Deictic Conceptualisation of Space, Time and Person (Pragmatics and Beyond New Series)

Deictic Conceptualisation of Space, Time and Person

Pragmatics & Beyond New Series Editor Andreas H. Jucker University of Zurich, English Department Plattenstrasse 47, CH-8032 Zurich, Switzerland e-mail: [email protected]

Associate Editors Jacob L. Mey University of Southern Denmark

Herman Parret Belgian National Science Foundation, Universities of Louvain and Antwerp

Jef Verschueren Belgian National Science Foundation, University of Antwerp

Editorial Board Shoshana Blum-Kulka Hebrew University of Jerusalem

Catherine Kerbrat-Orecchioni University of Lyon 2

Jean Caron Université de Poitiers

Claudia de Lemos University of Campinas, Brazil

Robyn Carston University College London

Marina Sbisà University of Trieste

Bruce Fraser Boston University

Emanuel Schegloff University of California at Los Angeles

Thorstein Fretheim University of Trondheim

Deborah Schiffrin Georgetown University

John Heritage University of California at Los Angeles

Paul O. Takahara Kansai Gaidai University

Susan Herring University of Texas at Arlington

Sandra Thompson University of California at Santa Barbara

Masako K. Hiraga St.Paul’s (Rikkyo) University

Teun A. Van Dijk Pompeu Fabra, Barcelona

David Holdcroft University of Leeds

Richard J. Watts University of Berne

Sachiko Ide Japan Women’s University

Volume 112 Deictic Conceptualisation of Space, Time and Person Edited by Friedrich Lenz

Deictic Conceptualisation of Space, Time and Person Edited by

Friedrich Lenz University of Passau

John Benjamins Publishing Company Amsterdam
/
Philadelphia

8

TM

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences – Permanence of Paper for Printed Library Materials, ansi z39.48-1984.

Library of Congress Cataloging-in-Publication Data Deictic Conceptualisation of Space, Time and Person / edited by Friedrich Lenz. p. cm. (Pragmatics & Beyond, New Series, issn 0922-842X ; v. 112) Includes bibliographical references and indexes. 1. Grammar, Comparative and general--Deixes. 2. Space and time in language. I. Lenz, Friedrich. 1956- II. Series. P299D44 D44 2003 415-dc21 isbn 90 272 5354 4 (Eur.) / 1 58811 375 2 (US) (Hb; alk. paper)

2003043712

© 2003 – John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa

Contents

Deictic conceptualisation of space, time and person: Introduction Friedrich Lenz

vii

Part 1: Space

1

‘Addressee effects’ in demonstrative systems: The cases of Tiriyó and Brazilian Portuguese Sérgio Meira

3

Deictics in the conversational dyad: Findings in Spanish and some cross-linguistic outlines Konstanze Jungbluth

13

Non-deictic uses of the deictic motion verbs kommen and gehen in German Claudio Di Meola

41

Origo, pointing, and conceptualization — what gestures reveal about the nature of the origo in face-to-face interaction Ellen Fricke

69

Part 2: Time

95

Two ways of construing complex temporal structures Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

97

“Look here, what I am saying!”: Speaker deixis and implicature as the basis of modality and future tense Thomas A. Fritz

135

The ‘subjective’ effects of negation and past subjunctive on deontic modals: The case of German dürfen and sollen Tanja Mortelmans

153

vi

Contents

Part 3: Person and Text

183

Politeness distinctions in second person pronouns Johannes Helmbrecht

185

Deictic use of demonstrative pronouns in the Rigveda Katharina Kupfer

203

Towards a unified model of domain-bound reference Manfred Consten

223

Deixis and speech situation revisited: The mechanism of perceived perception Heiko Hausendorf

249

Index

271

Deictic conceptualisation of space, time and person Introduction Friedrich Lenz University of Passau

Deixis is a universal phenomenon. Every language has a set of deictic expressions, most of which are frequently used and highly grammaticalised. Typical examples are demonstratives, personal pronouns, tenses, certain place and time adverbials and some verbs such as come and go. They cannot be described without recourse to “aspects of the communication act” (Fillmore 1975, 39), “properties of the extralinguistic context” (Anderson/Keenan 1985, 259) or “features of the context of utterance or speech event” (Levinson 1983, 54). Given that all standard deªnitions of pragmatics somehow hinge on the notion of language use in context, there is no doubt that the study of deixis genuinely belongs to pragmatics. But the study of deixis must go beyond a narrow pragmatic perspective. Deixis is at the interface of pragmatics and semantics. Though it does not lend itself to a traditional truth-semantic approach and it is problematic to give the utterance-independent sense or denotation of a deictic expression,1 a deictic clearly has referential meaning. Whenever a deictic is employed, the speaker refers to a certain entity. In this light it does not come as a surprise that the presumably most frequently quoted deªnition of deixis originates from a book on semantics. Lyons (1977, 637) deªnes deixis as “the location and identiªcation of persons, objects, events, processes and activities being talked about or referred to, in relation to the spatiotemporal context created and sustained by the act of utterance and the participation in it.” Deixis must therefore be studied on the level of discourse and can be regarded as “a special variety of reference” (Hanks 1992, 46). Successful deictic reference is not only

viii Friedrich Lenz

embedded in a speciªc speech event, but depends on the participants’ awareness of its constituents. Empirical data of place, time and participant roles must be related to those of the deictic referent. In this respect, the study of deixis calls for empirical investigations. Although the study of deictic reference necessarily has an empirical component, deictic research must also take account of perceptual and cognitive processes. In order to identify the referent, the attention of the addressee must be guided towards the referent. We need not only know the constitutive elements of the speech event, but also how they are conceived. Bühler’s (1934/ 1990) notion of the Zeigfeld (deictic ªeld) and his concept of the origo, the egocentric deictic zero point, show his conviction that the meaning of deictics2 is inseparably interwoven with aspects of perception and cognition. This insight is taken up in modern cognitive linguistics (not always giving Bühler credit though) where meaning is ultimately understood as a matter of conceptualisation (cf. Langacker 1987). It can be claimed that deixis is conceptualised in terms of an idealised cognitive model and “a deictic expression is one that builds up a mental space in which the speaker and the addressee are co-present at a given moment in time..., the mental space evoked by a deictic expression involves the conceptualisation of the deictic centre” Marmaridou (2000, 100). In local deixis the notion of mental space has a direct perceptual parallel. The referent is identiªed from a vantage point on the basis of ªgure-ground relations in a visual ªeld. In this type of deixis we can literally point in order to direct the addressee’s attention (deixis means pointing). For some local deictics visual perceptibility of the deictic centre is even a prerequisite for successful communication as they require the indication of a direction by gestures or equivalents (gaze for instance). In local space proximity-distality distinctions are immediately relevant, whereas in the other deictic dimensions the underlying centre-periphery image schema (cf. LakoŸ 1987) and its diŸerentiation can only be applied through conceptual metaphor. In time deixis the abstract domain of time is conceptualised by recourse to the concretely perceivable domain of space (cf. Radden 1997). In a linear order schema the referents are understood as ‘located’ at a certain ‘distance’ from the deictic centre. The onedimensional space can be construed along identity, overlap, proximity and distality relations between ªgure and ground. Since discourse happens in time and consists of temporal entities, discourse deictic referents are similarly understood as linearly ordered with a distance to the deictic centre (cf. Lenz 1999). In person deixis a space with social entities is built up. In this social space the vantage point is again the speaker. He or she is central in terms of

Introduction

participant roles, but social roles and interpersonal ‘distance’ can additionally be encoded. Deixis is based on the conceptualisation of the speech event and its cognitive categorisation is re¶ected in the grammatical systems of each language.3 Place, time and person are the constitutive elements of the speech event and all languages have deictic means to refer to entities in the corresponding mental space. Although there are remarkable similarities across languages, there are also diŸerences in the way in which local, temporal and personal entities are conceptualised and encoded. This book presents the results of theoretical and empirical investigations into the pragmatics, semantics and grammar of deictic expressions in several languages. Special emphasis is given to contrastive studies which take cognitive and cultural context into account.4 The volume is structured along the dimensions of the mental deictic space. In the ªrst part of the book the focus is on the conceptualisation of local space. The ªrst two contributions investigate the deictic use of demonstratives in speciªc languages. Demonstrative systems are traditionally classiªed according to the number of demonstratives that belong to the system and whether they are distance or person oriented. Sérgio Meira points out that demonstrative systems which are classiªed as distance oriented are not always entirely speaker grounded. His analysis of the use of demonstrative pronouns in Tiriyó, a Cariban language spoken in the Amazon region by about 2000 people, is based on ªeld work using a questionnaire as an elicitation tool. Tiriyó had been described as a typical distance oriented system with a proximal, medial and distal term, but Meira ªnds a tendency to favour medial terms over distals when the addressee is close to the referent. The demonstratives open up a “conceptual area” where distance is perceived relative to the position of the addressee. Meira argues that the participants share a conceptual proximity, even when they are physically far from each other. The same questionnaire was given to Brazilian speakers of Portuguese. Spoken Portuguese is assumed to have developed into a two-term distance oriented system (esse vs. acquele) as the third term of the written variety has apparently been dropped. The analysis of the questionnaires suggests that the addressee still plays a role in the mental space which Meira subdivides into a speaker-and addressee area and distal-from-speaker-andaddressee area. The next contribution deals with the Spanish demonstrative pronouns ese, este, aquel. There has been a long-running controversy as to whether the Spanish demonstrative system is to be interpreted as person or distance oriented. Constanze Jungbluth is not satisªed with the traditional interpretations

ix

x

Friedrich Lenz

on the grounds that they are all based on a limited conception of the relevant space. She argues for a conversationally constituted deictic space which she calls the dyad of conversation. This space is cognitively structured according to the position and the direction of both speaker and hearer. In face-to-face, faceto-back and side-by-side conversation the focal object is conceptualised in speciªc ways and accordingly encoded. On the basis of observation and elicited data the Spanish demonstrative system is reclassiªed for each dyad type and contrasted with other three term systems. Motion can be described as the change of the location of an entity through time. The conceptualisation of motion requires a complex train of cognitive events as both temporal and spatial concepts are implied. Though the expression of motion is often deictically grounded, deictic motion verbs have attracted less research attention than the highly grammaticalised demonstratives. Claudio di Meola’s study is a corpus-based investigation of the German motion verbs kommen and gehen applying LakoŸ’s source-path-goal image schema to describe their deictic core meaning. Meola, however, concentrates on their non-deictic variants, the meaning of which is based on metonymical extensions. Non-deictic gehen is used to focus on unrestricted motion, whereas nondeictic kommen refers to cases where external force in¶uences the movement performed by the human trajector. The identiªcation of entities in space by deictics can and in some cases even must be accompanied by gestures. Both the verbal and the gestural acts are grounded in the speech event. Ellen Fricke’s article examines the role of gestures in deictic reference acts on the basis of video recordings of route descriptions. Special emphasis is given to the interactive co-ordination of the origo which must be regrounded with every speaking turn. Fricke elaborates on Bühler’s distinction between perceivable deixis (Deixis ad oclulos) and imaginary deixis (Deixis am Phantasma) and takes up Bühler’s metaphor of the imaginary wanderer through (mental) space. Her data suggest that there are two types of origo involved, and therefore the perspective is only partly shifted in the course of the interaction. On the gestural level the speaker is perceived as the origo, but on the verbal level the origo is attached to the addressee projecting the future as an imaginary wanderer. The second section of the book deals with temporal and modal mental space. The article by von Stutterheim, Carroll and Klein is a contrastive study of English and German ªlm retellings in which speakers must construe complex temporal structures. Tense and time adverbials are both typical means to build up a referential ground. In the authors’ theoretical framework tenses

Introduction

express temporal relations between the temporal anchor, often but not necessarily the moment of speech, and the topic time, the time being talked about (a redeªnition of Reichenbach’s reference time). Their data suggest that English speakers take a reportive perspective where the temporal anchors proceed with the individual utterances corresponding to the experiencer’s now. They employ a deictic ‘now-I-see strategy.’ The German speakers favour an anaphoric strategy which links the current topic time to the preceding situation time. The authors claim that grammaticalised categories trigger the processing format for conceptual substance. The preference for the ‘now-I-see strategy’ can then be seen in the light of the English aspectual system. English has a grammaticalised progressive aspect. An event can be viewed as ongoing from an external reference point, for which the experiencer’s now is a good candidate. Thomas Fritz’s contribution deals with modality and the future tense. He applies pragmatic concepts to explain the utterance dependent meaning of mood and modal verbs arguing that every sentence contains a deictic element for the factuality of a proposition and the speaker’s believe in it are inherently expressed in every communicative act. The amount of factuality to which the speaker is committed is interpreted on the basis of scalar implicature. As the speaker’s commitment about forthcoming events cannot be modelled in a traditional logical way either, the future tense is interpreted along similar lines. Tanja Mortelmans also focuses on the expression of modality. She investigates German mood markers adopting Langacker’s basic epistemic model and dynamic evolutionary model. Mood markers such as the German Konjunktiv or dürfen and sollen are considered deictic expressions because they locate a state of aŸairs in an epistemic region which can be described by its relative distance to the deictic centre. As root and epistemic modals interact with negation and the past subjunctive in a similar way both showing a tendency towards a speaker-oriented interpretation, she argues for an integrated deictic approach. The last four contributions treat the ways in which the participants of the speech event refer to each other and how they create a textual space, the entities of which can themselves become the focus of attention. Formally, the articles are primarily concerned with pronouns. Johannes Helmbrecht describes politeness distinctions in second person pronouns. First and second person pronouns are prototypical deictic expressions, but only second person pronouns can also function as honoriªcs. In addition to participant roles they construe social roles that are based on social distance between hearer and addressee either in terms of social rank, power and prestige or in terms of the degree of

xi

xii Friedrich Lenz

intimacy and familiarity. In Helmbrecht’s sample of 100 languages only a quarter encode politeness in their pronominal system. His typological study shows that politeness construals can diŸer considerably across languages. The largely geographically conditioned distribution of honoriªc types points to contact induced borrowing. Katharina Kupfer investigates the pronominal system of Old Indic using the Rigveda, a collection of the oldest preserved texts, as a corpus. She focuses on the most frequently used demonstrative pronouns and shows that in this language there is a clear distinction between the accentuated forms which are used deictically and the enclitic pronouns which have a text-cohesive, i.e. anaphoric function. In addition to the deictic pronouns expressing proximity and distality (idám and adás) there is a third pronoun (tád) which is neither proximal nor distal. As tád is not anaphorically used, it is considered to be deictically neutral. It builds up a mental space in which localisation by other means can take place. Manfred Consten’s article looks into the distinction between anaphora and deixis from a theoretical perspective. Traditionally, the two notions are strictly kept apart on the grounds that deictic expressions refer to new entities identiªed from an origo in the speech event whereas anaphoric expressions take up what has already been identiªed by co-referring antecedents in the pretext. Consten claims that indirect anaphora is paralleled by indirect deixis. In both cases the referent is not immediately accessible. In indirect anaphora the pretext does not contain the referent, but only an associated anaphoric anchor. In indirect deixis the referent cannot directly be perceived in the situation, but is still anchored in the physical environment. Consten analyses cross-references between text and pictures in newspapers and suggests that there is a gradual transition from direct to indirect reference on the one hand and from anaphora to deixis on the other. As textual and non-textual information interact in various ways, he argues for a uniªed model. It is a commonplace — and all contributions in this book have taken it for granted as well — that deictics can only be understood when situational information is available. Heiko Hausendorf focuses on the notion of situation itself. He does not treat it as given, but as socially constructed and takes the stance of conversation analysis asking how the interacting participants achieve the situation. He holds that the transition from single perceiving activities to shared perception seems to be the crucial element for a social situation. By means of deictics and pointing gestures that can be heard and seen by the participants

Introduction xiii

perception itself becomes perceivable. Hausendorf thus locates deixis at the interface between language as a semiotic code and sensory perception as a channel of communication.

Notes 1. It is interesting to see how lexicographers deal with this problem. They usually replace one deictic by another, for instance here by in this place (cf. Lenz 2001). This way of paraphrasing deictics does not lead to a context-independent meaning, but is of course legitimate for a lexicographer. Philosophical attempts to entirely eliminate deicticity or indexicality have failed, see Bar-Hillel (1954) or Lyons (1977, 639–646). 2. According to Bühler deictics acquire their semantic value only in the deictic ªeld that constitutes itself on the basis of how the three deictic dimensions - space, time, person - are perceived in the speech event, whereas naming words function as symbols that have a stable meaning in the symbolic ªeld. 3. According to Langacker (1999, 93f.) it is also re¶ected psychologically. Recurrent conceptualisation in usage events leads to the entrenchment of meaning. This should be particularly relevant for deictics. 4. The papers are based on contributions to an international workshop on deictic conceptualisation held at the University of Leipzig in 2001. The workshop was part of the annual conference of the German Linguistics Society (DGfS) which in that year was dedicated to “Language and Cognition.” I would like to thank the organisers for their support.

References Anderson, Stephen/Edward Keenan 1985 “Deixis”. In: Language Typology and Syntactic Description, T. Shopen (ed.), 259–308. Cambridge: CUP. Bar-Hillel, Yehoshua 1954 “Indexical Expressions”. Mind 63, 359–379. Bühler, Karl 1934 Sprachtheorie. Die Darstellungsfunktion der Sprache. Stuttgart: Fischer. 1990 Theory of Language. The Representational Function of Language (transl. by Donald Fraser Goodwin). Amsterdam: Benjamins. Fillmore, Charles 1975 Santa Cruz Lectures on Deixis 1971. Bloomington: Indiana University Linguistics Club. Reprinted as: 1997 Lectures on Deixis. Stanford: CSLI Publications.

xiv Friedrich Lenz

Hanks, William 1992 “The Indexical Ground of Deictic Reference”. In Rethinking Context, Language as an Interative Phenomenon. A. Duranti/C. Goodwin (eds), 43–76. Cambridge: CUP. Lakoff, George 1987 Women, Fire, and Dangerous Things. What Categories Reveal about the Mind. Chicago: University of Chicago Press. Langacker, Ronald 1987 Founndations of Cognitive Grammar I. Theoretical Prerequisites. Stanford: Stanford University Press 1999 Grammar and Conceptualization. Berlin: Mouton de Gruyter. Lenz, Friedrich 1999 “The Temporal Dimension of Discourse Deixis”. Arbeiten aus Anglistik und Amerikanistik 24/1, 3–14. 2001 “Here is hier, and there is dort, but where is da? Contrastive Reflections on German and English Demonstratives“. Arbeiten aus Anglistik und Amerikanistik 26/1, 39–51. Lyons, John 1977 Semantics (Vol. 1 & 2). Cambridge: CUP. Levinson, Stephen 1983 Pragmatics. Cambridge: CUP. Marmaridou, Sophia 2000 Pragmatic Meaning and Cognition. Amsterdam: Benjamins. Radden, Günter 1997 “Metaphorisierte Zeit”. In Sprache im Fokus. Festschrift für Heinz Vater zum 65. Geburtstag, C. Dürscheid et al. (eds), 427–442. Tübingen: Niemeyer.

Part 1: Space

‘Addressee eŸects’ in demonstrative systems The cases of Tiriyó and Brazilian Portuguese

Sérgio Meira Max-Planck-Institut für Psycholinguistik, Nijmegen Rijksuniversiteit Leiden, Leiden

1.

Introduction

An important distinction is often made in the literature on demonstratives between so-called distance- and person-oriented systems (Anderson and Keenan 1985: 282–286; Diessel 1999: 39). These terms are somewhat misleading: for instance, a distance-based system is still person-oriented in that it is speaker-centered, i.e. its ‘proximal’, ‘medial’ and ‘distal’ terms refer to the position of the speaker. There are other problems with this distinction (cf. Dunn and Meira (in prep.) for further discussion). The main goal of the present paper is precisely to address one of these problems: the eŸects that the location of the addressee may have on demonstrative choice even in “distance”-oriented systems. For this purpose, the demonstrative systems of two unrelated languages, Tiriyó (a Cariban language of Northern Amazonia) and spoken Brazilian Portuguese, are presented, with special attention to the role played by the position of the addressee. For Tiriyó, in which there is a typical three-term “distance”-oriented system, the data show that, despite the absence of any addressee-centered terms, the position of the addressee is still relevant for the choice of the “medial” and “distal” terms. In Brazilian Portuguese, the spoken language (unlike the written standard) has apparently a two-term system (with a distal aquele opposed to a proximal esse); the data, however, show that the position of the addressee is of crucial importance for the use of esse and aquele. The Tiriyó and Brazilian Portuguese cases

Sergio Meira

will be presented in the next two sections, which are followed by a concluding discussion of ‘addresse eŸects’ in the ªnal section.

2.

Tiriyó

Tiriyó is a Taranoan (Cariban) language spoken by approximately 2,000 people in several villages scattered along both sides of the Brazil-Surinam border in the Amazon region. For a reference grammar of Tiriyó, see Meira (1999a). This section contains a brief summary of the Tiriyó demonstrative system (a more detailed account is forthcoming as Meira [forth]),1 followed by an examination of the facts that concern the in¶uence of the position of the addressee in the usage of demonstratives. Tiriyó has basically a three-term “distance”-based demonstrative system, with two additional terms — one for referents that can be heard but not seen, and one with only anaphoric uses — which belong to the same formal class as the demonstratives; the whole group can be termed simply “third-person pronouns”. There are diŸerent forms for animate and inanimate referents, and there is also a number distinction (traditionally termed collective, i.e. “all”, vs. non-collective, i.e. “less than all”). All forms are listed in Table 1. “Proximal”, “medial”, and “distal” forms are typically used as expected, i.e. with the speaker as the reference point, and the referent at diŸerent distances from the speaker. This is illustrated in the diagram in Figure 1. In order to assess the “distance” parameter for proximal, medial and distal forms in Tiriyó, an elicitation tool developed at the Max-Planck-Institut für Table 1.

Tiriyó third-person pronouns (after Meira 1999a: 154)*

Anaphoric Demonstrative ‘Visible’ Proximal Medial Distal ‘Invisible’

Inanimate Non-Collective Collective

Animate Non-Collective

Collective

irë

irëto(mo)

nërë

namo

se(nï) serë mërë ooni më(nï)

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

mëe

mëesa(mo)

mëërë ohkï mëkï

mëëja(mo) ohkïja(mo) mëkïja(mo)

* Symbols and IPA equivalents: ï = [G]~[Y], ë = [6]~[%], j = IPA [j], s = [‰]~[w], r = [q]~[ ], n = [n] ([]] if word-ªnal or before k, [m] before p, [\] before j). Parentheses mark reducing syllables (cf. Meira 1999a: 77–94, 1999b).

o

4

‘Addressee effects’ in demonstrative systems

se(nï)/serë, mëe (Proximal) mërë, mëërë (Medial)

SPEAKER

ooni, ohkï (Distal)

Figure 1. A diagram of the three “distance” levels of Tiriyó demonstratives

Psycholinguistik in Nijmegen (David Wilkins’ “Demonstrative Questionnaire”; see Wilkins 1999, or Dunn & Meira forth.) was employed to obtain exophoric uses of demonstratives with a total of 10 native Tiriyó speakers with respect to a number of situations in which speaker, addressee, and referent occupied various diŸerent positions. Table 2 below summarizes the results. A ªrst inspection of Table 2 suggests that Tiriyó demonstratives are not sensitive to the position of the addressee. For instance, the third row shows that medials can be used for objects in contact with the addressee’s body or in his/ her vicinity, but also for objects that are farther away from both speaker and addressee. Also, the last two rows suggest that, for referents far away from the speaker, the proximity of the addressee to the referent or to the speaker is not relevant. One would therefore have an almost prototypical distance-based system, in which the position of the speaker is the only relevant factor. However, closer inspection of the actual answers of speakers shows that the position of the addressee still has some in¶uence. For the situations in which the referent was far away from the speaker (the last two rows of Table 2), the fact that the addressee was close to the referent (e.g. the addressee was sitting next to it, or holding it in his hands) was enough for nearly all spontaneous utterances (from 9 out of 10 speakers) to contain medial terms. Distal forms were accepted by 6 of the speakers when suggested by the interviewer, but always with apparent hesitation, and there is the possibility that they involved a diŸerent interpretation of the situation.2 In other words: when the addressee is far away from the speaker, and the referent is close to the addressee, a certain preference for medials can be felt (though distals cannot be excluded). This tendency is not present when the addressee is close to the speaker (i.e. in the situations in which the speaker and addressee are close to each other and the referent is far away from both of them, there seemed to be no signiªcant

5

6

Sergio Meira

Table 2. Exophoric use of Tiriyó demonstratives (a summary of Meira [forth]) Demonstratives

Proximals only

Proximals much better than Medials Proximals and Medials more or less equivalent

Situations – referent is a part of speaker’s body (tooth, ªnger, etc.); – referent is a small objects in immediate contact with speaker’s body (insect/stone on shoulder, etc.); – referent is an object in the immediate vicinity of speaker and addresee, and closer to speaker than to addressee; – referent is touched by speaker – referent is a part of addressee’s body, or a small object on addressee’s body; speaker points at it * – referent is a part of addressee’s body, or a small object on addressee’s body, and speaker does not point at it; – referent is halfway between speaker and addressee (who are not far from each other: referent is in their immediate vicinity); – referent is in front of speaker and addressee (distance up to a few meters)

– referent is a small object on the addressee’s body, and the speaker does not point at it at all; Medials better than – referent is an object in the immediate vicinity of speaker and Proximals and than addressee, and closer to addressee than to speaker; Distals – referent is more than a few meters away from speaker and addressee – referent is far from speaker and addressee (a few hundred meters); Medials and Distals – referent and addressee are both far from speaker (a few more or less hundred meters; e.g. addressee is far from speaker and sits equivalent next to referent) – referent is far from speaker and addressee (a few hundred meters), and is blocked from view; Distals much better – referent and addressee are both far from speaker (a few than Medials hundred meters), and addressee blocks referent from view – referent is far away from speaker and addressee (e.g. mountain many kilometers away) * If the speaker touches the referent, the proximal forms are obligatory.

preference for medials over distals; the actual score in the spontaneous reactions of the 10 speakers was 6 medials vs. 4 distals). Thus, there is a tendency to favor medial terms over distals when the addressee is close to the referent. This tendency does not mean that the medial terms should be analyzed as addressee-centered: medials are perfectly adequate when referring to objects that are far from both speaker and addressee (cf. ªfth row of Table 2). Rather, the presence of the addressee seems to bring the referent “closer” to the speaker. One might say that, since speaker and ad-

‘Addressee effects’ in demonstrative systems

dressee are both participants of the speech act, they feel “close” to each other, i.e. they share a certain “conceptual proximity”, even when they are physically far from each other. This “conceptual proximity” then apparently leads the speaker to favor medial terms for referents situated near the addressee. It is as if the addressee caused the “local spacetime” to “bend”, so that the referent seems to the speaker to be closer than it would otherwise seem to be. This possibility is represented graphically in Figure 2.

se(nï)/serë, mëe (Proximal) mërë, mëërë (Medial)

SPEAKER

ADDR

ooni, ohkï (Distal)

Figure 2. The eŸect of the position of the addressee on the basic system of Figure 1

3.

