the addressee than the speaker, and visible to the addressee or both) orig.: тот (ближе к слушающему, чем к говорящему, видимый обоим или только слушающему) ‘that there’ (distant from speaker and addressee, and visible to both) вон тот (в отдалении от говорящего и слушающего, видимый обоим)

‘this’ MASC., FEM., NEUT.

(Bogoras wotqen) ənqen

ŋanqen

that

that there (not very far away)

orig.: этот, это, эта 1. ‘that’ (located nearby); 2. ‘this’ orig.: 1. тот (находящийся поблизости); 2. этот –

316

Michael J. Dunn

(cont.) Bogoras, 1922

Skorik, 1961–1977

Moll & Inenlikej, 1957

ŋaanqen

that yonder

ŋoonqen

that yonder

‘that there’ (located far from the speaker) orig:. вон тот (находящийся вдали от говорящего) ‘that’ (located far from the speaker) orig.: тот (находящийся вдали от говорящего)

ɣanqen (Bogoras)

there (quite far)

‘that yonder’ (further than ŋanqen, visible to both) во-он тот (дальше, чем ŋanqen, видимый говорящему и слушающему) ‘that yo-o-onder’ (further than ŋaanqen, visible to both) во-о-о-н тот (дальше, чем ŋaanqen, видимый говорящему и слушающему) ‘that yo-o-o-o-onder’ (very far, further than ŋoonqen, visible to both) во-о-о-о-н тот (очень далеко, дальше, чем ŋoonqen, видимый говорящему и слушающему) ‘this here’ (with a gesture towards a distant object visible to both speaker and addressee) вот этот (с указанием на отдаленный предмет, видимый говорящему и слушающему) ‘that there’ (located behind the addressee) вон тот (находящийся за слушающим) ‘that there’ (located behind the addressee) вон этот (находящийся за говорящим) –

ɣaanqen (Skorik) wajənqen

that there (midway to some object)

rajənqen

that behind the person addressed

ŋotənqen

that behind the person speaking

ŋunqen

that apart from the speaker

–

1. ‘this there’; 2. ‘that there, that one’ orig.: 1. вон этот; 2. вон тот, вон он ‘this here; this here behind you’ orig.: вот тот; вот тот, за тобой –

–

References Bogoras, W. (1904–1909). The Chukchee. Leiden [New York]: Jesup North Pacific Expedition. Publications vol. VII, E. J. Brill [Memoirs of the American Museum of Natural History in New York. Vol. XI, Parts 1–3, G. E. Stechert & Co.]. (1922). Chukchee. In F. Boas, ed., Handbook of American Indian languages. Washington, DC: Government Printing Office, p. 2. Dunn, M. (1999). A grammar of Chukchi. Doctoral thesis, Department of Linguistics, Australian National University, Canberra.

Chukchi Demonstratives

317

(2000a). Planning for failure: The niche of standard Chukchi. Current Issues in Language Planning. Special issue: Language planning and language ecology, 1 (3), 389–399. (2000b). Chukchi women’s language: A historical-comparative perspective. Anthropological Linguistics, 42(3), 305–328. Inenlikej, P. I. (1976). Chukotsko-russkij russko-chukotskij slovar’ [Chukchi-Russian Russian-Chukchi dictionary]. Leningrad: Prosveshchenie. Levinson, S. C., Brown, P., Danziger, E., De León, L., Haviland, J. B., Pederson, E. & Senft, G. (1992). Man and tree & space games. In S. C. Levinson, ed., Space stimuli kit (version 1.2): November 1992. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 7–14. Moll, T. A. & Inenlikej, P. I. (1957). Chukotsko-russkij slovar’ [Chukchi-Russian dictionary]. Leningrad: Uchpedgiz. Skorik, P. J. (1961–1977). Grammatika chukotskogo jazyka [Grammar of the Chukchi language]. Moscow/Leningrad: Izdatel’stvo akademii nauk SSSR. Vakhtin, N. (2001). Yazyki narodov Severa v XX veke: ocherki yazykovogo sdviga [Languages of the peoples of the North in the 20th century: Essays on language shift]. St Petersburg: Dmitrii Bulalil. Vdovin, I. S. (1954). Istorija izuchenija p-aleoaziatskix jazykov [History of the study of the paleo-asian languages]. Moscow/Leningrad; Academija nauk SSSR. Wilkins, D. P. (1999; this volume). The 1999 demonstrative questionnaire: ‘This’ and ‘that’ in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24.

15

Yélî Dnye: Demonstratives in the Language of Rossel Island, Papua New Guinea Stephen C. Levinson

1

Introduction

The demonstrative system of Yélî Dnye, the ‘Papuan’ language of Rossel Island, is interesting in a number of respects. Although it could be readily misconstrued as a standard ‘this, that, yonder’ system based on distance, the system is in fact a complex multi-dimensional one that invokes a number of semantic dimensions, including epistemic and anaphoric parameters, but also other contributory factors like attention, touch, evidentiality and unmarkedness. Despite the many parameters actually involved in the meaning of these demonstratives, the field instrument used here (see Chapter 2, this volume), which focuses primarily on spatial contrasts, nevertheless proved very useful in unravelling the semantics of these terms. 2

The Language and Its Speakers

Yélî Dnye, literally ‘Rossel island sounds’ (earlier known variously as Yele, Yela, Yelentye, or simply Rossel Island language) is a Papuan – i.e. nonAustronesian – language spoken on Rossel Island in the Louiseade archipelago, Milne Bay Province, which lies 450 km off the coast of mainland Papua New Guinea. Yélî Dnye is an isolate whose affiliation to any other languages has not been definitively established. There are about 5,000 inhabitants of Rossel, including a few married-in native speakers of Austronesian languages (especially the languages of Sudest, Misima and Nimowa). The people are physically distinct from surrounding peoples, and the working assumption is that they, at least partially, represent genetically, historically, culturally and linguistically the pre-Austronesian peoples who presumably filled the whole of near Oceania before Austronesian expansion (Levinson, 2006d). Recent interdisciplinary work shows that the prehistory of near Eastern Papua is complex, with Austronesian peoples arriving about 3,000 years ago in the

318

Yélî Dnye Demonstratives

319

Solomons to the north-east of Rossel, finding earlier occupants going back some 30,000 years. On top of these migrations, there have been migrations of non-Austronesian peoples (e.g. in the Santa Cruz islands) and much language shift to Austronesian languages, with a consequent mosaic of languages, genes and cultural items (see, e.g., Hunley et al., 2008; Delfin et al., 2012). Genetic samples have been collected on Rossel Island but remain, frustratingly, largely unpublished; however, reports suggest that the mtDNA is special to the area (Friedlaender et al., 2005; 2007), and the Y-chromosome appears also distinctive of Rossel in particular, with a large dose of Asian genes representing Austronesian contact (Van Oven et al., 2014). Recent unpublished archaeological work on Rossel Island by Ben Shaw (2015) has found pre-Austronesian material but of no deep antiquity. Linguistically, Yélî Dnye has less than 5 per cent Austronesian loans, but interestingly these connect to proto-Oceanic not the current surrounding languages. Since there are no clear cognates with other Papuan (i.e. nonAustronesian) languages, we conducted a phylogenetic analysis based on morphosyntax and phonology with the other offshore East Papuan languages: on this analysis Yélî Dnye remains an outlier (Dunn et al., 2005; 2007; 2008). Yélî Dnye is the single predominant language on the island (little or no Tok Pisin is spoken in the area – English being the provincial linga franca), although many younger people also know a considerable amount of English through schooling (in English) or outside employment. Rossel is a remote island surrounded by difficult seas, served by few vessels and no air strip and is quite isolated. Before my own research, the only substantial work on Rossel Island language is the brief but invaluable grammar sketch by James Henderson (1995), an SIL linguist who, together with Anne Henderson, translated the New Testament, produced word lists (Henderson and Henderson, 1987) and in numerous ways encouraged literacy in the language. In what follows I shall employ the practical orthography described in that grammar, and my glosses are based roughly on Henderson’s analysis of the verb complex. Although surrounded by Austronesian languages, Yélî Dnye shows little evidence of influence by them, and with its huge phoneme inventory and complex grammar is scarcely ever mastered by outsiders. The language has 90 phonemes (a record for the Pacific), including phonetic distinctions unique in the languages of the world (Levinson, in prep.). The syntax exhibits free phrase order but a predominant SOV pattern, with postpositions and adjectival modifiers following nouns, but no systematic leftbranching (modifiers tend to come after heads). The language is ergative and is a rarity in exemplifying ergative syntax, so that all major syntactic operations are organized on an ergative/absolutive basis, not on a nominative/accusative

320

Stephen C. Levinson

one. Nominals are marked in an ergative/absolutive pattern, but the crossreferencing on the verb follows a partially nominative/accusative pattern. The cross-referencing is manifest in pre- and post-verbal clitics, which are portmanteau morphs carrying tense, aspect, mood, person, number information, as well as a number of other features including conditionality, counterfactuality and – pertinent to this chapter – deictic information. Consequently, there are well over a thousand such morphemes (or form–meaning pairs, although complex syncretism reduces the actual number of forms). Argumentstructure alternations are confined to object incorporation, other alternations (e.g. of transitivity) being marked by lexical doublets. Verbs generally supplete but in an irregular way on aspect, tense, mood and sometimes on person. The language taken as a whole, like many other Papuan languages, is a dauntingly complex system at almost every level and awaits full description (Levinson, in prep.). Publications other than Henderson (1995) explore, for example, the colour terms (Levinson, 2000a), landscape terminology (Levinson, 2008), positional verbs (Levinson, 2000b), the kinship terms (Levinson, 2006a), spatial and temporal description (Levinson, 2006c; Levinson and Majid, 2013), body parts (Levinson 2006b), verbal semantics (Levinson, 2007a; Levinson and Brown, 2012), reciprocal constructions (Levinson, 2011) and, most relevantly, person reference (Levinson, 2007b). 3

The Demonstratives in the Context of Yélî Dnye Grammar

Yélî Dnye has extensive deictic systems, covering, for example, reference to the person, social statuses, time and place of speaking. The person system makes distinctions between three persons and three numbers (singular, dual, plural). Social deixis is expressed, for example, in a taboo vocabulary reserved for speaking to or in the presence of in-laws. Temporal deixis includes six absolute tenses in indicatives (on the pattern of events earlier today, yesterday or before, later today, tomorrow or later) and two in imperatives (do it now versus later), and partings include a (non-compositional) specification of how many days from today one expects to meet the addressee again (Levinson and Majid, 2013). Spatial deixis includes not only demonstratives (the focus here) and adverbials but also deictic morphemes attached to pre-verbal clitics indicating ‘hither’ and associated motion markers indicating ‘do while going’ (Levinson, 2006c). Demonstratives belong to a special part of speech, namely determiners preposed to the head nominal phrase, in a slot in which only possessive pronouns compete. Since there are no clear articles in Yélî Dnye, certainly in that slot, the demonstratives exhaust that form-class. The NP is built according to the following template (Levinson, in prep.):

Yélî Dnye Demonstratives

Demonstrative Possessor

321

(N) N-Specifier-Adjective-Classifier-Indef-Quantifier-Plural-Postpostion

Nearly every element of this phrase is optional, including (surprisingly enough) the head N, marked in bold here, although some co-occurrence constraints exist. Demonstratives and possessive pronouns are the only items that can occur before the head, apart from nouns (here marked in brackets) which form a compound with the head (bare nouns without demonstratives or possessors tend to be interpreted indefinitely). In addition, demonstratives trigger a specifier, a bound affix following the head, irregularly formed or formed according to three classes of noun that may be distinguished in this respect. The first class takes no specifier (as in ala yââ ‘this leaf’); the second class adds ni to the noun (as in ala pi-ni ‘this man’), which may have irregular repercussions on the stem, often a vowel raising as in mbwaa ‘water’ ➔ ala mbeeni ‘this water’. A third class of nouns has a suppletive form of the root for the specified form (often built with one of the forms -li/-pi/-pu/-mi/-mu) as in pyââ ‘woman’ ➔ ala pywópu ‘this woman’. The possessive pronouns do not trigger these specifiers (cf. k:ii ‘banana’, a k:ii ‘my banana’, ala ki-ni ‘this banana’). Therefore, on formal grounds alone the demonstratives constitute a formclass. In addition, the demonstratives are determiners, not pronouns, so in order to play a role as pronouns they must occur before the pronoun n:ii (which otherwise occurs as a relative pronoun), replacing the position of the head noun, as in: (1)

ala tpile versus ala n:ii ‘this thing’ ‘this one’

Thus one might ask: ló n:ii? ‘which one?’, and receive the answer ala n:ii ‘this one’. A further complex property of demonstratives is that they also occur in the pre-verbal clitic slot, where they irregularly fuse with the portmanteau morphs indicating tense/aspect/mood/person/ number (of subject). In this role they seem to play epistemic functions, indicating, for example, certainty or uncertainty, as well as deictic roles (see Henderson, 1995: 48–55), an issue taken up in section 4. To appreciate the subtleties of the Yélî Dnye demonstratives, it is useful to approach them through a series of approximations or insufficient analyses, each of which can be shown to be inadequate. Suppose we took as a working assumption that the four basic demonstratives were organized primarily through a hybrid of interlocutor anchoring and spatial distance measure.

322

Stephen C. Levinson

Table 15.1 An (inaccurate) first approximation: The core set of spatial demonstratives

Proximal Medial (neutral) Distal

Speaker-based

Addressee-based

ala kî mu (far from Spkr)

ye – (potentially equivalent to ye)

Table 15.2 Demonstrative pronouns and corresponding adverbs, according to the first approximation in Table 15.1

Proximal Medial (neutral) Distal Anaphoric

Pronouns

Adverbs

ala n:ii kî n:ii mu n:ii yi n:ii

al:ii k:ii mw:ii y:i

‘here’ ‘there’ ‘yonder’ ‘there as mentioned’

We would then propose an analysis like that in Table 15.1, and we would find it consistent with much functional usage. The terms have no internal morphological structure, and are thus semantically opaque, with the possible exception of the proximal ala, which could possibly be diachronically related to the first person possessive (cf. a-la ‘my-bit’). The proposed distance metric in the speaker-based series needs to be understood as relative: all the terms can be employed to distinguish, for example, objects on a table top (see below) but equally to refer, for example, to a string of villages at some kilometre distances along the coast. These terms only form a core: in addition, there is an anaphoric determiner yi, and a number of others to be mentioned below. Corresponding to these demonstrative determiners are a set of demonstrative adverbs, as in Table 15.2, again here analysed according to a simple distance metric. Incidentally, it is possible to combine the pronominal n:ii and the adverbial series, as in al:ii n:ii ‘the here one’, but such uses hardly seem to occur. To test this first approximation, data were collected using the Pederson and Wilkins (1996) ‘table-top placement’ task.1 In that task, one, two or three objects (e.g. cups) were placed in various arrangements in front of a speaker, with the investigator beside him or her. The results can be illustrated diagrammatically as in Figures 15.1–15.3. 1

See also the entries in http://fieldmanuals.mpi.nl/projects/demonstratives/.

Yélî Dnye Demonstratives

323 ala

ala

kî

Figure 15.1 Demonstratives used (with pronoun n:ii) for single objects on a table ala kî

ala

mu

kî mu

Figure 15.2 Demonstratives used (with pronoun n:ii) for two objects on a table ala

kî

mu

Figure 15.3 Demonstratives used (with pronoun n:ii) for three objects on a table

This kind of data supports the first-order approximation as presented in Table 15.1, with a series of three speaker-distances distinguished, where the relevant distance is partly a function of the contrasts to be made. Incidentally, these distance distinctions are neutralized for an array of three objects in transverse order across the speaker’s line of gaze: then the same three terms as in Figure 15.3 are employed, arbitrarily starting to left or right (unlike the pattern in Figure 15.3 where order of reference is not at stake, only distance). If the speaker turns his back on the array, the same three terms are employed as in Figure 15.3 (showing that visibility is not a necessary feature for these three terms). If the array is vertically arranged, with one object on the floor, another at navel height and another at head height, the system is neutralized and ala (proximal) is used for the object on the floor, kî for the object at navel height and mu (distal) for head height. If the addressee is not beside the speaker as is presumed in Figures 15.1– 15.3, but at the other end of the table, as in Figure 15.4, the same terms are employable, except that the object denoted mu in Figure 15.3 can (but need not) be equally well designated ye ‘near addressee’:

324

Stephen C. Levinson

Table 15.3 Henderson’s (1995: 46) analysis of the deictics Deictic term

Referring use

Anaphoric/cataphoric use

kî wu ala ye yi mu

‘in sight’ ‘out of sight’ ‘close to speaker’ ‘close to hearer (addressee)’ ‘previously discussed’ ‘the other’

anaphoric cataphoric anaphoric anaphoric cataphoric

ala

kî

mu/ye

speaker addressee

Figure 15.4 When addressee is opposite speaker

The reader should now have a good idea of the basis for the generalization in Table 15.1 above. This pattern can be repetitively elicited, but it is not in fact an adequate analysis of the system, which actually involves additional semantic parameters. For there are in fact six, not just four demonstrative determiners – Henderson (1995) gives (without any further details) the table above (Table 15.3). Table 15.3 correctly suggests that the terms have additional functions in anaphora, and also that there may be epistemic issues in play, not just distance from speaker. I will dispute some of these glosses, but the insistence on additional factors is correct. In fact, to preview some of the main results, I will argue that the correct analysis is semantically multi-dimensional, requiring distinctions on three dimensions: spatial distance, discourse location and epistemic basis. The analysis can be sketched in the diagram opposite (Figure 15.5), which treats each of these dimensions as the side of a cube. Take the spatial dimension first, here shown vertically. As we have seen, ala and ye both indicate proximity (here marked ‘+ Close’), in the first case to the speaker, and in the second case to the addressee. At the other extreme, mu indicates non-proximity (here marked ‘− Close’). On a different dimension is kî, which will be argued below to be in fact neutral or unmarked for distance (hence marked ‘+/− Close’). On the horizontal dimension, we have the unfolding of discourse in time. For referents that are behind in discourse time, the special anaphoric pronoun yi is mostly employed; for referents that are ahead in time cataphoric reference can be made with spatial series (as in any

Yélî Dnye Demonstratives

cataphoric

DISCOURSE

mu -Close +Cert

325

anaphoric

yi +anaphoric

SPACE

wu -Cert kî ±Close +Cert

ala OR ye +Close to +Close to Addressee Speaker +Cert +Cert

IC

EM

T PIS

E

Figure 15.5 The three dimensions of Yélî Dnye deictic determiners

multi-dimensional plot, yi therefore contrasts with all the items to the left of the space). The final dimension involves an evidential or epistemic parameter. On this parameter the spatial series contrasts with an additional deictic determiner wu ‘that (indirectly inferred)’. The evidence for much of this analysis will be presented in the sections below. 4

Results of the Wilkins Demonstrative Questionnaire (1999)

The Wilkins questionnaire (Chapter 2, this volume) was run twice with my main assistant (Isidore Yidika) on different occasions, and with two other consultants, with individual scenes being checked with yet further consultants. Where practical we acted out the scenes – and thereby found that, for example, if the speaker in scene 7 is sitting on a chair (a rare object on Rossel island), then ala ‘this proximal’ is relatively unnatural in reference to a book on the floor in front of him, whereas if sitting on the floor it would be fine. Many variants of each scene were therefore enacted or scenarios set up by verbal description. The main results of the questionnaire are given in Table 15.4 below.

A

A

Ye Referent is in contact with addr; prior attention by addr essential

(2) (could be ala if spkr touching!)

S

A

Ye / Kî Choice depends on whether addr is attending to referent; kî is default option

(5) (kî if not noticed by addr)

S

S

S

(17)

S

(14)

S A

(12)

Kî

A

A

(4) (kî if not noticed by addr)

S

(9) (ye only if addr touching referent)

A

Ala Ala / Kî S pointing at or ala if spkr is pointing touching referent,or close to or touching so close it is in spkr’s referent, otherwise space kî

S

(7) (ditto)

S

A

(4) (must have been noticed by addr)

S

A

S

(19)

S

A

(6) (ala only if pointing or touching)

S

(16)

S

(8) (ala still poss, if pointing and close)

(3)

(1)

S

A

(11) (wu possible)

S

A

A

A

A

S

A S

Kî

(20)

S

(23)

A

(22)

Ye / Mu Depending on whether addr is holding it or not

(18) (ye if addr holding, otherwise mu)

A

S

A

Kî / Mu (kî requires pointing, mu requires referent distant or Spkr can’t see it)

(13) (mu is OK here for small thing far away)

S A

(10) (kî unless spkr can’t see referent)

S

S

Mu distant

(25) (mu required if can’t see referent)

A S

(21) mu

A

(24) mu

A S

Table 15.4 Yélî Dnye demonstratives used in the scenes in the Wilkins (1999) Demonstrative Questionnaire

A

Wu For unseen object which addr doesn’t know about

(11) wu or kî/ala

S

(15) wu

S A

Yélî Dnye Demonstratives

327

Table 15.5 Another arrangement of Yélî Dnye demonstratives (The arrows below the table show the direct contrast between close to speaker or addressee versus far from both, and the indirect opposition between those marked forms and the unmarked kî.) +CLOSE, +ATTENTION

SPKR

+CLOSISH, +ATTENTION

Ala

UNMARKED

Ala/Kî Kî

ADDR

Ye

-CLOSE FROM SPEAKER

+DISTANT FROM SPEAKER

Kî / Mu

Mu

(or Ye)

(or Ye)

+UNCERTAIN (to S & A)

Wu

Ye/Kî

Mu

Ala OR Ye

Kî

Table 15.4 should be understood as a simple grouping of scenes under the main deictic determiners. It is not in fact a revealing grouping, which is hard to achieve on a single page complete with columns of thumbnail pictures, because the Yélî Dnye system is, as mentioned above, multi-dimensional, so resisting a single close-to-far array of scenes. A more revealing grouping is schematically indicated in Table 15.5, where closeness to speaker versus closeness to addressee are treated in parallel, and an additional column at the end picks up the epistemically marked form. In Table 15.5, some attempt has been made to bring the multi-dimensional Rossel system into the same format as the other systems described in this book, with ‘proximal’ forms to the left, and ‘distal’ to the right. But clearly, this is artificial, since mu is distal only from the speaker, and sometimes a referent can be far from speaker and close to addressee, so alternatively mu or ye. Figure 15.5 above is a better representation of the multi-dimensional character of the system, but nevertheless, this table is a good point of departure for a description of the main findings from the questionnaire. The major findings of the study are summarized in the following sections. 4.1

The Nature of the Speaker-centered Distance System: Ala ➔ Kî ➔ Mu

From the direct responses and following discussions of each scenario, it became clear that the ‘medial’ term kî is employable in most scenarios. Kî

328

Stephen C. Levinson Kî (unmarked)

Ala (+Close to S)

Mu (+Dist)

Figure 15.6 Unmarked vs marked oppositions in the speaker-oriented series

becomes pre-empted under specific conditions, essentially when the referent is close to or touching speaker or addressee, or when it is very far from both (a statement to be refined below). Following the analysis I first proposed for English that (Levinson, 1995; 2000c: 94; 2004), it seems that kî is actually unmarked for distance – a ‘neutral’ demonstrative in the terminology of this book. This predicts that in fact kî could logically occur for referents at any distance, and the fact that ala and mu occur at close and far distances respectively is a matter of pragmatic pre-emption. Such pre-emption would follow from Grice’s (1975) first maxim of Quantity, ‘Make your contribution as informative as is required’, which enjoins a speaker to use the most informative description that applies – with the result that the addressee interprets a less informative description (here kî) as suggesting (implicating) that the more informative description (say ala or mu) does not apply. In this way, a division of labour between alternative forms arises, without an actual lexical specification of a contrastive meaning. The advantage of such an analysis is that it accounts for the fact that kî is acceptable nearly everywhere, although sometimes misleading – in other words, it accounts for the flexibility of demonstrative use. Since this kind of marked/unmarked opposition is probably very general in demonstrative series (cf. English this marked with that unmarked), it is worth sharpening the theory. Diagrammatically, we can represent the speaker-based series as having overlapping semantics, as shown in Figure 15.6. The division of labour between the three terms then arises by pragmatic principle: where terms are in privative opposition in this way, a ‘Horn-scale’ arises under which the use of an informationally weaker term systematically invites the inference that the stronger does not apply. The proximal forms ala and the distal mu pragmatically pre-empt kî, and thus the use of kî Q-implicates ‘not next to S and not distant from S’ (see Levinson, 2000c; 2004 for the theory). A further advantage of this analysis is that it is not just distance from speaker that lies behind the division of labour between the terms. In fact, the conditions for use of the two marked terms are quite specific:

Yélî Dnye Demonstratives

329

Ala: requires touching, or close pointing – even scene 4 where the speaker points to a ‘bug’ (or insect) on the addressee’s shoulder is marginal, unless the speaker is almost touching the bug. A possible interpretation is that ala can be used only if the scenario meets two conditions: (a) Referent must be in the speaker’s reaching area. (b) Referent must be the current focus of the speaker’s attention (hence pointing or manipulating with the hand or directional gaze seems an obligatory precondition for the use of ala). Notice we have now introduced yet another dimension, not captured in Figure 15.5, namely focal attention, a parameter known to play a systematic role in e.g. Turkish demonstratives (Küntay & Özyürek, 2006). Mu: requires certainty of location but lack of direct contact.2 Sheer distance is just one manifestation of ‘lack of direct contact’ in the intended sense. For example, if some object intervenes between speaker and referent, that creates the relevant kind of ‘virtual distance’. Hence scene 10 is a potential mu scene, where speaker and addressee are sitting down side by side, and the addressee is in between the speaker and the referent. In contrast, in scene 11, when the object is at a similar distance and similarly not visible to the speaker, but nothing intervenes between speaker and referent, mu is not possible. It follows from the fact that ‘virtual distance’ may involve occlusion that pointing is not obligatory with mu, unlike with ala (which requires pointing, gaze or touch). Because of these specific conditions, a semantic division of labour between ala, mu and kî is not a tenable analysis: kî would then need to be semantically specified with a detailed ad hoc list of negative conditions of the sort ‘not involving close pointing or touching with attentional focus’, ‘not involving lack of direct contact’ and so forth. And it would be hard then to account for the overlapping scenes diagrammed in Table 15.4. These problems are averted by the pragmatic analysis, where kî picks up the residue from the more specific demonstratives by pragmatic (and thus defeasible) opposition. 4.2

Addressee-centred Deictic: Ye

The addressee-centred deictic ye appears to have similar conditions to ala, except they now apply to the addressee rather than the speaker of course. 2

The ‘certainty’ feature may itself be implicated by contrast to wu, ‘referent with lack of direct perception, or uncertain location’, to be described below.

330

Stephen C. Levinson

Ye: Scenarios for the use of ye must meet two conditions: (a) The referent must be touched by the addressee, or at least within the addressee’s reach. (Note that these conditions could apply to an event, e.g. if I hear you fall down the stairs, I could say Ye lukwe? ‘That-near-you what? i.e. what’s happened to you?’). (b) The Addressee must already be attending to the referent (cf. the corresponding condition on the speaker for the use of ala). Hence the referent in scene 4 (bug on addressee’s shoulder) is more likely to be described with kî than ye – for the most natural interpretation of the scene is that the speaker is drawing the attention of the addressee to a perhaps dangerous insect on his or her own body. Both conditions seem to be generally necessary, although if you are unconsciously fiddling with a pen, I could perhaps designate it by ye, as an exploitation of the conventions (‘that pen that you should be aware of’). Generally it was felt that the most natural scenario for the use of ye was scene (16), where the speaker is distant and the addressee has the referent right in front of him. Notice that ye does not necessarily pre-empt the speaker-based system – thus scene (18) can be mu (distant from S) or ye (close to A), with mu preferred because of the ‘virtual distance’ caused by visual occlusion, while ye would be the more informative choice if the addressee is actually grasping the referent. 4.3

Wu Marks Epistemic Uncertainty (Henderson’s ‘Invisibility’)

From the point of view of spatial distance, wu is ‘off-scale’ (hence the relevant column is placed at the end of Tables 15.4 and 15.5), as it is clearly specialized to referents with uncertain epistemic conditions, regardless of distance. It seems to have two rather different and specific conditions of use, at least one of which must apply to the scenario of use. The first condition listed below is especially relevant for the opposition with mu ‘that-distal’ (as in scene 15), and the second for opposition to the proximal and neutral forms ala and kî (as with scene 11). Wu: Scenarios for the use of wu must meet at least one (or both) of two conditions: Condition 1. ‘Indirectly ascertained referent’: (a) Suppose we see a ship’s lights far out to sea – the ship can be referred to as mu (distal); but if the lights now disappear (e.g. behind a mist bank or if they are switched off), the relevant demonstrative is wu (cf. scene 25). In contrast, for an off-shore islet invisible in the mist but whose location is well known, mu is

Yélî Dnye Demonstratives

331

the relevant term, not wu, showing that invisibility is not the issue, but indirect inference is. (b) Suppose there is a noise outside: Wu lukwe ‘What’s that?’ is the natural query. (c) Suppose there is a book behind me, which I can indirectly infer from the feel of it in my back, wu would be possible in this version of scene 11, otherwise we would expect ala or kî. Condition 2. ‘Referent not part of shared common ground’: (a) If speaker knows the referent but the addressee doesn’t (e.g. for scene 1 the following is a natural interchange): a: I hurt this ala tooth b: Which one? a: This wu one (demonstrates invisibility not the criterion) Similarly for scene (15), the speaker would refer to a distant book behind the 3rd person as wu puku dmi if the addressee cannot see it and does not know about it or its location (b) If neither speaker nor addressee know for sure where the referent is (e.g. in scene 21 if someone told us that the referent is out there somewhere beside a tree but we have no direct evidence for its location). Henderson (1995) proposed (as noted in Table 15.3) that kî and wu are opposed as ‘visible’ versus ‘invisible’ respectively. We have now seen that wu may not be used for known but invisible locations (like the invisible but familiar offshore island mentioned in condition 1 as a variant of scene 25). Equally, it may be used for visible locations – like the tooth the speaker is pointing at in scene 1 – where there is some doubt about which referent is at issue. What about the other side of the opposition: is kî used only if the referent is visible? We have already seen that in scene 11 the speaker can refer to a book behind himself, for example, as kî puku dmi. Similarly, for someone who has just left, kî n:uu ‘That is who?’ is as good as wu n:uu ‘That-indirectlyascertained is who?’. If an object is hidden under a cloth, kî or ala are just as good as wu (in fact, in the ‘Walnut Game’ – see Pederson and Wilkins (1996) – in which the referent was hidden under three pots, kî/ala/mu all occurred multiple times, and wu just once). On the other hand, in scenes 18 and 25 where a speaker points at a distant object hidden from view, kî seems not usable, perhaps because both indirect access and far distance are combined. In sum, as suggested already in the ‘cuboid’ model for the Rossel demonstratives (Figure 15.5), the data requires a broader gloss than visibility: kî marks relatively direct perceptual access (‘certainty’), and wu marks indirect or

332

Stephen C. Levinson

inferential access to the referent, or difficulties of access for addressee or both speaker and addressee (‘uncertainty’). Such an analysis, unlike the visibility analysis, is fully consistent with the meanings of these same morphemes in a pre-verbal position to be described below.3 4.4

Yi: Dedicated Anaphoric Determiner

The questionnaire is not an instrument for the investigation of anaphora. Nevertheless, since the scenarios called forth little dialogues, anaphoric usages occur. Henderson (1995: 46) notes (see Table 15.3 above) that all the demonstratives except kî can have anaphoric (or cataphoric) usages, but in the scenario-driven dialogues it is the determiner yi that appears, as illustrated: Scene 14 (referent in midst of large space like a football field) Speaker: kî puku dmi njimi u puku dmi that-unmarked is Jimmy’s book. Addressee: nyââ yes Speaker: yi puku dmi u yi a nga ka kwo that-anaphoric book I want (lit. its desire is standing to me)

Yi is a determiner that can have only anaphoric (and not cataphoric) usage, and the possession of such a term probably has fundamental effects on the structure of any demonstrative system. For one thing it breaks the normal grammaticalization chain from demonstrative to complementizer via cataphora (‘John said that: he will come’ ⇾ ‘John said that he would come’; see Heine et al., 1991: 180). 4.5

Gesture, Attention, Mutual Knowledge and Choice of Demonstrative

There are a number of further conditions on the use of the terms, some mentioned already, but best compared together. First, note the correlation with gesture in Table 15.6. The obligatory gesture associated with ala and ye – the two proximal forms – is not necessarily or even protoypically a pointing gesture: the essential ingredient is that the referent is in some sense presented, either by picking it up, touching it or close pointing. Now, there are exceptions to these gestures, when, for example, ala is used ‘symbolically’ rather than ‘gesturally’ (as Fillmore, 1997 puts it), as in ala dyámê ‘this island, this world’ said in the middle of Rossel Island, for example, or when it is used temporally, as in ala ngwo ‘now, 3

It may be wondered whether, if wu is marked ‘uncertain’, kî is only in pragmatic or privative (unmarked) opposition. At least in the pre-verbal clitic position, it seems to do so, as discussed below.

Yélî Dnye Demonstratives

333

Table 15.6 Basic collocations with gesture + Gesture

ala, ye

± Gesture – Gesture (unnecessary, redundant)

kî, mu, wu yi (anaphoric)

lit. this time’. But for reference to physical objects that can be visually individuated, a gesture or presentation is necessary. The gestural prescription for the proximal deictics is further grounds for considering these terms to be the semantically marked terms, with the most preconditions on their use. (For a study of Yélî Dnye deictic reference in action, with details of gesture and gaze, see Levinson, 2007b.) Pointing or presenting is a way of getting the addressee to focus his/her attention on the referent and thus serves to individuate it. For the speaker to refer entails that the speaker has the referent in his or her attention, but obviously there is no such entailment for the addressee – part of the job of a demonstrative is to achieve this mutual attention. Nevertheless, there seems to be a curious precondition on the use of the addressee-proximal ye, namely that the addressee must already be attending to the referent. Scene 4 was particularly revealing here: as noted above consultants said that the natural way to point out the presence of an insect on the addressee’s shoulder is to say kî not ye: A: kî lukwe? ‘What’s that-neutral?’ B: oo! ‘Oh!’ A: mg:ee! ‘(It’s a) centipede!’

Suppose B is a child or ignorant foreigner, who has noticed the insect and grabbed it so it is now in his/her hand, then A could have said the following: A: ye lukwe? ‘What’s that-near-you?’ B: oo! A: ye mg:ee, kéé ngi! ‘That-near-you is a centipede, throw it away immediately!’

It is noteworthy that for every (so-called) addressee-based term for which we have careful records, special constraints arise concerning the addressee’s attentional state. Thus in Turkish the term şu (which has been described as an addressee-proximal form) is in fact a term used to get the addressee to focus on the referent (so presupposing that the addressee’s attention is elsewhere), and as soon as mutual attention is achieved the term is replaced with bu (speakerproximal) or o (speaker-distal) (Küntay and Özyürek, 2002; 2006). This is of course the reverse of the Rossel pattern, where ye signals ‘I’m talking about the one you are focused on’, but in both cases it is the state of the addressee’s

334

Stephen C. Levinson

Table 15.7 Conditions on the deictic center’s attention with respect to referent Prior attention of Spkr and Addr: Prior attention of Spkr only: Prior attention of Spkr and Addr:

ye ala, kî, mu yi

Table 15.8 Epistemic certainty + Certainty − Certainty

ala, ye, kî, mu wu (indirectly ascertained)

attention that is wholly (Turkish) or partly (Yélî Dnye) at stake (see also Burenhult, this volume). By assimilating the anaphoric yi into the demonstrative paradigm, as presupposing mutual prior attention in discourse, we can oppose the terms as in Table 15.7. Notice that these attentional and gestural prerequisites do not align with the previously discussed epistemic conditions, here in table form for comparison (Table 15.8). 5

Some Further Facts about Yélî Dnye Deictics

There are a number of further points that should be mentioned. The Wilkins questionnaire concerns singular as opposed to contrastive uses of the demonstratives in which one referent is picked out and opposed to another (as in ‘This one, not that one’). English usage makes clear that under contrastive uses distance is effectively neutralized – the first-mentioned referent usually receives this, the second that. It is therefore worth recording that there is no neutralization of the Yélî Dnye terms in contrastive usage. The questionnaire was supplemented by a special contrastive usage task in which two diagrams, hidden on the reverse of paper squares, are aligned sagitally away from the speaker, and the speaker is asked ‘Which is the circle, and which is the square?’ or the like, and the speaker replies ‘I guess this is the circle, and that is the square’, as appropriate (Wilkins 1999).4 The order of mention with respect to distance can then be manipulated. The result for Yélî Dnye is that the order of mention makes no difference to usage, which follows the conditions already established for singular non-contrastive reference as described above. 4

See http://fieldmanuals.mpi.nl/volumes/1999/eliciting-contrastive-demonstratives-personalspace/

Yélî Dnye Demonstratives

335

Table 15.9 Adverbial (manner) uses of demonstratives Proximal to Spkr ?Proximal to Addr Anaphoric Uncertain

‘like this’ ‘like this’ ‘like that-mentioned’ ‘like that-uncertain’

ala nté ya nté yi nté wu nté

Order of Mention 2 wu n:ii

1 wu n:ii

1 ala n:ii

2 ala n:ii

Figure 15.7 Contrastive use of demonstratives (numerals indicate order of mention)

Figure 15.7 illustrates the pattern under which order of mention makes no difference to the choice of demonstrative. Note that because the diagrams were hidden under paper squares, we get the ‘indirectly ascertained’ form wu for the more distant referent, but ala (speakerproximal) for the close referent. This shows once again the pre-emptive character of the proximal demonstrative, over-riding the marking of epistemic uncertainty. In fact, if the speaker touches the more distant referent, then ala can (indeed should) be used for that referent too. Other demonstratives would only become pertinent if the speaker was to step back and point from a metre or two distance: then kî could be employed for the closer referent, and mu for the more distant. Note here too that contrastive use to an already mentioned referent would invoke the dedicated anaphoric determiner yi. Another general point is that in addition to the pronominal and adverbial series tabulated in Table 15.2, there is an additional adverbial series for demonstrations. This set of manner demonstratives is presented in Table 15.9. The form ya nté is presumably based on the addressee-proximal ye, but its usage conditions need further examination and are not pursued here. The ‘uncertain’ form wu nté is used, for example, as follows: A: ala nté ‘(he did it) like that’ B: wu nté? ‘How (I missed your demonstration)?’