Brazilian Portuguese

Portuguese is a Romance language spoken by approximately 190 million people in the world. Although a relatively well-known language, with a large bibliography of studies, the ªeld of space and spatial language remains still relatively unresearched (but see Batoréo 2000). Demographically, the Brazilian dialect of Portuguese is the most important, with more than 160 million speakers in Brazil. The use of demonstratives in Brazilian Portuguese is an area of signiªcant divergence between actual spoken usage and the normative usage codiªed for the written language. Tables 4 and 5 contrast the normative system of written Brazilian Portuguese with the system found in the spoken language. Parenthetical s’s indicate the existence of the corresponding plural forms. A detailed discussion of this analysis is given in Meira and Guirardello (forth.), from which the tables below were taken. One of the conspicuous diŸerences between Tables 3 and 4 is the absence of the speaker-centered term este, used in written Brazilian Portuguese but generally not in the spoken language.3 The distinction ‘near speaker’ vs. ‘near addressee’ is still maintained in the set of locative adverbs (aqui ‘here, near speaker’, aí ‘there, near addressee’, ali and lá ‘there, far from speaker and

7

8

Sergio Meira

Table 3. The demonstrative system of written (normative) Brazilian Portuguese Masculine this (near speaker) this (near addressee) that (far from both)

Variable Feminine

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

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

Non-Variable isto isso aquilo

Table 4. The demonstrative system of spoken Brazilian Portuguese* Local Unmarked S-centered Variable Masc. esse(s) Fem. essa(s) Non-Var. isso

A-centered

esse(s) aqui esse(s) aí essa(s) aqui essa(s) aí isso aqui isso aí

Non-Local Unmarked Proximal Distal aquele(s) aquela(s) aquilo

aquele(s) ali aquele(s) lá aquela(s) ali aquela(s) lá aquilo ali aquilo lá

* A modiªed noun is placed between the two parts of the marked (combined) forms: esse livro aí ‘this (near addressee) book’, aquela faca lá ‘that (far from speaker and addressee) knife’.

addressee’), which can combine with esse and aquele to form the marked forms of Table 4. Concerning the loss of este, it has already been suggested that spoken Brazilian Portuguese has moved from a three-term, speaker-and-addresseebased system to a simpler two-term, distance-based system (esse ‘proximal’ vs. aquele ‘distal’; cf. e.g. Câmara Jr. 1971), following a marked tendency towards the loss of addressee terms in the Romance family (e.g. Italian codesto, Catalan aqueix; cf., for the latter, Badia i Margarit 1981: 311). However, data obtained from 10 speakers of Brazilian Portuguese with the same elicitation tool (Wilkins’ “Demonstrative Questionnaire”) mentioned in the previous section contradicts this hypothesis. The clearest cases were those in which the speaker and the addressee were very far from each other (a few hundred meters) and the referent was placed in the immediate vicinity of the addressee (who held it in his/her hands, or sat next to it): the use of the putatively proximal form esse (usually in combination with the speaker-centered adverbial aí) was obligatory. All speakers refused the distal form aquele in this case. Note that, when the addressee was moved to a position close to the speaker while the referent remained a few hundred meters away from both, all speakers spontaneously used aquele (usually in combination with the distal adverbs ali and lá). In other words: an object placed at a large distance from the speaker is referred to with esse if the addressee is close to it, and with aquele if the addressee is not close to it. In fact, all speakers

‘Addressee effects’ in demonstrative systems

agreed that aquele-forms are impossible as long as the referent is close to the addressee, regardless of the distance between speaker and addressee. Since esse is also used for referents close to the speaker, one may speak of a speaker-andaddressee area, which is opposed to a distal-from-speaker-and-addressee area covered by aquele, as in Figure 3 (cf. Diessel 1999: 40–41 for other cases of speaker-and-addressee terms). Thus, the merger of este and esse in spoken Brazilian Portuguese has not led to the total obliteration of the speaker-addressee opposition: even if one disregards the locative adverbs (in which the opposition is still alive in the terms aqui and aí), the position of the addressee continues to be important for the correct usage of esse vs. aquele. lá A Q U EL E

ali SPEAKER

aqui

ADDRESSEE

aí

E S S E

Figure 3. The conceptual areas covered by spoken Brazilian Portuguese demonstratives.

4.

Conclusion

In the two cases discussed in the previous sections, the position of the addressee plays an unexpected role. In Tiriyó, what looks like an almost prototypical distance-based demonstrative system shows signs of recognizing the importance of the position of the addressee. In spoken Brazilian Portuguese, the merging of the speaker- and addressee-centered terms of the written language has not resulted in the loss of the position of the addressee as a relevant feature for demonstrative usage; rather, an opposition between a speaker-and-addressee area and the rest of the space has arisen. In both cases, the relevance of the position of the addressee is obvious. Based on such examples, it may be suggested that the position of the addressee is a good candidate for a cognitively obvious reference point: even for

9

10

Sergio Meira

systems in which it has not given rise to any special dedicated terms, or in which it has not become a general semantic feature of the system, it still remains an obvious place of interest. Since the addressee is obviously relevant to the speaker, his/her position becomes more salient than most other positions in the surrounding space and is thus a more likely target for communication. In other words, it seems plausible that the speaker should frequently worry about whether or not objects are accessible (close) to the addressee, even if the language they speak has no grammatical means of indicating this fact directly. This “salience of the addressee” is obvious in the very existence, and relative frequency, of addressee-centered terms in the languages of the world; the fact that it may have eŸects even in systems that lack such terms should not come as a surprise. It is probably the case — though empirical tests must still, of course, be carried out — that most, perhaps all “distance”-based systems would show similar eŸects. The similarity between the Brazilian Portuguese speaker-and-addressee area (Figure 3) and the Tiriyó “deformed proximal” area (Figure 2) suggest that person- and distance-oriented systems may diŸer less essentially than the terminology suggests. It may be that, for distance-based systems, the position of the addressee is pragmatically important (it is an “unmarked salient place” for reference), and that this importance or salience has simply grammaticalized into a semantic feature in person-based systems. This suggests a possible path of evolution from distance- to person-based systems: a medial term might perhaps evolve into an addressee-centered term by incorporating or “semanticizing” the “relevance/salience of the position of the addressee” into a “near addressee” feature. It would be interesting to check, for the language families whose history is better known, if there is any evidence of this kind of diachronic change.

Notes 1. Working versions of forthcoming papers by Meira (and Meira and Guirardello) are available on the internet, from the PrePrint Server of the Max-Planck-Institut für Psycholinguistik at http://www.mpi.nl/cgi-bin/preprint/preprint_public.pl (select the letter M for author and click on “Do Search”). 2. Two of the speakers who refused the distal forms claimed that, in these circumstances (i.e. addressee far away from speaker), the use of distals to refer to an object close to the addressee would imply that it is “behind” the addressee, that the addressee has to “turn around” to see it. (In the last row of Table 2, one sees that distals were considered much better than medials when the referent is blocked from view by the addressee, i.e., for instance, when it is behind the addressee.)

‘Addressee effects’ in demonstrative systems

3. A small number of occurrences of the speaker-centered term este in the spoken language data are attested, usually as a result of the in¶uence of the written norm; cf. Meira and Guirardello (forth.) for details.

Bibliography Anderson, S. R., and E. L. Keenan 1985 “Deixis”. In Language typology and syntactic description, vol. 3, T. Shopen (ed.), 259–308. Cambridge: Cambridge University Press. Badia i Margarit, Antoni. 1981 Gramàtica històrica catalana [Biblioteca d’Estudis i Investigacions, núm. 4]. Benetússer (València): Graffitti XXI. Batoréo, Hanna Jakubowicz 2000 Expressão do espaço no português europeu. Lisboa: Fundação Calouste Gulbekian and Fundação para a Ciência e a Tecnologia, Ministério da Ciência e da Tecnologia. Câmara Jr., Joaquim Mattoso 1971 “Uma evolução em marcha: a relação entre êste e êsse”. In Sprache und Geschichte: Festschrift für Harri Meier zum 65. Geburtstag, Eugenio Coseriu and Wolf-Dieter Stempel (eds), 327-331. München: Wilhelm Fink Verlag. Diessel, Holger 1999 Demonstratives: form, function, and grammaticalization [Typological Studies in Language (TSL), vol. 42]. Amsterdam, Philadelphia: John Benjamins. Dunn, Michael, and Sérgio Meira (eds) In prep. Demonstratives in cross-linguistic perspective. Meira, Sérgio 1999a A grammar of Tiriyó. Houston, Texas: Rice University Ph.D. dissertation. 1999b “Syllable reduction and ghost syllables in Tiriyó”. In XXV LACUS Forum, Shin Ja J. Hwang and Arle R. Lommel (eds), 127-133 . Fullerton, California: The Linguistic Association of Canada and the United Status (LACUS). Forth “Non-contrastive exophoric use of demonstratives in Tiriyó”. In Demonstratives in cross-linguistic perspective, Michael Dunn and Sérgio Meira (eds). Meira, Sérgio, and Raquel Guirardello. Forth. “Non-contrastive exophoric use of demonstratives in spoken Brazilian Portuguese”. In Demonstratives in cross-linguistic perspective, Michael Dunn and Sérgio Meira (eds.). Wilkins, David 1999 The 1999 Demonstrative Questionnaire: “THIS” and “THAT” in comparative perspective. In “Manual” for the 1999 field season, David Wilkins (ed), 124. Nijmegen, The Netherlands: Max-Planck-Institut für Psycholinguistik.

11

Deictics in the conversational dyad Findings in Spanish and some cross-linguistic outlines Konstanze Jungbluth Tübingen

1.

Introduction

In this paper I will propose a dyad-oriented system of Spanish deictic pronouns and show how this supercedes traditional person-oriented or distance-oriented accounts. This approach focuses on the physical orientation of the speaker and hearer when they are engaged in conversation. It will be shown that the use of demonstratives in Spanish is determined by the position and orientation of speaker and hearer relative to each other as well as the position of the object referred to, relative to the shared space of speaker and hearer. This conversational dyad is established by the speaker and the hearer while talking with one another. My research is based on ªeldwork data collected in Spain: observing language use in the routine speech practice of native speakers was crucial for the development of the proposed dyad-oriented system. The system is demonstrated in the contrasting use of demonstratives in three diŸerent types of constellation: face-to-face conversation, face-to-back conversation and side-by-side conversation. With these three diŸerent arrangements I hope to have covered the most basic conceptualizations of relative position. Language is directed towards the other (the alter). Speaker and hearer/s share a jointly established aim (ªnal) while interacting (for alterization and ªnalization, see Schlieben-Lange 1983). Since deictics are based on context and context is multidimensional1, deictics may refer simultaneously to aspects of diŸerent contextual ªelds (e.g. discourse context, physical context).

14

Konstanze Jungbluth

The paper is organized in the following way. The initial section introduces the Spanish three-term paradigm of demonstratives, and the traditional person- and distance-oriented interpretations are discussed. In the next section, the core of the paper, the dyad-oriented system is presented in detail. The concluding section outlines the potential the dyadic approach opens for crosslinguistic research.

2. Spanish demonstratives: discussing person- and distance-oriented interpretations Spanish occupies a large multinational space on both sides of the Atlantic and it is clear that the way people speak Spanish in this vast territory is by no means homogeneous. In spite of the national eŸorts in Latin America, the Real Academia Española and the Cervantes Institute, founded in 1991, still consider themselves to be the custodians of the shared Castilian language and the protagonist for its prestige, competing with the other international languages. In most places where Spanish is spoken, it is not the only language in use: there are many bilingual and multilingual speakers in the diŸerent countries. This is not only true for the New World, but for the Iberian Peninsula too, where, apart from Castilian Spanish, Galician, Catalan and Basque are also spoken. In order to investigate the use of the sensitive three-term demonstrative system of Castilian, I therefore chose a small town in the countryside to the south of Toledo, the heartland of Castilia, to collect my data. The villa called Consuegra was already known to the Romans, as Consaburensis. Most inhabitants have several generations of contact with the area and although the majority are directly or indirectly involved in agriculture, there is some industry too. Far more people move away, typically to the provincial capital Toledo and the national capital Madrid, than move into the area. Thus language use in this area is relatively free from external in¶uences. People were observed in their everyday activities at home and work (Jungbluth 2001, forthcoming). The data were collected in the ªeld by participant observation and sometimes by elicitation. These data form the empirical basis of my research, for which qualitative methods of interpretation were used.

Deictics in the conversational dyad

2.1 The Spanish demonstrative pronouns Demonstratives in Spanish are used as determiners and as pronouns and agree in number and gender with the noun(s) they refer to. Although Spanish has only two genders, the demonstratives also have a neuter form, as do the articles. This form is restricted to pronominal use, and is often in phraseinitial-position. The following table shows the forms and their equivalents in English. On the left is the three-term Spanish system and on the right the twoterm English system. The Spanish “middle” term will sometimes be translated as this and sometimes as that: Table 1. The paradigm of demonstratives in Spanish and English Term 1. 2. 3.

Singular Plural Singular Plural Singular Plural

Spanish Masculine

Feminine

Neuter

English

Term

este estos ese esos aquel aquellos

esta estas esa esas aquella aquellas

esto

this

1.

eso

this that

aquello

that

2.

I will refer to all of the forms using the masculine singular forms este, ese, aquel. There seems to be no diŸerence in deictic content between their uses as determiners and prominals2 or any relevant diŸerence among the genders. 2.2 Is the Spanish system person-oriented? Traditionally the Spanish system is described in grammars as being parallel to the grammatical persons. First person I relates to este, second person you to ese and third person he, she, it to aquel (see Table 2). Modern accounts of the system reinterpret the grammatical persons pragmatically, emphasizing the roles they play in discourse. From this perspective the speaker is combined with este, the hearer with ese, and the object of discourse, for example people not present, with aquel. This system has been called person-oriented (Anderson/Keenan 1985, Diessel 1999, for the pragmatic reinterpretation see Hausendorf in this volume).

15

16

Konstanze Jungbluth

Table 2. The traditional person-oriented interpretation person - oriented personal demostrativo

yo este, esta, esto

tú ese, esa, eso

él aquel, aquella, aquello (Alonso 1968)

Researchers have been uncomfortable with this interpretation for several reasons; the most important being that the expected use in face-to-face conversation is not regularly found. The meanings usually attributed to the demonstratives (…) suggest that reference to a speaker’s or hearer’s region of proximity is, if not the only, at least the most typical and elementary use of demonstratives in the situation of ‘demonstratio ad oculos’. (Hottenroth 1982:138)

The pronoun este refers to the objects that are located between the speaker and the hearer, e.g. inside the conversational dyad, without any distinction between the region of the speaker and the one of the hearer. The following example shows that the expected use of ese for the hearer’s region is not found (see also 3.1.1. Example No. 2). Example 1. Use of este for objects which are close to the hearer Mujer (w, 2030)

¿Te gusta este radio? Do you like this radio?

The woman refers to the radio with este although it is in front of the hearer. BRATOLI/TüDeiKo No/09

We therefore reject the person-oriented interpretation of the system of Spanish demonstratives. 2.3 Is the Spanish system distance-oriented? Investigating possible uses of demonstrative pronouns and related locational adverbs via intuitional data, Hottenroth (1982) proposes an interpretation of the Spanish system as distance-oriented (repeated in: Diessel 1999: 39). Imagine the speaker, for instance, as a patient pointing to diŸerent places on his own body in order to show the doctor where he feels the pain. In this case he can use este (…) as well as ese (…) and even aquel (…), although taken objectively, all places referred to are on his own body and therefore undoubtedly within his region of proximity. (Hottenroth 1982: 141)

Deictics in the conversational dyad

In this context, the distance-oriented perspective leaves out the hearer‘s perspective and three diŸerent regions are modelled around the speaker-ego. The regions of este and aquí, of ese and ahí, and of aquel and allí are taken to form concentric circles around the ego, i.e. the speaking subject. (Hottenroth 1982: 142).

The diŸerentiation is made between este for proximal, ese for medial and aquel for distal areas and thus a sequence is established.

X

X Speaker este ese aquel

Concentric circles around the speaker (Hottenroth 1982)

Figure 1. Distance-oriented interpretation

This view is problematic because two objectively equal distances may not be conceptualized in the same way. A speaker refers to the regions in front and behind conversation participants diŸerently (see 3.1.2.1. Example 3 below). Objects behind are not visible and subjectively seem to be further away than objects in front of the speaker. Similar to the person-related account, this model embodies the idea that the three demonstrative pronouns should co-occur, which is rarely found empirically. As was shown by Josse de Kock (1988), the Spanish demonstrative pronouns occur isolated in written texts: Salvo algunos ejemplos en los que aquel se opone a este, este, ese y aquel no aparecen de costumbre agrupados, relacionados o confrontados en ninguna unidad sintáctica o semántica, sino surgen dispersos, es decir, independientes unos de otros. (de Kock 1988: 423) Except for some examples in which aquel is contrasted with este, the demonstratives este, ese and aquel do not usually appear forming a group, a relation or an opposition in the same syntactical or semantical union; on the contrary they occur dispersedly, e.g. independently of each another. (translation K.J.)

Fortunately our contextual data show some co-occurrences of the demonstratives, but their use does not follow either interpretation. These two arguments

17

18

Konstanze Jungbluth

strengthen our view that the distance-oriented system is not a fully adequate account of the use of Spanish demonstratives. Taking the aspect of alterization seriously, e.g. integrating the perspective of the hearer, I propose the dyadoriented system.

3.

Dyad-oriented interpretation

Weinrich (1988) condemns most of the linguistic theory based on Karl Bühler as ego-fallacious. He calls for a transformation of the former monadic deixis into a dyadic one, thus allowing the analysis to be a proper part of linguistics and communication theory in general. The origo of the system is not the speaker-ego alone, but it is formed by two persons: the speaker and the hearer (Weinrich 1988: 82–83). Der Text ist das “sprachliche Primärzeichen”. (...) Als solches ist es folglich auch dem Menschen als Sozialwesen am nächsten: die hier gegebenen anthropologischen Implikationen haben zu einer dialogisch (dyadisch, partnerbezogen), kommunikativ, situativ (kontextuell) ausgerichteten, zu einer “leiblichkeitsorientierten” Linguistik vom Text-in-Situation (und Funktion) geführt, was methodologisch die Sprache wieder zum Menschen zurückgebracht hat. (Kalverkämper 2000: 11) The text is the linguistic primary datum. (...) As such it relates most closely to people as social beings. These anthropological implications have led to a dialogue-related (dyadic, hearer-oriented), communicative, situational (contextual) “bodyoriented” linguistics of text-in-situation (and text-in-function), which has methodologically returned language to people. (translation K.J.)

Two people talking with one another are the minimal pair for conversation.3 In natural contexts there are many ways in which speaker and hearer may be situated relative to each other. The vast number of situational parameters which determine the choice of one demonstrative form from the three-term Spanish paradigm can be reduced to three diŸerent constellations: – face-to-face conversation – face-to-back conversation – side-by-side conversation While the ªrst is well known in linguistics, the other two are usually not taken into account. Maybe this fact is due to the object of research itself. If the research is based on the observation of routine speech practice in a language community (see Hanks 1996, 2001), there are a number of situations where

Deictics in the conversational dyad

people establish a dialogue without looking at one another. For example, when somebody instructs another person, the so-called “teacher” is often standing behind his/her “student”. Everyday activities such as choosing products in supermarkets may create situations where people have a conversation without looking at one another. At work people are often situated one behind the other. In all these situations speaker and hearer are turned away from each other. In contrast to the classical face-to-face conversation these constellations are called face-to-back conversations. Another turned-away position which is quite common may be observed when the participants of conversation are not situated one behind the other, but next to each other. Speaker and hearer are standing or sitting side by side, perhaps sharing a panoramic view. But they are not always looking into the distance, they are often ordering objects or orienting themselves in the space they share. Talking to one another in this position is called side-by-side conversation. It is the only constellation where speaker and hearer may be assumed to share roughly the same view, because their perspectives are adjoining. In the other two constellations the perspectives of the two participants diŸer. 3.1 Face-to-face conversation 3.1.1 Objects inside the conversational dyad People in face-to-face conversation treat their shared conversational space as uniform. Everything inside the conversational dyad is treated as proximal without any further diŸerentiation. Therefore este is used even when objects are nearer to the hearer or belong to him. The following table shows an example taken from our corpus BraToLi (open collection of European and American Spanish, European and American Portuguese examples of action embracing speech of BRAsileiros, TOledanos and LImeños (in prep.) http:// www.sfb441.uni-tuebingen.de). Figure 2 shows the abstraction following the conventions of the MPI-Nijmegen (Wilkins 1999 “Space-Project”). Note that the data were not produced in order to fulªll the necessities of a previously elaborated questionnaire. On the contrary, data were observed in natural context and the abstraction was developed on this base. Figure 3 shows a model of the conversational dyad. The elliptic space represents the conversational space, which is limited by the speaker and the hearer. The position of the organs of communication on the human body are the reason why this space is naturally closed with regard to the outside world. People talking to one another form something like a circle. This “ideal” form is easier to attain when more

19

20

Konstanze Jungbluth

than two persons are involved as is often the case in natural contexts. All hearers forming the circle are treated as active, regardless of whether they themselves have taken or will take the role of the speaker at some moment during the ongoing conversation. Two persons, the speaker and the hearer, form the minimal pair. Their perspectives are represented by open angles (see Hanks 1992, 1996). The space is not the same in front of and behind people. Most models fail in this respect (see for ex.: Langacker 1990: 42, Rubba 1996: 231), situating the speaker and the hearer “somewhere” in the observed space. In order to understand the use of deictic elements one has to slip into the role of the speaker and the hearer relinguishing the perspective of an outside observer. He would not be able to understand the changing choices the participants make when they use the diŸerent forms of the paradigm. Example 2. Use of este for objects inside the conversational dyad Vendedora: ¿Que más quiere? (f, 3040) What else do you need? Cliente: (m, 4050)

The market woman is surrounded by lots of crates with diŸerent kinds of vegetable and fruit. Las (2 sec.), estas (4 sec.). The customer refers to the let¿Como se llaman? Las lechugas. tuces with estas although they are The (2 sec.), these (4 sec.). nearer to the woman and they what’s their name? The lettuces. belong to her. BRATOLI CS021

Figure 2. Inside the conversational dyad: in face-to-face conversation objects even close to the hearer are usually referred to by este (see Asenjo Orive 1990)

Figure 3. Inside the conversational dyad este is dominant

Deictics in the conversational dyad

The person-oriented interpretation cannot explain the use shown above (Example 2). Following the person-oriented interpretation, one would expect the speaker to use the second term ese, because the object is close to the hearer. Using ese in contrast to este would presuppose the existence of two opposite regions inside the conversational dyad. But there is a certain inhibition about establishing a border line in the shared space. The divergent perspectives of the speaker and the hearer must be abstracted over in order for them to interact socially. While talking together, people harmonize the two perspectives. Not alterization, but ªnalization of language use (see Section 1) is the reason why even three-term systems prefer to use one term for the whole space between the speaker and the hearer, thus establishing the conversational dyad as a homogeneous space. The aim of acting together overrules the individual diŸerences and treats the conversational space, here called the conversational dyad, as a shared space of proximity which is shaped in opposition to the outside world. 3.1.2 Objects outside the conversational dyad In contrast to the space shared by the speaker and the hearer, which is addressed as este, the outside space is referred to by aquel. This outside space is negatively deªned in contrast to the inside-space and is conceptualized as endless (see Langacker 1990: 52). The outside space may include present and absent spaces. While the former may be visible to one of the participants, the latter is identical to an invisible space. The following data show, ªrst, a situation where the speaker is talking about an object behind the hearer, and second, a situation where the speaker is talking about an object behind himself. 3.1.2.1 Behind the hearer Collaborating in selling and buying situations, the speaker is directing the attention of the hearer to the objects he/she desires. The interactive work includes locations which are behind the hearer. The outside term aquel is used to refer to objects which are behind the hearer regardless of how near or far away they may be. Example 3. Use of aquel for objects behind the hearer Cliente: (m, 3040)

¡Dame de las manzanas, (5 sec.) aquellas! Give me some of the apples (5 sec.), those over there!

The customer is looking at the apples, which are situated behind the market woman BRATOLI CS0210

21

22

Konstanze Jungbluth

Figure 4. Outside the conversational dyad: in face-to-face conversation when talking about objects behind the hearer the speaker uses aquel

Figure 5. Inside and outside of the conversational dyad este contrasts with aquel

While the objects are visible to the speaker, they are not visible to the hearer. The outside world begins behind the participants in conversation. 3.1.2.2 Behind the speaker This last situation is quite unusual and not easy to ªnd. People prefer to turn around and face the object they are talking about. Professionals like tourist guides take “the other fellow’s point of view” (Fillmore 1971: 41) shifting the deictic center behind them (Rauh 1983: 46, see Rubba 1996: 230). They are instructed to speak as if they were standing alongside their audience.

Deictics in the conversational dyad

Example 4. Tourist guides adopt their audience’s perspective Guia:

¡Usted puede ver aquí en frente la famosa catedral de Burgos! Esta fechada tiene las

The tourists are looking towards the cathedral.

siguientes características…

guide is standing in

You can see here in front of you the famous

of them looking towards his audience. He speaks

The front

as cathedral of Burgos. This façade has the to tics…

if he was standing next following characteristhem.4

I had the chance to observe a manufacturer talking to his audience without shifting his deictic center to the opposite side. Example 5. Use of aquel to refer to the region behind the speaker Empresario: (m, 4050)

Lo hacemos en aquel recipiente. We make it in that receptacle.

A gesture with the head accompanies the words of the manufacturer. The receptacle is situated in his back, but the people visiting

Figure 6. Outside the conversational dyad: in face-to-face conversation when talking about objects behind the speaker aquel is used

23

24

Konstanze Jungbluth

Figure 7. Inside and outside of the conversational dyad este is opposed to aquel

the place are looking at it. BRATOLI CS0410

Situations where speakers are talking about objects behind them without turning around are limited to immobile objects or objects in well-known rooms, where the owners of the respective space are familiar with the pictures or the furniture surrounding them. The hearer must be able to focus on the objects. As the speaker in these situations often blocks the hearer‘s line of vision, the objects have to be of a certain size for the addressee to catch sight of part of them. This con¶ict is not limited to the face-to-face conversation while the focus is on objects behind the speaker, but happens also in the following type of conversation where the speaker is standing behind his/her hearer while ordering him/her to do something. 3.2 Face-to-back conversation The situations subsumed in the constellation called face-to-back conversation are especially important in research on three-term systems. The position of the hearer turned away from the speaker creates a separate hearer-only space. Limited by the hearer’s own back this hearer-side space is distinct from the speaker’s. The following situation is observed in a supermarket. The adult speaker ªrst refers to his own space in front of him and then to the space in front of his son, which is the hearer-side space referred to as ese.

Deictics in the conversational dyad

Example 6. Use of ese to refer to the region of the hearer Hijo: (m, 0810)

¿Este? This?

The son is about to take a juice pack out of the shelves in a supermarket.

Padre: (m, 3040)

¡No (2 sec), este (8 sec), ese! No (pause), this (pause), that !

The father is holding the advertising lea¶et in his hands. He ªrst points to the picture (este), then he looks up and instructs his son, who is squatting in front of him at some distance. Both are looking in the same direction towards the juices oŸered in the shelves. He refers to the space in front of the hearer as ese. BRATOLI CS0810

Figure 8. Face-to-back conversation: the space in front of the hearer is referred to as ese

Figure 9. Ese refers to the hearer-side space

25

26

Konstanze Jungbluth

Of course recognizing the hearer-side space as diŸerent from the speaker-side one is not restricted to face-to-back conversation. There are certain topics and situations which force the speaker in face-to-face conversation to make the distinction, too. But ªrst the borderlines between the two areas are a lot more di¹cult to describe in face-to-face conversation, because they are ¶exible and subject to change. And secondly they only come into play when the topic or the attitudes of the participants make them necessary. Usually in face-to-face conversation the space between speaker and hearer is treated as homogeneous (see 3.1.1.). 3.3 Side-by-side conversation The third type of constellation is the side-by-side conversation, where speaker and hearer are standing or sitting close to one another looking in the same direction. Their perspectives are nearly the same. The large or small scale space in front of them can be divided into several regions. The three-term system permits speakers of Spanish to establish three regions in sequence: a proximal one belonging to este, a medial one belonging to ese and a distal one belonging to aquel. Example 7. Este, ese and aquel co-occur when referring to three regions in sequence Carpintero: Esto es una refresadora. Eso es (m, 2030) una cierra. Aquello es una escuadradora. This is a milling. That is a saw. That over there is a cutter.

The carpenter makes his message clear with eye-gesture and the assumption of knowledge of speciªc lexical items. The machines are in the same hall at diŸerent distances. BRATOLI CS0810

Deictics in the conversational dyad

Figure 10. In side-by-side conversation three regions are subsequently diŸerentiated: the proximal one by este, the medial one by ese and the distal one by aquel.