336

Stephen C. Levinson

There are other candidate demonstrative forms, for example mwada, which basically means ‘other’ but gets used as a distal demonstrative, as in mwada pee ‘the other/far side’, mwada y:ii ‘over there, lit. other mentioned-place’. As noted in Table 15.3, Henderson (1995: 46) considers the form mu essentially an ‘other’ term. This may be diachronically correct, but synchronically mu acts now like a systematic distal. But this suggests a general way in which distal demonstratives may be recruited, through the use of an ‘other’ term where three contrasts have to be made in a two-term system, and then the ‘other’ term gets gradually included in the system. The temporal uses of the system of deictic determiners would take us far afield (see Levinson and Majid, 2013), but the following facts are perhaps revealing about the spatial meanings. Future time reference uses ala or kî, as in ala/kî Sunday ngê ‘this coming Sunday’. For reference in the past, days are counted backwards (e.g. m:ii tuwó ‘day before yesterday’) or forward (e.g. tómê ‘nine days from now’) with dedicated forms that do not involve demonstratives. However, if one wishes to refer to some specific day backwards but can’t specify precisely, one could say mu Sunday meaning ‘that Sunday’, ‘the other Sunday’ (as in English the other day), while wu Sunday would mean ‘that particular Sunday whenever it was (when we went fishing etc.)’. The uncertainty content of wu makes it appropriate for remote time, as in the frozen expression wu-nê ‘long ago’. Similarly, the proximal meaning of ala determines its use for the present, as in ala ngwo ‘right now’ (lit. ‘this time’). A final and highly complex aspect to the deictic determiners is their role outside the NP in the pre-verbal clitic slot, which Henderson (1995) calls the Pre-verbal Nucleus. This pre-verbal slot is typically filled by a Tense-AspectPerson- Mood (TAMP) marker, a portmanteau morpheme indicating 144 basic contrasts on these dimensions. However, the slot also absorbs other grammatical categories, e.g. negative, conditional, counterfactual and deictic categories, and in this case there may be either a monomorphemic portmanteau morph expressing these additional features along with one of the 144 TAMP distinctions, or in some cases recognizable, separable multi-morphemic instantiations of the distinctions. These separable multi-morphemic variants allow one to analyse this pre-verbal slot as in fact consisting of an ordered series of micro-slots, roughly as in Table 15.10. As the arrows make clear, the deictic determiners are distributed through the various slots in this series, flanking, for example, negation and the basic TAMP clitic (note all of these precede the verb). This partly reflects the special roles they acquire inside the verb complex (nicely described in Henderson, 1995: 48–55). Thus kî (unmarked, certain) and wu (uncertain) come to have evidential functions: they mark certainty of the event or state and uncertainty respectively. Thus if you ask ‘Where’s the bottle?’, and I’ve recently seen it, I can say

mye

kî

wu

Addition ‘also’

Epistemic

mu

Distal Deictic

mê

daa

Negation

yi

Anaphoric, Cleft

Deictics from demonstrative series

Repetition

Table 15.10 Ordered slots in the pre-verbal nucleus Basic TAMP Proclitic

mî n:aa

Motion

a nê

Proximal Deictic

338

Stephen C. Levinson

the following (where the k- morpheme is analysable as derived from the kî demonstrative): (2)

pód:a tapil mbêmê k-a kwo bottle table on CERT-3sPresentContinuousIndicative stand ‘The bottle is on the table (I’m sure, I saw it).’

Henderson thus tries to maintain the ‘visible’ condition which he posited for kî in its noun-determiner uses, but as noted above, this is too specific a meaning. Similarly, wu carries evidential meaning in the pre-verbal slot. It occurs typically in future tenses and questions (Henderson, for reasons unclear to me, suggests that it carries ‘definiteness’ meaning, but uncertainty is the clearer gloss). The deictics mu and yi retain their referential functions in the pre-verbal slot – mu refers to a distant place where the event occurs, and yi to the anaphoric subject of the clause. Henderson (1995: 54) mentions that yi may have cleft-like force, but I think this is in a special construction with yinê, a complex form also presumably derived from yi: (3)

yinê dê d:uu ngmê Those-are-the-ones dual did PolyfocalSubject.3sObject.ProximatePast ‘Those two are the ones who did it (earlier today).’

Finally, the deictic a (from ala ‘proximal’) acts like a hither-particle. Yélî Dnye has no lexicalized ‘come’/‘go’ or ‘bring’/‘take’ oppositions, but marks the contrast with the -a morpheme at the end of the pre-verbal clitic. In this case there is no exact overt counterpart for ‘thither’: the unmarked form therefore implicates ‘motion away’ by pragmatic opposition to the marked form that has not been used. The implicature can be reinforced by use of the morpheme mî/n: aa, a form with many allopmorphs including d:uu, which indicates associated motion (‘go and VERB’), and thus, in contrast to -a, implicates motion away. The following examples illustrate the interpretive oppositions that can be obtained just by varying these elements (the TAMP element is in second position, and in (c) we have a portmanteau morph covering TAMP and Motion): (4) a. ngomo dê kee house 3SubjectPunctual.ImmediatePast enter ‘He entered the house’ implicates ‘away from here.’ b. ngomo d-a kee house 3SubjectPunctual.ImmediatePast-Proximal enter ‘He entered hither the house’ entails ‘towards here.’ c. ngomo d:uu kee house 3SubjectPunctual.ImmediatePast+Motion enter ‘He went and entered the house’ strongly implicates ‘away from here.’

Yélî Dnye Demonstratives

339

d. ala ngomo dê kee this house 3SubjectPunctual.ImmediatePast enter ‘He entered this house’ entails ‘towards here’, so defeating the implicature in (a).

The fact that in the pre-verbal position the deictics perform varied functions – evidential (kî/wu), referential (yi), adverbial-referential (mu) and deictic direction of motion (-a) – accounts for their different positions in the pre-verbal nucleus, and for the fact that, apart from the opposed evidentials, they can all co-occur together. 6

Conclusions

From all the evidence accumulated here, the best analysis of the Yélî Dnye deictics is that already offered as the three-dimensional cube in Figure 15.5, supplemented with notes about pre-emptive behaviour, gesture and focal attention. The fact that the very same demonstrative can play a role on three dimensions accounts for some of the complexities of the system. Thus kî can be a neutral or unmarked spatial demonstrative, a cataphoric element in endophora, and a ‘certain’ or ‘directly ascertained’ marker on the epistemic or evidential dimension. On each dimension it can play a role not only through its coded content but also through its pragmatic opposition to the other elements. The terms exophoric and endophoric, as used elsewhere in this book, do not really capture the essential dimensions here – a single usage can be both deictic and referential (‘exophoric’) and anaphoric (‘endophoric’) at the same time5 (as when a speaker uses yi to refer to something right in front of him which has already been mentioned); similarly a term can be directly referential and evidential at the same time (as with wu said pointing to a ship lost in the mist). Nor, however, are the dimensions themselves adequately captured by my terminology. Take the essential dimension we have called ‘spatial’ or ‘distance’. We have already seen that none of the terms, with the possible exception of mu, could be said to have its semantics or usage conditions adequately characterized in purely spatial terms. The proximal forms ala and ye basically indicate bodily involvement of one of the two deictic centres (speaker or addressee respectively) – for the use of these terms the relevant deictic centres must be attending to the referent, and either have it within reach or be actually holding it or pointing at it closely. They could be said to have haptic prototypes. In this way, Kemmerer (1999) too hastily rejected the relevance of the peri-personal space which is known to 5

Halliday and Hasan (1976: 36–37) also state that ‘(a)ny given INSTANCE of reference may be either (exophoric) or (endophoric), or it may even be both at once’ (original capitals).

340

Stephen C. Levinson

play such an important role in the neurocognition of space. This leaves only kî on the ‘spatial’ dimension, and we have already established that it does not actually encode any spatial discrimination – it is a spatially neutral term which, by picking up the residue between distal mu and the proximal terms, tends to have medial functions by implicature. Perhaps its epistemic function (‘certainty’) is also only implicated by the fact that it contrasts with wu (‘indirectly ascertained’) by being directly referential: the referent must be an object in space which one can directly vouch for (hence the specialized interpretation ‘in sight, visible’ offered by Henderson, 1995). Deictic determiners are proverbial for their Protean semantics, and the difficulty with which their meanings are pinned down. We now have a good idea of why this should be so: they are – in the Rossel case at least – multidimensional, and the pragmatic (and therefore defeasible) oppositions they make are at least as important as their coded content. In fact, it is the sheer emptiness or semantic generality of their content that gives them the wide functions they enjoy. Their core content is to suggest to the addressee that using whatever little semantic constraints they impose, and by monitoring the speaker’s gaze and gesture, the addressee will be able to find the intended referent. That generality allows for contextual enrichment of many different sorts. Even a system of just half a dozen demonstratives like the Rossel one does nothing to moderate this complex picture. References Delfin, F., Myles, S., Choi, Y., Hughes, D., Illek, R., van Oven, M., Pakendorf, B., Kayser, M. & Stoneking, M. (2012). Bridging near and remote Oceania: MtDNA and NRY variation in the Solomon Islands. Molecular Biology and Evolution, 29 (2), 545–564. doi: 10.1093/molbev/msr186. Dunn, M., Terrill, A., Reesink, G., Foley, R. & Levinson, S. C. (2005). Structural phylogenetics and the reconstruction of ancient language history. Science, 309, 2072–2075. Dunn, M., Foley, R., Levinson, S. C., Reesink, G. & Terrill, A. (2007). Statistical reasoning in the evaluation of typological diversity in Island Melanesia. Oceanic Linguistics, 46(2), 388–403. Dunn, M., Levinson, S. C., Lindström, E., Reesink, G. & Terrill, A. (2008). Structural phylogeny in historical linguistics: Methodological explorations applied in Island Melanesia. Language, 84(4), 710–759. Fillmore, C. A. (1997). Lectures on Deixis. Lecture Notes 65. Stanford, CA: CSLI Publications. Friedlaender, J., ed. (2007). Genes, Language, and Culture History in the Southwest Pacific. Oxford: Oxford University Press. Friedlaender, J., Schurr, T., Gentz, F., Koki, G., Friedlaender, F., Horvat, G. & Babb, P. (2005). Expanding Southwest Pacific mitochondrial haplogroups P and Q. Molecuar Biology and Evolution, 22(6), 1506–1517. doi: 10.1093/molbev/msi142.

Yélî Dnye Demonstratives

341

Grice, P. (1975). Logic and conversation. In P. Cole & J. Morgan, eds., Syntax and Semantics. Vol. III: Speech acts. New York: Academic Press, pp. 41–58. Halliday, M. A. K. & Hasan, R. (1976). Cohesion in English. London: Longman. Heine, B., Claudi, U. & Hünnemeyer, F. (1991). Grammaticalization: A conceptual framework. Chicago: University of Chicago Press. Henderson, J. (1995). Phonology and the grammar of Yele, Papua New Guinea. Pacific Linguistics, Series B–112. Canberra: Australian National University. Henderson, J. & Henderson, A. (1987) [revised 1999]. Dictionaries of Papua New Guinea. Vol. IX. Ukarumpa: Summer Institute of Linguistics. Hunley, K., Dunn, M., Lindström, E., Reesink, G., Terrill, A., Healy, M. E., Koki, G., Friedlaender, F. R. & Friedlaender, J. S. (2008). Genetic and linguistic coevolution in Northern Island Melanesia. PLoS Genet., 4(10), e1000239. Kemmerer, D. (1999). ‘Near’ and ‘far’ in language and perception. Cognition, 73, 35–63. Küntay, A. C. & Özyürek, A. (2002). Joint attention and the development of the use of demonstrative pronouns in Turkish. In B. Skarabela, S. Fish & A. H. Do, eds., Proceedings of the 26th annual Boston University Conference on Language Development. Somerville, MA: Cascadilla Press, pp. 336–347. (2006). Learning to use demonstratives in conversation: What do language specific strategies in Turkish reveal? Journal of Child Language, 33(2), 303–320. doi:10.1017/S0305000906007380. Levinson, S. C. (1995). Three levels of meaning. In F. Palmer, ed., Grammar and meaning: Festschrift for John Lyons. Cambridge: Cambridge University Press, pp. 90–115. (2000a). Yélî Dnye and the theory of basic color terms. Journal of Linguistic Anthropology, 10(1), 3–55. (2000b). H. P. Grice on location on Rossel Island. In S. S. Chang, L. Liaw & J. Ruppenhofer, eds., Proceedings of the 25th annual meeting of the Berkeley Linguistics Society. Berkeley, CA: BLS, pp. 210–224. (2000c). Presumptive meanings. Cambridge, MA: MIT Press. (2004). Deixis. In L. Horn, ed., The handbook of pragmatics. Oxford: Blackwell, pp. 97–121. (2006a). Matrilineal clans and kin terms on Rossel island. Anthropological Linguistics, 48(1), 1–43. (2006b). Parts of the body in Yélî Dnye, the Papuan language of Rossel Island. Language Sciences (special issue, ed. A. Majid, N. Enfield & M. van Staden), 28, 221–240. (2006c). The language of space in Yélî Dnye. In S. C. Levinson & D. Wilkins, eds., Grammars of space. Cambridge: Cambridge University Press, pp. 157–204. (2006d). Introduction: The evolution of culture in a microcosm. In S. C. Levinson & P. Jaisson, eds., Evolution and culture. Cambridge, MA: MIT Press, pp. 1–41. (2007a). ‘Cut’ and ‘break’ verbs in Yélî Dnye, the Papuan language of Rossel Island. Cognitive Linguistics (special issue, ed. A. Majid & M. Bowerman), 18(2), 207–217. (2007b). Optimizing person reference: Perspectives from usage on Rossel Island. In N. Enfield & T. Stivers, eds., Person reference in interaction. Cambridge: Cambridge University Press, pp. 29–72.

342

Stephen C. Levinson

(2008). Landscape, seascape and the ontology of places on Rossel Island, Papua New Guinea. Language Sciences, 30(2/3), 256–290. (2011). Reciprocals in Yélî Dnye, the Papuan language of Rossel Island. In N. Evans, A. Gaby, S. C. Levinson & A. Majid, eds., Reciprocals and semantic typology. Amsterdam: John Benjamins, pp. 177–194. (in preparation). A Grammar of Yélî Dnye. Levinson, S. C. & Brown, P. (2012). Put and take in Yelî Dnye, the Papuan language of Rossel Island. In A. Kopecka & B. Narasimhan, eds., Events of putting and taking. Amsterdam: John Benjamins, pp. 273–296. Levinson, S. C. & Majid, A. (2013). The island of time: Yélî Dnye, the language of Rossel Island. Frontiers in Psychology, 4, 61. doi:10.3389/fpsyg.2013.00061. Pederson, E. & Wilkins, D. P. (1996). A cross-linguistic questionnaire on ‘demonstratives’. In S. C. Levinson, ed., Manual for the 1996 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–14. Shaw, B. (2015). The archaeology of Rossel Island, Massim, Papua New Guinea: Towards a prehistory of the Louisiade Archipelago. Unpublished Doctoral thesis, Australian National University, Canberra. van Oven, M., Brauer, S., Choi, Y., Ensing, J. Schiefenhövel, W., Stoneking, M. and Kayser, M. (2014). Human genetics of the Kula Ring: Y-chromosome and mitochondrial DNA variation in the Massim of Papua New Guinea. European Journal of Human Genetics, 22(12), 1393–1403. Wilkins, D. P. (1999; this volume). The 1999 demonstrative questionnaire: ‘This’ & ‘that’ in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24. Wilkins, D. P. (1999). Eliciting contrastive use of demonstratives for objects within close personal space (all objects well within arm’s reach). In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 25–28. (Available online at http://fieldmanuals.mpi.nl/volu mes/1999/eliciting-contrastive-demonstratives-personal-space/.)

16

Tidore: Non-contrastive Demonstratives Miriam van Staden

1

The Language and Its Speakers

To preview the main findings of this chapter, Tidore has a multi-term demonstrative system with seven basic demonstrative roots that can indicate either the referent’s distance from speaker/addressee, or its position relative to landmarks used in the wider system of absolute landmark orientation. We distinguish interactionally anchored terms and geographically anchored terms. It is argued that instead of different domains of orientation, the relevant parameters for selecting the appropriate demonstrative are the salience of interactional space versus the salience of geographical space. Tidore is spoken on the island of Tidore and on a few adjacent, smaller islands in the North Moluccas of Indonesia. It is classified as a Papuan language, but more than 1,000 years of contact with neighbouring Austronesian languages has clearly left its mark on the lexicon, syntax and morphology. There are at present approximately 40,000 people for whom Tidore may be considered their first language, although all speakers of Tidore are fully bilingual in the local Malay vernacular, North Moluccan Malay. The Tidore language is clearly endangered. The influence from North Moluccan Malay as well as the national language, Indonesian, is increasing, and Tidore is riddled with Malay loans and code switches. Interestingly, there is also transfer from Tidore to North Moluccan Malay (as spoken on Tidore), in particular in the use of demonstratives, directional and locational verbs in locative descriptions. The system that is described here for Tidore is used identically in the North Moluccan Malay spoken on Tidore. The North Halmahera family to which Tidore belongs is generally characterised by SOV constituent order, subject and object marking on the verb, and a possessor-possessed order. Of these characteristics, Tidore has retained only subject marking on the verb and the order in the possessive construction. Tidore is now an SVO language. It is predominantly an isolating language with little inflectional or derivational morphology. Main word classes are nouns, verbs and adjectives; some adverbs are found, but they are typically verbal in origin; there are few prepositions, and no postpositions. 343

344

Miriam van Staden

2

Description of Form Classes and Previous Claims

Tidore has a set of seven demonstrative enclitics, which can be attached to the personal pronouns, to give demonstrative pronouns (1) or to (proper) nouns and noun phrases for exophoric reference (2):1 (1)

Mina=tina mo-kora bahaya2 3F=landward 3F.A-lie dangerous ‘She that is landward (“she landward”) lies terribly.’

(2)

Ngoto foli 1SG.N.A buy ‘I don’t buy that bread.’

roti=ge bread=GEN

ua NEG

The first column in Table 16.1 gives the forms of the demonstrative enclitics. They can follow all pronominal forms, but only when used in conjunction with the third person non-human pronoun ena do the demonstrative pronouns show phonological reduction. The reduced forms are presented in the second column of Table 16.1 Although these forms are considered informal and not appropriate in formal written records, they are most frequent in speech and may have human referents as well. The demonstrative pronouns may constitute the noun phrase as in (3) or they may themselves follow (pro)nouns as in (4). Table 16.1 Demonstratives in Tidore

interactionally anchored

geographically anchored

demonstrative enclitic

3NH dem. pronouns

proximal generic nonproximal distal

re ge

nde ngge

ino

ta

nta

ia

seaward landward upward downward

tai tina tau tahu*

ntai ntina ntau ntahu

hoo isa ine tora

directional verb

* The demonstrative tahu is also found reduced to tau, giving a rather curious homophony in the demonstratives for ‘up’ and ‘down’. I will not add to the confusion by using tau for ‘down’ anywhere in this chapter 1

2

In the examples in this chapter, the verbs may or may not show first argument cross-references in the form of prefixes. The reason is that the examples here are taken from spontaneous speech and the cross-referencing of the first argument on the verb is optional. This means that there are environments in which the verb may not be marked, but there are no environments where it must be marked; see Van Staden (2000) for a discussion of this optionality. The transcription follows the conventions for the spelling of modern Standard Indonesian. The symbols y and w represents glides, j and c are palatal plosives.

Tidore Demonstratives (3)

(4)

Oro nde fetch 3NH.PROX ‘Just take this one!’

345

ma MIT

Lobino=re mansia nta yo-gahi night=PROX people 3NH.?DIST 3PL.A-make ‘Tonight those people will perform the jinn ritual.’

jin jinn

I have not yet been able to pinpoint the difference in use between the demonstrative enclitics and these reduced demonstrative pronouns. It is a subject for further research whether perhaps discourse factors, regional or even personal preferences play a role here. Two of the demonstrative enclitics, =re and =ge, and their corresponding demonstrative pronouns nde and ngge have the additional function of ‘binding’ information, that is to say, to create units of information that are treated as a single entity. In this function they may cliticise to all phrase types and to clauses. They are used, for instance, to mark topics and extra-clausal information (5), or they give relative clause readings, as in (6): (5)

Ngori nde to-gahi ngam 1SG.N 3NH.PROX 1SG.A-make food ‘As for me I’ve already prepared food.’

(6)

Ona tagi yau=ge yo-wako rai 3PL go fish=GEN 3PL.A-return already ‘They who had gone fishing have already returned.’

rai already

In addition, they are used in temporal expressions, as in lobino=re (dark=prox) ‘tonight’ or wange nde (day 3nh.prox) ‘present times, nowadays’. Attached to ka-, an erstwhile essive verb, the demonstrative enclitics give locational verbs. (7)

Pati ka-tai rai P. PRED-seaward already ‘Pati is already in a seaward location.’

(8)

No-ten besi-besi ka-re 2SG.A-put RED-iron PRED-PROX ‘Just put the frying pan here!’

ma! MIT

In addition, there is a corresponding set of directional verbs, given in the third column of Table 16.1. The demonstratives, the locational verbs and the directional verbs are very frequently used in all genres of speech and are highly emblematic of the Tidore language: The directional verbs are among the 10 most frequently used verbs in the language, along with ‘say’, ‘take’, ‘make’ and ‘move/go’, and talking about their own language people will immediately point out the large number of locationals and directionals. In fact, such a complex

346

Miriam van Staden

system of spatial deixis is highly characteristic for all the languages of the North Halmahera language family (see, e.g., Taylor, 1984 and Shelden, 1991). The demonstratives are used for exophoric reference, but so are the directional and locational verbs. These are preferred even when referents are identified among a number of similar objects: (9)

No-oro cici toma besi-besi ma-gumuru hoo nde 2SG.A-fetch knife LOC RED-iron INAL-side seawards 3NH.PROX ‘Get (me) the knife that is seawards of the fryingpan (lit. that is at the side of the frying pan and then seawards).’

Such descriptions as these are deemed much more precise than the use of demonstratives even when combined with pointing. However, only the demonstrative enclitics are deictic in the sense that they rely on the position of speaker and/or addressee in actual space for their interpretation. At first blush, it appears that the three forms =re, =ge and =ta are anchored to the relative distance to the speaker/addressee position and may be glossed ‘proximal’, ‘medial’ and ‘distal’ respectively, while the other four forms bear on the shared position of speaker/addressee in relation to the landmarks that are used in an absolute landmark system of orientation. We shall refer to the former as ‘interactionally anchored’ and the latter as ‘geographically anchored’ terms. It is not immediately obvious whether ta should be analysed as interactionally or geographically anchored. It is used to refer to objects at some distance to the deictic centre (the position where the lines in Figure 16.1 cross), but also for objects located ‘along the coast’, as the orthogonal term to the land/sea terms illustrated in Figure 16.1.

LAND

TINA

TA

deictic centre

TAI TA

SEA

Figure 16.1 Simplistic analysis of ta

Tidore Demonstratives

347

There is evidence that =ta historically forms a paradigm with =tai, =tau, =tahu and =tina rather than with =re and =ge. In related Northeast Halmaheran languages, a proximal/distal constrast for the geographically anchored terms shows up in a /t/ – /d/ alternation of the first consonant. The proximal terms all have initial /t/; the distal terms have initial /d/. The equivalent of ta in these languages features in this paradigm. We make a provisional distinction between two meanings of ta1 ‘that far distal’ and ta2 ‘that along the coast’. In the remainder of this chapter, I shall show that this is indeed an oversimplification, and although the use of =ta is rather complex, it has just one meaning. In order to understand the use of the geographically anchored demonstratives, it is first necessary to expound a little on the geography of Tidore and the system of absolute orientation used. On Tidore, a small island dominated by a large volcano, the use of the land/sea axis does not refer to any cardinal directions, but to movements and locations towards the sea and the volcano. The up/down axis refers to vertical up/down and movements downhill and uphill, but also to movements towards (up) or away from (down) the site of the former sultan’s palace (cf. the spot indicated up in Figure 16.2). The ultimate ‘down’ location is the port of Rum (cf. the spot indicated down in Figure 16.3).

UP

Figure 16.2 Royal up/down on Tidore. All objects or movements located in the direction of the sultan’s palace are ‘up’

DOWN

Figure 16.3 Royal up/down on Tidore. All movements and objects towards the port to Ternate are ‘down’

348

Miriam van Staden

We will refer to this as ‘royal up/down’. As Figure 16.2 shows, a downhill movement towards the sultan’s palace is at the same time ‘seawards’, ‘down’ and ‘royal up’. But the terms cannot be used interchangeably. The royal up/down and the land/sea axis are both ‘primary’ in Palmer’s (2002) terms. This means that they both have an independent motivation so that neither is defined solely as the orthogonal to the other. The interpretation of the royal up/down axis as the orthogonal to the land/sea axis is a pragmatic implication, and not part of the meaning. It is not the case that royal up/down refers to all movements and locations along the coast, i.e. on the cross-axis from the land/sea axis, but rather it refers to movements and locations towards or away from the sultan’s palace. When an object is located both in the direction of the site of the former sultan’s palace and seawards, the land/sea axis, being in most cases the more salient of the two, will generally be preferred. Yet, if such reference is made in the direct vicinity of the site of the sultan’s palace, up/ down may be preferred. Interestingly, local politics play some part here such that ‘royalists’ are more likely to use the up/down axis than ‘republicans’. I have even observed some republicans use the royal up/down for locations and movements towards the governor’s offices, some 10 kilometres away from the sultan’s palace. Although in practice the up/down axis appears to be used on the cross-axis from the land/sea axis, this is in fact only an implicature. In Van Staden (2000), the use of the demonstratives, locationals and directionals is related to three distinct domains in which reference is established: ‘in and around the house’, ‘on the island’ and ‘beyond the island’. The first domain is the primary domain for the interactionally anchored terms =re ‘proximal’, =ge ‘generic’ and =ta ‘distal’. The land/sea geographically anchored terms are also used in this domain, but the royal up/down axis is not normally used. By contrast, for reference to locations and movements in the second domain, on the island, the geographically anchored terms are in general preferred over the interactional ones. As was mentioned earlier, there are many situations in which two or even three axes of orientation, even within the geographical domain, may be applicable at the same time. While often there is no one axis that is strictly speaking wrong, various locations, such as the market or the house of a relative, have an associated axis of orientation, so that going to the market will always be ‘seawards’ when coming from mountain village Gurabunga (A), although ‘downwards’ would also have been correct, while coming from Gurabati located along the coast (B), ‘royal up’ is used, even though the market is somewhat past the sultan’s palace (Figure 16.4). Beyond the island, the third domain identified, sea currents, land size and power relations come into play. Halmahera, visible as a long stretch of land, is considered ‘landward’, while ‘brother island’ Ternate, of roughly equal size and political import is ‘distal’. Papua is ‘upwards’ because of the sea currents and the historical trading routes, but almost all other islands and countries will be considered ‘seawards’. Somebody coming home from Halmahera will be

Tidore Demonstratives

349

A SEAWARD B

UP Market

Figure 16.4 Going to the market

referred to as ‘going seawards’, while someone returning from Java will be travelling ‘landwards’ (see also Figure 16.5). 3

The Wilkins Questionnaire

This system of absolute orientation meant that a few complications arose in the application of the Wilkins Demonstrative Questionnaire (1999; Chapter 2, this volume). The first is that the Wilkins questionnaire aims very specifically at a number of contrasts, such as visibility, position of addressee versus speaker, boundary crossing and distance, that are relevant for many different languages but pertain mostly to ‘interactionally anchored terms’. When applied to a system that also uses distinctions such as relative height and environmental parameters, the entire questionnaire must be carried out in the various geographic directions that are relevant, so that instead of the original 25 scenes for Tidore, we end up with at least 75: 25 scenes along the land/sea axis, 25 along the ‘royal’ up/down axis and 25 along the up/down axis, for instance on the slope of a hill. This multiplication of scenes ran into practical problems as speakers became confused or, worse, came up with ad hoc analyses of the system that got in the way of native-speaker intuitions. This happened even when we tried only a few scenes in one session. Thus, where most researchers found that in the performance of the task certain descriptions were accepted – which in spontaneous speech are pre-empted by the existence of a more specific term (cf. Meira, this volume) – for Tidore it was found that possibilities that were actually encountered in spontaneous speech could be rejected during elicitation sessions. The explanation is perhaps that, for speakers of Tidore, the system of spatial orientation is highly emblematic of the language. It was not uncommon to find that after introducing myself as a learner of the language speakers started to list all the directional and locational terms and how they might be used in conjunction. Discussions on the meanings of these terms would, however, invariably lead to confusion, and it is quite possible that when questioned about the system as a system, even indirectly as in the Wilkins task, speakers applied it

350

Miriam van Staden

Figure 16.5 Tidore region

wrongly or more rigidly than in spontaneous speech. The Wilkins task was therefore carried out systematically in a linguist-consultant setting only twice, with additional data from these two consultants and several other people collected ‘on the fly’ every time an appropriate situation presented itself. In this manner, speakers were focused less on the system, and more on the individual situation. The drawback of this approach is that statistics cannot be given, and data could not be recorded on video. Finally, as soon as objects are physically higher or lower than the speaker, vertical up/down is always an option, and so for reasons of transparency we have chosen to present only the distribution of the terms on the horizontal plane. This horizontal plane may be

Tidore Demonstratives

351

considered an ‘unmarked’ setting to the extent that with a few exceptions all villages and houses are located along the coast on a reasonably level surface. Because of the interaction between interactionally and geographically anchored terms, a few scenes had to be added to the questionnaire, and a few had to be repeated with modifications to apply to different situations. The following three scenes were added:

S

A

Landmark orientation is different for speaker and addressee. The object is between the speaker and the addressee, but speaker and addressee are facing different directions, so that the object is ‘downwards’ for the 27 speaker, but ‘landwards’ for the addressee.

A S

S

Relative distance between speaker/ object and speaker/addressee. The speaker is between the object and the addressee. The object is out of reach of the speaker, while the distance between the speaker and the 26 object is shorter than between the speaker and the addressee.

A

Increasing distance from speaker/ addressee. This scene is a variant on scene 12, but the object is further removed from speaker and addressee; it is well out of reach towards the edge of the area in which speaker and 28 addressee are the most salient points of orientation

352

Miriam van Staden

Scenes were modified where the questionnaire could not test the three different domains of orientation identified for Tidore (Van Staden, 2000). For reference to objects at a distance from the shared speaker/addressee position, a distinction must be made as to whether this object was still located on the island itself, e.g. high up on the mountain, or whether it was located on a different island. For this reason pictures 24 and 25 occur twice: 24 and 25 refer to situations on the island itself, 24’ and 25’ refer to situations beyond the island. Finally, scene 19 was duplicated because it was found that a difference occurred when the object was (19) or was not (19’) within reach of the speaker. Table 16.2 gives the use of demonstratives in non-contrastive, exophoric use, on the horizontal plane. It conflates the scenes set up in different geographic orientations. The interactionally and geographically anchored terms are presented on different rows. While for the geographically anchored terms different geographic orientations would yield different forms only, for the interactionally anchored terms the different geographic orientations actually gave different results. In particular the distance from the speaker at which ta became appropriate was much greater in the land/sea direction than it was in the orthogonal direction along the coast. For instance for scene 12 it was found that when speaker and addressee faced the sea, =ta was not correct, but when oriented in the orthogonal direction it could be used. 4

Discussion

The most important finding is that only one of the three domains identified in Van Staden (2000) is confirmed, namely the original third domain: beyond the island. Within this domain a conventionalised term is chosen on the basis of land size, historical relations, etc. A distant column of smoke, evidence of a small fire in a garden, will be referred to as ena=tina ‘landward’ when it is on Halmahera and ena=ta ‘distal’ when it is on Ternate. In Table 16.2 these situations are given in column 8. The other two domains identified earlier ‘in and around the house’ and ‘on the island’ may be conflated into only one domain if we take into consideration the salience of landmarks and the positions of speaker and addressee. It appears that in this domain speakers set up an interactional space in which the position, especially of the speaker – but to some extent also that of the addressee – is prominent. As we move further away from this deictic centre, the landmarks ‘sea’, ‘land’ and ‘the sultan’s palace’ become more salient, and at some point the deictic centre has lost so much of its salience that the interactionally anchored terms are no longer appropriate. This region in which the geographically anchored terms are used, we refer to as ‘geographic space’. Clearly, the original identification of two domains relates to this distinction: the interactionally anchored terms were used typically in the domain ‘in and around the

re

1

S

S

A

S

S

A

(26)

(6)

(3)

(1)

S

S

S

A

S

S

re / ge

2

S

A

A

A

A

A

(20)

(19)

(8)

(7)

(4)

(2)

ge

3 S

S

S

A

A

A

(11)

(9)

(5)

S

A

S

A

S

A

A

(23)

(22)

(19’)

(10)

ta, tai, tina, tau & tahu

(18)

(16)

S

(ta, tai, tina, tau & tahu)

A

(12)

5

ge

A

S

S

A

ge (preferred)

S

4

Table 16.2 Application of scenes from the Wilkins questionnaire

A

S

S

A

A

(28)

(27)

(17)

ta, tai, tina, tau & tahu

ge & ta

S

6

A S

A S

A

S

S A

SA

S A

(25)

(24)

(21)

(15)

(14)

(13)

ta, tai, tina, tau & tahu

7

A S

A S

(25’)

(24’)

conventional ta, tai, tina, tau & tahu

8

354

Miriam van Staden TINA

TA/TINA TAU

GE/TINA

TAHU

TA

RE

SPKR ADDR

TA

GE TA/TAI

TAI

GE/TAI

Palace Sea

Figure 16.6 Use of demonstratives in geographical and interactional space

house’, or at closer proximity to speaker and addressee, and the geographically anchored terms elsewhere ‘on the island’. What it means is that some geographically anchored terms do occur in and around the house. Even within the house people can refer to objects as in being in a seaward or landward location. And in a house located next to the site of the sultan’s palace, we found that even ‘royal’ up/down was used ‘in and around the house’. Also the apparent boundaries of the two domains ‘in and around the house’ and ‘on the island’ could not capture the puzzling use of ta. In and around the house, this demonstrative can be used for objects located at some ‘remote’ distance. However, on the land/sea axis the distance must be significantly greater than on the orthogonal for ta to be appropriately used. This situation is sketched in Figure 16.6. At the same time, on the island the distal demonstrative is replaced by the geographically anchored terms after a certain distance. Relative salience captures the use of these terms much better than the more rigid notions of domains that suggest clear-cut boundaries. The domain of interactional space includes the space of everything within reach of speaker and addressee, plus some distance beyond.3 For the smallish inanimate objects that we primarily deal with in this chapter, this is some 2 to 3 3

I have excluded from this discussion the use of demonstratives in, for instance, telephone conversations when speaker and hearer are exceptionally far away from each other.