Figure 11. Proximal, medial and distal regions may be distinguished in side-by-side constellations: este, ese and aquel co-occur in this context only.

The importance of the distinction between the diŸerent relative positions which hearer and speaker can occupy while talking to each other must be emphasized. Space is not always conceptualized in the same way. People looking at one another contrast the inside and the outside space. The inside space ends with the bodies of the speaker and the hearer(s) and is naturally closed relative to the outside world. They may establish further subregions, but they also may not. When speaker and hearer are in line, standing or sitting one behind the other looking in the same direction, there are two diŸerent regions: one in front of the speaker, another in front of the hearer. The back of the person in front forms the boundary of the inner space. People standing or sitting side-by-side conceptualize the space in a series of regions, and the speaker and the hearer share roughly the same perspective, both forming the origo of the side-by-side system. It is clear that the use of the demonstratives in Spanish, which shows a paradigm of the three terms este: ese: aquel, is determined by the diŸerent kinds of spaces, which are relevant in face-to-face conversation, face-to-back conversation and side-by-side conversation. Speak-

27

28

Konstanze Jungbluth

ers in the heartland of Castilia use this system without any hesitation when coordinating joint activities. My data show that the diŸerent kinds of constellations are all frequent in everyday conversation and it is hardly possible to decide whether one of them has to be taken as basic.5 Conversation can take place with two or more partners in any of the identiªed positions. Speakers conceptualize space with respect to these positions and hearers conªrm the given choice of reference or express disagreement. In any case, it seems impossible to take one of them as primary.6

Figure 12. The dyad-oriented system: face-to-face, face-to-back and side-by-side conversation

4.

Inside and outside: outlines for cross-linguistic research

The Spanish system of demonstratives was described as person-oriented until the 1980s and thereafter as distance-oriented. Without any doubt the Spanish system has the potential to be used in face-to-face conversation as if it were person-oriented, but in side-by-side-conversation it may be used as if it were

Deictics in the conversational dyad

distance-oriented. It may well be that other three-term or multi-term-systems are similarly chameleon-like adjustable. Only the integration of the two perspectives in the dyad-oriented system allows us to account for all data representing the pragmatically basic conceptualizations of space. As the Spanish system re¶ects pragmatic universals, I will compare basic features of the use of demonstratives in some other languages in order to outline the possible value of the dyad-oriented system for cross-linguistic research. First, I will discuss the use of demonstratives in Finnish and Japanese, two other three-term systems. I will then contrast two- and multi-term systems (English, German, European and Brazilian Portuguese). Finally I will focus on the combinations of certain features like inside-outside, visibility, and accessibility, comparing my results with the description of the locative particles in Cora, an Amerindian language (Langacker 1990). Finnish exhibits the inside/ outside contrast (Laury 1996). While tämä is used for inside, tuo is used for the opposite meaning: In their most basic use, the Finnish demonstratives include or exclude the referent in question in the speech participants (speaker’s or addressee’s) current, socially deªned sphere. Thus, in my view, tämä presents a referent or referents which the speaker includes within his or her current sphere, while tuo points to referents which the speaker considers to be outside his or her current sphere (Laury 1996: 307, my emphasis).

The third demonstrative se refers “to something in the addressee’s perceptual sphere” (Laury 1996: 306). This form shows similarities to the Spanish “middle” term ese. Finnish being a non-article language, the use of demonstratives is comparable to other languages only to a certain degree, because some functions seem to be more extended than in Spanish. But the analysis of this nordic three-term system is close to the described system in Spanish in face-toface conversation. Though Diessel (1999) contrasts Spanish with Japanese, Spanish (…) has a distance-oriented system, consisting of the demonstratives este ‘proximal’, ese ‘medial’ and aquel ‘distal’ (…), while Japanese has a personoriented system, in which the middle terms (based on the deictic root so-) refer to a location near the hearer: sore ‘that (near hearer)’, soko ‘there (near hearer)’ etc. (…) (Diessel 1999: 39).

my interpretation shows what the two languages have in common. Diessel follows Kuno (1973: 27) in interpreting the three-term system of demonstratives in Japanese, and paraphrases kono as ‘near speaker’, sono ‘near hearer’ and

29

30

Konstanze Jungbluth

ano ‘away from speaker and hearer’ (Diessel 1999: 59) . Similar to the Spanish data, in Japanese either the ªrst and the second terms, or the ªrst and the third terms co-occur. The three-term systems considered contrast ªrstly the insideoutside space, and then they may further divide the inside-space into a speaker- and a hearer-side region. Table 3. Three-term systems compared (face-to-face conversation only) Spanish

Inside este Speaker-side este

Finnish

Outside aquel Hearer-side ese

Inside tämä Speaker-side tämä

Outside tuo Hearer-side se

Inside Japanese7 kono Speaker-side kono

Outside ano Hearer-side sono

Table 3 only shows the use of the demonstratives in face-to-face conversation. There is an asymmetric structure which includes a sequence of two binary contrasts. The ªrst contrast distinguishes the inside and outside regions, and the second makes the additional division within the inside-region into the speaker’s and the hearer’s region. Spanish in this ªgure seems to be personoriented, but compare the system with Table 4 (below), where side-by-side conversation is shown. I do not have access to descriptions of the use of the demonstratives in Finnish or Japanese with regard to distance. It seems as if only some of the terms are used in the distance-oriented context, e.g. in the side-by-side constellation, and the others are replaced by other locative terms. The four possible regions — inside, outside, speaker’s region, hearer’s region — are also considered by other researchers on deixis who work on languages which are not typologically related. Sternberg (1983) proposes the distinction between logocentric and exocentric use, and further divides the former into egocentric and hearercentric ªelds. The ªrst contrast is equivalent to the contrast between inside and outside, the second one between speakerand hearer-side regions. His research focuses on personal pronouns in English.

Deictics in the conversational dyad

This suggests that the primary opposition is not between the egocentric and the nonegocentric but between what we may call the logocentric (speech-oriented, within the deictic bound uniting speaker and addressee) and the exocentric (oriented to the outside world, where the nonparticipant belongs). (Sternberg 1983: 284). Logocentric before exocentric, within the logocentric, egocentric before addrecentric, [which centers on the addressee] (Sternberg 1983: 311, my emphasis and interpretation in brackets).

Ehrich (1992), working on German locational adverbs (hier : da : dort) extends the three-term contrast by adding the default of the terms (Ø : hier : da : dort) to implicitly refer to the shared space. Carvalho (1976: 247–251), observing Portuguese demonstratives (este : esse : aquele) and locational adverbs (aqui : aí : lá), considers the same regions. For purposes of comparation he discusses the English two-term system, proposing that both English demonstrative terms are used “heterotopically”, e.g. in more than one way. While this may be ‘this table near you and me’, it can also be used to refer to ‘this table near me but far away from you’. While that can be used for ‘that tree in the distance, far away from us both’, it can also refer to a region belonging to the hearer in contrast to that of the speaker (‘that tree, far away from me, against which you are leaning’). The three-term system of European Portuguese has only one heterotopic term este, which shows multiple functions like the English data, because it may refer to the shared space or to the egocentric space. Esse refers to the hearer-side space and aquele to the outside, which is negatively determined ‘where you and I are not’. Note that not all two-term systems are like the English system in diachronic perspective, the English two-term system has developed from an earlier threeterm system by cutting the third term yonder. The Brazilian Portuguese twoterm system has developed from an earlier three-term system by cutting the ªrst term este. The Croatian two-term system spoken in the City of Zagreb (Zi¦c Fuchs 1996) has developed from an earlier three-term system by cutting the second or middle term. To contrast European and Brazilian Portuguese (Meira/Guirardello forthcoming), the latter uses the ªrst term (esse) for three regions: for the inside space, for the speaker-side and for the hearer-side region, thus keeping the other term aquele for just the outside world. As in German and English, misunderstandings are avoided by the addition of locational adverbs: esse aqui for the speaker-side region, esse aí for the hearer-side region. This raises the question whether it should be called a four-term system or whether it has to be considered still as a two-term system.8 By comparing the description of the Spanish system with that of the Amerindian language Cora, it becomes apparent that certain combinations of

31

32

Konstanze Jungbluth

features are language-speciªc and not universal. In Cora, an Uto-Aztecan language “spoken by more than 10 000 people in the state of Nayarit, Mexico” (Langacker 1990: 346), the inside/outside contrast is represented by the locative particles u/a. The prototypical use is illustrated by two sentences which contrast the space inside and outside a house (1990: 36). This characteristic use may develop further to involve physical enclosure. In this kind of context the contrast between u and a is that between inside and surface or between deep or shallow penetration. The side-by-side constellation is also discussed where people are locating persons in the panoramic view at diŸerent distances. Here further morphemes come into play. While proximal is represented by y + u, medial distances are addressed by m + a and distal ones by a only. As in Spanish, the “inside” term is used for proximity, the “outside” term for distance. The threeterm system of Spanish has a special term for the medial distance, while Cora uses the distal term for this region, too. In order to make a third term, the distal term merges with another locational morpheme m. A similar pattern is found in Brazilian Portuguese. As in Cora this language has only two terms. Brazilian Portuguese uses the “outside” form for the distal and for the medial region, combining the latter with the locational adverb aí (see Table 4). Accessibility is another factor to be considered in Cora: u, the inside term, encodes inaccessiblity, while a the opposite: It stems from the pragmatic consideration (…), namely that an enclosure often serves as a barrier to perception, or to access more generally. From an external vantage point, an entity that bears an ‘inside’ relation to another is commonly inaccessible to view, while one that bears an ‘outside’ relation remains accessible. It is therefore natural that the u/a contrast should be extended to mean ‘inaccessible/ ‘accessible’, primarily (though not exclusively) with respect to vision. (Langacker 1990: 45)

This feature is extended to the cases where objects are located in front or behind speaker or hearer, becoming u ‘inaccessible/behind’ and a ‘accessible/ in front’ (Langacker 1990: 48). While in its basic use the inside/outside contrast is similar to that described for Spanish, Cora requires us to introduce combinations absent from Spanish. Before looking at the Cora system it did not seem necessary to mention that in Spanish objects inside the conversational dyad are of course visible, in the line of vision of the speaker and the hearer, accessible and in front of the two or more participants of conversation. This combination is true for Finnish too:

Deictics in the conversational dyad

(…) the demonstratives express social and cognitive accessibility, and not concrete accessibility such as the ability to reach and manipulate something, although such ability may of course coincide, and often does, with social and cognitive accessibility.” (Laury 1996:307, my emphasis)

It appeared to be natural (unlike in Langacker above) that objects placed in the outside are inaccessible, out of sight, invisible to at least one of the two participants, behind speaker or hearer, e.g. outside the conversational dyad. The basic contrast is comparable in the two languages while the further uses are diŸerent. For both systems and maybe for all natural languages it is true that “the ‘outside’ region is much more extensive than the ‘inside’ region”, but it is a particular feature of Cora that it “includes even those areas where the possibility of an ‘inside/outside’ contrast does not arise.” (Langacker 1990: 52). The ¶exibility of the borderlines — sometimes drawn on a small scale, in other occasions extended to the horizon or the universe — is another universal. The scope of prediction is far greater in talking about the position of planets in the solar system than in discussing the arrangement of furniture in a room, even though one might use the same locative predictions in each. (…) Obviously there is nothing absolute or fully predictable about these matters, but it is quite possible for natural tendencies like these to become frozen in linguistic convention in various ways. (Langacker 1990: 52)

In Cora the outside term functions as the unmarked member in the two-term system, unlike in Spanish. Reference is exclusive: the inside-term may not be used for objects outside nor the outside-term for objects inside. Since Cora is a two-term system the relative position of face-to-back is not very likely to throw further light on the system. Certain information9 on Cora would be interesting: Is the inside-term u judged as inside from the perspective of the speaker and the hearer who are in the house or from an observer who is outside? This may change the whole interpretation. Our dyadic approach requires information about the two participants of conversation. It is possible that the concept of “inside” in Spanish and Cora are diŸerent, but perhaps the diŸerence only appears in theory because the descriptions of the two languages start from diŸerent points. Crosslinguistic research is based upon the descriptions accumulated by the community of investigators, but the lacunas found require further ªeldwork. Careful descriptions which take into consideration all possible constellations and both participants are desirable. Only on the basis of this information can we be sure of comparing identical terms. The following table, which shows face-to-face conversation and side-by-side conversation but lacks face-to-back conversation, summarizes the discussion:

33

34

Konstanze Jungbluth

Table 4. Inside and Outside: Spanish, Brazilian Portuguese and Cora compared face-to-face conversation Inside Spanish: este Brazilian Portuguese: esse

Outside Cora: u

Cora: a

Spanish: aquel Braz.Portuguese: aquele

enclosed/hidden

visible accessible in front of in sight

invisible inaccessible behind out of sight

side-by-side conversation

Spanish Cora Brazilian Portuguese

proximal

medial

distal

este (y +) u esse

ese m+a esse + aí

aquel a aquele (lá)

The last table shows the potential distance-oriented interpretation of the Spanish system in contrast to Table 3, where the face-to-face conversation is shown and which distinguishes between a speaker- and a hearer-side region. The basic contrast between inside and outside seems to be universal. Two-term systems encode this contrast, three-term systems use their ªrst and third term for this purpose. Two-term systems overcome their seeming deªciency with regard to three-term systems in distance-relevant contexts by adding further locative elements, for example locational adverbs.10 In order to allow possible crosslinguistic research, it is important to describe not only the speaker and the object but also the position and perspective of the hearer.

5.

Conclusion

It is no coincidence that in more recent descriptions (Meira/Guirardello forthcoming, Laury 1996, Vann 1992) the hearer is found to be relevant. The claimed diŸerence is not inherent in the languages themselves, but possibly due to an ongoing change in the scientiªc approach to them (Kuhn 1962/ 1976). There is a general tendency to make systems comprehensible from both

Deictics in the conversational dyad

sides: from that of the speaker, who encodes the utterance, and from the side of the hearer, who has the task of decoding the same utterance. In this context I feel encouraged to propose a new system for deictic research which I call dyadoriented. For research on deictics it is important to take into account the diŸerent relative positions of speaker and hearer(s). Not only face-to-face conversations, but also face-to-back conversations and side-by-side conversations have to be observed. Person-oriented and distance-oriented interpretations of the Spanish system are not satisfying because each of them generalizes features which are only valid for a part of the multifaceted conversational reality (see 3). It is surprising that even three-term systems do not regularly distinguish between the speakerand the hearer-side region. This observation undermines the person-oriented interpretation in face-to-face conversation. Thus languages with three terms as well as languages with two terms use the ªrst term for the inside region in the conversational dyad as a whole. Another term for the endless outside space is of course obligatory. This contrast has to be considered as a pragmatic universal. The “middle” or second term of three-term systems comes into play when the hearer-side region becomes relevant, for instance when a teacher stands behind his student. In this face-to-back constellation it is very likely that the turned-away position of the hearer forces the speaker to conceptualize a region belonging to the hearer only, which is diŸerent from his own speaker-side region. Three-term systems oŸer a special term for this hearer-side region. In side-by-side conversation people are inclined to establish three diŸerent regions: a proximal, a medial and a distal one (see Wilkins 1999, “deictic sequence as a three-in-one” Sternberg 1983). This constellation is the only one where the diŸerent perspectives of the two (or more) persons involved nearly coincide. Spanish uses the three terms of the demonstrative pronouns to encode the distance-relevant distinctions. But there is no reason to generalize this view for the use of the demonstratives as proposed by the distanceoriented interpretation. The relevant concept of distances is not a measurable one. Thus ‘proximal’ can be everything, even stars in the sky, depending on the topic and the subjectively established borderlines between the diŸerent regions. Objects in front and others behind the speaker or the hearer may be equidistant, but only the objects inside the conversational dyad are considered to be ‘proximal’. The dyad-oriented interpretation takes into account the three constellations which create diŸerent conceptualizations of space and which are at the same time basic for conversational reality.

35

36

Konstanze Jungbluth

The proposal of this dyad-oriented system is valid for the description of the use of demonstrative pronouns in Spanish. The situations modelled this way are, however, not language speciªc. Instead they cover general conditions of language use in conversation. As I have shown, comparing diŸerent languages, the dyad-oriented system oŸers new insights into how particular languages refer to conceptualizations of space with respect to the position and orientation of the speaker and the hearer in relation to one another. Hopefully further research on deictic systems in other languages based on the dyad-oriented system will deepen our knowledge about language use in general at the universal level and of language use in particular at the historical level of the diŸerent languages of the world.

Notes 1. The term context is used in a very broad sense. Bühler (1934) proposes diŸerent contextual ªelds: the deictic, symbolic, synsemantic, symphysical and sympractical ªeld (see also Goodwin 1990). This concept is further diŸerentiated by Eugenio Coseriu. He distinguishes between direct and indirect situation, language speciªc territory, geographic region, cultural context, context within the language system, direct and indirect context of discourse, physical outside-discourse-context, empirical, natural, historical and cultural outside-discourse-context, practical outside-discourse-context and universe of discourse or domain (Coseriu 1955-56, Schlieben-Lange 1983: 23). 2. Weinrich ªnds no diŸerence, either, when he describes the use of demonstratives in Portuguese: “Note-se, de passagem, que a explicação que proponho dos fenómenos deícticos não é afectada, no que diz respeito à signiªcação e função destas formas, pela diferença estrutural existente entre as formas presas (bound forms) como este (3), esse (12), aquele (9) e as formas livres (free forms) que surge no texto, por exemplo isto (4), isso (19) e aquilo (2).” (Weinrich 1993: 21) 3. See “Conversation as a social system (…) Experiment 1: individuals versus dyad”;. In: Bavelas/Chovil/Lawrie/Wade 1992: 476. 4. This example is invented. Years ago while working as a tourist guide I remember my Spanish colleagues speaking like that. 5. Certainly the diŸerent relative positions cannot be integrated in only one model. Ritva Laury, who is working on Finnish demonstratives, argues that “(…) the socially and interactionally based speaker/hearer contrast is more likely to be the basis for the acquisition of deictics than is spatial distality. (…) In the absence of diachronic data on the use of demonstratives in face-to-face conversation, we might look to child language acquisition to see whether the spatial meanings are indeed basic. Experimental studies have shown that children learn the contrastive meaning of this and that as well as here and there before they learn the contrastive meaning

Deictics in the conversational dyad

of near and far, (…) but after the my and your contrast (…) it is the ªrst/second person contrast which is the elementary deictic contrast (…).” (Laury 1996: 305). 6. Further research may show whether the face-to-back conversation may be integrated in face-to-face conversation (see section 4., esp. Table 4). Since face-to-back conversation often elicits the use of the middle terms, these situations show data which are of particular theoretical interest. 7. The analysis follows Diessel 1999, reinterpreted with the help of the colleagues at the Japanese department of the EKU Tübingen (Martina Ebi). 8. I would propose a four-term characterization for the system in Brazilian Portuguese. First the basic contrast between inside and outside represented by esse : aquele. In a second step the inside region may be divided into a speaker-side and a hearer-side region encoded by esse aqui : esse aí (see Jungbluth forthcoming). Or should we leave behind the infelicitous counting of terms? When using “demonstratives” in the way proposed by Diessel (1999) who includes locational adverbs, the paradigm of many languages will be in¶ated to multiterm systems. We have to be careful to control whether the items we are comparing are similar or whether they are diŸerent from one another. Diachronic studies show that the paradigms are not closed, but certain deictic terms may “¶ow” from one to another: Cavalcante (2000) considers that the terms have a certain deictical force: the locational adverbs are more powerful than the demonstratives, and these are more powerful than articles. When the paradigm shows lacunas, terms from the “upper” step may become part of the “lower” paradigm thus restoring lost contrasts. 9. In some sentences (e.g. in Langacker 1990: 52-53: 2a and b, 13 a and b) sometimes the inside-term is used and at others the outside-term. I am not familiar with Cora or any related language, but it may be that the di¹culty to explain why is due to the lack of consideration given to the hearer’s position. I am not surprised that in speaking about the town one has to decide whether this entity is “inside” or “outside”. In the ªrst case the speaker conceptualizes himself as being part of the town, that is, the speaker’s own identity is built upon living in this place (see Mondada 2002). In the second case there is a contrast expressed between one’s own place, e.g. the ¶at or house where one is living, and the outside world. In this kind of context we would never expect the term for medial distance. 10. In other contexts where the “middle” term refers to the hearer-side region, two-term systems may use possessive pronouns where the three-term systems still encode a similar concept by using their “second” demonstrative term.

References Alonso, Martin 1968 Gramatica del español contemporaneo. Madrid: Guadarrame. Anderson, Stephen R. and Edward L. Keenan 1985 “Deixis”. In Language typology and syntactic description. Vol. III: Grammatical categories and the lexicon, T. Shopen (ed.), 257–308. Cambridge: CUP.

37

38

Konstanze Jungbluth

Asenjo Orive, Maria Rosa 1990 Los Demonstrativos. (Problemas fundamentales del español 5). Salamanca: Colegio de España. Bavelas, Janet Beavin, Nicole Chovil, Douglas A. Lawrie and Allan Wade 1992 “Interactive Gestures”. Discourse Processes 15: 469–489. Benveniste, Émile [1958] 1966 Problèmes de linguistique générale. Paris: Gallimard. Bühler, Karl [1934] 1982 Sprachtheorie - Die Darstellungsfunktion der Sprache. Stuttgart: Gustav Fischer. Spanish translation: Julián Marías [1950/1961/1967/1979] 1985 Teoría del lenguaje. Madrid: Alianza. English translation: Donald Fraser Goodwin 1990 Theory of Language. Amsterdam: Benjamins. Carvalho, José G. Herculano de 1976 “Systems of deictics in Portuguese”. In Readings in Portuguese Linguistics, J. Schmidt-Radefeldt (ed.), 245–266. Amsterdam: Benjamins. Cavalcante, Mônica Magalhaes 2000 Expressões indicais em contextos de uso: por uma caracterização dos dêicticos discursivos. Tese de doutorado, Recife (UFPE), unpublished. Cheshire, Jenny 1996 “That jacksprat: An interactional perspective on English that”. Journal of Pragmatics 25: 369–393. 1997 “Involvement in ‘standard’ and ‘nonstandard’ English”. In Taming the vernacular: from dialect to written standard language, J. Cheshire and D. Stein (eds), 68–82. Harlow: Longman. Coseriu, Eugenio 1955–56 “Determinacíon y entorno: dos problemas de una lingüística del hablar”. Romanistisches Jahrbuch 7: 29–54. Diessel, Holger 1999 Demonstratives. Form, function and grammaticalization. Amsterdam: Benjamins. this volume “Demonstratives and the conceptualization of space: A cross-linguistic study”. Ehrich, Veronika 1992 Hier und Jetzt. Tübingen: Niemeyer. Fillmore, Charles [1971] 1975 Santa Cruz Lectures on Deixis. Bloomington: Indiana University Linguistic Club. Hanks, William F. 1992 “The indexical ground of deictic reference”. In Rethinking context: language as an interactive phenomenon. Studies in the social and cultural foundation of language, A. Duranti and C. Goodwin (eds), 43–76. Cambridge: Cambridge UP. 1996 “Language form and communicative practices”. In Rethinking linguistic relativity, J. J. Gumperz and St. Levinson (eds), 232–270. Cambridge: CUP.

Deictics in the conversational dyad

Hausendorf, Heiko this volume “Deixis and Speech Situation revisited: an Interactionist Approach”. Hottenroth, Priska 1982 “The System of Local Deixis in Spanish”. In Here and There. Cross-linguistics Studies on Deixis and Demonstration, J. Weissenborn and W. Klein (eds), 133–153. Amsterdam: Benjamins. Jungbluth, Konstanze 2001 “Binary and ternary deictic systems in speech and writing”. Philologie am Netz 15: 1–24, http://www.fu-berlin.de/phin/phin15/p15t1.htm. 2002 “Der Gebrauch von Deiktika in alltäglichen Handlungszusammenhängen. Bericht einer Feldstudie aus der Provinz Toledo”. In Deixis. Zeitschrift für Literaturwissenschaft und Linguistik 125, W. Klein and K. Jungbluth (eds), 56–78. Stuttgart: Kohlhammer. Jungbluth, Konstanze forthcoming Pragmatik der Demonstrativpronomina in den iberoromanischen Sprachen. Tübingen: Niemeyer. Kalverkämper, Hartwig 2000 “Vorläufer der Textlinguistik: die Rhetorik.” In Text- und Gesprächslinguistik, [Handbücher für Sprache und Kommunikation, Band 16,1 Textlinguistik], K. Brinker et al. (eds), 1–17. Berlin: de Gruyter. Kock, Josse de 1988 «Este, ese y aquel en el español escrito». In Historia de la lengua: el español contemporáneo, homenaje a Alonso Zamora Vicente, vol. I, Pedro Peira et al., 411–424. Madrid: Castalia. Kuhn, Thomas S. [1962] 21976 Die Struktur wissenschaftlicher Revolutionen. Frankfurt/Main: Suhrkamp. Kuno, Susumu 1973 The Structure of Japanese Language, London: MIT Press. Langacker, Ronald W. 1990 Concept, Image, and Symbol: The Cognitive Basis of Grammar. Berlin/New York: de Gruyter. Laury, Ritva 1996 “Conversational Use and Basic Meaning of Finnish Demonstratives”. In Conceptual Structure, Discourse and Language, A. E. Goldberg (ed.), 303– 319. San Diego: University of California. Lavric, Eva 1998 “Este, ese y aquel en posición determinativa”. In Atti del XXI Congresso internazionale di linguistica e filologia romanza, Bd. III: Lessicologia e semantica delle lingue romanze, G. Ruffino (ed.), 405–418. Tübingen: Niemeyer. Meira, Sergio and Raquel Guirardello Forthcoming Non-Contrastive Exophoric Use of Demonstratives in Spoken Brazilian Portuguese. Nijmegen (MPI).

39

40

Konstanze Jungbluth

Mondada, Lorenza 2002 “Die Indexikalität der Referenz in der sozialen Interaktion: diskursive Konstruktionen von «ich» und «hier»”. Deixis. Zeitschrift für Literaturwissenschaft und Linguistik 125, W. Klein and K. Jungbluth (eds), 79–113. Stuttgart: Kohlhammer. Rauh, Gisa 1983 “Aspects of Deixis”. In Essays on deixis, G. Rauh (ed.), 9–60. (Tübinger Beiträge zur Linguistik 188) Tübingen: Narr. Rubba, Jo 1996 “Alternate Grounds in the Interpretation of Deictic Expressions”. In Spaces, Worlds and Grammars, G. Fauconnier and E. Sweetser (eds), 227–261. (Cognitive Theory of Language and Culture) Chicago: Chicago University Press. Schlieben-Lange, Brigitte 1983 Traditionen des Sprechens. Stuttgart: Kohlhammer. Portuguese translation: 1993 Historía do falar. Campinas: Unicamp. Sternberg, Mair 1983 “World, Language and Convention”. In Essays on deixis, G. Rauh (ed.), 277– 316. (Tübinger Beiträge zur Linguistik 188) Tübingen: Narr. Vann, Robert E. 1995 “Constructing Catalanism: Motion Verbs, Demonstratives, and Locatives in the Spanish of Barcelona”. Catalan Review IX, Nº 2: 253–274. Weinrich, Harald [1988] 21995 “Über Sprache, Leib und Gedächtnis”. In Materialität der Kommunikation, H.-U. Gumbrecht (ed.), 80–93. Frankfurt/Main: Suhrkamp. Weinrich, Harald 1993 “Análise textual dos demonstrativos em português”. In Semiótica e linguística portuguesa e românica, J. Schmidt-Radefeldt (ed.), 15–23. Tübingen: Narr. Wilkins, David 1999 “Manual” for the 1999 Field Season, Version 1.0., Nijmegen: Language and Cognition Group of the Max Planck Institute for Psycholinguistics, mimeo. Zi¦c Fuchs, Milena 1996 “‘Here’ and ‘there’ in Croatian: a case study of an urban standard variety”. In The Construal of Space in Language and Thought, M. Pütz and R. Dirven (eds), 49–62. Berlin: de Gruyter.