Tidore Demonstratives

355

metres out from the circle that contains both speaker and addressee. In practice, this space will be approximately the size of the interior of a small house or a room in a large house, a porch or an outside fireplace. In this interactional space, the speaker and addressee positions are highly salient, but geographic landmarks may be salient too. Interactional space includes both speaker and addressee, and this accounts for ‘addressee’ effects, for instance the possibility of using =ge for scene 18, where the distance between the speaker and the object is quite large. If the addressee had been closer to the speaker, but the object had remained in place, e.g. in scenes 13, 14 and 15, the use of =ge would not have been appropriate. Beyond interactional space, the geographic landmarks are more salient than the relative positions of speaker and addressee. First, we examine the clear-cut cases of interactionally anchored terms in interactional space (section 4.1) and geographically anchored terms in geographic space (section 4.2), then we turn to the more complex cases (section 4.3), where the two sets of terms appear in competition. 4.1

Re/nde, Ge/ngge and Ta/nta in Interactional Space

Column 1 in Table 16.2. shows that the proximals =re and nde are the only possibility when the speaker is touching the object, for example where the speaker points at his own teeth (scene 1), or when there is a bug on the speaker’s shoulder (scene 3). When the object is no longer within reach of the speaker, the range in which =re/nde can be used is partly dependent on the distance between the speaker and the addressee. When the speaker is directly between the object and the addressee, the proximal term can be used even for objects that are well out of reach of the speaker, approximately to a distance equivalent to the distance between the speaker and the addressee, as in scene 26. In Figure 16.6 this is represented by the oval for re extending much further to the speaker’s right than to the speaker’s left, where the addressee is located. Because the proximal term is clearly speaker anchored, the use of ge/ngge in this situation implicates that the object is close to the addressee. This is indeed an implicature only: when the object is removed further from the speaker, the preferred term will be =ge/ngge rather than =re/nde (see scene 11). If =ge/ngge had been addressee anchored, we would expect that objects closer to speaker than to addressee could never have =ge/ngge. Column 2 represents scenes in which an object is within reach of the speaker, but not touching, and both =re/nde and =ge/ngge can be used. When an object touches the addressee, =ge and ngge are used (column 3). Now the proximal terms are used only when the speaker is touching the object on the addressee or in the case of a close finger point (scenes 2 and 4 in column 2). The ‘distal’ form =ta/nta is used only when the object is distant from speaker, towards the outer boundaries of the interactional space, as in scene

356

Miriam van Staden

speaker’s proximal zone

close interactional space

outer interactional space

re ge ta (ta usually pre-empts ge in this zone)

Figure 16.7 Re, ge and ta in interactional space

28 represented in column 6, but not beyond. It is ‘distal’ then only within the interactional space. Like proximal =re/nde it is speaker anchored. The generic forms =ge/ngge may also be used here but are commonly pre-empted by the existence of the distal form. We will call =ge/ngge the generic non-proximal terms in interactional space, compatible with all but the proximal situations, i.e. when the speaker touches or closely points at the object. 4.2

Tau & Tahu, and Tai & Tina in Geographic Space

The geographically anchored terms =tau/ntau ‘upward’, =tahu/ntahu ‘downward’, =tai/ntai ‘seaward’ and =tina/ntina ‘landward’ are used when objects are well out of reach of speaker and addressee (columns 6–8 in Table 16.2). Like =re and =ta, the geographically anchored terms are speaker oriented, so that an object that is landwards from the speaker but seawards from the addressee will always be referred to as ‘landwards’. Typical situations are represented in scenes 13, 14 and 15, in column 7 where the speaker and addressee are close together on one end of the football ground and the object is located at a considerable distance from both, or in scene 21 where again speaker and addressee are together and the object is located at some distance outside the house, and 24 and 25 where the object is very far from both speaker and addressee. The other terms are inappropriate here always. 4.3

Geographically Anchored Terms in Interactional Space

In addition to their complementary distributions, Table 16.2 shows clearly how in columns 4 to 6 both the interactionally anchored and geographically anchored demonstratives may both be appropriate. These are the situations where the analysis of the two domains ‘in and around the house’ and ‘on the island’ in Van Staden (2000) runs into problems. The difference between column 4 and column 5 is that in situations represented in column 4 =ge appears to be preferred, while in the scenes represented in column 5 both are used. It would seem also that salience of the landmarks is to some extent

Tidore Demonstratives

357

gradable and not the same for all speakers. In general, however, the land/sea axis has relatively high salience even for objects that are just out of reach of the speaker or that are close to the addressee but at some distance from the speaker. As mentioned earlier, the volcano of Tidore is visible from almost every part of the island. Houses are typically oriented along the coast with either the porch, front door and windows, or the kitchen door and windows facing the sea, which means that the volcano or the sea is visible almost always. The result is that the land/sea axis is preferred over the interactionally anchored terms at a much closer distance from speaker than the up/down axis. The immediate effects are twofold: first, interactional space appears not to radiate from speaker/addressee position to form a circle, but to take an oblong shape. Second, the distal form =ta is appropriate on the land/sea axis for reference to objects at a distance that would still be referred to by the generic term =ge on the orthogonal. This is captured by the ovals in Figure 16.6. The royal up/down axis is in general much less salient, but it can have more salience under the circumstances described in section 2. If we now think of geographic and interactional space as radiating from the respective landmark or deictic centre that are their sources, it is easy to see how the two meet and how the salience of geographic space may push the interactional space back, or conversely how a large distance between speaker and addressee may create a very large interactional space, pushing geographical space into the background. The more salient a landmark is, the further it will ‘invade’ interactional space. This system is easily interpreted as having a primary orientation based on the land/sea opposition and a secondary orientation ‘across’ or ‘along the coast’ whereby objects in both directions along the coast can be referred to by the distal form. In practice, this is how most native speakers would initially analyse their own system of spatial reference, and it is possible that it will eventually develop into such a system. However, the results here clearly indicate that it is more accurately described as a system that uses both geographically anchored and interactionally anchored terms in a complex system of spatial reference that takes into consideration the relative salience of the landmarks and the speaker/ addressee position in relation to the object to be located. 5

Additional Observations

A final note concerns the facts as presented in Figure 16.6. If it is indeed relative salience of landmarks that determines use rather than a system of demarcated domains, we might expect differences depending on actual location. This is indeed the case. Most of the data elicited for this project were collected in two villages along the coast at some distance from the palace. However, choosing another site gives different shapes for the interactional and geographical

358

Miriam van Staden

spaces. In a village on the higher slopes of the volcano, the vertical up/down axis is more salient than the land/sea axis. In Kalaodi, a village located on a mountain ridge north of the volcano, the sea is visible on two sides of the island, and the use of demonstratives there is again slightly different. It was not possible to carry out this questionnaire in all conceivable circumstances, but some impressionistic observations show that in Kalaodi, ‘landward’ refers primarily to locations in the direction of the volcano, and ‘seaward’ could in principle refer to any other direction. It is not surprising that the interactionally anchored terms, in particular ta, were used much more frequently than in villages along the coast. These additional data support the intensional semantic analysis developed. We conclude with two illustrations of how the various terms are used in Gurabati, a village located along the coast some 8 miles away from the sultan’s palace. I. A speaker is seated in front of a house on the terrace, facing the sea. A balloon is on the terrace some 2 metres away from the speaker in a seaward direction. Both the distal term =ta and seaward =tai can be used to refer to the balloon, and even non-proximal =ge is possible, though not offered spontaneously. Then the balloon is blown to a distance of about 4 metres in the direction of the sea. The seaward term =tai is now the only appropriate term; =ta and =ge can no longer be used. The balloon is picked up and placed at a distance of some 2 metres to the side of the speaker, in the direction of the sultan’s palace. The palace is, however, far away and not salient. The royal up term =tau is possible, but somewhat marked, the distal form =ta is just possible, but more commonly one might expect =ge. Now the speaker kicks the balloon to a distance of some 4 metres in the direction of the sultan’s palace and remarks “now it’s =ta!”. II. A speaker (right) and the interviewer (left) are standing in the doorway of a house, facing the sea, as depicted in Figure 16.8. The sea is visible in the distance. A dog (the black spot) is about a metre from speaker and addressee (left picture). Both proximal =re and generic =ge are fine, since the dog is more or less within reach. The fact that speaker and addressee are inside the house, i.e. there is a social and physical boundary, is irrelevant here. The seaward term =tai is also possible, but it would perhaps be suggestive of a larger distance. The dog trots off to some 5 metres from speaker and interviewer in a ‘diagonal’ direction (Tidore: seisei) that is neither clearly seawards, nor clearly downwards (away from the sultan’s palace); see right picture. It is now referred to as either =tai ‘seaward’ or =ta ‘distal’. The dog has now reached the small road and follows it along the coast. Soon it is almost out of sight, and now it is referred to as either =ta or =tau. The latter form is somewhat marked apparently because the object here is a dog – an animal not held in very high esteem by the Islamic inhabitants

Tidore Demonstratives

359

Figure 16.8 A dog trotting off into geographical space

of Tidore and consequently associating it with the sultan’s palace, even indirectly through the use of the demonstrative forms is problematic. 6

Conclusion

In this chapter we have seen that the demonstrative system of Tidore not only relies on the speaker/addressee position, but in addition uses the landmarks ‘sea’, ‘land’ and ‘(site of the former) sultan’s palace’. Beyond the island, the terms are used in a conventionalised system, but on the island a number of interacting factors determine the appropriate use of the demonstrative enclitics and pronouns. The appropriateness of the terms depends on the relative distance between speaker and addressee (as this determines the size of the interactional space and the appropriate use of the speaker-proximal term), the relative distance of the object to the speaker, and the relative salience of landmarks. Ultimately, even the nature of the object may affect the use of the enclitics, as suggested by the example in which =tau was not preferred when the referent was a dog, but this is a topic that would need to be further investigated. Abbreviations a dist fgen inal loc m mit

actor distal feminine generic inalienable locative masculine mitigating

360 n neg nh pl pred prox red sg

Miriam van Staden neutral negation non-human plural predicate proximal reduplication singular

References Palmer, B. (2002). Absolute spatial reference and the grammaticalisation of perceptually salient phenomena. In G. Bennardo, ed., Representing space in Oceania: Culture in language and mind. Canberra: Pacific Linguistics, pp. 107–157. Shelden, D. (1991). Setting a course in Galela: An orientation system of North Halmahera. In T. Dutton, ed., Papers in Papuan Linguistics, no. 1. Canberra: Pacific Linguistics A-73, pp. 147–160. Taylor, P. (1984). Tobelorese Deixis. Anthropological Linguistics, 26, 102–122. Van Staden, M. (2000). Tidore: A linguistic description of a language of the North Moluccas. Doctoral thesis, Leiden University. Wilkins, D. P. (1999; this volume). The 1999 demonstrative questionnaire: ‘This’ and ‘that’ in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24.

17

The Jahai Multi-term Demonstrative System: What’s Spatial about It? Niclas Burenhult

1

Introduction

Demonstratives are considered to be a well-defined and cross-linguistically pervasive (even universal) class of words dedicated to coordinating the interlocutors’ focus of attention (Diessel, 2006). Yet there is a growing body of work suggesting that the semantic parameters operating within this well-established category vary greatly and are notoriously difficult to pin down. For example, the descriptively entrenched distance-based analyses of the commonly occurring “here versus there” type of distinctions have been challenged for a number of systems in recent years. The assumed spatial encoding has been shown to be heavily influenced or overridden by pragmatic factors pertaining to the attentional relationship between addressee and referent (Özyürek, 1998; Burenhult, 2003; Küntay and Özyürek, 2006), or by the social or interactional context of the speech situation (Enfield, 2003; Jungbluth, 2003; Hanks, 2005). There is also a budding theoretical interest in multi-term demonstrative systems with parameters which are cross-linguistically uncommon but add a range of diverse semantic dimensions to the class, such as visibility, elevation, and motion of the referent (Burenhult, 2008; Schapper and San Roque, 2011). Existing typologies of demonstrative systems tend to regard such parameters as “special” (Diessel, 1999: 42–47) or “ancillary” (Levinson, 2006: 117), implying that they somehow form a qualitatively distinct phenomenon tangential to the core theoretical issues. However, their functional relationship with, and relevance to, the more commonly theorized parameters of distance, participantanchoring, and attention remains largely unexplored. Are there, in fact, major insights to be gained into demonstrative function generally from in-depth analysis of multi-term systems? In this chapter I investigate a system full of such seeming exoticness, namely that of Jahai, an Austroasiatic language spoken by groups of subsistence foragers in the Malay Peninsula. Jahai has no less than nine demonstrative distinctions, each with both a nominal and an adverbial reflex. In addition to the 361

362

Niclas Burenhult

more familiar distance-related distinctions, this system contains a number of unusual (even unique) distinctions pertaining to the referent’s status: cognitively accessible to the addressee; cognitively inaccessible to the addressee; located above the deictic center; located below the deictic center; located outside the speaker’s side of the speech dyad; located outside the addressee’s side of the speech dyad; and perceived through its emissions (smell, sound, temperature, etc.) rather than direct visual or tactile experience of its inherent physical properties. Thus, semantic dimensions of accessibility, elevation, exteriority, and perception are intricately cross-cut by binary contrasts in the form of participant, attentional, and directional oppositions, producing considerable distinctional richness (see Table 17.1 below). I present detailed data and analysis of exophoric usage of demonstrative forms – based on Wilkins’ Demonstrative Questionnaire (1999; this volume) and supporting data from previous studies – to show that a multi-term demonstrative system with unusual distinctions can offer unique opportunities for exploring and understanding indexical principles in languages at large. I will argue that the fine-grained semantic encoding of the distinct Jahai forms partly unpacks functional dimensions proposed to be relevant by current theory but which are typically obscured in systems with less elaboration. In particular, the Jahai system sheds light on what is fundamentally versus symptomatically spatial in demonstratives; it helps to characterize and typologize the speaker’s management of the addressee’s attention; and it demonstrates that, far from being tangential phenomena, crosslinguistically unusual demonstrative parameters are tremendously relevant to core issues in deictic theory. 2

Background

2.1

Jahai: People and Language

The Jahai are approximately 1,000 mobile subsistence foragers in the mountain rainforests of northern Peninsular Malaysia and southernmost Thailand. Their economy is based on the hunting and gathering of wild animals and plants, supplemented by trade in forest products as well as occasional wage-labor and slash-and-burn horticulture. Traditionally, bands of between 10 and 50 individuals dwelled in temporary camps of lean-tos or huts, moving every few days or weeks (sometimes months) depending on the sustenance circumstances. Nowadays most Jahai are resettled in regroupment villages, but some groups still pursue a nomadic existence. The majority of Jahai uphold their traditional animistic belief system, rituals, and egalitarian ideology.

Jahai Demonstratives

363

The Jahai language belongs to the Northern Aslian clade of the Aslian branch of the Austroasiatic language family. Like other Aslian languages, it has a fairly rich sound system and displays complex paradigms and processes of derivational morphology (Burenhult, 2005). The lexicon is imbued with unanalyzable monolexemic forms encoding fine semantic detail in diverse domains like anatomy, biology, perception, motion, and ingestion. Such lexical specificity is a systematic principle of the language. While many children today receive basic schooling in Malay (the surrounding Austronesian majority language), most Jahai are non-literate, and Jahai is an unwritten language. 2.2

Previous Analyses

Multi-term demonstrative systems, and arguably any demonstrative system, hardly lend themselves to easy functional description. If my own experiences are anything to go by, anyone aspiring to come close to understanding a complex demonstrative system should be prepared to be in it for the long haul. Despite 15 years of research on the language, my comprehension of the Jahai demonstrative system is still developing. Now and again it teasingly dishes up new surprises. New methodological approaches (and not necessarily ones that target demonstratives) suddenly challenge what seemed to be wellestablished principles. And during the course of my publishing of results, I have twice had to amend the number of distinctions upwards, due to the discovery of new forms. For example, this chapter introduces a ninth distinction to the paradigm: the emission-perceptible. It was only in 2009 that the formal morphological evidence for the demonstrative status of this distinction emerged from a new type of data. The present account is therefore to be considered as work in progress, and the data reported represent just another piece of a complex puzzle. An early hint at the functional complexity of the Jahai system was provided by Schebesta (1928), who postulated a basic system of three distance-based distinctions and an additional pair of locatives signifying location upstream and downstream, as well as four forms corresponding to the four cardinal points.1 1

Although linguistically well ahead of his time, Father Paul Schebesta clearly did not pursue Jahai demonstratives in any depth. Judging from his short description, he did discern their basic formal properties but did not identify the paradigmatic systematicity, functional characteristics, or magnitude of the system. Six of the forms given clearly bear a resemblance to forms described in the present work. One of them provides a revealing accidental glimpse into an undocumented elicitation situation of nearly a hundred years ago. Thus, Schebesta glossed the form aɲʊ as ‘south’. This form is undoubtedly the addressee-anchored exterior ʔɲɨʔ, meaning ‘there, outside your side of our speech dyad’ (see section 3 and Figure 17.2). Cardinal directions are never used for everyday spatial reference by present-day Jahai speakers, and there are no indigenous terms for ‘north’ and ‘south’. Schebesta likely asked for the Jahai equivalent of the Malay term for ‘south’ (selatan), and, thinking he was asking for the direction, the unknown Jahai language consultant probably gestured beyond the researcher and replied

364

Niclas Burenhult

My own initial analyses of the system (Burenhult, 2002), based on interviews and participant observation, identified seven distinctions and involved four distance-based definitions (speaker-anchored proximal, addressee-anchored proximal, medial, and distal), two forms encoding elevation (superjacent and subjacent), and one form erroneously defined as encoding invisibility of the referent (later redefined as a speaker-anchored exterior).2 My subsequent analyses were based on tasks featured in the field manuals of the Language and Cognition Group at the Max Planck Institute for Psycholinguistics, such as the “Hidden colour-chips task” (Enfield and Bohnemeyer, 2001), “Demonstrative Questionnaire” (Wilkins, 1999; this volume, explored in section 4), and “Shape Classifier Task” (Seifart, 2003). These prompted a fundamental reinterpretation of the distance-based parts of the system (Burenhult, 2003) and provided further hints as to the function of the elevation- and exteriority-encoding forms. I also targeted the exterior distinctions, unique to Jahai, by means of specially designed elicitation tasks and analyses of usage in natural conversation captured on video (Burenhult, 2008; see also Terrill and Burenhult, 2008). My current interpretation of the system, based on this diverse set of data types, is explained in detail in section 3 and summarized in Table 17.1. 2.3

Definitions, Terminology

Demonstratives form the most prominent exponent of spatial deixis in language. Following Burenhult (2008: 100–101), demonstrative is defined generally here as any member (in the form of a word or a bound morpheme) of a closed grammatical class of expressions serving to narrow the contextually relevant search domain in the locational relativization of a referent to the deictic center (the speech situation or either of its two components, speaker and addressee). This definition incorporates not only concrete spatial (situational or exophoric) uses of demonstratives, but also abstract discourse-internal (or endophoric) uses. Thus, locational relativization may pertain to a referent’s location in actual space, or its location in discourse. The definition includes both nominal and adverbial demonstratives (e.g. English ‘this’ versus ‘here’). The formal, language-specific definition of Jahai demonstratives is given in section 3.

2

“ʔɲɨʔ!”, “There, away on your side of our speech dyad!”. Provided his consultant was familiar with cardinal directions, we can be reasonably confident that Father Schebesta was sitting south of his Jahai teacher during this conversation! This possibility to reconstruct the spatial layout of Schebesta’s interview underscores the profoundly angular meaning of some of the Jahai demonstratives. A common location of referents of this form is behind the speaker’s back, which in face-to-face conversation means that the referent is typically out of sight for both speaker and addressee.

Jahai Demonstratives

365

Multi-term demonstrative system is a label sometimes given to systems which comprise more distinctions than the more commonly occurring two- or three-term systems (see, e.g., Jungbluth, 2003: 29; Levinson, 2006: 110; Schapper and San Roque, 2011: 387). Such systems typically exhibit crosslinguistically unusual distinctions, encoding parameters like elevation, visibility, and movement of the referent (Anderson and Keenan, 1985: 291–292; Hyslop, 1993; Diessel, 1999: 42–47, 50–51; Dixon, 2003: 89). A subset of such distinctions has the function of projecting angular search domains and thus invokes spatial frames of reference. These are referred to here as spatialcoordinate demonstratives (Burenhult, 2008). Diessel (2006) defines demonstratives functionally as attentioncoordinators, dedicated to creating (or manipulating) joint focus of attention. For the present purpose I prefer to characterize demonstratives as attentionmanagers, underscoring their role as tools for the speaker to manage the addressee’s attention. The data to be reported further require us to subclassify attention managers into three functional categories: attention-drawers, which draw the addressee’s attention to a referent; attention-correctors, which shift the addressee’s inaccurate attention to the correct referent; and attentionconfirmers, which verify that the addressee attends to the intended referent (see further below). 3

Form and Function: An Outline

Jahai demonstratives are a morphosyntactically uniform and easily defined form class. Nine roots, all of which begin with a glottal stop /ʔ/, are used adverbially, typically in adjunct adpositional phrases headed by prepositional proclitics encoding distinctions like location, source, and goal (‘at here’, ‘from there’, etc.). Each root can be turned into a nominal demonstrative through the replacement of the initial glottal stop with the voiceless alveolar stop /t/. This process of initial phonemic supersession is a morphological strategy unique to this set of roots. It does not exist in other parts of the grammar. The derived forms are either used adnominally as modifiers of nouns or pronouns and then occur in a single, post-nominal NP slot, or used pronominally and then represent full NPs themselves. Demonstratives cannot modify each other. This mutual syntactic exclusivity, together with the unique and identical morphophonemic behavior, is what makes the nine distinctions a straightforwardly identified form class. Examples (1) and (2) illustrate adverbial and adnominal usage, respectively: (1)

ʔoʔ cɨp ba=ʔə̃ h 3S to.go GOAL=DEM ‘S/he came here.’

366 (2)

Niclas Burenhult ʔoʔ deʔ hayɛ̃ ʔ tə̃ h 3S to.make hut DEM ‘S/he built this hut.’

Their uncomplicated structural classification notwithstanding, the nine demonstrative distinctions encompass a number of diverse semantic dimensions. Firstly, they can be broadly grouped into three semantic “supercategories”. One such category is accessibility: four of the distinctions encode the accessibility and inaccessibility of a referent in relation to speaker and addressee, respectively. These are akin to the distance-related distinctions commonly found in demonstrative systems (see further in section 3.1). The second semantic supercategory is spatial frames of reference (see Levinson, 2003). Four distinctions encode such frames. This means that they invoke full-fledged spatial coordinate systems, whereby the demonstrative signifies a Figure in the form of a referent which is located angularly in relation to a Ground in the form of the speech dyad (or a part of it). Such distinctions are cross-linguistically unusual and still underexplored. Burenhult (2008: 109–114) identifies two subtypes among existing systems: absolute distinctions, which rely on a spatial asymmetry external to the deictic Figure-Ground constellation to project search domains, and intrinsic distinctions, which rely on spatial asymmetries internal to the deictic Figure-Ground constellation. Jahai exhibits both subtypes: two of the distinctions involve the absolute frame of reference, signifying referents which are found in search domains above and below the speech dyad, respectively: ‘that/there, up’ and ‘that/there, down’. These distinctions typically invoke elevation as manifested geophysically – topographically (uphill/downhill) or hydrologically (upstream/downstream) – but can also involve more general search domains above or below the deictic center (a semantic subtype termed global elevation in the typology of Burenhult, 2008: 110). The other two distinctions signify referents found in search domains that project away from the speaker’s and addressee’s positions in the speech dyad, respectively: ‘this/here, outside my side of the speech dyad’ and ‘that/there, outside your side of the speech dyad’. They cannot be used for reference between the interlocutors, hence they are called exterior demonstratives. The distinctions are conveniently described as invoking an intrinsic frame of reference, since the speech dyad is conceptualized as a whole entity (Ground) with two facets in the form of speaker and addressee, and the search domains in which the referent (Figure) is found project away from those facets. The meaning, logic, and usage of these distinctions are described in detail in Burenhult (2008). The third supercategory pertains to perceptual modality. This contains only one form which encodes that the referent is perceived only through its

Jahai Demonstratives

367

emissions, not through direct visual or tactile experience of its inherent physical properties. This can involve a referent which at the time of the utterance is (or has just been) heard, smelled, or felt in the skin, or one which produces a vague body-internal sensation or pain experienced by the speaker. It brings to mind the cross-linguistically unusual distinction of “invisible” described for some demonstrative systems. However, it is not invisibility as such which is in focus in the Jahai form, but rather the indirect, emitted characteristics of the experience. Evidentiality also comes to mind as a potential parameter of relevance to the distinction between this and the other demonstrative forms (cf. Meira this volume; Levinson, this volume). However, there is no evidence at present that it encodes anything beyond experience through a particular set of sensory modalities other than vision and touch. Secondly, as is evident from the above, the semantic supercategories are partly cross-cut by a binary distinction, creating semantic opposites for each category involved. In the accessible/inaccessible and exterior categories, this opposition involves the parameter of participant-anchoring, which distinguishes speaker-anchored versions from addressee-anchored ones. The semantic opposition in the elevation category distinguishes between superjacency and subjacency. The perception-encoding category is the only one which does not make such a binary distinction and thus contains only one form. The forms and functions of the full system are summarized in Table 17.1. 3.1

The Accessibility Parameter

The four accessibility-related distinctions have presented particular challenges for the semantic analysis of Jahai demonstratives. In the tentative analysis in Burenhult (2002), I considered distance to be a key parameter on the basis of the prevailing spatial distribution of referents, as well as consultants’ typical judgments of meaning. However, it was clear already from the outset that actual usage was far more varied and complex. For example, it was evident from participant observation that the forms displayed considerable flexibility with respect to the location and distance of referents in relation to speaker and addressee. Accumulating additional data prompted me to reassess the fundamental parameters of the system. For example, an analysis of demonstrative usage in a Director-Matcher game designed to probe shape distinctions in classifiers revealed that the addressee-anchored accessible ton was exclusively associated with referents which had the addressee’s attention (Burenhult, 2003).3 3

The game in question was the Shape Classifier Task (ShaClaTa), developed by Seifart (2003). Another Director-Matcher game – Man & Tree, designed to elicit spatial frames of reference – similarly generated abundant use of demonstratives in the Jahai setting (Terrill and Burenhult, 2008).

Perceptual modality

Spatial frame of reference

+ACCESS

Accessibility

EMISSIONPERCEPTIBLE

ELEVATION

EXTERIOR

–ACCESS

Parameter

Supercategory

tũn

tadeh tɲɨʔ titɨh tuyih

ʔũn ʔadeh ʔɲɨʔ ʔitɨh ʔuyih ʔnɛʔ

Addressee-anchored

Speaker-anchored Addressee-anchored Superjacent Subjacent

–

tnɛʔ

taniʔ

ton

ʔon

Addressee-anchored

ʔaniʔ

tə̃ h

ʔə̃ h

Speaker-anchored

Speaker-anchored

Nominal

Adverbial

Oppositions

Table 17.1 The Jahai demonstrative system

Accessible to speaker (proximal, perceptible, reachable, approachable, etc.) Accessible to addressee (familiar, established, attended to, manipulated) Inaccessible to speaker (distal, imperceptible, unreachable, inapproachable, etc.) Inaccessible to addressee (unfamiliar, unestablished, unidentified, unattended to) Located outside speaker’s side of speech perimeter Located outside addressee’s side of speech perimeter Located above speech situation (overhead, uphill, or upstream) Located below speech situation (underneath, downhill, or downstream) Perceived only through its emissions, not through direct visual or tactile experience of its inherent physical properties (heard, smelled, felt in the skin or body)

Referent characteristics in exophoric use

Draw addressee’s attention to referent

Shifts addressee’s inaccurate attention to correct referent

Confirms that addressee’s attention is on referent

Attention-managing properties in exophoric use

Jahai Demonstratives

369

Specifically, the speaker used ton to indicate to the addressee that the attention of the latter was on the correct object among a pile of wooden shapes competing for reference in the game. Other demonstratives used in the task (which included speaker-anchored accessible, speaker-anchored inaccessible, addresseeanchored inaccessible, and addressee-anchored exterior) always had the function of diverting the addressee’s attention elsewhere. This and some further observational data also indicated that the addresseeanchored inaccessible form tũn (then thought to encode medial distance) was consistently used to divert the addressee’s attention to an object not attended to, irrespective of distance from the addressee.4 Thus, in these studies the addresseeanchored forms ton and tũn were interpreted to function primarily as attention-confirmer and attention-drawer, respectively. However, other observational data show that ton may also involve referents about which the addressee is considered to have prior knowledge but is not necessarily currently attending to, e.g. his or her dwelling, or an object recently manipulated by him or her. Furthermore, additional data suggest that the attentiondrawing function of tũn is restricted to situations in which the addressee’s attention is on the wrong entity (from the speaker’s perspective) or where he or she is explicitly asking for guidance. For example, tũn is frequently used in reply to location- and item-questioning. It is thus not simply an attentiondrawer but rather attention-corrector, intended to help the addressee shift attention to the intended referent. Taken together, these results and observations showed that the addresseeanchored forms handle the distinction between “shared knowledge” and “new information”. They were therefore reinterpreted to encode the speaker’s construal of the addressee’s cognitive accessibility to the referent (cf. Hanks, 2011: 330–331). The speaker-anchored forms tə̃ h and taniʔ similarly required a reappraisal. Here, the observational data suggested that physical distance from the speaker is an important factor in the choice of demonstrative form but hardly one that fully predicts which one is used. Several other semantic parameters play a role, including visual access, reachability, approachability, possession, and social relationships. For example, the presence of an obstacle between speaker and referent tends to occasion use of speaker-anchored inaccessible taniʔ at the expense of accessible tə̃ h, even in proximal table-top space (a pattern evident in the Director-Matcher game described above; Burenhult, 2003). This has clear parallels in large-scale geographic space, where, for example, referents located in the same river valley as the speech situation, although physically distant, may be denoted by speaker-anchored accessible tə̃ h. Referents located inaccessibly on the other side of a topographical divide 4

The reanalysis and re-interpretation of the medial was initially inspired by its spatial flexibility, as revealed by the results of the Demonstrative Questionnaire, explored in section 4.

370

Niclas Burenhult

are more likely to be referred to by means of inaccessible taniʔ, whether or not they are physically distant. The opposite side of a stream or river is referred to as sir taniʔ ‘that side’, as opposed to sir tə̃ h ‘this side’. So whereas the parameter of physical distance is unable to fully predict referent location, the wider (and vaguer) notion of physical accessibility seems to be able to account for most of the analyzed instances of the speaker-anchored forms. 4

The Questionnaire Study

As part of the refined analysis, I conducted an in-depth study of native speaker judgment of demonstrative meaning among Jahai speakers in 2002 with the help of the Demonstrative Questionnaire (Wilkins, 1999; this volume), an elicitation tool developed specifically for the detailed study of the extensional range of demonstrative pronouns in exophoric, non-contrastive spatial use. The study remains the only attempt I have made so far to characterize the full set of Jahai demonstratives with a single method; previously published work provides in-depth analyses of subsets of distinctions (Burenhult, 2003; 2008). The questionnaire data were collected with three adult male consultants. Interviewing took the form of enactments in Jahai of each of the 25 questionnaire scenes, where I asked the consultant to judge the acceptability of forms suggested by myself. I thus enacted the role of speaker. The role of addressee was enacted by a third person (native Jahai speaker). In other words, the consultant took on the participant role of observer and evaluator of a foreign language learner. This line of approach had practical advantages: me having greater control over the different parameters involved in the scenes and a better possibility to probe the full range of demonstrative distinctions in a systematic manner. It was also the approach I deemed easiest to comprehend and execute for the native speakers, who are not accustomed to prompted role-playing. The participants quickly grasped this format and the observing consultant produced judgments without difficulty. For each questionnaire scene, the eight known demonstrative forms were presented for judgment, and the consultant was asked (1) which forms were possible and (2) which form was the preferred one.5 Yes/no judgments were documented on paper in a pre-prepared form with a simple +/− value, each scene matched with the full set of demonstratives (creating a total of 200 judgments per consultant). Preferred forms were marked as such in the same form. Sessions were not audio- or video recorded. Interviewing could not be done in secluded settings, which meant that onlookers regularly took part in the elicitation. Thus, judgments by a consultant were frequently preceded by 5

The emission-perceptible tnɛʔ was not tested in this study. Note that at the time of data collection, this distinction had not yet been formally identified as a demonstrative.

Jahai Demonstratives

371

discussion among native speakers. However, the main consultant was always asked for a final say in each case. The questionnaire focuses on exophoric object-referential usage and the objects used here included a book (in two of the three elicitation sessions) and a jungle knife (in one elicitation session). Elicitation always involved the pronominal/adnominal demonstratives and suggested forms were presented in a sentence like “_jungle knife is a good one”. The questionnaire revealed a clear split in the demonstrative system in terms of the specificity and uniformity of responses. The subset of distinctions involving spatial frames of reference (including the speaker-anchored exterior, addressee-anchored exterior, superjacent and subjacent forms) exhibited clear spatial limitations as to the location of referents and a marked consistency in judgments between consultants. The accessibility-related distinctions (including the speaker- and addressee-anchored accessible and speaker- and addresseeanchored inaccessible) did not exhibit well-defined spatial limitations, and consultant judgments for these forms were less homogeneous. The following outline of the results reflects this division. 4.1

Accessible and Inaccessible Forms

The aim of the Demonstrative Questionnaire is to facilitate the differentiation and comparison of demonstrative distinctions pertaining to participantanchoring, distance and visibility, parameters which are considered central to demonstrative meaning. An interesting fact about the Jahai results is that the questionnaire encounters difficulties in uncovering clear patterns with respect to those parameters in forms where they would be most expected, namely the speaker- and addressee-anchored accessibles and inaccessibles. More specifically, it does not straightforwardly separate these four forms because they are all judged to be fully acceptable in most scenes of the questionnaire. As noted, judgment varies somewhat for the forms across the three consultants, but, as far as acceptability is concerned, the overall impression is that there is considerable overlap between them across the whole set of scenes. Judgments of preferred rather than acceptable forms reveal clearer patterns. To illustrate with two extremes: the first two scenes of the questionnaire pertain to demonstrative reference to body parts; in scene 1 the referent is one of the speaker’s teeth and in scene 2 one of the addressee’s teeth. Although consultants were occasionally hesitant with regard to some of the forms presented in this context, the overall pattern is clear: given the right circumstances, the full system of demonstrative distinctions may be used to refer to both the speaker’s and the addressee’s teeth. The only forms that display a clear, scenerelated restriction are the speaker-anchored and addressee-anchored exteriors (discussed in more detail in section 4.2), the former unacceptable for reference to the addressee’s teeth (scene 2), the latter unacceptable for reference to the

372

Niclas Burenhult

speaker’s teeth (scene 1). There are differences relating to which tooth is referred to: thus, whereas the speaker-anchored accessible, addresseeanchored accessible and inaccessible can be used to refer to any tooth of both speaker and addressee, the speaker-anchored inaccessible is judged acceptable only if referring to a back tooth (as are the speaker-anchored and addresseeanchored exteriors; also, the superjacent and subjacent forms can be used only to refer to a tooth in the upper and lower jaw respectively; see section 4.2). At the other end of the distance scale, scenes 24 and 25 pertain to reference to a distant object in large-scale geographic space. Here, the speaker- and addressee-anchored accessibles and inaccessibles can all be used. These two extreme examples just serve to illustrate that there is no clear correlation between the distance-related demonstrative distinctions and physical distance.6 Looking at the whole range of scenes, one pattern of divergence appears clearly discernible: the speaker-anchored inaccessible is judged to be unacceptable by all three consultants in most scenes in which an object is located closer to (but not on) the speaker than to the addressee (scenes 6, 7, 11, 19). The speaker-anchored inaccessible is also determined by two consultants to be inappropriate in some other scenes where the referent is located relatively close to the speaker, irrespective of proximity to addressee (scenes 3, 9, 20). However, it is acceptable in two scenes where the referent is located near a distant addressee (scenes 16, 18). The speaker-anchored accessible is judged by at least two consultants to be inappropriate if the referent is located near a third person far from speaker and addressee (scenes 13, 15) or near a distant addressee (scenes 16, 18). As far as the addressee-anchored accessible and inaccessible are concerned, both are generally judged by at least two consultants to be acceptable over the whole set of scenes. There are a couple of exceptions for each distinction: only one consultant accepts the addressee-anchored accessible for scenes 3 and 11, both of which involve a referent located close to the speaker; and only one consultant accepts the addressee-anchored accessible for scenes 13 and 21, two out of several scenes in which the referent is located away from both speaker and addressee. Consultants’ judgments of preferred (rather than acceptable) forms for each scene reveal a somewhat clearer picture. The speaker-anchored accessible is now preferred by at least two consultants for scenes in which the referent is located close to the speaker (scenes 1, 2, 3, 6, 7) or between speaker and addressee (scenes 8, 22); the addressee-anchored accessible is preferred by at least two consultants for scenes in which the referent is located close to the 6

My suggested initial reference for all demonstrative forms in the tooth examples was invariably an upper front tooth, but consultants spontaneously explained which teeth would go with each form, providing interesting contrastive data.