Non-deictic uses of the deictic motion verbs kommen and gehen in German* Claudio Di Meola University of Rome “La Sapienza”

The German motion verbs kommen ‘to come’ and gehen ‘to go’ are generally considered to be deictic verbs describing motion towards an observer and motion away from an observer respectively. The existence of non-deictic uses of these two verbs, however, has received little attention in the literature. Studies of deictic motion verbs — in German as well as in other languages — have commented very little on it.1 Studies of the semantic ªeld of German motion verbs only examine the verb gehen coding a speciªc form of locomotion in physical space (with the meaning of ‘to walk’ as opposed to other verbs such as laufen ‘to run’, schleichen ‘to creep’ or stolzieren ‘to strut’).2 Studies of the so called Funktionsverben (empty verbs) are limited to the verb kommen in some abstract uses (ins Erzählen kommen ‘to start talking’, zu einer Entscheidung kommen ‘to come to a decision’ etc.).3 In this paper I will try to give a systematic account of the non-deictic uses of the two motion verbs starting from my previous work on kommen and gehen (Di Meola 1993, 1994). The analysis is based on 16,000 examples from a corpus of newspaper texts kindly put at my disposal by the “Institut für Deutsche Sprache” in Mannheim. The examples have mostly been simpliªed for the sake of conciseness and clarity; only in some cases have they been reported verbatim. I will concentrate on the uses of kommen and gehen having a human being as moving entity (trajector).

1.

Deictic core meaning and non-deictic variants

Motion can be deªned as “change through time in the location of some entity” (Langacker 1987: 167) and can be associated with a SOURCE-PATH-GOAL

42

Claudio Di Meola

image schema (LakoŸ 1987: 275, Johnson 1987: 113–117) as exempliªed in (i):

(i)

SOURCE GOAL I---------------------->I PATH General motion schema

The deictic use of gehen describes motion away from the observer (viewpoint = SOURCE), the deictic use of kommen describes motion towards the observer (viewpoint = GOAL) (ii): gehen:

SOURCE GOAL I----------------------> I PATH O

kommen: SOURCE GOAL I----------------------> I PATH O

(ii) Deictic motion schemas The deictic origo is typically deªned by the speaker’s location at coding time (Example 1): (1) Martin wirft die Partnerin einer Nacht raus: “Geh!” und ruft ihr im selben Atemzug nach: “Komm wieder!”. (“Die Zeit”; February 21, 1986; p. 53) ‘Martin throws his partner of one night out: “Go!” and shouts after her in one breath: “Come back!”.’

But there are several forms of origo displacement. The following sentences with kommen exemplify identification with the speaker’s location at reference time (2), identification with the hearer’s location either at coding time or at reference time (3a/b), identification with a home base (4) or with a particular viewpoint (5): (2)

Bitte komm morgen in mein Büro. ‘Please come to my o¹ce tomorrow.’

(3) a.

Ich komme jetzt in dein Büro. ‘I’m coming to your o¹ce now.’

Non-deictic uses of the deictic motion verbs kommen and gehen in German

b. Ich komme morgen in dein Büro. ‘I’ll come to your o¹ce tomorrow.’ (4)

Morgen bin ich nicht in meinem Büro. Du kannst trotzdem kommen und meinen Computer benutzen. ‘Tomorrow I will not be at my o¹ce. You can come all the same and use my computer.’

(5)

Alle waren im Sitzungssaal versammelt. Dann kam der Hauptredner. ‘They were all gathered in the conference hall. Then the main speaker came.’

In this last example the viewpoint is not taken by a speech act participant, but by the narrator. By now we have exclusively considered uses of the two verbs in which no element of the SOURCE-PATH-GOAL schema is explicated: the only spatial reference system is the deictic viewpoint. In other uses a second, external coordinate system can be added in the form of an explicit indication of the GOAL, the PATH or the SOURCE (cf. 6 and 7): (6) a.

Der Redner kam in den Saal. ‘The speaker came into the hall.’ b. Der Redner kam durch den langen Gang. ‘The speaker came down the long corridor.’ c. Der Redner kam aus dem Treppenhaus. ‘The speaker came out of the stairwell.’

(7) a.

Der Redner ging in den Saal. ‘The speaker went into the hall.’ b. Der Redner ging durch den langen Gang. ‘The speaker went down the long corridor.’ c. Der Redner ging aus dem Treppenhaus. ‘The speaker went out of the stairwell.’

Such sentences are linked to non-deictic uses because they can also be interpreted in a non-deictic sense. Take examples (7a–c): in a deictic reading they can be seen in opposition to kommen, in a non-deictic reading in opposition to other (non-deictic) motion verbs. The position of a potential observer now becomes irrelevant (8): (8) a.

Der Redner ging / schlich / schritt / lief / humpelte ... in den Saal. ‘The speaker walked / crept / stalked / ran / hobbled... into the hall.’

43

44

Claudio Di Meola

b. Der Redner ging / schlich / schritt / lief / humpelte ... durch den langen Gang. ‘The speaker walked / crept / stalked / ran / hobbled... down the long corridor.’ c. Der Redner ging / schlich / schritt / lief / humpelte ... aus dem Treppenhaus. ‘The speaker walked / crept / stalked / ran / hobbled... out of the stairwell.’

2.

Non-deictic gehen as a verb of unrestricted motion

The deictic uses of kommen and gehen considered so far are examples of the most typical and natural form of human locomotion. From a trajector equipped with intelligence, will and autonomous mobility we expect a completely free and unrestricted form of motion. Unrestricted motion can be characterized as unimpeded (there are no obstacles disturbing the regular course of the movement), intentional (the trajector moves in direction of a goal set up by himself), and active (the trajector performs a self-propelled movement). The non-deictic uses of gehen, too, refer to instances of unrestricted motion. More speciªcally, in its basic non-deictic use gehen describes a movement performed “on foot” (9): (9) Sie ging in die Oper. ‘She walked to the opera.’

Sentence (9) could be an answer to a question such as Wie gelangte sie in die Oper? (‘How did she get to the opera?’). The same sentence, however, also has an alternative reading (10) and can be thought of as an answer to a question such as Was machte sie gestern abend? (‘What did she do yesterday evening?’): (10) Sie ging in die Oper. ‘She went to the opera.’ (i.e. ‘She went to see a performance at the opera.’)

Now the attention is focused on what happens after the physical dislocation of the trajector: motion towards the GOAL metonymically stands for activity at the GOAL (the precondition of an action stands for the action itself). This particular kind of metonymy can be considered a subtype of the general metonymic principle according to which the place stands for the activity performed at the place. The manner of dislocation loses its relevance since the

Non-deictic uses of the deictic motion verbs kommen and gehen in German

movement could be performed on foot as well as by a means of transport. There are many other similar examples (11): (11) a.

ins Kino / ins Theater / in die Disco / ins Stadion / in den Zirkus gehen ‘to go to the cinema / to the theater / to the stadium / to the circus’ b. ins Restaurant / in eine Kneipe / in einen Stehausschank gehen ‘to go to a restaurant / to a pub / to a stand-up bar’ c. zur Kirche gehen ‘to go to church’ d. zum Arbeitsamt / zum Sozialamt / zur Post / zur Bank / zum Arzt gehen ‘to go to the job center / to the social security o¹ce / to the post o¹ce / to the bank / to the doctor’ e. in den Kindergarten / zur Schule / zur Universität gehen ‘to go to the kindergarten / to school / to university’

The trajector sets oŸ for a GOAL with the intention of performing the normal activity for that kind of place. The type of activity is determined by our social conventions: the expression ins Kino gehen means ‘to go to a cinema and watch a ªlm’, ins Restaurant gehen means ‘to go to a restaurant and eat a meal’ etc. It is interesting to note that in the functional uses of gehen the activity at the GOAL is so strongly foregrounded that the physical GOAL itself can become irrelevant. Take for example the expression auf die Straße gehen (lit. ‘to go onto the street’, functionally ‘to demonstrate’): the activity of demonstrating could be performed in other places than a street, and we would still say auf die Straße gehen. Take the expression ins Bett gehen (lit. ‘to get into bed’, functionally ‘to go to sleep’ or ‘to have sexual intercourse’): the relevant activities could also be performed in other places than a bed. Many of the above expressions are potentially ambiguous in their functional reading since they can refer both to a unique (12a) and to a regular activity (12b): (12) a.

An dem Morgen ging mein Sohn zur Schule, obwohl er krank war. ‘That morning my son went to school although he was ill.’ b. Mein Sohn ging zur Schule, obwohl er bereits zwanzig war. ‘My son went to school although he was already 20 years old.’

In other cases the trajector moves to a distant place with the intention of settling down (Examples 13):

45

46

Claudio Di Meola

(13) a.

Er ging nach Amsterdam und begann dort eine Kaufmannslehre. (“Mannheimer Morgen”; June 19, 1986; p. 34) ‘He went to Amsterdam and began training as a sales assistant there.’ b. Er war nach der Promotion als Lektor des “Deutschen Akademischen Austauschdienstes” nach Hongkong gegangen. (“Die Zeit”; March 14, 1986; p. 66) ‘After he got his Ph.D. he went to Hongkong as a lector of the German Academic Exchange Service.’ c. Wer dafür nicht die nötige Sensibilität aufbringen kann, der sollte vielleicht nicht unbedingt als Aufbauhelfer nach Nicaragua gehen. (“Die Zeit”; September 5, 1986; p. 63) ‘Those who haven’t got the required sensitivity shouldn’t necessarily go to Nicaragua as a reconstruction helper.’ d. Heinrich Meng ging nach Basel [...], Erich Fromm und Frieda Fromm-Reichmann gingen ebenso wie Horkheimer in die USA. (“Die Zeit”; January 17, 1986; p. 57) ‘Heinrich Meng emigrated to Basle [...], Erich Fromm and Frieda Fromm-Reichmann emigrated to the USA as did Horkheimer.’

In this context the verb gehen always refers to a long-term stay at the destination. In the case of a short visit gehen is in fact barely acceptable (14a); instead, other verbs such as fahren ‘to go by a means of transportation’ or reisen ‘to travel’ are used (14b): (14) a. ??Er ging für einen Kurzurlaub / zu einem Kongreß in die USA. ‘He went to the USA for a short holiday / for a conference.’ b. Er fuhr für einen Kurzurlaub / er reiste zu einem Kongreß in die USA. ‘He went to the USA for a short holiday / for a conference.’

Interestingly gehen can here be combined with a time adverbial indicating duration. This adverbial refers not to the movement, but to the activity at the destination (15): (15) Er ging für drei Jahre in die USA. ‘He went to the USA and stayed there for three years.’

In other cases gehen indicates a prohibited activity performed at the GOAL (16): (16) a.

Wer ist an den Kuchen gegangen? ‘Who has touched the cake?’

Non-deictic uses of the deictic motion verbs kommen and gehen in German

b. Wer ist an meine Briefmarkensammlung gegangen? ‘Who has touched my stamp collection?’ c. einer Frau an den Busen / an die Wäsche / unter den Rock gehen ‘to touch a woman’s breasts / underwear / skirt’

In the (a)-example somebody has eaten a piece of cake and should not have; in the (b)-example somebody has messed up the speaker’s stamp collection; the (c)-examples are clear cases of sexual harassment. There are ªnally some non-deictic uses of gehen describing motion without any coordinate system serving as spatial reference frame (17–19): (17) a.

Im Alter von einem Jahr lernt ein Kind gehen. ‘At the age of one a child learns to walk.’ b. Nach dem Unfall konnte er monatelang nicht gehen. ‘After the accident he couldn’t walk for months.’ (18) a. Sie nimmt nicht den Bus, sie geht lieber. ‘She doesn’t take the bus, she prefers to walk.’ b. Sie stürzt mehr, als sie geht. ‘She races rather than walks.’ (19) a. Rot bedeutet stehenbleiben, grün bedeutet gehen. ‘The red light means stop, green means go. b. Wo er geht und steht... ‘Wherever he goes or is...’

In these contexts, gehen indicates the human ability to walk (17), the choice of walking in opposition to other forms of locomotion (18) or a general condition of locomotion in opposition to a condition of standstill (19). In some uses the specific way of walking is explicitly described (20): (20) a.

auf Zehenspitzen / auf Plattfüßen / barfuß gehen ‘to walk on tiptoe / ¶at-footed / barefooted’ b. in Turnschuhen / auf Socken gehen ‘to walk with sneakers / on socks’ c. am Stock / auf Krücken / auf Stelzen gehen ‘to walk with a stick / on crutches / on stilts’

3.

Non-deictic kommen as a verb of restricted motion

The non-deictic uses of kommen, in contrast to those of gehen, refer to cases where an external force in¶uences the movement performed by the human

47

48

Claudio Di Meola

trajector. We can distinguish three main cases of restricted motion: obstacle (the movement is no longer unimpeded), accident (the movement is no longer intentional), and transport (the movement is no longer active). (1) Obstacle. The movement performed by the trajector is restricted by the presence of an obstacle that could disturb the regular course of the movement and the arrival at the planned GOAL. In the following examples the di¹culties to be overcome in order to reach the GOAL are mentioned in the sentence (21): (21) a.

Man kommt kaum über den Platz, ohne daß einen die Autos über den Haufen fahren! (“Die Zeit”; November 7, 1987; p. 60) ‘You can hardly cross the square without the cars running you over!’ b. Sie stecken im Stau auf der Autobahn oder kommen nur schrittweise durch die City. (“Mannheimer Morgen”; February 21, 1987; p. 24) ‘They are caught in a tra¹c jam on the expressway or make their way through the city center at walking speed.’ c. Wer ist nicht auch schon völlig orientierungslos durch eine fremde Stadt geirrt und nur auf Umwegen an sein Ziel gekommen? (“Mannheimer Morgen”; March 6, 1987; p. 22) ‘Who hasn’t completely lost his way in an unknown city and has got to his destination only by a roundabout route?’ d. Um zur “Bravo-Batterie” des 1. Hawk-Luftabwehrbatallions zu kommen, muß man vom Hauptquartier [...] rund 20 Kilometer über kurvenreiche, holprige Nebenstraßen fahren. (“Stern”; September 17, 1987; p. 32) ‘In order to reach the “Bravo battery” of the 1. Hawk anti-aircraft defense battalion one has to drive from headquarters for about 20 kilometers on bendy, bumpy by-roads.’

In some cases the relevant difficulties are not explicitly mentioned, but can easily be inferred from the context (22): (22) a.

Sie hatten Mühe, zu ihren Plätzen zu kommen. ‘They had di¹culties in reaching their seats.’ b. Ich komme mit dem Besen nicht unter den Schrank. ‘I can’t get under the cupboard with the broom.’

In (22a) we immediately think of an overcrowded hall, in (22b) we imagine that there is little space between the bottom of the cupboard and the floor. In the following examples the notion of an obstacle has somehow faded (23):

Non-deictic uses of the deictic motion verbs kommen and gehen in German

(23) a.

Auf seinen Reisen ist er bis nach Grönland gekommen. ‘On his travels he has got as far as Greenland.’ b. Entschuldigen Sie, wie komme ich zum Hauptbahnhof? ‘Excuse me, how do I get to the central station?’

At first sight, kommen seems only to indicate arrival at the destination after a series of intermediate stops. On closer inspection, however, it is clear that the GOAL is not that easy to reach. In fact, Greenland is a distant and extremely unusual destination for a German traveller (23a). And even an apparently easy task such as reaching the central station can be difficult for somebody who doesn’t know his way around the town (23b). The verb kommen is also used when the obstacle is merely potential (24): (24) a.

Sie kamen ohne Probleme durch den Zoll. ‘They got through customs without problems.’ b. Er ist leicht über den hohen Zaun gekommen. ‘He easily got over the high fence.’

In the following examples the trajector tries to move away, but an external force nails him down (25): (25) a.

Er wollte weglaufen, aber er kam nicht von der Stelle / vom Fleck. ‘He tried to run away, but he was rooted to the spot.’ b. Ein jubelndes Publikum sorgte dafür, daß die Musiker nicht so schnell von der Bühne kamen. (“Mannheimer Morgen”; January 16, 1987; p. 36) ‘A cheering audience made sure that the musicians couldn’t leave the stage so quickly.’

(2) Accident. The movement is restricted in so far as it is no longer intentional. The trajector moves by himself, but his movement takes an unforeseen turn so that he reaches a GOAL he hasn’t intended to (26): (26) a.

Er ist an den Schreibtisch gekommen und hat sich weh getan. ‘He ran into the desk and hurt himself.’ b. Beim Überqueren der Straße ist er unter ein Auto gekommen. ‘Crossing the street he got run over by a car.’ c. Er ist mit seinem Finger in den Rasenmäher gekommen. ‘He caught his ªnger in the lawn-mower.’ d. Das Kind ist an die Blumenvase gekommen, die dann vom Tisch gefallen ist. ‘The child bumped into the ¶ower vase, which fell oŸ the table.

49

50

Claudio Di Meola

These sentences are examples of metonymical extensions. It is not the movement itself which is focused on, but the consequences arising for the trajector or for other entities in the spatial conªguration. (3) Transport. In quite a number of uses kommen describes a movement the trajector has completely lost control of (27): (27) Vier junge Frauen sterben, 36 Opfer kommen mit lebensgefährlichen Verletzungen ins Krankenhaus. (“Stern”; July 16, 1987; p. 8) ‘Four young women die, 36 victims go to hospital with critical injuries.’

The trajector does not move “under his own steam” any more. S/he cannot decide on his/her own fate but has to submit to the authority of other people (in this case for example a doctor). The GOAL is the place of destination for the trajector — a place designated by the conventions of our society: a seriously injured or ill person should be sent to an appropriate medical institution (28), a person accused of a crime to a court of justice (29), a convicted person to a place of detention (30), an orphan or a problem child to a reformatory (31), an old and infirm person to a place where s/he can be looked after (32): (28) a.

in die Klinik / ins Sanatorium kommen ‘to go into a clinic / into a sanatorium’ b. auf die Intensivstation / in Quarantäne kommen ‘to be sent to an intensive care unit / to be put in quarantine’ c. an den Tropf / auf den OP-Tisch kommen ‘to be put on a drip / on the operating table’ d. in die Nervenklinik / in eine psychiatrische Anstalt / in die Psychiatrie kommen ‘to be sent to a psychiatric clinic’ (29) vor Gericht / vor den Richter / auf die Anklagebank kommen ‘to be taken to court’ (30) a. ins Gefängnis / in den Knast / ins Zuchthaus / in eine Justizvollzugsanstalt / in eine Strafanstalt kommen ‘to go to prison’ b. hinter Gitter / hinter schwedische Gardinen kommen ‘to be sent behind bars’ c. auf den elektrischen Stuhl / in die Gaskammer kommen ‘to go to the electric chair / to the gas chamber’ (31) in ein Kinderheim / in eine Erziehungsanstalt / in ein Internat kommen ‘to be sent to a children’s home / to a reformatory / to boarding school’

Non-deictic uses of the deictic motion verbs kommen and gehen in German

(32) ins Altersheim / ins P¶egeheim kommen ‘to go into an old peoples’ home / into a hospice’

These expressions are characterized by a metonymy: the movement towards the GOAL is of minor importance compared to what happens after reaching the GOAL: the precondition for an event (motion) stands for the event itself (medical cure, trial, detention or execution etc.). Interestingly, these expressions can be combined with a time adverbial indicating duration. The adverbial refers not to the movement, but to the event at destination: (33) a.

Er kam für drei Wochen ins Krankenhaus. ‘He went into hospital and stayed there for three weeks.’ b. Er kam für drei Jahre ins Gefängnis. ‘He went to prison and stayed there for three years.’

We can describe the non-deictic uses of kommen as a progressive deviation from the characteristics of prototypical human locomotion as coded by gehen. For the gehen-movement and for the variants of the kommen-movement (obstacle, accident, transport) we will consider three different aspects: (1) the role played by the trajector; (2) the role played by the external force; (3) the relative strength of the two entities involved. The contrast between the internal force residing in the trajector and the external force can be seen in terms of force dynamics (cf. Talmy 1988). Consider table (iii): (iii) Non-deictic uses of kommen and gehen: role of trajector vs. role of external force gehen: kommen: obstacle kommen: accident kommen: transport

role of trajector

role of external force

stronger entity

plan and performance plan and performance

non-existent /irrelevant hindrance

trajector trajector

disparity between plan and performance neither plan nor performance

diversion

external force

propulsion

external force

With prototypical gehen-movement, the trajector has full mental and physical control of the movement: s/he performs the movement in the precise way s/he has planned it. Potentially adverse external forces do not exist or are considered totally irrelevant. In the obstacle-case the trajector still performs the movement s/he has in mind, but s/he has to overcome resistance. The trajector proves to be stronger

51

52

Claudio Di Meola

than the opposing external force. In the accident-case there is no correspondence between plan and performance of the movement: the trajector has set up a certain GOAL, but the movement s/he actually performs brings him/her to another, undesirable GOAL. Now the external force proves to be stronger. In the transport-case the trajector has no role in the movement at all: s/he has not planned it and s/he doesn’t perform it. The supremacy of the external force is even more evident. With the various kinds of kommen-movement, we can thus observe a decreasing importance of the trajector and an increasing importance of the external force. The gradual power shift from trajector to external force becomes clear when we ask ourselves: what would have happened without the intervention of an external force? There is in fact an increasing diŸerence between the real scenario (presence of external force) and the hypothetical scenario (absence of external force). With gehen, there would have been no diŸerence for the alternative scenario: the movement would have been exactly the same. In the obstacle-case the movement would have been easier and/or quicker; in the accident-case it would have taken another direction (towards the planned GOAL); in the transport-case it would not have taken place at all.

4.

The opposition of non-deictic kommen and gehen in comparable contexts

It has been claimed that restriction/non-restriction is the crucial factor for distinguishing the non-deictic uses of kommen and gehen: for the former, the in¶uence of an external force is relevant, for the latter it is not. In the literature, on the other hand, the diŸerence between the two verbs is generally interpreted in terms of focalization: focus on the SOURCE (gehen) as opposed to focus on the GOAL (kommen). But two important facts contradict this hypothesis: 1. The existence of the functional, purpose-focusing variant of gehen (e.g. zur Schule gehen ‘to go to school’). How can we say that gehen focuses the SOURCE when the activity at the GOAL is more important than the movement itself? If there is a focus at all, it seems to be primarily on the GOAL (place of activity) and only secondarily on the SOURCE (place where movement and activity have been planned).

Non-deictic uses of the deictic motion verbs kommen and gehen in German

2. The existence of kommen-uses with a SOURCE/PATH-indication. Uses such as er kommt nicht von der Stelle ‘he is rooted to the spot’ or er kommt über den Zaun ‘he got over the fence’ do not focus on the GOAL, but clearly on the SOURCE and PATH of the movement respectively. Moreover, the GOAL is not even explicitly mentioned. We now look at some “minimal pairs” (sentences in which kommen and gehen are placed into the same linguistic context). These examples will show that restriction is indeed the fundamental diŸerence between the two verbs in the non-deictic sphere. Consider the following sentences (34): (34) a.

Ich weiß nicht, ob ich heute nachmittag noch zur Post gehe. ‘I don’t know if I will go to the post o¹ce this afternoon.’ b. Ich weiß nicht, ob ich heute nachmittag noch zur Post komme. ‘I don’t know if I can get to the post o¹ce this afternoon.’

In the gehen-example the speaker doesn’t know yet whether s/he will be in the mood to go to the post office or not. In the kommen-example it is understood that the speaker has so many things to do that s/he might not have the time to go to the post office. The following pair of examples (35) is similar: (35) a.

Er geht selten aus dem Haus. ‘He seldom leaves his home.’ b. Er kommt selten aus dem Haus. ‘He can seldom leave his home.’

In the gehen-sentence it is the decision of the trajector to stay at home, in the kommen-sentence the trajector would like to go out but is inevitably detained. Consider the following examples (36): (36) a.

Nach einem Herzanfall ging sie ins Krankenhaus. ‘After a heart attack she went into hospital.’ b. Nach einem Herzanfall kam sie ins Krankenhaus. ‘After a heart attack she was sent to hospital.’

In (36a) the trajector has the choice between out-patient and in-patient treatment (she presumably had a slight heart attack), in (36b) the decision is taken by somebody else (she presumably had a serious heart attack und wasn’t even in a condition to make a decision). A similar contrast is shown in the following pair of sentences (37): (37) a.

Der Beamte ging nach Berlin. ‘The civil servant went to Berlin.’

53

54

Claudio Di Meola

b. Der Beamte kam nach Berlin. ‘The civil servant was sent to Berlin.’

In the gehen-sentence the civil servant decides to work in Berlin (for professional or personal reasons), in the kommen-sentence he is transferred without having an influence on his career. In the following context gehen refers to a trajector who is the plaintiff (38a), kommen to a trajector who is the defendant (38b): (38) a.

Sie ging vor Gericht. ‘She went to court.’ b. Sie kam vor Gericht. ‘She was taken to court.’

5.

A semantic network for kommen and gehen in physical space

The following network (iv) gives an overview of the most important uses of kommen and gehen in physical space: kommen – transport [G only: event at Goal] – accident [G only: event at Goal] – obstacle [S/P/G] centripetal deictic

centripetal →

restricted →

deictic + S/P/G

centrifugal

centrifugal

***

non-deictic + S/P/G ← → non-deictic unrestricted manner-focusing – manner-determined [S/P/G] – purpose-focusing [G only: activity at Goal]

gehen

viewpoint ○

Legend:

○

○

no viewpoint ○

○

○

○

○

○

○

○

○

○

○

○

○

○

○

○

○

○

○

○

○

external coordinate system S = SOURCE P = PATH G = GOAL

(iv) The polysemy of kommen and gehen in physical space

In the upper part of the network are the uses of kommen, in the lower part those of gehen; on the left are the deictic uses of both verbs and on the right the non-

Non-deictic uses of the deictic motion verbs kommen and gehen in German

deictic uses. As regards the deictic uses, the diŸerence between kommen and gehen is the direction of the movement: centripetal for kommen (i.e. directed towards an observer), centrifugal for gehen (i.e. moving away from an observer). This opposition is valid for “simple” deictic uses and for those that — in addition to the deictic viewpoint — are characterized by a supplementary coordinate system involving a SOURCE/PATH/GOAL-indication. For the non-deictic uses containing an indication of SOURCE/PATH/GOAL, the two verbs diŸer in the restricted (kommen) vs. unrestricted (gehen) character of the movement. For the non-deictic uses without an external coordinate system, gehen is the only possible verb (non-deictic kommen is ruled out). I would like to draw attention to some important issues regarding this semantic network: 1. Basicness of deictic uses. We can assume that deictic uses of the two verbs are basic with respect to non-deictic uses. Several facts support this claim: (a) deixis is a fundamental human experience (egocentric spatial categorization is more immediate and less abstract than an “objective” external coordinate system); (b) interlinguistic comparison shows more similarities for deictic variants than for non-deictic ones; (c) speakers intuitively characterize the opposition between the two verbs as deictic; (d) children seem to acquire deictic variants before non-deictic ones. 2. Derivation of non-deictic uses from deictic uses. The connection between deictic and non-deictic uses does not seem to be straightforward: it is not immediately clear what the link is between centripetal and restricted motion and between centrifugal and unrestricted motion. In my opinion two stages have to be postulated: ªrst from viewpoint to focus (via attention shift), then from focus to restriction/non-restriction (via implicature). With the ªrst step the place where the (deictic) observer was positioned now becomes the place the (non-deictic) attention is focused on: kommen: gehen:

centripetal > GOAL-focus centrifugal > SOURCE-focus

The second step leads from focus to restriction/non-restriction: kommen: gehen:

GOAL-focus > restricted SOURCE-focus > unrestricted

This step is based on two fundamental assumptions the speaker/hearer makes: a. a human trajector usually performs an intentional and active movement; b. some external factor in¶uencing the movement can always arise.