Jahai Demonstratives

373

addressee (scenes 4, 5, 9, 10, 16, 23) or, in one case, close to both speaker and addressee (scene 20). The addressee-anchored inaccessible is never chosen as the sole preferred form; in the few cases where it is judged as preferred, it is always equally preferable as some other form, and then only judged as such by one consultant (these instances appear to be associated with equidistance from speaker and addressee: scenes 12, 14, 17, 19, 20). The speaker-anchored inaccessible, finally, is the preferred form for at least two consultants for scenes in which the referent is located at a distance from speaker and addressee (scenes 13, 14, 15, 24, 25). It should be noted that if speakers are asked to state their reasons for choosing one of these distinctions as a preferred form, they will justify their choice with reference to proximity/distance. The speaker-anchored accessible is then associated with referents which are pdəh ‘to be near’ and the inaccessible is associated with referents which are mɲjiʔ ‘to be distant’. The addresseeanchored accessible and inaccessible do not have such distance-related connotations. 4.2

The Exteriors, Superjacent and Subjacent

In the second set of demonstrative forms – which consists of the frame-ofreference-encoding speaker- and addressee-anchored exteriors, the superjacent and the subjacent – consultant judgments of acceptability reveal significant restrictions. These forms encode angular distinctions which are not targeted by the questionnaire; yet their distribution is clearly identifiable by it, albeit not exhaustively. Take the exteriors, for example: the speaker-anchored exterior is considered fully acceptable in scenes 6 and 11; it is also acceptable in scenes 1 and 3 if the referent is located on the side of the speaker’s body that is facing away from the addressee. For other scenes it is generally judged as unacceptable. Similarly, the addressee-anchored exterior is considered fully acceptable in scenes 10 and 18, but in scenes 2, 4, and 5 only if the referent is located on the

1

3

6

11 S

S

A

A

S S

Figure 17.1 Scenes 1, 3, 6, and 11: Speaker-anchored exterior acceptable. The added arrow indicates the required position of the referent, if different from that indicated in the questionnaire.

Niclas Burenhult

374 2

4 S

A

5 S

10

S

A

A

18 S

S

A

A

Figure 17.2 Scenes 2, 4, 5, 10, and 18: Addressee-anchored exterior acceptable. The added arrow indicates the required position of the referent, if different from that indicated in the questionnaire.

side of the addressee’s body that is facing away from the speaker. For other scenes it is generally judged unacceptable. These results conform to the spatial distribution of exteriors outlined in section 3. As for the superjacent and subjacent forms, these are judged acceptable in scenes 1 and 2 if the referent (a tooth) is located in the upper or lower jaw respectively, and in scenes 3, 4, and 5 if the referent (a leech) is located on the participant’s head or lower extremities respectively. In other cases, acceptability is dependent on features of the local environment. One of the elicitation sessions was conducted in a slightly inclined setting with the enacted speaker at the higher end of the incline, and in scenes where the referent was located at the lower end of the incline the subjacent form was judged acceptable. Another elicitation session took place beside a stream with the enacted speaker frequently in an upstream position in relation to the referent. Again, the subjacent form was judged acceptable in these situations, although there was no perceptually recognizable difference in elevation between speaker and referent. Finally, the superjacent form was judged acceptable for distant referents in

Jahai Demonstratives

375

large-scale geographic space (scenes 24 and 25) if they were located saliently higher than the speaker. These results conform to the spatial definitions of the superjacent and subjacent given in section 3. Another feature that appears to set these forms apart from the accessibilityrelated forms described in section 4.1 is that the three consultants are more unanimous in their judgments of acceptability of them. This is consistent with their restricted and well-defined spatial distribution. Also, they are only sporadically judged to be a preferred form: two consultants give the speakeranchored exterior as the preferred form for scene 11; one consultant gives the addressee-anchored exterior as the preferred form for scene 18. 5

Discussion

5.1

Speaker-anchored Accessible and Inaccessible

The questionnaire data provide ample support to a number of earlier assumptions and hypotheses about the speaker-anchored accessibility distinctions. Firstly, they show that the speaker-anchored inaccessible taniʔ is acceptable in scenes where the referent is addressee-proximal but not speaker-proximal (e.g. scenes 16 and 18). This provides further support for the notion that this form is clearly participant-anchored to the speaker alone and not to the speaker and addressee as a dyadic unit. Secondly, the data confirm that the speaker-anchored forms are suitably defined in terms of “accessibility” rather than, say, “distance”, with a binary distinction between what is accessible and what is not. Accessibility is to be understood in a very wide sense as the speaker’s notion of the current relationship between participants and referent, taking a number of factors into account: distance, visual access, reachability, approachability, possession, manipulation, familiarity, etc. It is to be taken as devoid of primary spatial meaning. Positing accessibility as key meaning has several advantages. Firstly, in relation to the questionnaire data, it does a good job at encompassing the full range of acceptability these terms enjoy with respect to the spatial location of their referents. Given the right circumstances, most referents can be conceptually accessible or inaccessible irrespective of their spatial location. Yet the restrictions in their acceptability that do exist are also explicable: for example, the speaker-anchored inaccessible may be judged as unacceptable for speakerproximal referents simply because consultants have a hard time conceptualizing such referents to be out of access; similarly, an object located near a distant third person may be difficult to conceptualize as accessible and therefore not considered a suitable referent of the accessible form. The accessibility-based approach is also successful in coping with the inconsistencies in consultant judgments of acceptability: the many factors that form notions of accessibility

376

Niclas Burenhult

are exceedingly difficult to control for within the framework of the questionnaire. Small and unintentional differences in the elicitation setting are certain to influence judgments, and variation in responses to what is meant to be an identical scene then comes as no surprise. In a wider perspective, accessibility-based definitions have no trouble explaining the spatial patterns that do exist in the usage of these forms. Recall that the distinctions show some spatial tendencies as to locations of referents. Recall also that consultant judgments of preferred forms show a clearer spatial distribution than acceptable ones, and that consultants typically justify their choice of form with reference to proximity or distance. But these are symptomatic in the sense that they reflect the fact that some locations are conceptually more typical and more commonly encountered than others in relation to the notion of accessibility. Thus, the most typical location for a referent which is conceptually speaker-accessible is one in close proximity to the speaker; the most typical location for a referent which is speakerinaccessible is one distant from the speaker. Note again that I do not consider distance to be a primary semantic dimension in the speaker-anchored forms. But it is evident that notions of distance are a significant factor in consultants’ conceptualization of whether a referent is accessible or not. The apparent spatial patterning of referents is a most notable consequence of the wider notion of accessibility. Possibly, the impalpability of the general notion of accessibility is the reason why native speakers and researcher alike are tempted to resort to this much more tangible spatial symptom in their metalinguistic endeavor to define demonstrative meaning. 5.2

Addressee-anchored Accessible and Inaccessible

The addressee-anchored accessible ton behaves like the speaker-anchored distinctions in displaying considerable flexibility with regard to spatial location of its referent while still having a pattern of spatial bias, now centering on locations near the addressee. However, as noted in section 3.1 and Burenhult (2003), the addresseeanchored accessible exhibits pragmatic restrictions: it is associated with referents considered by the speaker to represent “shared knowledge”. Hence it is dedicated to referents conceptualized by the speaker as cognitively accessible to the addressee. Why this pragmatic specialization? An important factor to consider is that an addressee-anchored form encodes the speaker’s real-time impression of the addressee’s relation to the referent. Since the speaker does not have first-hand experience of the many factors that may determine whether a referent is considered accessible or not by the addressee, the functional nature of the addressee-anchored form

Jahai Demonstratives

377

is certain to be different from that of the speaker-anchored counterpart. Speakers use the addressee-anchored form to indicate what they perceive the addressee’s cognitive relationship to the referent to be like. A referent interpreted as perceptually and cognitively accessible to the addressee is then likely to be referred to with the addressee-anchored accessible form. A natural consequence of this is that the addressee-anchored accessible form is used for referents considered by the speaker to have the addressee’s attention or prior knowledge. This goes a long way in explaining why the addressee-anchored form appears to have such strong pragmatic characteristics related to “shared knowledge”. The addressee-anchored inaccessible tũn presents a partly different picture. Its distribution is very flexible in relation to the questionnaire scenes, and neither preferred nor accepted uses display any distinguishable spatial patterning. Referents spread out along the distance dimension from proximal to distal locations, and there is no observable spatial anchoring to either speaker or addressee. As noted in other types of data (see, e.g., Burenhult, 2003), it patterns with the addressee-anchored accessible in that it has pragmatic restrictions: it is associated with referents considered by the speaker to represent “new information” and to be “cognitively inaccessible” to the addressee (not known, not previously introduced, not attended to). This is in accordance with its function as attention corrector (see section 3.1). Although the questionnaire data do not expose attention correction as such as a parameter of tũn, its revealed spatial flexibility is certainly consistent with such encoding. 5.3

Summary

The Demonstrative Questionnaire study provides the first data set and in-depth portrayal of the Jahai demonstrative system as a whole. As employed here, it lays bare the fundamental distinction within the Jahai demonstrative system between forms which show flexibility with respect to spatial parameters (represented by the speaker- and addressee-anchored accessible and inaccessible) and those which are rigidly associated with certain angular contexts (represented by the speaker-anchored and addressee-anchored exteriors, the superjacent and the subjacent). The questionnaire does not effectively differentiate members of the former group from each other on spatial grounds, although judgments of preference show a higher degree of spatial differentiation than judgments of acceptability. The results support the hypothesis that accessibility, not physical distance, is the primary semantic dimension in this part of the system.

378

Niclas Burenhult

6

Conclusions

Exophoric demonstratives manage attention in actual space. Yet neither attentional nor spatial factors, alone or in combination, can fully account for all exophoric demonstrative usage in Jahai. Their ability to predict choice of demonstrative form or the location of a referent is partial, at best. Instead, a predictive model ought to set out from the language-specific semantic dimensions that emerge from the analysis, and it is such a model that I propose here. While this model has yet to be put to the test on a large-scale corpus, the semantic unpacking done so far on the system does allow us to draw some conclusions of wider significance. Firstly, as argued in Burenhult (2008), space can be a fundamental and inviolable semantic parameter in demonstratives (see Schapper and San Roque, 2011). The “spatial-coordinate demonstratives” of the Jahai system project angular search domains from the deictic center on the basis of spatial asymmetries (and thus invoke spatial frames of reference). These distinctions are directional indicators with as much precision as a pointing gesture. The spatial inflexibility of the Jahai forms comes out clearly in the questionnaire data. This is in stark contrast to the accessibility-related forms, whose indistinct distance patterning is an effect of the primary dimension of accessibility. The Jahai data here offer an illuminating distinction between primary spatial encoding and secondary spatial symptoms, of potential relevance cross-linguistically. Secondly, the Jahai system provides new albeit tentative clues to the attentionmanaging role of demonstratives, and it seems to underscore the need for a more thorough typology of such management. In particular, it suggests that attention-managing can be broken down into at least three functional subtypes: attention-drawing, attention-correcting, and attention-confirming. In Jahai, attention-drawing is the general purpose of seven of the demonstratives. Attention-correcting and attention-confirming, on the other hand, are associated with dedicated individual forms. Again, however, the pragmatic specialization of these forms does not represent primary encoding but rather a secondary effect resulting from the addressee-anchoring of the primary semantic dimension of accessibility. Thirdly, as is evident, cross-linguistically unusual distinctions are inextricably entwined in issues of core concern to current theorization of demonstratives. Apart from elucidating the role of space and helping towards a typology of attention management, they further highlight the proposed significance of binary contrast in demonstrative systems (Diessel, 2006: 469). They show that such contrast can structure a variety of dimensions beyond the well-known parameters of distance and participant: accessibility can be contrasted with inaccessibility, attention with non-attention, and directional opposites can surface in both absolute and intrinsic form.

Jahai Demonstratives

379

It is not the multitude of distinctions as such that makes the Jahai demonstrative system remarkable, nor is it the semantic parameters themselves (although some of those have yet to be documented elsewhere). For example, other languages add a variety of more or less optional auxiliary devices to demonstratives (e.g. adverbs, particles) to create angular contrasts akin to those in Jahai. But what is noteworthy is the way the language packages these functional dimensions into distinct and semantically compact forms, where each form encodes a complex configuration of space, access, or perception. Recall also the formal distinctiveness of the set and its internal equivalence: all nine distinctions converge with mutual exclusivity on a single syntactic slot. Their functional properties are thus formally equal and all deeply grammaticalized. These characteristics make Jahai demonstratives particularly helpful in our efforts to disentangle the semantic and pragmatic dimensions that matter in human attention-managing. Acknowledgments I am grateful to Jürgen Bohnemeyer, Sarah Cutfield, Michael Dunn, Nick Enfield, Alice Gaby, Bill Hanks, Nicole Kruspe, Steve Levinson, Loretta O’Connor, Eric Pederson, Angela Terrill, and David Wilkins for valuable discussions and comments on earlier versions. I wish to express my acknowledgements to the Economic Planning Unit, Putrajaya; the Department of Aboriginal Affairs, Kuala Lumpur; and the Department of Speech and Hearing Sciences, National University of Malaysia, Kuala Lumpur. The longterm research on which this chapter is based was carried out with generous support from the Max Planck Society, a European Community Marie Curie Fellowship, the Volkswagen Foundation’s DOBES program, the Swedish Research Council (421–2007–1281), and the European Research Council (LACOLA-263512). References Anderson, S. R. & Keenan, E. L. (1985). Deixis. In T. Shopen, ed., Language typology and syntactic description. Vol. III: Grammatical categories and the lexicon. Cambridge: Cambridge University Press, pp. 259–308. Burenhult, N. (2002). A grammar of Jahai. Unpublished Doctoral thesis, Lund University. (2003). Attention, accessibility and the addressee: The case of the Jahai demonstrative ton. Pragmatics, 13, 363–379. (2005). A grammar of Jahai. Canberra: Pacific Linguistics. (2008). Spatial coordinate systems in demonstrative meaning. Linguistic Typology, 12, 99–142. Diessel, H. (1999). Demonstratives: Form, function, and grammaticalization. Amsterdam: John Benjamins.

380

Niclas Burenhult

(2006). Demonstratives, joint attention, and the emergence of grammar. Cognitive Linguistics, 17, 463–489. Dixon, R. M. W. (2003). Demonstratives: A cross-linguistic typology. Studies in Language, 27, 62–112. Enfield, N. J. (2003). Demonstratives in space and interaction: Data from Lao speakers and implications for semantic analysis. Language, 79, 82–117. Enfield, N. J. & Bohnemeyer, J. (2001). Hidden colour-chips task: demonstratives, attention, and interaction. In S. C. Levinson & N. J. Enfield, eds., Manual for the field season 2001. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 21–28. Hanks, W. F. (2005). Explorations in the deictic field. Current Anthropology, 46, 191–220. (2011). Deixis and indexicality. In W. Bublitz & N. R. Norrick, eds., Foundations of pragmatics. Berlin/Boston: de Gruyter, pp. 315–346. Hyslop, C. (1993). Towards a typology of spatial deixis. Unpublished bachelor’s thesis, Australian National University, Canberra. Jungbluth, K. (2003). Deictics in the conversational dyad: Findings in Spanish and some cross-linguistic outlines. In F. Lenz, ed., Deictic conceptualization of space, time and person. Amsterdam: John Benjamins, pp. 13–40. Küntay, A. C. & Özyürek, A. (2006). Learning to use demonstratives in conversation: What do language-specific categories in Turkish reveal? Journal of Child Language, 33, 303–320. Levinson, S. C. (2003). Space in language and cognition. Cambridge: Cambridge University Press. (2006). Deixis. In L. Horn & G. Ward, eds., The handbook of pragmatics. Malden, MA/Oxford: Blackwell Publishing, pp. 97–121. Özyürek, A. (1998). An analysis of the basic meaning of Turkish demonstratives in faceto-face conversational interaction. In S. Santi, I. Guaïtella, C. Cavé & G. Konopczynski, eds., Oralité et gestualité: Communication multimodale, interaction. Paris: L’Harmattan, pp. 609–614. Schapper, A. & San Roque, L. (2011). Demonstratives and non-embedded nominalisations in three Papuan languages of the Timor-Alor-Pantar family. Studies in Language, 35(2), 380–408. Schebesta, P. (1928). Grammatical sketch of the Jahai dialect, spoken by a Negrito tribe of Ulu Perak and Ulu Kelantan, Malay Peninsula [trans. and adapted by C. O. Blagden]. Bulletin of the School of Oriental Studies 1928, 803–826. Seifart, F. (2003). Encoding shape: Formal means and semantic distinctions. In N. J. Enfield, ed., Field research manual 2003 part I: Multimodal interaction, space, event representation. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 57–59. Terrill, A. & Burenhult, N. (2008). Orientation as a strategy of spatial reference. Studies in Language, 32, 93–136. Wilkins, D. P. (1999; this volume). The 1999 demonstrative questionnaire: “This” and “that” in comparative perspective. In D. P. Wilkins, ed., Manual for the 1999 field season. Nijmegen: Max Planck Institute for Psycholinguistics, pp. 1–24.