55

56

Claudio Di Meola

In the case of the verb kommen we know that the trajector has reached the GOAL. On the basis of conversational implicature (which becomes conventional) the speaker/hearer argues as follows: if we focus on the GOAL, the fact of reaching it must be relevant (non-obvious); therefore an external force must have in¶uenced the movement (hindering the trajector, pushing him/her towards an unforeseen GOAL, transporting him/her towards the GOAL). In the case of the verb gehen we know that the trajector leaves the SOURCE. We recognize that s/he is willing and able to perform the planned movement. Since we do not yet know anything about possible restrictions coming up we see the movement as (potentially) unrestricted. 3. Manner of motion. The verb gehen, in many of its uses, is underspeciªed with respect to the manner of motion. It is not relevant whether the movement is performed using a means of transport or simply by ‘walking’, because other aspects are foregrounded: in the deictic uses the direction relative to the observer, in the functional non-deictic uses the purpose (the activity to be performed at the GOAL). The focus on the purpose is so strong that in some cases we have gehen even if a movement ‘on foot’ is excluded (er ging nach Australien ‘he emigrated to Australia’ —which is only possible by plane or ship). Only in those non-deictic uses in which the purpose is backgrounded is the manner of motion (pre-)determined: gehen means ‘to walk’ (sie geht die Straße entlang ‘she walks along the road’). In the non-deictic uses without an external reference system, it is precisely the manner of motion which is focused on (sie kann wieder gehen ‘she can walk again’): it is explicitly stated that the movement takes place ‘walking’ and not in some other way. 4. Asymmetry between the two verbs. The verb kommen seems to have one prototypical center in the network: the (centripetal) deictic use. On the other hand, gehen seems to have two prototypical centers: the (centrifugal) deictic use and the (manner-focusing) non-deictic use.

6.

Non-deictic kommen and gehen in abstract space

Abstract uses of kommen and gehen describe the performing of an action or the course of events as motion in space. A change of state — i.e. the transition from state A to state B — is categorized as movement from location A to location B. LakoŸ (1990: 57) mentions several conceptual metaphors that determine the mapping from the source domain (motion in physical space) to the target

Non-deictic uses of the deictic motion verbs kommen and gehen in German

domain (event structure) (v): (v) Motion metaphors CHANGES ARE MOVEMENTS INTO OR OUT OF BOUNDED REGIONS PROCESSES ARE MOVEMENTS ACTIONS ARE SELF-PROPELLED MOVEMENTS PURPOSES ARE DESTINATIONS MEANS ARE PATHS TO DESTINATIONS

It is now interesting to note that non-deictic kommen and non-deictic gehen preserve their distinctiveness in abstract uses: the verb gehen refers to an unrestricted change of state (intentionally and actively pursued without impediments), the verb kommen to a change of state which is in some way restricted. The differences in the semantics of the two verbs become clear when we insert kommen and gehen in similar contexts. Consider the following pair of examples (39): (39) a.

Der Bergsteiger ging an die Grenzen seiner Möglichkeiten. ‘The mountain climber went to the limit of his possibilities.’ b. Der Bergsteiger kam an die Grenzen seiner Möglichkeiten. ‘The mountain climber reached the limit of his possibilities.’

In the gehen-sentence the mountain climber wanted to put his ability to the test, in the kommen-sentence he overestimated his powers and lost control of the situation. In the following pair of examples gehen indicates that the trajector had a choice (40a), kommen that he had no choice (40b): (40) a.

Er ging zur Bundeswehr. ‘He joined the [German] army.’ b. Er kam zur Bundeswehr. ‘He was sent to the [German] army.’

Typically the first sentence refers to a volunteer who wants to become a professional soldier, the second to a conscript. Somewhat surprisingly, the gehen-sentence can also be used for a conscript: in this particular context the verb gehen emphasizes the fact that the trajector had some power of decision (military service is seen as an alternative to community service). The following examples also refer to professional life (41): (41) a.

Sie ging zum Theater. ‘She went into the theater.’ b. Sie kam zum Theater. ‘She ended up on the stage.’

57

58

Claudio Di Meola

In the gehen-sentence the trajector decides to start an acting career following her vocation. The kommen-sentence, on the other hand, generally implies that she had other professional plans and it was only by chance that she became an actress. Compare the following two sentences (42): (42) a.

Sie ging auf Konfrontationskurs. ‘She headed for a confrontation.’ b. Sie kam auf Konfrontationskurs. ‘She got into a confrontation.’

The gehen-example refers to a deliberate, risk-taking decision, the kommenexample to an accidental development or to a situation where the trajector has to submit to the powerful course of events. The following two sentences describe totally different circumstances (43): (43) a.

Er ging an sein Geld. ‘He took his money.’ b. Er kam an sein Geld. ‘He got his money.’

In the gehen-sentence the trajector has saved money and decides to take his savings to buy something (e.g. a new car), in the kommen-sentence he ªnally gets the money he is entitled to (e.g. a sum his employer has owed him for a long time). In some cases the gehen/kommen-constructions diŸer in the choice of the preposition in the PP indicating the abstract GOAL. The two sentences constitute “quasi-minimal” pairs (44): (44) a.

Sie ging in den Tod. ‘She went to her death.’ b. Sie kam zu Tode. ‘She met her death.’

The gehen-example refers to a suicide or to a deliberate acceptance of an extremely dangerous situation that could lead to her death, the kommenexample to a deadly accident (e.g. a car crash). Consider also the following sentences (45): (45) a.

Sie ging an eine schwere Aufgabe. ‘She tackled a di¹cult task.’ b. Sie kam zu einer schweren Aufgabe. ‘She was assigned a di¹cult task.’

Non-deictic uses of the deictic motion verbs kommen and gehen in German

In the gehen-sentence the trajector willingly confronts a problem, in the kommen-sentence she is compelled to deal with it. It is interesting to note that the use of gehen or kommen is also motivated in idiomatic expressions (where only one of the two verbs can appear). In the following pairs of examples (46–48) both expressions describe a comparable development; the fundamental diŸerence is that with the gehen-expression the new situation is seen as actively and consciously pursued by the human trajector, while with the kommen-expression it is seen as an inevitable event experienced by the trajector: (46) a.

in den Untergrund gehen ‘to go underground’ b. auf die schiefe Bahn kommen ‘to leave the straight and narrow’ (47) a. auf Tuchfühlung gehen ‘to get in contact’ b. in Kontakt kommen ‘to come into contact’ (48) a. in den Kampf gehen ‘to go to war’ b. ins Gefecht kommen ‘to get into a battle’

7.

Outlook: kommen and gehen with inanimate trajectors

The verbs kommen and gehen can also be used with inanimate trajectors. Interestingly, the fundamental opposition based on restriction/non-restriction is still preserved in many uses. Compare sentences (49a/b) to (50a/b): (49) a.

Der Wind geht durch den Pullover. ‘The wind goes through the pullover.’ b. Die Röntgenstrahlen gehen durch den Körper. ‘The X-rays go through the body.’ (50) a. Der Wind kommt durch die Windjacke. ‘The wind gets through the windcheater.’ b. Die Röntgenstrahlen kommen durch den Schutzschirm. ‘The X-rays get through the shield.’

59

60

Claudio Di Meola

The gehen-examples indicate an unrestricted movement, the kommen-examples the overcoming of an obstacle — the function of a windcheater is to protect against the wind, the function of a shield is to protect against X-rays. We also find an opposition between the two verbs (51) in cases of subjective motion (cf. Langacker 1986): (51) a.

Die Straße geht bis zum Bergdorf. ‘The road goes to the mountain village.’ b. Die Straße kommt bis zum Bergdorf. ‘The road reaches the mountain village.’

The verb gehen indicates the “normal”, unhindered scanning performed by the conceptualizer of the spatial scene, the verb kommen implicates the presence of some obstacle. It is emphasized that the road indeed runs as far as the barely accessible mountain village. In the following example (52) an unrestricted potential motion (gehen in the sense of ‘to ªt’) is opposed to a passive motion (kommen in the sense of ‘to be transported to the place of destination’): (52) a.

Die Teller gehen in den kleinen Schrank. ‘The plates ªt into the small cupboard.’ b. Die Teller kommen in den kleinen Schrank. ‘The plates go into the small cupboard.’

The non-deictic opposition for inanimate trajectors is preserved even in abstract uses of the two verbs (53): (53) a.

Ich werde dafür sorgen, daß alles in Ordnung geht. ‘I will make sure that everything will be put right. b. Ich hoŸe, daß alles in Ordnung kommt. ‘I hope that everything will turn out all right.’

The use of gehen highlights the active character of the change of state (the course of events is seen as influenced by humans) and the use of kommen the passive character (the course of events is seen as inevitable or as independent of human intervention). The opposition in the next two pairs of examples (54/ 55) is similar: (54) a.

Die Produktion geht in die zweite Phase. ‘The production is set for the next phase.’ b. Die Produktion kommt in die zweite Phase. ‘The production goes into the next phase.’

Non-deictic uses of the deictic motion verbs kommen and gehen in German

(55) a.

Das Flugblatt ging in Umlauf. ‘The lea¶et was put into circulation.’ b. Das Flugblatt kam in Umlauf. ‘The lea¶et went in circulation.’

However, gehen is rarely used in abstract space when referring to an inanimate trajector (precisely because the trajector is not equipped with autonomous will and mobility). On the other hand, kommen is quite frequent. These kommenexamples have been thoroughly investigated in the literature on the so called Funktionsverben. These “empty” verbs have (partially) lost their original lexical meaning and have acquired the grammatical function of signaling the Aktionsart of the verbal complex. Interestingly, two different Aktionsarten are generally assigned to the verb kommen — terminative in expressions such as (56), inchoative in expressions such as (57): (56) zum Halten / zum Stehen / zum Stillstand kommen ‘to come to a halt / to come to a stop / to come to a standstill’ (57) in Bewegung / in Fahrt / in Schwung kommen ‘to start moving / to get going / to gain momentum’

Terminative Aktionsart can be equated with focus on the GOAL of an abstract movement, inchoative Aktionsart with focus on the SOURCE. The fact that kommen seems to be compatible with both types of focus thus provides further evidence for my claim that focus cannot be the crucial factor in distinguishing kommen and gehen in the non-deictic sphere.

8.

Potential ambiguity: deictic vs. non-deictic interpretation

We have seen that the verbs kommen and gehen can take on a deictic as well as a non-deictic value. The deictic function of these verbs is therefore not an intrinsic property of the two lexical items, but a quality assigned by the situational and linguistic context. Consider ªrst the verb gehen (58): (58) Jetzt kann ich gehen. ‘Now I can go.’ / ‘Now I can walk.’

In this use there are no explicit spatial indications. The sentence has two different possibilities of interpretation: ‘now it is possible for me to leave my present location’ (deictic) vs. ‘now I have recovered the ability to move’ (nondeictic). It is the context that has a disambiguating function suggesting a

61

62

Claudio Di Meola

directional or a manner-focusing reading (59–60) respectively: (59) Ich habe alles erledigt. Jetzt kann ich gehen. ‘I’ve settled everything. Now I can go.’ (60) Eben hatte ich einen stechenden Schmerz im Bein. Jetzt kann ich gehen.

‘A minute ago I had a stabbing pain in the leg. Now I can walk.’ In the deictic use gehen is in opposition to kommen, in the non-deictic use it is in opposition to non-deictic motion verbs such as humpeln ‘to hobble’, rennen ‘to run’ etc. The gehen-examples containing an explicit SOURCE/PATH/ GOAL-indication can also be potentially ambiguous. It is the context which determines whether the movement is directed away from the observer or whether there is no observer at all (61): (61) Er geht aus dem Zimmer. ‘He leaves the room.’

Consider now the verb kommen. Sentences lacking an explicit spatial indication are unambiguous because they can only be interpreted in a deictic sense (e.g. er kommt ‘he comes’). On the other hand, sentences containing a spatial indication may always have two different readings (62–64): (62) Das Kind kam in den Garten. ‘The child came / got into the garden.’ (63) Sie kam an den Schreibtisch. ‘She came to / bumped into the desk.’ (64) Er kam ins Krankenhaus. ‘He came / was taken to the hospital.’

In the deictic use, kommen describes a movement approaching the observer; in the non-deictic use, it refers to the overcoming of an obstacle (62), to an accidental movement (63) or to a passive movement (64). The possible alternative interpretations diŸer in two important points. The ªrst diŸerence concerns the question of whether the prepositional phrase containing the spatial indication can be deleted or not. Consider example (64) again. If kommen has a deictic reading, the prepositional phrase is a free adverbial whose only function is to describe the spatial region around the observer more precisely (65): (65) Er kam hier ins Krankenhaus / auf die Krankenstation / auf das Krankenzimmer / ... . ‘He came here to the hospital / to the ward / into the hospital room / ... .’

Non-deictic uses of the deictic motion verbs kommen and gehen in German

If kommen has a non-deictic reading, however, the prepositional phrase becomes an obligatory complement indicating the conventional destination where the trajector is to be transported to. The second question is connected to the value of a possible time indication (66): (66) Um 10 Uhr kam er ins Krankenhaus. ‘At 10 o’clock he came to the hospital.’ [deictic] ‘At 10 o’clock he was admitted / was taken to the hospital.’ [non-deictic]

In the deictic reading of the verb kommen, the indication um 10 Uhr ‘at 10 o’clock’ refers to the arrival at the hospital (time at GOAL). In the non-deictic reading it can refer — depending on the context — to the time of arrival (time at GOAL) as well as to the time of departure (time at SOURCE).

9.

Summary

(1) Non-deictic uses of kommen and gehen are by no means isolated examples, but are a fundamental part of the semantic range of these highly polysemous lexical items. Kommen and gehen constitute a complex semantic network in which non-deictic uses can be derived from deictic ones and in which there are systematic connections between the non-deictic uses themselves. These links are primarily based on metonymical extensions and on a progressive deviation from the characteristics of the prototypical human movement. (2) Both verbs preserve their distinctiveness in the non-deictic sphere. The distinguishing factor, however, is not focus but restriction: gehen describes an unrestricted movement, kommen a restricted movement where the trajector has to confront an external, opposing force. Several variants can be distinguished for non-deictic kommen in accordance with the increasing importance of the external force (obstacle, accident, transport). The distinctiveness of kommen and gehen is preserved even in abstract uses that can be traced back to basic ones via conceptual metaphors. (3) Deixis is not an intrinsic characteristic of kommen and gehen, but is context-dependent: in many cases one and the same sentence can have both a deictic and a non-deictic reading.

63

64

Claudio Di Meola

Notes * I wish to thank Günter Radden for insightful comments on a previous version of this paper. 1. Cf. the contrastive and typological studies including, among others, the German language: Gathercole (1977, 1978), Schlyter (1979), Rauh (1981), Fleischman (1982), Radden (1988, 1989, 1995), Ricca (1991, 1993), Rösner (1993), Abraham (1994) and Roßdeutscher (1997). The studies of the English motion verbs come and go also focus on deictic uses: not only the pioneering work by Fillmore (1966, 1972, [1971] 1997), but also subsequent studies by other autors (e.g. Cinque 1972, Clark 1974 and Taylor 1988). Some recent studies at least emphasize the semantic complexity of the two verbs (Wilkins and Hill 1995, Goddard 1997). Some remarks on kommen and gehen can ªnally be found in studies concerning the categorization of space in general (e.g. Janzen and Katz 2000). 2. Cf. Baumgärtner (1967), Wotjak (1971), Diersch (1972) and Krohn (1975). 3. Beside general work on Funktionsverben (e.g. Helbig 1979) two speciªc studies of kommen should be mentioned: Heringer (1968) and Persson (1984).

References Abraham, Werner 1994 “Von ‘hinichen’ und honorablen Sprachen: Variationen, ihre Linguistiken und verbale Deixissysteme”. American Journal of Germanic Linguistics and Literatures 6: 211–250. Baumgärtner, Klaus 1967 “Die Struktur des Bedeutungsfeldes”. In Satz und Wort im heutigen Deutsch, H. Moser (ed.), 165–197 [Sprache der Gegenwart 1]. Düsseldorf: Schwann. Cinque, Guglielmo 1972 “Fillmore’s semantics of ‘come’ revisited”. Lingua e stile 7: 575–599. Clark, Eve V. 1974 “Normal states and evaluative viewpoints”. Language 50: 316–332. Di Meola, Claudio 1993 “Espressioni deittiche spaziali usate in contesti non deittici: I verbi di movimento ‘kommen’ e ‘gehen’ in tedesco”. Nuovi Annali della Facoltà di Magistero dell’Università di Messina 11: 109–140. 1994 ‘Kommen’ und ‘gehen’. Eine kognitiv-linguistische Untersuchung der Polysemie deiktischer Bewegungsverben [Linguistische Arbeiten 325]. Tübingen: Niemeyer. Diersch, Helga 1972 Verben der Fortbewegung in der deutschen Sprache der Gegenwart: Eine Untersuchung zu syntagmatischen und paradigmatischen Beziehungen des Wortinhalts. Berlin: Akademie.

Non-deictic uses of the deictic motion verbs kommen and gehen in German

Fillmore, Charles J. 1966 “Deictic categories in the semantics of ‘come’”. Foundations of Language 2: 219–227. 1972 “How to know whether you’re coming or going”. In Linguistik 1971: Referate des 6. Linguistischen Kolloquiums 11.-14. August 1971 in Kopenhagen, K. Hyldgaard-Jensen (ed.), 369-379. Frankfurt a.M.: Athenäum. 1997 Lectures on Deixis. Stanford: Center for the Study of Language and Information Publications. First published Bloomington: Indiana University Linguistics Club [1971]. Fleischman, Suzanne 1982 “The past and the future: Are they ‘coming’ or ‘going’?” Proceedings of the 8th Annual Meeting of the Berkeley Linguistics Society: 322–334. Gathercole, Virginia C. 1977 “A study of the ‘comings’ and ‘goings’ of the speakers of four languages: Spanish, Japanese, English, and Turkish”. Kansas Working Papers in Linguistics 2: 61–94. 1978 “Towards a universal for deictic verbs of motion”. Kansas Working Papers in Linguistics 3: 72–88. Goddard, Cliff 1997 “The semantics of coming and going”. Pragmatics 7: 147–162. Helbig, Gerhard 1979 “Probleme der Beschreibung von Funktionsverbgefügen im Deutschen”. Deutsch als Fremdsprache 16: 273–285. Heringer, Hans-Jürgen 1968 Die Opposition von ‘kommen’ und ‘bringen’ als Funktionsverben: Untersuchungen zur grammatischen Wertigkeit und Aktionsart. Düsseldorf: Schwann. Janzen, Gabriele and Steffi Katz 2000 “Die Wahrnehmung von Eigen- und Fremdbewegung und ihr Ausdruck in der Sprache”. In Räumliche Konzepte und sprachliche Strukturen, Ch. Habel and Ch. v. Stutterheim (eds), 45-67 [Linguistische Arbeiten 417]. Tübingen: Niemeyer. Johnson, Mark 1987 The Body in the Mind: The Bodily Basis of Meaning, Imagination, and Reason. Chicago: University of Chicago Press. Krohn, Dieter 1975 Verbinhalt und semantische Merkmale: Studien zu paradigmatischen und syntagmatischen Relationen im Bedeutungsfeld der menschlichen Fortbewegung im heutigen Deutsch und Schwedisch. Göteborg: Acta Universitatis Gothoburgensis. Lakoff, George 1987 Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. Chicago: University of Chicago Press. 1990 “The Invariance Hypothesis: is abstract reason based on image-schemas?” Cognitive Linguistics 1: 39-74.

65

66

Claudio Di Meola

Langacker, Ronald W. 1986 “Abstract Motion”. Proceedings of the 12th Annual Meeting of the Berkeley Linguistics Society: 455-471. 1987 Foundations of Cognitive Grammar, vol. 1, Theoretical Prerequisites. Stanford (CA): Stanford University Press. Persson, Ingemar 1984 “Zu Konstruktionen mit ‘kommen’ und ‘gelangen’ vom Typ ‘zur Darstellung kommen’, ‘zu einer Lösung kommen/gelangen’”. Deutsch als Fremdsprache 21: 23-28. Radden, Günter 1988 “The concept of motion”. In Understanding the Lexicon: Meaning, Sense and World Knowledge in Lexical Semantics, W. Hüllen and R. Schulze (eds ), 380394. Tübingen: Niemeyer. 1989 “Das Bewegungskonzept: ‘to come’ und ‘to go’”. In Raumkonzepte in Verstehensprozessen: Interdisziplinäre Beiträge zu Sprache und Raum, Ch. Habel, M. Herweg and K. Rehkämper (eds), 228-248 [Linguistische Arbeiten 233]. Tübingen: Niemeyer. 1995 “Motion metaphorized: the case of ‘coming’ and ‘going’”. In Cognitive Linguistics in the Redwoods: The Expansion of a New Paradigm in Linguistics, E.H. Casad (ed), 423-458. Berlin/New York: Mouton De Gruyter. Rauh, Gisa 1981 “On coming and going in English and German”. Papers and Studies in Contrastive Linguistics 13: 53-68. Ricca, Davide 1991 “‘Andare’ e ‘venire’ nelle lingue romanze e germaniche: dall’Aktionsart alla deissi”. Archivio Glottologico Italiano 76: 159-192. 1993 I verbi deittici di movimento in Europa: una ricerca interlinguistica. Firenze: La Nuova Italia. Roßdeutscher, Antje 1997 “Perspective and propositional attitude in the semantics of ‘kommen’ (‘to come’)”. Paper presented at the 19th DGfS-Jahrestagung (Düsseldorf, February 26-28, 1997). Rösner, Jutta 1993 “Kommen und gehen im Deutschen und Französischen”. InStudien zum romanisch-deutschen Sprachvergleich, G. Rovere and G. Wotjak (eds), 215228 [Linguistische Arbeiten 297]. Tübingen: Niemeyer. Schlyter, Suzanne 1979 “Le verbe allemand ‘kommen’ et ses correspondances en français”. Moderna Språk 73: 251-267. Talmy, Leonard 1988 “Force dynamics in language and cognition”. Cognitive Semantics 2: 49-100. Taylor, Kenneth A. 1988 “We’ve got you coming and going”. Linguistics and Philosophy 11: 493-513.

Non-deictic uses of the deictic motion verbs kommen and gehen in German

Wilkins, David P. and Deborah Hill 1995 “When ‘go’ means ‘come’: Questioning the basicness of basic motion verbs”. Cognitive Linguistics 6: 209-259. Wotjak, Gerd 1971 Untersuchungen zur Struktur der Bedeutung. Hueber: München.

67

Origo, pointing, and conceptualization — what gestures reveal about the nature of the origo in face-to-face interaction Ellen Fricke Technische Universität Berlin

1.

Introduction

For understanding to be successful between communicating partners in face-toface interaction, both the speaker and the addressee need to be simultaneously engaged in cognitive processes within which they must be similarly oriented. Deixis assumes a special function in the coordination of cognitive representation: it can be understood as a communicative procedure in which the speaker focuses the attention of the addressee by means of verbal expression and gesture. Bühler (1982a: 28 and 107) understands deictics as signals with an appellative function that control the attention of the addressee’s inner and outer behavior. Ehlich (1982: 325) deªnes deixis as follows: “The deictic procedure is a linguistic instrument for achieving focusing of the hearer’s attention towards a speciªc item which is part of the respective deictic space (deiktischer Raum). The deictic procedure is performed by means of deictic expressions.” In addition, understanding can be only successful if the origos of speaker and addressee are coordinated with the respective cognitive representations. What is meant by origo? How are origos created? How are they structured?1The considerations in this paper aim to outline aspects of a plausible deixis-theoretical concept of origo by redeªning origo as a complex entity. This is to be achieved, on the one hand, by means of conceptual re¶ection and, on the other hand, by including concrete observation data for describing routes on Potsdamer Platz [Potsdam Square] in Berlin. The following questions are raised: – –

Is the origo a concrete or an abstract entity? Is it ªxed with respect to the speaker, or moveable?

70

Ellen Fricke

– – –

Is there more than one origo? If there are several origos, how are they related? Are they positioned on the same level or are they arranged in a hierarchy? Are origos within communication simply provided or are they, for example, actively placed by the speaker?

In what follows, “gestures” are deªned as “speech-associated gestures” or “cospeech gestures”. In accordance with Kendon (1980) and McNeill (1985, 1992, 2000), speech-associated gestures are spontaneous movements of the hands and arms that can be observed when someone speaks. They occur only during speech and are closely linked to the activity of speaking and the patterns of intonation. Gestures and speech are semantically and pragmatically coexpressive and have a constant relationship in time. That is, the gestures that accompany speech also present the same or closely related meanings semantically and perform the same functions pragmatically. According to Kendon and McNeill, such linkages imply that gestures should be viewed as manifestations of the same process of utterance. Gestures and speech are based on the same cognitive processes that access propositional and image-like representation at the same time. Above all, gestures reveal the speaker’s image-like representation during speech production, which to a certain extent could be understood as a “mental model” (cf. Johnson-Laird 1983) of the situation described. The investigation of speech-associated gestures provides an alternative empirical access to spatial cognition. The framework of cognitive psychology based on reaction-time analyses (e.g. the pattern of the availability of objects in working memory) includes characteristics of image-like representations (cf. Black, Turner, and Bower 1979; Bryant, Tversky, and Franklin 1992; Franklin, Tversky, and Coon 1992; Hörnig et al. 1996; Hörnig and Wiebrock 2000). The investigation of gestures complements this idea to a certain extent: aspects of image-like representation are inferred from observation data produced by video recordings of communication situations. It is assumed that aspects of these representations are projected into the three-dimensional gestural space and are, thus, accessible for direct observation (McNeill 1985, 1992). This empirical access reveals new aspects of interaction between verbal deictics, pointing, and the process of origo allocation (see Section 4).

Origo, pointing, and conceptualization

2.

The concept of origo according to Bühler

Bühler is generally accepted as the founder of modern deixis theory. Some later approaches have lost sight of distinctions introduced by Bühler, while other approaches have maintained conceptual uncertainties in Bühler’s framework. This is shown in the following section, where the problematic diŸerence between “deictic” and “intrinsic” (Miller and Johnson Laird 1976) is discussed. I shall ªrst outline the fundamental ideas that Bühler connects with the concept of origo. The term origo is introduced by Bühler (1934) within the framework of his Sprachtheorie and is conceptually understood as the origin of a coordinate system. He calls it “das Koordinatensystem der ‘subjektiven Orientierung’” [the coordinate system of ‘subjective orientation’] (Bühler [1934] 1982a: 102), which is used to organize the personal, spatial, and temporal structure of utterances. Let two perpendicularly intersecting lines on the paper suggest a coordinate system to us, 0 for the origin, the coordinate source: […] My claim is that if this arrangement is to represent the deictic ªeld of human language, three deictic words must be placed where the 0 is, namely the deictic words here, now and I. (Bühler [1934] 1990: 117)

In contrast to “symbols” or “naming words”, according to Bühler, “deictic words” or “pointing words” are characterized by the fact that they are only interpretable by recourse to an origo, which in the default-case is connected with the speaker (see the discussion in Section 4). Take the following situation for example: Anna and Beate do not only share an o¹ce, but also a waste-paper basket. Caroline comes to the door and asks Where is the letter? Two possible answers are: (1) a. The letter is in the waste-paper basket. b. The letter is there. (+ pointing gesture)

Answer (1b) is indisputably deictic whereas answer (1a) is not. Why? If the answer comes in the form of sentence (1a), then the referent of the noun phrase the letter can be found without great di¹culty using the prepositional phrase in the waste-paper basket. The knowledge about who utters this sentence is irrelevant. However, if the answer takes the form of sentence (1b), then this is not possible without information about the situational context. In sentence (1b) the prepositional phrase in the waste-paper basket is substituted by the

71

72

Ellen Fricke

local adverb there. In opposition to here, there refers to a distant area in relation to the speaker. Assuming that Anna and Beate were not sitting together but apart from each other on opposite sides of the room Caroline would begin to look for the letter in a diŸerent part of the room, depending on who utters sentence (1b). Now, information about who uttered this sentence and from where is not su¹cient to ªnd the letter in question. The utterance of there is obligatorily accompanied by a direction-indicating gesture: an extended arm, hand, or index ªnger, a glance or any other movement of a similar function. It is as if a straight line were drawn between two points, the tip of the indicating index ªnger and the point where the speaker is located, the origo. In order to ªnd the letter, attention must be focused on the direction of this straight line. Depending on who utters this sentence and performs the pointing gesture, Anna or Beate, and depending on where in the room the speaker is, the extension of the straight line would lead to diŸerent target points in the room. This dependency, observable in all direction-indicating gestures, is the reason why Bühler (and others) assume a similar feature for verbal elements. As tokens of the same type, these verbal elements refer to diŸerent situational context elements, depending on when, where and by whom they were uttered. As will be demonstrated below, Bühler’s deªnition of origo is not completely adequate to the phenomena associated with deixis, but it takes an important aspect into account, namely that the origo is under no circumstances to be identiªed with a concrete component of the situation. One argument against a concrete, physically deªned origo is Bühler’s idea of a shifting origo that resembles a mathematical volumeless point. A physical point cannot be shifted mentally. If it is possible to show that the concept of a shiftable origo is more adequate than the concept of an origo ªxed to the speaker, then the idea of a concrete origo is simultaneously rejected.

3. Is the origo ªxed or movable? The problem of distinguishing between deictic and intrinsic according to Miller and Johnson-Laird The assumption of a shiftable origo is not limited to Bühler alone and is widely accepted by later approaches. Nevertheless, the dichotomy of “deictic” versus “intrinsic” as proposed by Miller and Johnson-Laird (1976) is a common distinction in deixis theory although it is not compatible with a shiftable origo. I will show that if the deictic is categorically distinguished from the intrinsic

Origo, pointing, and conceptualization

this leads to inherent contradictions. Second, I will show an example which cannot be handled by Miller and Johnson-Laird’s dichotomy, but can be captured by Bühler’s concept of Versetzung [displacement].2 The concept of the intrinsic (without the actual use of this term) can be found in Bühler, who, however, deals with the intrinsic as part of deixis: […] we may view the important case, for example, of a vehicle (carriage, ship, locomotive, car) where one’s orientation immediately and not only conceptually, but of necessity perceptually, follows the conventional direction of movement of the object. Just as naturally as with animals and other humans. When a teacher of gymnastics facing a dressed line of gymnasts gives commands, the orders left and right are conventionally given and understood according to the gymnasts’ orientation. That is a paradigmatic case for whose explanation one must note the astonishingly easy translatability of all ªeld values of the visual system and the verbal deictic system from someone in another plane of orientation. (Bühler [1934] 1982b: 26 and 27; cf. 1982a: 131)

Bühler’s concept allows the shifting of the origo to other people, creatures and objects. In contrast, according to Miller and Johnson-Laird (1976) the origo can, by deªnition, only be located where the speaker is. Therefore, they exclude phenomena as non-deictic that fall under Bühler’s wider concept of deixis. Applied to local deixis and only this is considered the origo is provided via the visual orientation or perspective of the speaker. We will call the linguistic system for talking about space relative to a speaker’s egocentric origin and coordinate axes the deictic system. We will contrast the deictic system with the intrinsic system, where spatial terms are interpreted relative to coordinate axes derived from intrinsic parts of the referent itself. Another way to phrase this distinction is to say that in the deictic system spatial terms are interpreted relative to intrinsic parts of ego, whereas in the intrinsic system they are interpreted to intrinsic parts of something else. (Miller and Johnson-Laird 1976: 396)

Using the above quote as a starting point, the deictic system and the intrinsic system can be characterized in accordance with Miller and Johnson-Laird as follows: if a deictic system is present, then the speaker makes himself or his visual orientation a reference point for a linguistic localization. However, if an intrinsic system is present, then the reference point lies with an object that is not the speaker and is derived from the inherent features of the object. The object must have a clear front and back so that, to a certain extent, a spatial coordinate system can be constructed and ªxed. Examples of such objects are vehicles, chairs, and desks. Examples of non-intrinsic objects are balls, bushes,

73

74

Ellen Fricke

and single columns. According to this view, in linguistic localizations nonintrinsic objects are only used from a deictic perspective. Intrinsic objects can but do not have to be localized in intrinsic perspectives. This means that their intrinsic orientation can be ignored from a deictic perspective: if, for example, with reference to the inherent orientation of a car, a pair of of pliers lies behind it, a person may say: “From where I’m standing the pliers are to the left of the car.” (cf. Herrmann and Schweitzer 1998: 49). Sententces (2.1) and (2.2) are examples for a deictic localization whereas sentence (2.3) is an example for an intrinsic one: Deictic localization (2.1) The pliers are in front of me (2.2) The pliers are to the left of the car. Intrinsic localization (2.3) The pliers are behind the car.

speaker

Figure 1. Deictic and intrinsic localization

The deªnition given by Miller and Johnson-Laird is not convincing because the concept of the deictic system is not clearly distinguished from the concept of the intrinsic system but is rather derived from it. The opposition exists between the reference to the intrinsic coordinates of the speaker (deictic system), on the one hand, and the reference to the intrinsic coordinates of something else (intrinsic system), on the other hand. In other words: the actual opposition exists between the intrinsic systems of speaker and non-speaker.3 Indeed, cases can be constructed which can be explained with the help of Bühler’s concept of displacement but not with the diŸerentiation between deictic system and intrinsic system, as proposed by Miller and Johnson-Laird. Let us consider a chase scene in a suspense ªlm. Person A is directing person B through the town via mobile telephone. The following short dialogue ensues:

Origo, pointing, and conceptualization

(3) A: Where are you? Have you shaken them oŸ? B1: I am standing behind an advertising column. B2: I am standing in front of an advertising column.

The crucial point is the alternative answer from B. The sentences B1 and B2 are diŸerent solely because in answer B2 the preposition in front of is used instead of behind. Considering the answer B1, it is clear that the point of reference can neither be with the advertising column nor with the speaker. Why not? It cannot be with the advertising column because it has no intrinsic front or back. It is radially symmetrical and is, therefore, a non-intrinsic object as according to Miller and Johnson-Laird. The point of reference cannot be with the speaker because otherwise he would use the preposition in front of instead of behind as in answer B2. The point of reference in answer B1 is with the people searching and not with the speaker. In other words, it is in a third area that is not accounted for by Miller and Johnson-Laird.4 This area, however, can be covered by Bühler’s concept of displacement. According to Bühler, the origo can be shifted to other people, animals, or objects regardless of whether these are perceived or imagined. Consequently, the concept of origo, which starts with an abstract, shiftable origo, is clearly more adequate and is preferable to the concept of an origo ªxed to the speaker.

4.

Is there more than one origo?

If we turn again to the quote introduced at the beginning of this article, Bühler (1990: 117) seems to assume one single origo for all dimensions, a mutual coordinate starting point for personal, local, and temporal deixis. The question arises whether the assumption of one origo is su¹cient.5 If we could show that even when limited to the local dimension the assumption of a single origo is not adequate, then the concept of a single origo would be intra- as well as interdimensionally refuted. Aspects of Theo Herrmann’s model of local deixis (6H-Modell [six main-variants model]) provide the basis for further discussion. 4.1 Theo Herrmann’s model of local deixis (1990/1998) Herrmann’s 6H-model circumvents the listed disadvantages of the deicticintrinsic-dichotomy. Its features fulªll the requirements I have argued for. With reference to local deixis, the 6H-model is derived from a reference system that

75

76

Ellen Fricke

is analogous to a visual perspective. The abstract origo at the center can be instaniated by all kinds of entities, perceived or imagined, be it the speaker himself or other people, creatures, and objects that are not identical with the speaker. Let us consider the situation portrayed in Figure 2 and assume the speaker wants to inform the addressee in which spatial relation the pliers are located. Otto addressee

speaker

Figure 2. The connection of the intended object to diŸerent relatum objects (cf. Herrmann and Schweizer 1998: 43)

From the conditions shown in the illustration, he can produce at least the following equally appropriate utterances (cf. Herrmann and Schweizer 1998: 51): Three-point localizations: (4.1) The pliers are to the left of the car. (origo: speaker, relatum: car, intended object: pliers) (4.2) The pliers are in front of the car. (ORIGO: addressee, relatum: car, intended object: pliers) (4.3) The pliers are to the right of the car. (ORIGO: Otto, relatum: car, intended object: pliers) Two-point localizations: (4.4) The pliers are in front of me. (origo: speaker, relatum: speaker, intended object: pliers) (4.5) The pliers are in front of you. (origo: addressee, relatum: addressee, intended object: pliers) (4.6.a) The pliers are behind the car. (ORIGO: car, RELATUM: car, INTENDED OBJECT: pliers) (4.6.b) The pliers are in front of Otto. (origo: Otto, relatum: Otto, intended object: pliers)

Origo, pointing, and conceptualization

The object to be located, the pliers, is called the intended object and is the same in all utterances. This intended object is located with respect to a reference object, the relatum, and from a viewpoint, the origo. Relatum and origo can be instantiated by diŸerent entities, in this case either by the speaker, the listener, a third person named Otto, or the car. The utterances can be divided into two large groups, namely three-point localizations and two-point localizations. It depends on whether there are three diŸerent entities or only two which instantiate the position of the origo, the relatum, and the intended object. Utterances (4.1) to (4.3) are examples of three-point localization. Utterances (4.4) to (4.6) are examples of two-point localization. These six main variants of Herrmann’s 6H-model result from variation in the assignment of the origo to the speaker, to the addressee, or to a third party. 4.2 Intradimensional origos Bühler distinguishes between three dimensions of deixis, namely personal, local, and temporal deixis. Intradimensional origos are limited to one deictic dimension, in this case the local one. The question is: Does more than one origo exist within the local dimension? I will give an example that illustrates the phenomenon of two non-identical origos on the gestural and verbal level in German local deixis 4.2.1 The phenomenon of two non-identical origos on the gestural and verbal level in the dimension of German local deixis The following examples are taken from video recordings of route descriptions at Potsdamer Platz in Berlin. The recordings were made in December 2000 in Berlin. 33 informants, predominantly ªrst year students, participated in the data collection. There were three informant groups. Accompanied by the experiment leader, the members of informant group A each individually followed a certain pre-ªxed route on Potsdamer Platz, which started at the Potsdamer Platz underground station and ended at the Sony Center. Each informant was instructed to describe this route so precisely to an unknowing informant from group B that the latter would be in a position to describe the route, in turn, to a third person from informant group C, who should then be able to follow the route independently. In the situation in which the example sentences are elicited, both communication partners are standing at the Infobox-exit of the Potsdamer Platz underground station, the starting point of the route. Certain sections of the

77

78

Ellen Fricke

route are accessible to the mutual perception of speaker and addressee and can, therefore, be identiªed via demonstratio ad oculos. Others, however, can only be communicated by means of imagined space, thus, via imagination-oriented deixis. Let us now consider more closely the following conversation section and the conceptual spatial relation LEFT OF verbalized as links von dir [on your left] in German. The beginning and end of the pointing gesture is marked by square brackets. Other speech-associated gestures that appear are not considered in this context. (5) Dann soll irgendwann links ein Arka/ ein Eingang zum Arkadenzentrum kommen oder so was. Ich kenns leider nich, aber es soll dann also [links von dir] irgendwann Arkaden, zwischen den Häusern irgendwann stehen und da sollst du reingehen, nach links, ja? ‘Then a bit later on the left there should be an ent/ an entrance to the mall or something like that. I don’t know it, unfortunately, but you should see a mall some time later [on your left] between the buildings and that’s where you have to go in, to the left, okay?’

Caroline

Beate

Figure 3. Execution of the pointing gesture for the conceptual relation LEFT OF in Example (5)

The speaker belongs to informant group B. That means she has not followed the route on Potsdamer Platz herself. She tries to relay the route description given to her by a person from informant group A to a third person from informant group C as precisely as possible. She assumes that as a result of the conversation her addressee will then follow the route making as few mistakes as possible. Let us now analyse the utterances according to Herrmann’s model. The following table summarizes the results of the instantiation of the origo, the relatum, and the intended object in Example (5):

Origo, pointing, and conceptualization

Table 1. The conceptual relation LEFT OF in Example (5) according to Herrmann’s model LEFT OF (links von dir [on your left])

Origo (diŸers)

Relatum (diŸers)

Intended object (the same)

Verbal level Gestural level

addressee speaker

addressee speaker

mall mall

The intended object on the verbal and gestural level is the same, namely the entrance to the glass-covered shopping mall. On the verbal level the speaker localizes the mall with reference to the intrinsic coordinates of the addressee, who instantiates the origo as well as the relatum. However, if we include the gestural level in our analysis it is evident that, in contrast to the verbal localization, the speaker carries out the gestural localization of the intended object referring to her own current intrinsic coordinates. We can observe that while the intended objects at the verbal and gestural level are the same, the gestural and verbal origos and relata diŸer. A further similar example from the same conversation: (6) Und dann soll dort ein großer Platz sein, und [ rechts von dir] ist dies Stella-Musicaltheater/ähm/ der Eingang glaub ich vom/ also dieses Theater auf jeden Fall, und [links von dir] ist also viel Wasser und eine Brücke. ‘And then there should be a large square and [on your right] the Stella musical theater/umm/ the entrance I think of/ that theater at any rate and [to the left] there is a lot of water and a bridge.’

Caroline

Beate

Figure 4. Execution of the demonstrative gesture for the conceptual relation RIGHT OF in Example (6)

79

80

Ellen Fricke

Table 2. The conceptual relation RIGHT OF in Example (6) according to Herrmann’s model RIGHT OF (rechts von dir [on your right])

Origo (diŸers)

Relatum (diŸers)

Intended object (the same)

Verbal level Gestural level

addressee speaker

addressee speaker

theater theater

Caroline

Beate

Figure 5. Execution of the demonstrative gesture for the conceptual relation LEFT OF in Example (6) Table 3. The conceptual relation LEFT OF in Example (5) according to Herrmann’s model LEFT OF (links von dir [on your left])

Origo (diŸers)

Relatum (diŸers)

Intended object (the same)

Verbal level Gestural level

addressee speaker

addressee speaker

water and bridge water and bridge

In both examples it can be observed again that while the intended objects are the same, the gestural and verbal origos and relata diŸer. How is this ªnding to be interpreted? 4.2.2 Bühler’s three main cases of imagination-oriented deixis For the verbal and gestural levels in Examples (5) and (6) a demonstratio ad oculos can be excluded. On the verbal level an imagined addressee instantiates the origo as well as the relatum. Although there is a real perceptible demonstrative action on the gestural level, the intended object is located in the imagined space and not in the perceptual space. According to Bühler this should be considered a case of imagination-oriented deixis on both levels. Within this mode, Bühler distinguishes between three diŸerent main cases. The ªrst main case of imagination-oriented deixis, is characterized by the

Origo, pointing, and conceptualization

fact that the intended object is localized as an imaginary object within the actual perceptual space of speaker and addressee. The origo is instantiated by the speaker. “What is imagined, especially when movable things such as people are concerned, often comes to us, that is, into the given order of actual perception, within which it can be localized, though not quite ‘seen’.” (Bühler 1990: 150; cf. 1982a: 134) In the second main case, however, the given order of actual perception is eliminated. The origo is displaced to new positions within perceptual space or imagined space in the sense of Langacker’s (1987: 131) distinction between “mental transfer” and “cross-world identiªcation”. Thus, the uttered verbal deictics are not interpreted in relation to the current orientation of the speaker but rather in relation to another grounding system that is not the actual speaker himself (cf. Bühler 1982a: 135). The third main case is characterized by “the fact that the experiencer is in a position to indicate with his ªnger the direction in which something which is absent is seen from the mental eye” (Bühler 1982b: 29; cf. 1982a: 135). Bühler oŸers the following example from his teachings in Vienna: “If I ask, for example, the 500 students in my lecture ‘Where is Saint Steven’s Cathedral?’ about 300 index ªngers will be raised and point (with all kinds of interesting deviations) somewhere within the lecture hall.” (Bühler 1982b: 29; cf. 1982a: 135) This example illustrates what Bühler deªnes as the third case: neither does the speaker shift (as in the second main case) nor is the intended object localized as an imaginary object within the actual perceptual space of speaker and listener (as in the ªrst main case). The diŸerence to ocular demonstration is simply that the intended object Saint Steven’s Cathedral is not accessible to the visual perception of those present in the lecture room. If we transfer Bühler’s division to our examples, then for the gestural level only the ªrst main case can be appropiate, because the pointing gesture is carried out in relation to the intrinsic coordinates of the actual speaker and localizes the mall, for example, as an imaginary object. The purely verbal utterance fulªlls the criteria of the second main case: the origo is assumed by the imagined addressee, who is to follow the described imagined route after the current conversation. Thus, the circumstance of a divided origo or of two nonidentical origos is maintained. 4.2.3 McNeill’s model of gestures as indicators of certain perspectives Is our observation of a divided origo compatible with the characteristics of gestures interpreted as indicators of certain perspectives? According to

81

82

Ellen Fricke

McNeill (1985, 1992), the type of gesture execution can be used to determine whether the speaker assumes a protagonist’s viewpoint (“character voice”) or an observer’s viewpoint (“observer voice”), whereby the observer can be positioned inside or outside the described situation. If the role of the protagonist is assumed, the body of the speaker or parts of the body act like the protagonist in the described situation. The speaker behaves, to a certain extent, like an actor who imitates protagonists and plays a role in the conversation. At the same time, the speaker uses space when in the role of a protagonist: “With the character voice the space envelopes the narrator — it is a space for the enactment of the character, and includes the locus of the speaker at its center.” (McNeill 1992: 190) The role of the observer includes other body movements and a diŸerent use of space: “The narrator’s body is an onlooker […]”(McNeill 1992: 190). “With an observer’s voice, in contrast, the narrative space is localized in front of the narrator — as if it were an imaginary stage or screen […].”(McNeill 1992: 190) As an observer the speaker can be situated outside as well as inside the described situation. If he is situated inside the described situation, he adopts a position and perspective within the space in which the protagonist of the described situation acts. The speaker does this, however, without using hand and body movements of the protagonist. To a certain extent, the speaker imagines himself to be a “silent observer” in a ªctional situation without acting out a role himself. Let us transfer the ªrst possibility to our example: does the speaker place herself in the shoes of the imaginary addressee? With her pointing gesture to the left, the speaker does not act like the imagined addressee, who follows the route on Potsdamer Platz. Instead, the speaker acts like a passer-by who provides directional information. This behavior provides an argument against a protagonist’s viewpoint. Thus, it remains the perspective of the observer. The speaker localizes the intended object in relation to her own body coordinates, thus, making herself the origo as an intrinsically oriented object. This provides an argument against an observer’s perspective outside the described situation. An observer’s perspective outside the situation produces the expectation that, for example, one hand represents the “imaginary wanderer”, and in relation to this and its coordinates the other hand localizes the intended object (cf. Example (7) of this article). Most of our observations are compatible with the ªnding that the speaker adopts the role of an inside observer. The speaker integrates herself into the imagined space, assumes the origo herself and localizes the deictic object by

Origo, pointing, and conceptualization

abstract pointing within the gestural space. There is, consequently, no shift to the addressee on the gestural level. Thus, within the local dimension including the speech-associated gestures, at least two diŸerent origos coexist. In my opinion this division can be explained by two complementary achievements attained by the verbal and gestural levels: with iconic and deictic speech-associated gestures the speaker generates a temporary perceptible model of the described situation in the shared perceptual space of the speaker and addressee, in this case a model of Potsdamer Platz. The imagined addressee moves within this model, walks around in it and is oriented in it on the verbal level by means of verbal deictics. The aspect of the model formation is that which Bühler (1982a: 139) considers an important characteristic of imagination-oriented deixis (cf. the quote in Section 5 of this article). The assignment of an inside observer viewpoint to the gestural behavior is compatible to the assignment of Bühler’s ªrst main case. Thus, the contradiction of two diŸerent origos on the gestural and verbal level is not resolved so far. 4.3 Interdimensional origos The fact that not only several intra- but also inter-dimensional origos are present is shown by the following example taken from the ªrst conversation of the same series of dialogues. Anna belongs to informant group A, which follows the route together with the experiment leader. That means Anna can recall the route and in the conversation tries to explain the route as precisely as possible to Beate, who as a member of informant group B does not know the route. The route leads along the back of the Stella musical theater. Earlier on in the conversation, the theater had already been localized in the center of the gestural space such that Beate is facing the entrance. This position is maintained in the following directions given by Anna: (7) Beate: Also [ich bin hinter dem Theater langgelaufen]. ‘Beate: So then [I have walked behind the theater]’. (8) Anna: [Genau du bist hinter dem Theater lang]. ‘Anna: [Right, you have walked behind the theater]’.

The beginning and end of the gesture in relation to speech are marked by square brackets. The bold lettering indicates the climax of the gesture, the socalled “stroke”. In Beate’s utterance the stroke is preceded by a preparation phase (“preparation”) and followed by a stop phase (“post stroke hold”) (cf. McNeill 1992: 25). At the climax of both gestural utterances, the hands draw a

83

84

Ellen Fricke

line along the back of the theater. It seems to be possible that not only the path behind the theater is localized and depicted but also the protagonist herself as she walks along the back of the theater.

Beate

Anna

Figure 6. Execution of the demonstrative gestures for the conceptual relation BEHIND in Examples (7) and (8)

Which origos can now be assumed for the three classic dimensions, the personal, the local, and the temporal? The origo of the personal dimension lies indisputably with the actual speaker. Therefore, during Beate’s utterance the origo is with Beate and during Anna’s utterance, with Anna. The local dimension turns out to be more complex if we again consider the verbal and gestural levels separately. In both utterances on the verbal level the origo as well as the relatum of the conceptual relation BEHIND is instantiated by the imagined theater, which has an intrinsic orientation. On the gestural level, the origo is instantiated by each speaker. The intended object is the imaginary wanderer. The speaker localizes the imaginary wanderer in relation to the musical theater using a pointing gesture. The following analysis reveals the similarities between both utterances: Ad 7: Beate: Also [ich bin hinter demTheater langgelaufen]. ‘Beate: So then [I have walked behind the theater]‘. Local dimension: Spatial relation: BEHIND Verbal level: Deictic expression: hinter Intended object: imaginary wanderer, identiªed with Beate (= speaker) Relatum: musical theater (imagined space) ORIGO: musical theater (imagined space)

Origo, pointing, and conceptualization

Gestural level: Intended object: imaginary wanderer, identiªed with Beate (= speaker) Relatum: musical theater (imagined space) ORIGO: speaker (perceptual space) Ad 8: Anna: [Genau du bist hinter dem Theater lang]. ‘Anna: [Right, you have walked behind the theater]’. Local dimension: Spatial relation: BEHIND Verbal level: Deictic expression: hinter Intended object: imaginary wanderer, identiªed with Beate (= addressee) Relatum: musical theater (imagined space) Origo: musical theater (imagined space) Gestural level: Intended object: imaginary wanderer, identiªed with Beate (= addressee) Relatum: musical theater (imagined space) Origo: speaker (perceptual space)

In the temporal dimension, the origo is instantiated by the imaginary wanderer’s current spatial and temporal location. Why? Considering the examples, it becomes clear that the German sentences are expressed in the present perfect tense although the situation described is situated in the future. The “event time” of the described action precedes the current “reference time” which instantiates the temporal origo.6 The temporal origo is coupled with the imaginary wanderer and moves with him and his space-time-coordinates. This is possible because time, in general and in the case of this utterance, can be reconstructed as a dimension of the route followed.7 The reference time is not identical to the utterance time. For this reason, the temporal origo is not held by the current speaker and the time of her utterance but lies rather with the introduced wanderer, whose future actions are anticipated by the speaker. Let us summarize once again the results for the instantiation of the origos of the three dimensions: 1. Personal origo: 2. Local origo:

speaker (perceptual space) verbal: musical theater (imagined space) gestural: speaker (perceptual space) 3. Temporal origo: imaginary wanderer (imagined space)

85

86

Ellen Fricke

Our analysis shows that one single origo is not su¹cient for all dimensions. Consequently, we have to assume not only diŸerent intradimensional but also diŸerent interdimensional origos.

5.

What relations exist between the diŸerent origos?

If we are confronted with several origos, we need to ask which relations exist between them: – – – –

Are the origos arranged in a hierarchy between the dimensions (personal, local, temporal)? Are origos arranged in a hierarchy within a single dimension? How does the origo become allocated? Are origos in communication simply given passively or are they actively placed by the speaker?

Considering the observations made so far, it seems reasonable to distinguish between the following: – – –

a local origo as an abstract volumeless point, a concrete entity in the form of an intrinsically oriented object that instantiates this origo, and an agent who carries out the origo allocation.

If we assume an origo that is not ªxed to the speaker but shifts onto other people and objects, the origo allocation is not achieved by the simple attainment of the speaker-role while turn-taking. When origos are allocated, it can be presumed that there is an agent who carries out origo allocation and instantiation. Talking to somebody, the speaker acquires the speaker-role and with it, the right to allocate local origos or to provide the local origo with intrinsically oriented entities. Such an entity can also be the speaker himself. Therefore, it is important to distinguish between two diŸerent things: 1. The speaker who, in his role as speaker and as holder of a primary origo, allocates the secondary origos. 2. The speaker who, as an intrinsically arranged entity, instantiates a secondary origo. If we assume that the function of the origo allocation is connected with the speakerrole, then the personal dimension is the highest dimension in the hierarchy.8 Thus, the right to allocate origos changes with the communicative role.

Origo, pointing, and conceptualization

The distinction between primary and secondary origo eliminates the inherent semantic contradictions of verbal deictics that arise when the speaker is assumed to shift his origo to other entities. The semantic opposition between the meanings of I and you is determined by the opposition of the semantic features “origo-inclusive” versus “origo-exclusive” (cf. Diewald 1991: 212). I refers to the person who adopts the origo-inclusive role and, therefore, to the speaker himself since the origo lies with the speaker. You refers to the person who adopts the origo-exclusive role and, thus, to the listener or addressee. But how can the shifting of the origo to the addressee be explained without giving up the validity of the semantic opposition between the meanings of I and you? The following solution is proposed: we assume the strict validity of the semantic opposition between the meanings of I [+ origo-inclusive] and you [– origo-inclusive] in relation to the primary origo of the personal dimension, which is attained with the speaker-role. Although you has the semantic feature [– origo-inclusive] in relation to the primary origo of the personal dimension, the speaker, as bearer of the primary origo, can allocate a secondary local origo to the intrinsically oriented person of the addressee called you. Thus, although the addressee instantiates a secondary origo, the validity of the semantic opposition “origo-inclusive” versus “origo-exclusive” is maintained on the level of the primary origo.

6.