Index

absolute frame of reference, 27–28, 33, 366 accessibility, 14, 29, 31, 54 Dalabon, 114 discussed in the Demonstrative Questionnaire, 60–66 Goemai, 142 Jahai, 25, 362, 366, 367–370, 371–373, 375–379 Lao, 72, 80–81, 84–87 Yélî Dnye, 331–332 Yucatec, 188, 189, 195–196, 197–198 American English, 11, 129 anaphora, 10, 33–34 distal deixis and, 164 anaphoric pronouns, 4, 11, 119, 237, 251, 324 Anderson, S., 7, 9, 19, 25, 27, 186 apprehension, mode of, 30–31 arm’s reach, referral to things in and out of, 65 Arrernte, 69 indicative acts and choice of demonstrative, 67 spatial deictic demonstratives, 61–64 Aslian languages, 363 attention direction of, 196–199 joint, 2, 7, 9, 29, 35, 81, 176, 188, 201 role in function of demonstratives, 29–30 Auhelawa, 259 Austronesian languages, 319 Barral, B., de, 284–285, 287 Bhat, D. N. S., 4 Bininj Kunwok, 92 body referral to parts and places on, 64 referral to things in contact with, 64–65 Bogoras, W., 307–308 Bohnemeyer, J., 15, 31, 152 Brazilian Portuguese, 16, 18, 25, 35, 116–117, 229 accessibility distinction, 31

analysis of demonstratives and place adverbs, 126–130 compared with Trumai, 255–256 conclusions following analysis, 130–131 distal demonstratives, 20 esse acqui (this here), 19 similarities with Goemai, 141, 142 spoken demonstrative system, 119–122 use and results of the Demonstrative Questionnaire, 122–126 written demonstrative system, 117–119 Bühler, K., 66–67, 73 Buhutu, 259, 261, 269 Burenhult, N., 25, 27–29, 31, 364, 366, 367, 369, 376, 378 Câmara Jr., J. M., 120–122 Cariban, 222, 224 Carlin, E., 223, 224, 230 Chukchi, 25, 36, 314–315 attention in function of demonstratives, 29 contrast of two distals, 311 contrastive use, 32 deictic particles, 306–307 demonstrative adverbs, 306 demonstrative pronouns, 304–305 distal demonstratives, 310 earlier analyses, 307–309 form and syntax of demonstratives, 5 gesture, 311 invisibility of referent, 30, 312 language and its speakers, 303 naturalistic observation data, 312–314 neutral demonstratives, 310–311 proximal demonstratives, 310 waj=ənqen and raj=ənqen/caj=ənqen, 311–312 Cleary-Kemp, J., 258, 266 contrastive use, 24, 32. See also noncontrastive use Dalabon, 110–111 Trumai, 252

381

382

Index

contrastive use (cont.) Lavukaleve, 218 Saliba-Logea, 277 Warao, 294–296 versus non-contrastive use, 11, 14, 32 Yélî Dnye, 334–335 Cooper, R., 259, 269 Coventry, K. R. et al., 8, 31 Cutfield, S., 112 Dalabon, 35, 173 accessibility of referent, 31 contrastive uses of demonstratives, 32, 110–111 deictic gestures, 104–106, 110–111 Demonstrative Questionnaire discussion of findings, 113–114 methodology, 97–98 summary findings, 98–100 description of form classes and previous claims, 92–97 discourse status, 106–107 distal demonstratives, 23 emotionally-deictic uses data, 112 free pronouns distinct from demonstratives, 95 heuristics and implicatures in the paradigm, 112–113 identificational demonstratives, 95 language and its speakers, 92 non-spatial demonstratives, 104 ownership of referent, 109–110 pointing gesture, 32 politeness strategy, 108, 112 referents in the here-space referred to with nunda, 100–102 referents in the there-space referred to with djakih, 102–104 visibility of referent, 107–109 Dawuda, C., 258, 261 De Goeje, C. H., 223 deictic anchoring, 197 deixis, 2–11, 150, 164 discourse, 10–11, 76, 312 emotional, 112 neutral, 193 recognitional, 76 social, 320 spatial, 73–75, 177, 179–184, 188, 190, 196–199, 320, 346, 364 temporal, 24, 312–313, 320 ‘Demonstrative questionnaire: “This” and “that” in comparative perspective case study’ (Wilkins, 1999), 12–15, 206, 364 application and results

Brazilian Portuguese, 122–126 Chukchi, 309–312 Dalabon, 97–110 Goemai, 139–142 Jahai, 370–375, 377 Lao, 77 Lavukaleve, 213–217 Saliba-Logea, 270–276 Tidore, 349–352 Tiriyó, 226–231 Trumai, 242, 246–248, 255 Tzeltal, 161–166, 173 Warao, 282, 288–293 Yélî Dnye, 325–334 Yucatec, 188–190 compared with Hanks’ analysis of Yucatec, 176–178, 187 design and use, 44–45 design parameters, 45–55 discussion of distance/domain scale, 60–66 discussion of pointing, touching and presenting, 66–71 method of recording, 45, 87 number of consultants, 45 organization of data, 55 purpose, 43–44 demonstrative systems. See also individual languages generalizations, 34–38 demonstratives distinctions in pronominal vs. adverbial, 18–19 folk definitions, 63–64 form and function, 16–18 form and syntax, 4–5 identification of, 3–4 importance in theory of language, 2–3 non-spatial distinctions, 29–34 parameters for design space, 34–35 semantics of, 5–9, 15 uses, 10–11, 15. See also contrastive use; non-contrastive use Derbyshire, D. C., 224 Diessel, H., 3, 7, 9, 16, 25, 29, 186 definition of demonstratives, 365 demonstrative distributions, 92 distance distinctions, 19 morphology and syntax of demonstratives, 4–5 direction, 26–29 Director-Matcher game, 367, 369 distance, 7, 9, 19–25 speaker-centered system, 327–329 distance scale, 60–66

Index Dutch distal demonstratives, 20 indicative acts and choice of demonstrative, 69 ‘Eliciting contrastive use of demonstratives for objects within close personal space’ (Wilkins, 1999), 12, 32 Enfield, N. J., 8, 14, 72, 193, 300 English, 8, 257 anaphoric reference, 33 distal demonstratives, 20 this and that, 1, 2, 11, 21, 225 entailment scale, 194 Ewe, 68 eyebrow points, 104 eyebrow raise, 111 Fillmore, C. J., 7, 9, 10, 66 frames of reference, 26–29, 365, 366 Francois, A., 261, 267 Frei, H., 16, 19

383 differences in referential practice, 70 semantics of demonstratives, 7–8 head nod/point, 10, 33, 67, 105, 292 Heath, J., 60 Heinen, D., 297 Henderson, A., 319 Henderson, J., 319–320, 324, 330–332, 336–338, 340 Herrmann, S., 15 Himmelmann, N. P., 298 home space, referral to things within, 65–66 Horn scale, 194 idealized model of speaker-anchored radial spatial categories, 6 indexical ground, 7, 9, 14, 19, 21, 22 Indonesian, 343 Inenlikej, P. I., 308–309, 310 intrinsic frame of reference, 26–29, 366 Italian, 20

gaze, 6, 7, 10, 32–33, 45, 70 Dalabon, 104–105, 111 Yélî Dnye, 323, 329 German, 16 gestures, 6, 10–11, 15, 27, 32–33, 34. See also pointing Arrernte, 67 Chukchi, 310 deictic (in Dalabon), 104–106, 110–111 Yélî Dnye, 332–334 Goemai, 16, 25, 29, 34, 35, 134–135 diachronic origins of demonstratives, 135–139 evidence from naturalistic data, 142–147 function of demonstratives, 29 invisible referent, 30 proximal and distal demonstratives, 22 results from the Demonstrative Questionniare, 139–142 Grice, H. P., 24, 193–195, 328 Gricean conversational principles, 21 Gricean maxim of Quantity, 24, 328 Gricean mechanisms, 193–195 Guirardello-Damian (Guirardello) R., 142, 243, 245 Gunwinyguan language family, 92

Jahai, 9, 18, 29, 36, 361–362 accessibility demonstratives, 25, 31, 367–370 addressee-anchored accessible and inacessible, 376–377 and inaccessible forms, 371–373 speaker-anchored accessible and inaccessible, 375–376 conclusions from study, 378–379 definitions and terminology of demonstrative system, 364–365 elicitation using the Demonstrative Questionnaire, 370–371 exteriors, superjacent and subjacent forms, 373–375 form and syntax in demonstratives, 5 forms and functions of demonstratives, 365–367 frames of reference, 28 invisible referent, 30 language and its people, 362–363 previous analyses, 363–364 Japanese, 7, 29 indicative acts and choice of demonstrative, 67–69 Jawoyn, 92 Jungbluth, K., 14

Hanks, W. F., 15, 28, 31, 66, 152, 155 analysis of Yucatec demonstrative system, 159, 184–188 comparison with, 176–180, 190, 191, 195–196, 199–201

Kaplan, D., 66, 194 Keenan, E., 7, 9, 19, 25, 27, 186 Kemmerer, D., 8 Kirsner, Robert, 69 Kita, S., 104

384

Index

Korean, 61 proximal, medial and distal factors, 60–61 Kriol, 92 language acquisition, 2, 11, 38, 150 Lao, 8, 18, 19, 35, 72–73 conclusions from study, 87–88 contextual factors in selection of nan4 or nii4, 80–83 influence of addressee location arising from Schelling Game pragmatics, 84–87 larger set of demonstratives, 75–76 mechanisms of spatial deictic reference, 73–75 nan4, semantic distinction of place, not distance, 77–80 proximal and distal demonstratives, 22 responses from the Demonstrative Questionnaire, 77 semantics of nii4 and nan4, 76–77 unmarked nii4 when nan4 is not applicable, 83–84 Lavukaleve, 25, 28, 34, 36, 206 accessibility of referent, 31 conclusions on the use of exophoric demonstratives, 219–220 contrastive demonstrative usage, 32, 218 demonstrative morphology, 210–212 language features, 207–208 natural observational data, 217–218 pointing gesture, 32 pronoun and demonstrative forms, 208–210 proximal demonstratives, 24 similarities with Chukchi, 314 speakers of, 207 three-term system, 212 use of the Demonstrative Questionnaire, 213 Leavitt, C., 223 Levinson, S. C., 193, 194, 313 frames of reference, 26–27 lip point, 10, 33, 67, 104, 111, 292 Lithgow, D., 259–260 Lyons, J., 5, 156 Malay, 343, 363 Margetts, A., 258, 260, 275 Max Planck Institute for Psycholinguistics, 20, 176, 188, 222 Language and Cognition Group, 364 Mayan, 150–151, 159, 173, 176, 179 structure of demonstratives, 152 terminal deictics, 70 McNeill, D., 104 Meira, S., 142, 188, 222, 223, 224–225, 229 Moll, T. A., 308–309, 310

Mosel, U., 258 Mwotlap, 261, 267 N400, 9 Ngalakan, 92 non-contrastive use, 11, 14, 32. See also Brazilian Portuguese; Chukchi; Goemai; Tidore; Tiriyó; Trumai Dalabon, 92, 98, 113 Lavukaleve, 218 Saliba-Logea, 278 Warao, 294–296 North Halmahera language family, 343, 346 North Moluccan Malay, 343 Northeast Halmaheran languages, 347 Nunggubuyu, 28, 61 proximate, immediate and distal demonstrative roots, 60 Osborn, H., 284, 286–287, 293 Özyürek, A., 177 Palmer, B., 348 paradigmatic oppostion, 18–19 Pederson, E., 322 perceptual modality, 366–367 Peteliyaki, B., 259–260 Pidgin, 257 place adverbs, 179 Brazilian Portuguese, 117, 120, 123–125, 126–130, 131, 238, 255 Dalabon, 96 Saliba-Logea, 257, 258, 260, 265, 267–268, 269, 275, 276, 278–279 Trumai, 245, 254 pointing, 2, 10–11, 27, 32–33 Brazilian Portuguese, 125 Chukchi, 311 Dalabon, 104–106, 110–111 discussion in the Demonstrative Questionnaire, 66–71 proximity and, 27 Saliba-Logea, 274–275, 278 Tiriyó, 222, 228, 231, 235 Tzeltal, 162–164, 166–168 Warao, 34, 292–293, 299 Yélî Dnye, 329 Polian, G., 152, 160 politeness, 79, 87, 108, 112, 218 Portuguese, 243 Q-heuristic, 193 recipient design, 74 relative frame of reference, 26–29

Index Rembarrnga, 92 Romero-Figeroa, A., 284, 285–287 Rossel Island, 318–319 Russian distal demonstratives, 21 etot and tot, 225 Saliba-Logea, 25, 257–258 accessibility of referent, 31 adnominal demonstrative clitics, 264–267 attention and function of demonstratives, 30 clause-final demonstratives, 260–261 competition of demonstratives with other forms, 278–279 deictic distinctions, 273–274 Demonstrative Questionnaire elicitation procedure, 270–271 summary results, 271–273 distal terms, 23 extendability of demonstratives, 278 free demonstratives, 262–263 full place adverbs, 269 givenness and visibility, 275 interjections, 267–269 observational data, 276–277 ownership and control, 275–276 pointing and proximals, 32–33 pointing and touching, 274–275 social boundaries, 276 studies of, 258–260 visibility of referent, 30 Schebesta, P., 363 Schelling Game, 6, 73–74, 84–87 Senft, G., 178 Shape Classifier Task, 364 Shaw, Ben, 319 Sirlinger, E., 138 Skorik, P. J., 307–308 social space, referral to things within, 65 Sohn, H. M., 60 Spanish, 8, 283, 284 Spanish-Warao, 283 spatial bias, 7–8 spatial cognition, 9 spatial frames of reference, 26, 365, 366 spatial-coordinate demonstratives, 365, 378 Suau, 257 Suauic language family, 257, 259 table-top placement task, 322–323 Tenejapan Tzeltal, 151–152 forms and basic oppositions, 152 Terrill, A., 223 third-person pronouns, 34 Brazilian Portuguese, 116

385 Tidore, 344 Tiriyó, 224–225, 234–238 Warao, 284 Tidore, 25, 36 accessibility of referent, 31 application of the Demonstrative Questionnaire, 349–352 directional verbs, 345 form classes, 344–349 frames of reference, 28 geographically anchored terms, 347–349, 351–355 in geographic space, 356 in interactional space, 356–357 interactionally anchored terms, 348, 349–354 in interactional space, 355–357 language and its speakers, 343 observational data, 357–359 Tiriyó, 33, 34, 36 analyses of demonstratives, 224–225 animacy and number, 224 attention and function of demonstratives, 30 contrastive use, 23 Demonstrative Questionnaire application, 226 discussion of results, 226–231 demonstrative system, 222 exophoric demonstrative system, 234–238 language and its speakers, 222–224 observational data distal scene, 231 medial scenes, 232 proximal scenes, 232–234 pointing, 32–33 Tongan, 9 touching Brazilian Portuguese, 125 Chukchi, 310 discussion in the Demonstrative Questionnaire, 63–65, 66–71 Goemai, 147 Saliba-Logea, 274–275, 278 Tidore, 355–356 Tiriyó, 33, 222, 228, 231, 234–235 Trumai, 254 Tzeltal, 166–168 Warao, 292–293, 296, 299 Yélî Dnye, 328–329, 332, 335 Yucatec, 194 Trumai, 25 analysis of demonstrative system, 253–255 contrastive use of demonstratives, 252 data from the Demonstratives Questionnaire, 246–248

386

Index

Trumai (cont.) demonstrative system, 242–243, 255–256 function of demonstratives, 29 language and its speakers, 243 naturalistic/observational data, 248–252 studies of the demonstrative system, 243–246 Turkish, 7, 12 attention in function of demonstratives, 29 šu demonstrative, 177, 333 Tzeltal, 18, 36, 150 attention and function of demonstratives, 30 conclusions from study, 173 demonstrative adjectives, 155–157 demonstrative adverbs, 157–158 demonstrative pronouns, 155 demonstratives, directionals, and time reference, 170–172 dialects, 151–152 distal demonstratives, 23 form and function, 17 form and syntax of demonstratives, 5 gesture, 32, 162–164, 166–168, 173 initial and terminal deictic pairings, 159–161 language and its speakers, 150–152 motion and direction in demonstrative usage, 169 Oxchuc dialect, 162 presentational demonstrative, 158–159 prior attention versus calling attention to referent, 164 results from the Demonstrative Questionnaire, 161–166 spatial factors and the referent, 164–166 tey a mene (that over there), 19 visibility of referent, 30, 164 Van Staden, M., 348, 352, 356 Vaquero, A., 284 visibility/invisibility, 14, 30 Chukchi, 312 Dalabon, 107–109 Saliba-Logea, 275 Tiriyó, 229–230 Tzeltal, 164 Yélî Dnye, 330–332 Warao, 15, 25, 36 additional data types and variables, 294 attention on object, 296–297 consultants’ ideal definition of a term, 294 contrastive use, 294–296 demonstrative pronouns, 291

distance from speaker, 293, 299–300 form and class of demonstratives, 287–288 gesture, 33, 34, 292–293, 299 invisible referent, 30 language and its speakers, 283–284 results using the Demonstrative Questionnaire, 289 studies from the lower Delta region, 284–285 studies from the western Delta region, 285–286 use of demonstratives in discourse, 297–298 visibility of referent, 31, 292 Warlpiri, 63 West Greenlandic, 9, 28 Wierzbicka, A., 77 Wilkins, D. P., 32, 176, 188, 206, 322 World Atlas of Language Structures, 9 Yélî Dnye, 36, 173, 313, 339–340 accessibility of referent, 31 addressee-centered ye, 329–330 adverbial uses of demonstratives, 335–336 anaphoric determineter, yi, 332 anaphoric reference, 34 application of the Demonstrative Questionnaire, 325–327 attention and function of demonstratives, 30 contrastive use of demonstratives, 32, 334–335 demonstratives in the context of grammar, 320–325 distal and proximal demonstratives, 24 gesture, attention, mutual knowledge and choice of demonstrative, 332–334 language and its speakers, 318–320 pointing and proximals, 32 pre-verbal nucleus, 336–339 proximal and distal demonstratives, 23 speaker-centered distance system: ala ➔ kî ➔ mu, 327–329 temporal uses, 336 wu and visibility/invisibility of referent, 30, 330–332 Yucatec, 18, 36, 159, 176–179 accessibility of referent, 31 attention and function of demonstrative, 30 comparison with Hanks’ analysis of demonstratives, 199–201 contrastive use, 32 expression of spatial deixis, 179–184 form and function, 17 form and syntax of demonstratives, 5

Index impact of the addressee’s location on demonstrative choice, 190–192 invisible referent, 31 language and its speakers, 179–184 pointing gesture, 32 proximal and distal demonstratives, 22

387 role of attention direction in spatial deixis, 196–199 semantics of the non-immediate forms, 196 use and results of the Demonstrative Questionaire, 188–190 visibility of referent, 30