The origo-allocating act

Considering the dimension of local deixis, how can the phenomenon of two non-identical origos as demonstrated in Example (5) be explained? The following diagram illustrates the function of origo allocation as outlined in the previous section. COMMUNICATORS (communicative roles)

OBJECTS OF COMMUNICATION (entities)

ADDRESSEE

SPEAKER (primary origo)

origoallocating act

speaker (secondary origo, gestural level) sign-relation (iconicity) imaginary addressee (secondary origo, verbal level)

Figure 7. The origo-allocating act in Example (5)

87

88

Ellen Fricke

On the one hand, we have the communicators in their roles as speaker and addressee. On the other hand, we have the object of communication, those entities that are talked about. In the communication process, the communicators alternately assume the roles of the speaker and the addressee. With the role of the speaker, the primary origo is acquired and, thereby, the possibility to intentionally allocate secondary origos to intrinsically arranged entities, be these perceptual or imaginary. These intrinsically oriented entities can be any objects, people, and creatures, even the addressee and the speaker, in so far as they are objects of communication that are referred to in speech. With the hierarchical organization of the origos as primary and secondary, the discrepancy of the local origos on the verbal and gestural levels in Example (5) becomes clear: the speaker (primary origo) allocates a secondary local origo to the imaginary addressee on the verbal level. At the same time, on the gestural level she allocates a secondary local origo to her own body as an intrinsically oriented entity. Since the body of the speaker is analogous to that of the imaginary addressee, the speaker allows herself to be understood as a model that represents the imaginary addressee to the speaker as well as to the addressee present in the actual communication. Thus, on the gestural level the speaker does not shift herself to the addressee but rather instantiates the origo by her own body, which functions as an iconic sign of the imaginary wanderer. In other words: there is a mapping between the body of the speaker and the imaginary wanderer via sign relation. Upon initial consideration, this interpretation seems to correspond with what Klein (1978: 29) describes as “analogische Deixis” [analogical deixis]. The introduction of the concept “analogical deixis” is, however, based on a misconception of the essence of Bühler’s imagination-oriented deixis. Klein (1978: 29) considers the example Die Kugel traf ihn hier [The bullet hit him here]. During the utterance, the speaker points to his right shoulder. According to Klein, the body of the speaker in this utterance depicts by analogy the body of the person referred to in speech. Klein (1978: 39) presumes that, expressed in Bühler’s categories, this is a case of demonstratio ad oculos.9 This is an incorrect presumption. Bühler (1934: 139) actually considers the availability or the construction of an analogon as an important characteristic of the ªrst main case of imagination-oriented deixis: ‘Here I was — he was there — the brook is there’: the narrator begins thus with indicative gestures, and the stage is ready, the present space is transformed into a stage. We paper-bound people will take a pencil in hand on such occasions and

Origo, pointing, and conceptualization

sketch the situation with a few lines. For Example, I want to give a visual account of the course of the decisive battle between Caesar and Pompey with deictic clues as Plutarch describes it, and make a line drawing: ‘this is Caesar’s line of battle — here is the tenth legion — here is the cavalry — here he is himself. This is Pompey’s battle line’; and so on. A psychological study of the most elementary imagination-oriented deixis with linguistic resources requires that we proceed from that sort of basis. If there is no surface to draw a sketch on, then an animated speaker can temporarily ‘transform’ his own body with two outstreched arms into the pattern of the battle line. (Bühler 1990: 156; cf. 1982a: 139)

Sennholz (1985: 87), however, believes that Bühler’s quote ‘Here I was — he was there — the brook is there’ is wrongly interpreted as imagination-oriented deixis. In his view, it is a case of analogical deixis and, therefore, a demonstratio ad oculos.10 According to Sennholz, this is a demonstratio ad oculos simply because perceivable pointing gestures appear. These gestures are seen as an indicator for the classiªcation as demonstratio ad oculos, because demonstrative gestures cannot appear in imagined space.11 I would like to support Bühler against Klein and Sennholz and argue against analogical deixis as a special form of demonstratio ad oculos and for the ªrst main case of imagination-oriented deixis. The fact that perceptibility is considered a criterion for the classiªcation in demonstratio ad oculos, marginalizes the fact that a perceivable analogon is not only a perceivable object but rather a perceivable sign that depicts an absent person or object. It is an iconic sign in the sense of Peirce. According to Peirce, signs do not form classes of objects but rather emerge and exist only in the consciousness of an interpreter who produces these relations: “A representation is such only so far as it is conceived to be one.” (Peirce [1872–1878] 1986: 64) If we choose the sign function as the distinguishing criterion between demonstratio ad oculos and imagination-oriented deixis instead of the criterion of perceptibility, we arrive at the following delimitation: the demonstratio ad oculos is deixis in the perceptual space. The potential deictic objects are perceived objects that the speaker and the addressee do not interpret to be signs. The ªrst main case of imagination-oriented deixis is deixis in imagined space. The potential deictic objects are interpreted as signs that represent something else. In our examples, the perceivable body of the speaker is interpreted as a sign and, indeed, as a part of the imaginary model constructed by the partners in conversation. This dissolves the contradiction that emerges from two heterogeneous secondary origos on the gestural and verbal levels for the same conceptual relation. The body of the speaker functions as an iconic sign, as an analogon of the addressee projected into the future. The secondary origo on the

89

90

Ellen Fricke

gestural level, which is instantiated by the speaker, is connected via the sign relation to the secondary origo at the verbal level, which is instantiated by the imaginary wanderer.

7.

Conclusion

We must assume not only one single origo but also several origos, which have to be considered as abstract volumeless points. These origos are not ªxed to the speaker but rather can also be shifted to other entities. I have suggested a hierarchical structure beginning with a primary origo connected to the role of the speaker. Within the change of communicative roles, the primary origo is attained and with it the possibility of intentionally creating secondary origos by means of origo allocation. These secondary origos can be instantiated by perceptible and imaginary entities. This diŸerentiation allows us to eliminate a contradiction inherent in the origo instantiation of local deixis between the verbal and gestural levels. The contradiction demonstrated in my examples is that, for the same conceptual relation, the origo of the verbal level is allocated to the addressee projected into the future as the imaginary wanderer, whereas the origo of the gestural level is allocated to the body of the speaker herself. If, as I propose, the body of the speaker is understood as an analogon, i.e. as an iconic sign representing the addressee as the imaginary wanderer, the contradiction is resolved.

Notes 1. Fundamental questions and discussion points can be found, for example, in Klein (1978), Holenstein (1972), Harweg (1990), Ziegler (1989), Hörnig and Wiebrock (2000), and Fricke (in peparation). 2. Bühler’s concept of displacement covers what Langacker (1987: 131) calls “mental transfer” and “cross-world identiªcation”. 3. The inadequacy of this dichotomy is also criticized by Harweg (1990), Hörnig and Wiebrock (2000) and Herrmann and Schweizer (1998). According to Herrmann and Schweizer (1998: 49) the speaker’s intrinsic orientation raises the question of whether the utterance “The parking lot is in front of me” is not just as much from an intrinsic perspective as the utterance “The parking lot is in front of the main railway station”. 4. A detailed criticism of Miller and Johnson-Laird’s dichotomy can be found in Harweg (1990: 216–218).

Origo, pointing, and conceptualization

5. Sennholz (1985: 24) notes that there cannot be a single origo for all dimensions since in some circumstances several deictics used in one and the same speech sequence can each have their own origo. However, as the examples of local deixis show in this article, one and the same verbal deictic can also be connected with at least two diŸerent origos. Klein (1978: 33 and 34) diŸerentiates between “primary” and “secondary” origos. In the following these terms are adopted but are deªned diŸerently. 6. According to Reichenbach (1947), the distinctions of tense are distinctions between three points along a time-line. These three points are the utterance time or time of speaking, the reference time, and the event time. 7. Cf. Lyons (1977: 692). 8. Cf. Bühler (1982a: 113 and 1982a: 79), Lyons (1977: 574 and 575), Sennholz (1985: 141). Ziegler (1989: 202) also argues for the priority of the personal dimension. Following Diewald, (1991) we exclude the third person from the personal dimension. 9. “[…] ich weiß nicht, wie Bühler den Fall Hier traf ihn die Kugel behandeln würde; es ist, in Bühlers Kategorien, oŸenbar eine demonstratio ad oculos, aber eben eine andere als bei Hier traf mich die Kugel. […].”(Klein 1978: 39) [I don’t know how Bühler would deal with the case The bullet hit him here; it is, in Bühler’s categories, obviously a demonstratio ad oculos, but diŸerent from the bullet hit me here.] 10. Cf. Sennholz (1985: 87): [Bühler himself wrongly interpreted this type as imaginationoriented deixis. However, I interpret it as analogical deixis in the sense discussed above, whereby the analogon is a kind of ‘stage’ (as Bühler ªttingly says). Then, within this ‘stage space’ the same conditions exist as in reality so that, correspondingly, a completely normal deixis takes place, a demonstratio ad oculos in the truest sense of the word.] 11. For the most part, Sennholz (1985) investigates only demonstratio ad oculos and marginalizes imagination-oriented deixis, which for him only includes displacement.

References Black, John B., Terrence J. Turner and Gordon H. Bower 1979 “Point of View in Narrative Comprehension, Memory, and Production”. Journal of Verbal Learning and Verbal Behavior 18: 187–198. Bryant, David, Barbara Tversky and Nancy Franklin 1992 “Internal and external spatial frameworks for representing described scenes”. Journal of Memory & Language 31: 74–98. Bühler, Karl 1982a Sprachtheorie: Die Darstellungsfunktion der Sprache. Stuttgart/New York: Fischer. First published 1934. 1982b “The Deictic Field of Language and Deictic Words”. In Speech, Place and Action: Studies in Deixis and Related Topics, R. J. Jarvella and W. Klein (eds.), 9–30. Chichester/New York/Brisbane/Toronto/Singapore: John Wiley & Sons.

91

92

Ellen Fricke

1990

Theory of language. The representational function of language. Translated by D. F. Goodwin. Amsterdam and Philadelphia: John Benjamins. Diewald, Gabriele Maria 1991 Deixis und Textsorten im Deutschen. Tübingen: Niemeyer. Ehlich, Konrad 1982 “Anaphora and Deixis: Same, Similar, or Different?” In Speech, Place and Action: Studies in Deixis and Related Topics, R. J. Jarvella and W. Klein (eds.), 315–383. Chichester/New York/Brisbane/Toronto/Singapore: John Wiley & Sons. Franklin, Nancy, Barbara Tversky and Vicky Coon 1992 “Switching points of view in spatial mental models. Memory and Cognition 20 (5): 507–518. Fricke, Ellen in press “Origo, pointing, and speech: the impact of co-speech gestures on linguistic deixis theory”. Gesture 2 (2). in prep. Origo, Zeigen, Repräsentation. Eine deixistheoretische Grundlagenstudie unter Berücksichtigung empirischer Daten zur Interaktion von Gestik und Rede in Beschreibungen des Potsdamer Platzes in Berlin. Ph. D. dissertation, Technische Universität Berlin. Harweg, Roland 1990 Studien zur Deixis. [Bochumer Beiträge zur Semiotik 25]. Bochum: Brockmeyer. Herrmann, Theo 1990 “Vor, hinter, rechts und links: das 6H-Modell. Psychologische Studien zum sprachlichen Lokalisieren”. Zeitschrift für Literaturwissenschaft und Linguistik 78: 117–140. Herrmann, Theo and Karin Schweizer 1998 Sprechen über Raum: sprachliches Lokalisieren und seine kognitiven Grundlagen. Bern/Göttingen/Toronto/Seattle: Huber. Hörnig, Robin, Berry Claus, Daniel Durstewitz, Ellen Fricke, Ute Schmidt and Klaus Eyferth 1996 Object access in mental models under different perspectives induced by linguistic expressions. [KIT-Report 126]. Berlin: Technische Universität Berlin. Hörnig, Robin and Sylvia Wiebrock 2000 “Deiktisch vor und Konsorten: Projizierte Bezugssysteme oder wegbasiertes Lokalisieren?”. In Räumliche Konzepte und sprachliche Strukturen, Ch. Habel and Ch. von Stutterheim (eds.), 261–282. [Linguistische Arbeiten 417]. Tübingen: Niemeyer. Holenstein, Elmar 1972 “Der Nullpunkt der Orientierung: Eine Auseinandersetzung mit der herkömmlichen phänomenologischen These der egozentrischen Raumwahrnehmung”. Tijdschrift voor Filosofie 34: 28–78. Jarvella, Robert J. and Wolfgang Klein (eds.) 1982 Speech, Place and Action: Studies in Deixis and Related Topics. Chichester/ New York/Brisbane/Toronto/Singapore: John Wiley & Sons.

Origo, pointing, and conceptualization

Johnson-Laird, Philipp N. 1983 Mental models. Cambridge, MA: Cambridge University Press. Kendon, Adam 1980 “Gesticulation and Speech: Two Aspects of the Process of Utterance”. In The Relationship of Verbal and Nonverbal Communication, M. R. Key (ed.), 207–227. The Hague/Paris/New York: Mouton. Klein, Wolfgang 1978 “Wo ist hier? Präliminarien zu einer Untersuchung der lokalen Deixis”. Linguistische Berichte 58: 18–39. Langacker, Ronald W. 1987 Foundations of Cognitive Grammar. Vol. 1, Theoretical Prerequisites. Stanford: Stanford University Press. Lyons, John 1977 Semantics, Vol. 2. Cambridge/London/New York/Melbourne: Cambridge University Press. McNeill, David 1985 “So you think gestures are nonverbal?”. Psychological Review 92(3): 350–371. 1992 Hand and Mind: What Gestures Reveal about Thought. Chicago: Chicago University Press. 2000 “Catchments and contexts: non-modular factors in speech and gesture production”. In Language and gesture. [Language, culture, and cognition 2], D. McNeill (ed.), 312–328. Cambridge: Cambridge University Press. Miller, George A. and Philip N. Johnson-Laird 1976 Language and perception. Cambridge MA: Harvard University Press. Peirce, Charles Sanders 1986 Writings of Charles S. Peirce. A Chronological Edition, Volume 3. 1872–1878. Ch. W. Kloesel (ed.). Bloomington: Indiana University Press. Reichenbach, Hans 1947 Elements of Symbolic Logic. New York: Macmillan. Sennholz, Klaus 1985 Grundzüge der Deixis. Bochum: Brockmeyer. Ziegler, Jürgen 1989 “Die Origo und das Grundlagenproblem der Deixis”. Deutsche Sprache 17: 193–205.

93

Part 2: Time

Two ways of construing complex temporal structures Christiane von Stutterheim, Mary Carroll and Wolfgang Klein University of Heidelberg / University of Heidelberg / Max Planck Institute for Psycholinguistics, Nijmegen

1.

Introduction

When faced with a complex verbal task, such as giving an account of the most unpleasant event in your life, or describing your dentist’s face, or explaining why jogging is healthy, there are many options open to the speaker in accomplishing this task. Despite the wealth of variation which one may encounter, the result in each case is a text — a sequence of utterances which is structurally related. The shape of the text is to a large extent determined by the quaestio, that is the explicit or implicit question which the speaker sets out to answer. Thus, the eye witness of a car accident will come up with very diŸerent texts, depending on whether the quaestio is ‘What did you see?’, ‘What happened?’ or ‘Who, to your mind, caused the accident?’. The quaestio partitions the text into a main structure, i.e., utterances which directly answer the quaestio, and side structures, i.e., utterances which supply diŸerent types of associated information. It also determines the topic-focus organisation of main structure utterances and thus sets guidelines for the way information is selected and coded at utterance level (see, in detail, Klein and von Stutterheim 1987, von Stutterheim 1997). Although the quaestio clearly channels some of the main coding options available to the speaker, it does not take care of all decisions which the speaker has to make in constructing a coherent text. The information organisation required to meet such demands can be described with respect to diŸerent structural aspects, the two most crucial being thematic continuity (manifested in terms of causal and temporal relations in the case of dynamic situations) and

98

Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

referential framing.1 By referential framing we mean something which — in the basic sense of this notion — has been described in classical poetics as the unit of time and space. Phrased in modern semantic terms, propositions (predicateargument structures) have to be referentially anchored with respect to times and worlds and possibly also spaces. In order for a sequence of propositions to be coherent, these referential properties have to be related consistently across utterances. There is a long tradition of studies which deal with these features of text structure under such headings as macrostructural planning, anaphoric linkage, mental models, etc. (Kintsch and van Dijk 1978, Brown and Yule 1983, Garrod and Sanford 1988, Ochs et al. 1992, Goldberg 1996). So, in addition to the guidelines set by the quaestio, what determines the actual linguistic options chosen when constructing a coherent text — is it just personal preferences, cultural or social habits, or do we have to take into account the in¶uence of the speciªc encoding system, of the particular language which the speaker uses? The question is, do speakers of English and speakers of German, for example, having witnessed the same accident, systematically follow diŸerent strategies in text construction when they set out to answer the question ‘What happened?’. This is the issue we investigate in this paper. More speciªcally, we shall look at the way in which English and German speakers typically organise and convey temporal information when answering the text quaestio ‘What happened?’, since structuring temporal information is a crucial part of the complex verbal task. Roughly speaking, temporal information involves two components. First, the ‘entire event’, here the entire car accident, must be embedded in space and time. In the present context, we are only interested in the temporal side of this embedding or framing. This is where deixis comes in: tense and other means such as deictic adverbials relate the entire event to the here-and-now of the speech situation. Second, the ‘entire event’ has a complex temporal structure in itself: it consists of a number of sub-events and states which are temporally related: they may follow each other, be contained in each other, partly overlap etc. In order to convey its temporal make up, the speaker must split the entire event into sub-events and phases within these sub-events and encode these, and he or she must mark what comes when, or during what other event. This is where aspect and Aktionsart as well as various types of anaphoric linkage become relevant. The mere fact that English and German diŸer with respect to these linguistic devices might be seen as an answer to our initial question. German, for example, has no progressive, that is, no formal means such as the -ing (as in they are falling) to code systematically the ongoing aspect of an event

Two ways of construing complex temporal structures

on the verb, and this should have consequences for the way in which German speakers present temporal information. But ªrst, the absence of a progressive marker does not mean that German speakers cannot present events as ongoing (or whatever the overall function of the English progressive may be). And furthermore, it does not mean that within a given language, speakers cannot follow diŸerent strategies in the way in which they approach the temporal organisation of their complex verbal task. The ways in which languages encode time is surely one of the most researched areas in linguistics. But most of this work deals with the temporal structure of individual sentences, and here in particular with what verbal categories such as tense, aspect and Aktionsart contribute to this structure (Reichenbach 1947, Comrie 1976, Bäuerle 1979, Partee 1983, Fabricius-Hansen 1986). Research which goes beyond the individual sentence, for example in Discourse Representation Theory (Kamp and Reyle 1993, Roßdeutscher 2000), is still the exception rather than the rule. Moreover, it is usually conªned to the anaphoric relations between two or three adjacent sentences, and therefore does not address the temporal make up of a whole ‘event’ and how it is rendered in the entire text. There is a further major problem in the research tradition on temporality: it has a strong bias towards ‘individual events in reality’, and canonical deªnitions of tense such as ‘a verbal category which localises the event before, after or around the moment of speech’ simply do not work for many text types. This and some other problems of the research tradition will be discussed in more detail in the next section. In conclusion, there is as yet no satisfactory analytical framework for the kind of analysis we have in mind here. We will largely adopt the framework proposed in Klein (1994, 2000a), together with some adaptations to be explained in the next section. In Sections 3 and 4, we will report on two empirical studies, showing that there are substantial diŸerences in the way in which speakers of English and speakers of German approach the task. The concluding section relates the empirical ªndings to the grammatical structure of these two languages.

2.

Three temporal parameters

Ever since Aristotle, a tensed verb form is assumed to serve two functions: it describes a ‘situation’ (i.e. event, process, state), and it locates this situation in time by relating it to the moment of speech: the situation may be before, after, or — more or less — simultaneous with the moment of speech. Hence, there

99

100 Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

are two temporal parameters — the time of speaking and the time of the situation — and three temporal relations between them: before, after, simultaneous with. This idea may be right, but it is surely insu¹cient as an account of a fully developed tense system. Classical Greek, for example, has more than three tense forms, and so do most Indo-European languages. One way of dealing with this problem is to assume a third temporal parameter, thus allowing more than three temporal relations. This idea was very common in 19th century linguistics (see, for example, Paul’s in¶uential Principien der Sprachgeschichte 1882). Nowadays it is often associated with the name of Hans Reichenbach and his three time points: the ‘point of speech’ S, the ‘point of event’ E, and the ‘point of reference’ R (Reichenbach 1947). Reichenbach used his system for English; but it has found wide popularity in the analysis of other languages as well. This is remarkable since it raises many problems. One of them has to do with the idea that Reichenbach speaks of ‘points’ rather than of ‘time spans’ or ‘intervals’ which can occupy a certain stretch of time, as is obviously the case for events as well as for the time of speech (the case is less clear for the ‘reference time’). Most followers of Reichenbach adopt his S, R, and E, but interpret them as intervals, which can have a certain measurable duration, instead of points. A second set of problems has to do with the concrete analysis of tense forms in a speciªc language. Should the German Perfekt be analysed as ‘E before (S simultaneous to R)’, like the English present perfect, or should it be analysed as ‘(E simultaneous to R) before S’, like the German Präteritum? In the present context, we are less interested in these speciªc questions than in a number of problems that are quite independent of particular forms and what they mean. These problems have to do with the deªnition of the three temporal parameters (interpreted as intervals) and hence touch upon the very foundation of such an account. They will now be discussed in turn. 2.1 Problems with S, R, and E There are at least four substantial problems with the notion of S, the time of speech. a. In the ªrst place, S seems to be deªned for the ‘canonical speech situation’, in which speaker and hearer are present in a well-deªned spatio-temporal setting. But what counts as S in written texts — the moment of writing or the moment of reading? There is no uniform answer. In written instructions, for example, it is typically the time at which a speciªc reader reads the instructions

Two ways of construing complex temporal structures 101

(now, you turn the handle ...); on a postcard sent while on holidays, it is typically the time at which the postcard is written (tomorrow we will visit the Palazzo). b. What counts as ‘time of speech’, if the speech is very long — what, for example, was S in Lincoln’s ‘Gettysburg address’? Does each utterance have its own S? Even a single utterance may require two diŸerent temporal anchors, as illustrated by sentences such as From now, it is two seconds to now. This problem often goes hand in hand with the ªrst problem: which S has Proust in ‘A la recherche du temps perdu?’. Or take the case of complex instructions, such as route directions: Now, you take the ªrst street to the right, continue on to the ªrst tra¹c light, and now/then cross ... . Note that the problem arises for the deªnition of the tense forms as well as for the interpretation of temporal adverbials. c. There are many text types for which the actual moment of speech (or writing) apparently plays no role whatsoever. A good example are legal or administrative texts. Take a simple sentence such as The court will decide no later than two weeks after the application has been submitted. It includes a future tense form and a present perfect, none of which is related to the moment of writing or the moment of reading this sentence (for a detailed argument on this point, see Klein 2000b). An even more salient case are all sorts of ªctitious texts, from historical novels to science ªction stories, from fairy tales to ªlm retellings. Their plot is simply not ‘in this world’, and hence, it does not make sense to position them on the time line in relation to the moment of speech, whether told in the simple past or present tense. In either case, their temporal anchor cannot be the moment of speaking, writing or reading. d. Even with the ‘most canonical use type’, that is, when talking to someone about singular events in the past, the usual deªnition of tense forms is often wrong. Take for example the common deªnition of the English present tense: ‘time of situation at or around moment of speech’. There are many cases, in which the situation is clearly in the past, and still, the present tense is used. Two well-known cases are the ‘historical present’, as in And then along comes this guy and slumps to the ground just in front of the shop, and the ‘praesens tabulare’, as in In the year of 1809, he publishes ‘Die Wahlverwandtschaften’ and begins to work on the second part of ‘Faust’. These problems shed considerable doubt on the classical understanding of ‘deixis’ — the ‘here-and-now of the speech situation’ in the analysis of temporal expressions, be it tense forms or temporal adverbials. It works in some cases, but there are at least as many in which it does not. Hence, we must replace the

102 Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

notion of deictic speech time by a somewhat broader notion of temporal anchoring (see Section 3.2.). Let us turn now to R, the ‘time of reference’. The problems here are of a very diŸerent nature: Reichenbach never gave a deªnition of this notion; he considers it to be a time (in his case, a point), which is somehow given in context. This, incidentally, is in remarkable contrast to earlier work on temporality which operates with a third parameter, such as Paul, who considers it to be the time from which something is imagined. More surprising is the fact that the literature following Reichenbach did not bother to deªne it either. But it seems obvious that the potential role of the third temporal parameter cannot be evaluated without a meaningful characterisation of this notion; R cannot just be an arbitrary time span. The notion of ‘topic time’, to be discussed below, is such an interpretation. This leaves us with E, the ‘time of the situation’. Note, ªrst, that Reichenbach did not consider E to be the ‘event’ but its ‘time’. Tense forms do not locate events but temporal intervals, to which such an event is related. He did not explain the nature of this relation, except that the temporal interval associated with some event was considered to have no extension. In subsequent work on E this idea was for the most part abandoned. But the precise relation between ‘the situation’ and ‘its time’ is hardly ever addressed. In fact, it raises a number of problems, primarily because situations in themselves often have a complex internal temporal structure. Consider, for example, a sentence such as He planned to have left by ªve. Clearly, the entire situation, which is selectively described by this sentence, involves a number of diŸerent temporal intervals — the time of his planning (t1), the time of his leaving (t2), and the time at which this leaving is over (t3). There are the temporal relations between these intervals: t1 precedes t2 as well as t3, t3 follows t2. There is no single ‘situation time’. It is possible to deªne such a time, for example as the minimal interval ts, which includes t1, t2, and t3. But such a notion is of little use, tense markings as well as temporal adverbials do not operate on an ‘entire situation time’ but on its individual components. Thus, the temporal expression by ªve does not specify ts but t3, i.e., the time at which his leaving is completed. Note that this is diŸerent for At ªve, he planned to have left the city, where at ªve indicates the time of his planning, i.e., t1. It is also diŸerent in He planned to have left the city at ªve, which is a bit odd under all three possible readings, with a certain preference for t1 (since the prospective nature of to have left in this context generally requires use of by and not at). The same point can be made for the relation between the ‘situation time’ and the

Two ways of construing complex temporal structures 103

temporal anchor, here the moment of speech S: it is his planning which is related to the moment of speech; the temporal relation of t2 and of t3, and hence of ts, to S is completely opaque. Now, a sentence such as He planned to have left by ªve has a compound verb phrase. But very similar problems arise with simple verb phrases, as in She closed the window: There must be a time t1 at which she does something (e.g. turns a handle or pushes a button), a time t2 when the window is open, and a time t3, when the window is closed; t2 must precede t3, and t1 must overlap with t2 (but can also stretch into t3). In cases like this, that is, when the diŸerent time spans are packaged in a single word, it is much more di¹cult to ‘access’ them individually; but still, it is possible. Thus, a temporal expression such as for ªve minutes can indicate the duration of t3, but not, for example, the duration of t1, nor the duration of the minimal interval which includes t1, t2 and t3,, i.e., the full situation of ‘closing the window’. Summing up, the situation, as described by the sentence, regularly involves a whole array of temporal intervals: the ‘time of situation’ is not a single interval but a ‘temporal structure’, with a single interval as a borderline case. This applies whether the verb phrase is simple or complex. In this section we have shown that each of the three parameters S, R, and E raises massive problems. In the next section, we will try to accommodate some of them. 2.2 The characterisation of temporal structures In what follows, we shall adopt some ideas developed in Klein 1994 (and further reªned in 2000a). They make use of three time spans as well, called TU, TT, and TSIT (for ‘time of utterance’, ‘topic time’ and ‘time of situation’, respectively); hence, these ideas continue the tradition of three-parameter approaches. The crucial feature is a rigid distinction between ‘temporal structures’, that is, time spans and the temporal relations between them, on the one hand, and the characterisation of individual time spans which belong to such a temporal structure, on the other hand (see Klein 1994, Chapter 3). A temporal structure as such is, so to speak, ‘empty’. Temporal intervals have no properties, except duration and being before, after, or within other temporal intervals. But they can be assigned properties. This is either done explicitly, by the lexical content of verbs or other expressions associated with them, or implicitly, i.e., by all sorts of contextual information or world knowledge. Consider again a sentence such as He planned to have left by ªve. It minimally involves the three time spans t1, t2 and t3 together with the temporal

104 Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

relations described above. Each of the three time spans is ‘lexically characterised’ as the time of his planning, the time of his leaving and the time of his having left. Contextual knowledge may enrich this characterisation (for example the knowledge that ‘his leaving’ concerns ‘Peter leaving Heidelberg’ rather than ‘Napoleon leaving Moscow’). The situation, as characterised by the lexical content of the sentence, regularly involves a more or less complex cluster of time spans. In other words, the ‘time of situation’ is normally not a single interval but a ‘temporal structure’ (with a single interval as a borderline case)2. When we speak in the following of ‘the time of situation (TSIT)’, this term is to be understood in this sense: normally, it is not a single interval (let alone a point), but a temporal structure which may involve a number of sub-intervals. This applies whether the verb phrase is simple or complex. But compound constructions regularly add temporal intervals which are accessible to further speciªcation by adverbials and other devices. Consider, for example, an action such as digging a hole — it clearly involves various sub-phases: it has, for example, an initial phase and a ªnal phase. These, however, are not directly accessible to further speciªcation: no adverb added to this verb can specify just its initial phase. But it is possible to decompose the very same action into a more ªne-grained description and thus create further ‘accessible time spans’. One such possibility is to use ‘phasal verbs’, such as to start digging, which express an initial phase. Thus, in He started digging at ªve, the expression at ªve does not specify the time of digging but the time at which it was initiated, its initial phase. Another important possibility is the nucleus or intermediate phase, was digging, which gives an inner sub-interval. Hence, he was digging relates an inner sub-interval to the temporal anchor, thus giving the impression of somehow being within the action: the action of digging is seen as ongoing, as not completed. This explains the notion of ‘imperfective aspect’, which is typically associated with the English progressive. TSIT is a temporal structure, which is more or less ªlled with content by the words of the sentence (and by contextual information). It is not ‘embedded in time’ in itself, i.e., it is not related to some other time span outside TSIT. Such an external time can be provided by the moment of speech. But as discussed in Section 2.1, this notion is insu¹cient. We may now sum up the argument in three points: 1. Temporal expressions are often linked to an interval which is not given in the text itself but must be taken from context. Let us call this interval the ‘temporal anchor’ (abbreviated as TA).

Two ways of construing complex temporal structures 105

2. Depending on a number of factors, TA can have diŸerent interpretations. One of these is ‘the moment of speech’; but there are many other possibilities. 3. The interpretation of TA may vary within a text, for example, from one utterance to the next; but it may also remain constant over some stretch of discourse. Hence, for a whole text, we typically have a whole array of TAs, which can be related to each other in various ways. A single TA is just a borderline case of such a ‘TA-structure’. It is a matter of terminology whether any type of temporal anchoring in this broad sense should be called ‘deictic’.3 Note that the broad notion of temporal anchoring is not at variance with a deictic deªnition of tense (and temporal embedding in general). It is just a special case of a much broader phenomenon. For the sake of argument, let us assume in the following discussion that we indeed have only one utterance and a relevant ‘deictic origin’. How are such a TA and a TSIT brought together? Languages have various devices to this end. One of them is temporal adverbials. Another important device, found in many but not in all languages, is ‘tense marking’ on the ªnite verb. This is the point where the third temporal parameter, called here ‘topic time’ (abbreviated TT), comes in. A speaker, when formulating an utterance, must not only decide which descriptive content is to be included in the sentence; a decision must also be made on the time which he or she wants to talk about. These two decisions must be kept apart. Suppose the speaker, having witnessed an accident, decides to describe a scene in which a cyclist takes a left turn. In reality, this scene may have occupied a certain time span of, say, ten seconds; this is TSIT. For whatever reason, the speaker may choose to assert something not about these ten seconds as such but about a time which follows them, or about a time which precedes them, or a time which is properly included in them, or a time which includes them. English has grammaticised these options. The speaker may say The cyclist has turned left, The cyclist was going to turn left, The cyclist was turning left, The cyclist turned left, respectively. As these examples show, English solves the task by creating a more sophisticated temporal structure ‘TSIT’, which also involves post-times, pre-times, sub-intervals of the time which it takes the cyclist to turn left (the ‘original TSIT’). For instance, have left diŸers from the bare leave by adding a ‘post time’. The relation between the time talked about and the ‘original TSIT’ re¶ects various ASPECTS or ways of looking at the situation: thus, aspect is the grammaticalised relation between TT and TSIT. Other languages do not grammaticise the relation between TT and TSIT. In German, for

106 Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

example, there is no counterpart to the distinction between the progressive and simple form: German simple forms are neutral with respect to a ‘perfective’ and an ‘imperfective’ relation, just as the English copula to be in It was green. It is arguable whether the German Perfekt expresses a ‘perfect aspect’ in the same way as the English perfect — an issue to which we shall return in Section 5. It is the TT chosen by the speaker which is eventually related to the temporal anchor. In The cyclist had turned left, for example, TT precedes the temporal anchor, and in The cyclist has turned left, it includes the temporal anchor. In the construction The cyclist is turning left, an assertion is made about some time which includes the moment of speech (or some other temporal anchor), which is a sub-interval of the ‘taking a left turn’. In The cyclist was turning left, the time talked about is also a sub-interval of the ‘taking a left turn’, but it is said that this sub-interval precedes the temporal anchor. Summing up, we have three temporal parameters. These are – the temporal anchor TA, which can, but need not be the moment of speech; – the topic time TT, that is, a long or a short time span about which the speaker decides to speak; – the time of situation TSIT; as a rule, TSIT is not a single interval but a complex temporal structure, i.e., a set of intervals which are temporally related to each other; each of these intervals is characterised by the lexical content of the sentence (and by contextual information; many, though not necessarily each, of these intervals can be subject to further grammatical processes, such as adverbial modiªcation). All of this applies to individual utterances. But what about the entire text to which such an utterance belongs, say the full account of the accident? As for ‘situation time’, we must obviously distinguish between the time of the entire event and the times of the various sub-events which the speaker has chosen to map into a single utterance. In other words, we have a ‘big TSIT’ and many ‘small TSITs’ each of which is normally not a single interval but a ‘temporal structure’ in the sense explained above. Similarly, we must distinguish between the topic time of the entire text and the topic time of the individual sentences, and, ªnally, between the overall temporal anchor and the temporal anchors of individual sentences. Let us call the ‘text time parameters’ TA*, TT*, and TSIT*, respectively. The crucial question is, of course, how these are related to the corresponding parameters of the individual utterances. If the entire text consists of just one sentence, they are simply identical. If the text is more

Two ways of construing complex temporal structures 107

complex, there are many possibilities. In the case of TSIT*, the individual TSITs are coherent sub-structures of TSIT*, chosen at the speaker’s discretion. The case is more di¹cult for TT* and TA*. It may well be that the topic time is kept constant throughout the text, as in a picture description or in a mathematical argument, for example. The claim that 911 is a prime number is not just conªned to a short time span, it is claimed to hold forever, and is thus constant for all such arguments. But the speaker may also choose to change topic times while constructing the text. The same is true for the temporal anchor: there may be one temporal anchor which remains constant throughout the text; normally, however, there is a steady shift. This is clearly the case when ‘the moment of speech’ is chosen as a temporal anchor, thus giving rise to what in research on narratives is sometimes called ‘the order of mention’. It is these choices which lead to diŸerent temporal perspectives for one and the same complex verbal task, as deªned by the initial quaestio. We may now return to our ªrst question: taking English and German speakers, are there systematic diŸerences in the choices made in the same communicative situation? Before turning to the empirical data, we will take a look at some of the possibilities.

3.

Variation in perspective

The complex verbal task which we investigated was the retelling of a short silent ªlm, called Quest (see, in detail, Carroll and von Stutterheim 2003). In this setting, TSIT* is the highly complex temporal structure of the ªlm from which the speaker has to select sub-structures that will form the individual TSITs of each utterance.4 There is also a well-deªned ‘text speech time’ —the temporal interval during which the speaker produces the text. But there is no obvious way in which this speech time could be related to TSIT*, since the situation, the plot of the ªlm, is not in the real world. The events, processes and states shown in the ªlm are neither in the past nor in the present nor in the future, as deªned by the time of speech. Now, under the analysis of tense suggested above, there is no direct relation between TSIT and TA anyway. It is always mediated by TT* (and the individual TTs, respectively). But the problem remains: no matter how large the entire TT*, as set by the speaker, or how it is deªned for the individual utterances, the times talked about cannot be related directly to the time of speech, on the one hand, and the temporal structure deªned by TSIT*, on the other. There are various ways in which the

108 Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

speaker can solve this particular problem, as well as the general task of retelling the ªlm; we will brie¶y sketch three of them. First, the speaker may treat TSIT* as well as the individual situations as if they had really happened. This allows a ‘deictic’ treatment in the usual sense. The actual production time is considered to be the temporal anchor TA*; this time is split into subsequent smaller ‘times of utterances’, which serve as the temporal anchors for the various topic times. These precede their temporal anchor, hence the tense is the past. Depending on how the individual topic times are chosen in relation to the individual situations described by a sentence, various aspects are possible. In the typical case, all main-structure utterances, — those utterances which describe what happened rather than giving background information — are in the simple past. Another factor which may play a role here is the ‘length’ of some topic time: if the speaker considers it to be very short, the ‘time window’ on the alleged reality is very small, and only part of the situation is visible. In this case, the speaker might be forced to ‘zoom in’ on a special phase of the situation, for example by saying He started to dig a hole, instead of He dug a hole. Second, the speaker may indeed treat the ªlm events as ªctitious. Then a diŸerent type of temporal anchor TA* must be chosen. One way of doing this is to take ‘the time of experiencing the ªlm’ as the temporal anchor. Note that this time, in turn, is related to the moment of speech. But this need not have any relevance for the way in which the text is embedded in time: the entire story, as well as each of its sentences, is related to the time of experiencing the ªlm, reactivated, so to speak, by the speaker. This time must be split up for each utterance and results in a form of presentation in which each utterance is preªxed by ‘and now I/we see’. The content unfolds again in front of the mind’s eye. Thus, the appropriate tense is the present. The anchor might be called the ‘experiencer’s now’. As with the ªrst strategy, aspect depends on ‘window size’, i.e., on how large the TTs are set: if the full event is visible, then the simple present is chosen. Otherwise, the speaker must either use the present progressive or some other, more complex form, such as He stops digging. A third strategy treats the events of the ªlm as ªctitious entities and maintains the time of speech as temporal anchor. The entire text is, so to speak, preªxed by ‘these are the things I have seen in the ªlm’. In other words, there is a very big TT*, in which the ªctitious events are visible like a string of pearls. Each of these can then be described as a self-contained unit within this big topic time. An essential element of this way of evoking the material is the temporal order in which these entities stand in relation to each other. The tense

Two ways of construing complex temporal structures 109

form is the present, since the temporal anchor is included in TT*; the aspect should normally be perfective, and there should be devices which indicate the temporal relation of the sub-events, such as then, next, etc. This results in a linking strategy which relates TT to TSIT in an anaphoric relation. These are three ways of presenting the temporal structure of the ªlm — three perspectives under which it can be retold. There may be others, but we will not speculate on them here. Let us now turn to the question as to what English and German speakers indeed do when they try to solve the task.

4. Temporal perspectives in English and German texts: A ªrst experiment (Quest 1) The ªlm used is the silent ªlm Quest (duration 7 minutes) which shows the adventures of a clay ªgure who wakes up in a desert-like world and embarks on a search for water. His ‘quest’ takes him through four diŸerent worlds (a desert world, a paper world, a stone world, and ªnally a world dominated by robots and other industrial machines). The English and German speakers were for the most part university students in their mid-twenties. The data collection proceeded as follows: They ªrst viewed the entire ªlm without interruption and then viewed it again, but this time the ªlm was stopped at the end of each ‘world’. Each speaker was asked to tell what happened in the episode or ‘world’ they had just seen. The texts varied in length in both groups. The average length for the English group is 173 utterances, for the German corpus it is 175 utterances. In the following section we will summarise those ªndings which are relevant for the question of temporal perspective taking (for a detailed description see Carroll and von Stutterheim 2003).

4.1 English ªlm retellings English speakers adopt a referential frame in which the temporal anchor is the now of the experiencer, that is, as the ªlm is experienced in unfolding in front of the mind’s eye. We call this a ‘deictic strategy’ in which the temporal anchor ‘now’ is projected. In other words, TA* is the ‘ªlm experience’, rather than the time at which the ªlm is retold, and the temporal anchors of individual utterances are sub-intervals of this TA*. This corresponds to the ‘and now I see’

110 Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

strategy mentioned above. The result resembles a series of ‘screen shots’, each with the ‘experiencer’s now’ and a short time window related to this ‘now’, the topic time of a given utterance. Note that the ‘experiencer’s now’ and the TT of two subsequent utterances need not diŸer: two or even more subsequent utterances may be used to describe the same ‘screen shot’; metaphorically speaking, several situations may be visible through the same window. One feature of this referential frame is that circumstances are viewed as leading in some way to an event or sets of events, thereby indicating progression, as illustrated in the following example (see also Carroll and von Stutterheim 2003). The opposition in English between the simple present (sits) and the progressive (is sitting) is crucial in this context. Example 1. Circumstances: use of progressive (-- ing)) form (1) # he starts trying to scoop the water up (2) but the water is being absorbed into the surface # (3) and he cannot get any water from that switch to simple tense: (4) so he sits there (5) and looks up (6) to where he thinks they are (7) but there’s no more drips (8) they have stopped dripping (termination of events linked circumstantially)

In Example 1 the now of the experiencer, perceiving the events as they unfold before the mind’s eye, provides the topic time. The time of situation of the utterances (1–3) is linked to this TT, and is either simultaneous with TT, as in 1, or includes TT, as in 1 and 3. This means that TT (as well as the ‘experiencer’s now’) is maintained across the first three utterances: they show, so to speak, the same screen shot. The temporal interval introduced by the situation water being absorbed into the surface functions as TT-interval for the following events. The TSIT of (4–7) are situated within this frame/timeinterval, and nothing further is stated about their temporal location or the way in which they are linked. The present perfect form in (8) requires a point in time as TT as the point at which an event is terminated and the post state of the event holds as the point of reference. This point in time can only be given in relation to the ‘experiencer’s now’. The pattern of topic time management for this piece of text is as follows:

Two ways of construing complex temporal structures

(1) (2) (3) (4) (5) (6) (7) (8)

TT TT maintained TT maintained TT (interval) = TSIT2 TT maintained TT maintained TT maintained TT (point in time) = experiencer’s now

TT overlaps TSIT1 TT ⊂ TSIT2 TT ⊂ TSIT3 TSIT4 ⊂ TT TSIT5 ⊂ TT TSIT6 ⊂ TT TSIT7 ⊂ TT TT after TSIT8

Example 2 is another example which exhibits the same pattern. The two utterances (1) and (2) are linked to the deictic centre as their TT. They establish a temporal interval by their TSIT, which functions again as TT interval for the events reported in the following utterances. Example 2. Circumstances: progressive forms (1) # there’s no water to be found (2) and he’s walking through this area # Embedded actions which happen while walking: simple forms (3) and he hears the drip again (4) and sees this moist area on this piece of paper (5) so he gets excited (6) and he runs over

Events such as hears, sees, and possibly gets excited occur within the established time span. The termination of the time span walking through x is given by the event and he runs over. This yields the following general pattern of TT management: TA = TT ⊇ TSITx TSIT = TT ⊂ TSITx+n

In sum, the pattern followed by the English speakers is temporally complex. The initial TT provides the anchor for the introduction of a circumstantial frame, characterised by an imperfective view on the activities or processes reported. The frame includes a set of events, and these are accorded some implicit portion of time within the given time span. The events described may overlap or follow each other in time. Frames of this kind are marked by a general absence of temporal adverbs such as then. The temporal relations between the various TSITs remain implicit. The link between the frame and its associated events is often of a causal

111

112

Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

nature (there is circumstance x : so this leads to y). But causal links can be more or less prominent, however, as shown in the two examples above. As to event construal, no speciªc temporal property is required. All types of situation can be hooked up to the temporal anchor, there is no constraint with respect to the composition of the event. The speakers adjust the speciªc type of situation in its relation to the temporal anchor by means of phasal decomposition. We observe a high level of aspectual marking (inchoative with to start and ongoingness) which serves the purpose of expressing the ¶ow of the events in relation to the viewing point. Occasionally, speakers select another option in text construal, similar to that found in German, but these are not the preferred principles typical of the majority of ªlm retellings (Carroll and von Stutterheim 2003). The existence of these options, however, demonstrates that the strategy described above is not forced upon speakers. The structure of the language allows for diŸerent options (exploit the principles of construal allowed by grammaticised aspect, or not), but the data provide evidence of what can be described as a preferred form of temporal organisation which correlates with grammaticised meanings (ongoingness) and the type of referential frame they allow. 4.2 German film-retellings In contrast to speakers of English, the German speakers represent a dynamic situation as a well-delimited entity which can ªll a slot on a structured time line. The pattern exhibited in the retellings follows the strategy of linking the current topic time TT to the preceding time of situation. This is what we call an ‘anaphoric strategy’. This involves the explicit marking of a relation between diŸerent temporal intervals, usually of the type “posteriority”. This strategy evokes the notion of a big window in which the individual events are presented as a well-ordered string (cf. Section 3). The following example illustrates this type of temporal organisation which is followed by all German speakers in the sample. Example 3. (1) und wacht dann so langsam auf ‘and wakes then slowly up’ (2) und schaut sich um ‘and looks around’ (3) und sieht dann die ¶asche vor sich liegen ‘and sees then the bottle lying in front of him’

Two ways of construing complex temporal structures

(4) nimmt die ¶asche ‘takes the bottle’ (5) und guckt ‘and looks’ (6) ob da wasser drin ist ‘if there is water in it’ (7) dann steht er so langsam auf ‘then he gets slowly up’ (8) und kniet so ‘and kneels down’ (9) und guckt sich um ‘and looks around’ (10) und dann donnert es plötzlich ‘and then there is suddenly thunder’ (11) und er hört ‘and he hears’ (12) wie wassertropfen in den sand fallen ‘how water drops fall in the sand’

The organisation principle in the German retellings is characterised by two main features: a. Explicit marking of temporal relations by temporal adverbials in the context of the sequence of events. Adverbials are used to establish TT and to link them explicitly to the preceding TSIT. b. A holistic view on the events, including points of completion or results of an event, which provides the anchor point for the shift-relation of the TT interval. 4.3 Summing up The following two patterns were observed in organising temporal coherence on the basis of the unfolding topic time: TU = TT ⊇ TSITx

TSITx ≥ TT ⊂ TSITx+n (main structure)

TSIT= TT ⊂ TSITx+n

TSITy ⊂ TT (side structure)

(dominant pattern in the English texts)

(exclusive pattern in the German texts)

Given their systematicity the contrasts observed require an explanation. We assume (see Carroll and von Stutterheim 2003) that the basis for these diŸerent

113

114

Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

patterns of referential framing can be found in the structure of the languages involved, more speciªcally, in the diŸerent forms of aspectual marking, as indicated above. Note, however, that the diŸerences cannot be reduced to the fact that English has a systematic aspectual opposition between an imperfective and an unmarked form. Even where German has traces of aspect, the opposition encoded diŸers. Taking the linguistic means at the level of verbal morphology, the marked form in German is the perfective. These diŸerences have clear consequences for the relational potential of sentences. In order to explore this further, and in order to exclude the possibility that the diŸerences observed are only a stylistic convention in retelling ªlms (more reportive in English, see Smith 2000, and more narrative in German), we started a second study and checked what happens if the task is changed.

5.

Perspective shift and its structural roots

The design of the second study (Quest 7 — the seventh experiment with this ªlm) was essentially the same, except that the speakers were now asked to retell the ªlm in the past. Our expectations in selecting this design run as follows: the English speakers — bound to their ‘and now I see’ anchor — should encounter problems in transferring the content into a frame of reference located in the past. For them, a high degree of transformation is required with respect to the selection of event components, on the one hand, and with respect to the type of relation required between utterances. Broadly speaking, they have to switch to an anaphoric strategy. The latter implies the provision of boundaries for the event times by either verbalising the endpoints of situations or by closing the event time intervals by explicit temporal information (adverbials). German speakers have an easier task to solve. With an anaphoric strategy, it should not be particularly di¹cult for them to switch to a past referential frame. Since the general type of linkage between the situations reported is that of a ‘shift-in-time’ with states related to the event times, nothing with respect to the construal of events nor type of relation would have to be changed. We therefore expect English speakers to run into di¹culties, and the German speakers to have few, if any, in solving the task. The analyses of the data proceed in the following steps. First the English texts Quest 1 and Quest 7 will be compared with respect to the temporal frame selected and the patterns of event construal. Use of tense and aspectual forms will be analysed quantitatively (frequency of occurrence and tense switch) as

Two ways of construing complex temporal structures

well as qualitatively in relation to situations which trigger their use. The same analysis will then be carried out for the two German data sets. In a third step, English and German Quest 7 texts will be compared with respect to patterns of temporal organisation. In the concluding section we will relate structural properties of the languages involved to principles of information organisation which guide the production of complex texts. 5.1 Comparing English Quest 1 and Quest 7 The basic strategy selected with Quest 1 cannot be maintained by speakers under the second condition given with Quest 7, where speakers were asked to take a ‘past’ perspective on the situation, as mentioned above. One solution is to switch to the framework of a past time narrative, but this is typically found in English in narratives of personal experience (cf. Labov 1972)5. The other possibility is to produce some kind of hybrid between a deictic strategy based on ‘the experiencer’s now’ and the requirements of a temporal structure which is anaphorically linked. In this case we expect texts to be inconsistent with respect to the frame of reference and with respect to the formal devices used. The Quest 1 texts — with a few exceptions — are anchored in terms of the ‘now of the experiencer’. The present tense is used for both aspectually unmarked and marked forms. The second set (Quest 7) is characterised by the dominant use of the past tense form, showing, however, a considerable number of tense switches, corrections and inconsistencies in topic time (TT) management. The following quantitative and qualitative analyses substantiate these observations. We analysed the relative frequency of the tense forms used (Table 1), switches to the present tense and corrections from present to past Table 1.

English Quest 7: Relative frequency of tense forms

present simple

progressive

past simple

progressive

165 10.4%

35 2.2%

1187 74.6%

204 12.8%

Table 2.

English Quest 7: Switch to present tense

total number narrative utterances

present tense forms

past tense forms

200/1591 12.6%

1391/1591 87.4%

1591

115

116

Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

within one utterance (Table 2), and contexts that trigger switches (Table 3). The data analysed covers the ªrst three ‘worlds’ of the ªlm, which amounts to 2/ 3 of the material. The numbers show that the perspective selected in constructing the texts is relatively unstable. In 12.6% of the narrative utterances speakers switch to a present tense form. This type of tense shift forms the majority of tense shifts in the texts. In order to see if these switches are motivated by a strategy such as the historical present, or whether they result from an interference with a deictic strategy (‘now of the experiencer’), the contexts which trigger a switch in tense were analysed. If the tense switches are based on a clash in perspective, then switches to the present tense should occur in contexts which attract an anchor involving now. Given the possible pattern of TT management outlined above, a state or an ongoing action requires an external reference point (now). These are informational units (a) which establish a circumstantial frame at the beginning of a scene (see above) and (b) which present events as ongoing within sequences involving phasal segmentation (he started pushing, he is/was pushing). As presented in Table 3, there are ªve types of contexts in which switches into the present systematically occur; clearly indicating that the contexts which trigger tense switches are not random. Table 3.

English Quest 7: Contexts for switches and corrections

narrator viewpoint

phasal segmentation

states

situation reoccurs

activity protagonist environmental force

15.8%

15.8%

28.5%

4.7%

34.9%

Switches where the situation includes an external anchor (narrator viewpoint, phasal segmentation, states) add up to 64.8%. To use a term introduced by Smith, these are situations which need ‘capture’ (Smith 2000). Let us brie¶y illustrate the ªve cases. a. Narrator viewpoint, i.e., explicit references to the speakers activity of watching and interpreting the ªlm (e.g., predicates of seeing, hearing, interpreting), as well as comments and qualiªcations on the events: suddenly these boulders began to move and we see it is actually a human like ªgure; or, the clay man seems to wake up. b. Phasal segmentation, e.g.: the ªgure started falling and he’s pushing sand around

Two ways of construing complex temporal structures

c. State type predicates relating to circumstances, e.g.: and it is windy d. Reoccurrence of a situation, e.g.: and again the creature hears water dripping. The remaining category ‘activity’ is more problematic. It involves cases with no obvious contextual motivation and consists of references to events involving the protagonist or inanimate forces which he encounters. The switches to the present may be linked to inattention on the part of the speaker to the task requirements, since they tend to occur at the beginning of a new scene where speakers may be more inclined to fall back into the ‘preferred perspective’ for ªlm retellings, as described above. These switches in referential continuity indicate that speakers do not establish an anaphorically organised time-eventstructure. There is no TT link between event times and states related to these events. Local planning leads the speakers to take the ‘preferred’ frame of reference, that is, relating those predicates which ‘need capture’ to the viewpoint now. The tense switches suggest that speakers have superimposed the ‘past strategy’ onto an underlying ‘deictic strategy’ which is still conceptually dominant in text construction. Further evidence for the maintenance of at least basic aspects of a deictic strategy comes from parallels in the processes of information selection, in particular the level of granularity, which involves the degree of segmentation of the situations shown in the ªlm into sub-events and event phases. Looking at the data (see examples below) we ªnd that the resolution chosen by the Quest 7 speakers corresponds to that found with a deictic strategy. This is supported by the average number of utterances produced per text: 173 in Quest 1 and 182 in Quest 7 (20 speakers in each group) which do not diŸer signiªcantly. Events reported in the utterances are frequently segmented into phases, and the inchoative and imperfective aspect is selected for events within the story line. So this high degree of resolution is very similar to the retellings in the present. A further indicator of the maintenance of a deictic frame is the comparison with the referential frames found in the narratives of Quest 1 speakers (two in all), who spontaneously chose the past tense, and not the present. In contrast to the narratives of speakers who were asked to adopt a past tense perspective (Quest 7), there are markedly fewer events which relate to the ‘now of the experiencer’; in other words there are fewer events which are segmented into phases, and ongoing events (sheets of paper were blowing around) are not integrated into the sequence of narrative events, i.e., anaphorically related events, but are marked as secondary by devices such as subordination, attribution, etc. (cf. Example 11 below).

117

118

Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

In comparing Quest 1 and Quest 7 texts qualitatively in the next section we will begin with evidence for the maintenance of some basic features of a deictic strategy in the English Quest 7 narratives. The indicators found will be illustrated by examples. 1. The ªrst indicator concerns processes of information selection: Example 6 from the Quest 1 corpus gives an example of the referential frame found in narratives told in the present tense. Situations with a high degree of resolution or event segmentation were selected in order to illustrate and compare the relevant principles. The temporal aspects of the events relate some sub-interval of TSIT to TT and therefore do not need reference to an endpoint or boundary of TSIT. The event time intervals and their intrinsic relation remain implicit: Example 6. (1) (2) (3) (3) (4) (5) (6) (7)

so he starts digging in the sand towards where he thinks he hears the water and as he digs and digs the sand / through the sand he starts digging a hole and he starts sliding into what looks like dry quicksand and he starts falling into the hole ...

(ongoing)

(ongoing) (ongoing) (ongoing)

There is a marked contrast to Quest 7 texts in that the narratives in the past tense, with a similar degree of resolution, consistently present ongoing events as reaching a point of closure. This can be achieved at diŸerent levels of the macrostructure and by diŸerent means. The comparison between the Quest 1 and Quest 7 narrations shows that in the latter case speakers attempt to express closure not just for the coda or ªnal phase (ended up falling) but very often for events which form the onset and nucleus (began digging but after a very short time ...; kept digging and eventually ...; etc.). This diŸerence is illustrated by the following set of examples. In Example 7 speakers unfold a scene which corresponds in perspective to that in Quest 1, presenting the events as ongoing. At the end of the scene, however, they close the sequence by explicitly anchoring the last event either as a temporal endpoint (Example 7, 7) or as a causal result of the preceding discourse (Example 8, 3–4). Note that events which form part of the story line are mapped into a subordinate clause in this function (until he sunk), a packaging phenomenon which is due to the hybrid character of the perspective selected and which is not found in the spontaneous past time narratives.

Two ways of construing complex temporal structures

Example 7. (1) (2) (3) (4) (5) (6) (7)

he looked up at the sky to see if there were rain clouds but didn’t see any so then he dug in the sand he kept digging and digging until he sunk (closure)

Example 8. (1) (2) (3) (4)

he started to dig around and like a cat kind of eh throwing up the sand behind him and he dug so hard that he fell through into a diŸerent kind of equally desolate world (closure)

2. Other traces of the pattern of adaptation can be observed where utterances which are aspectually presented as ongoing are closed by aspectual termination (to end up) or temporal quantification, either in the utterance itself or by the following utterance. Example 9. (1) (2) (3) (4) (5) (6) (7) (8)

and he started to dig but he didn’t ªnd anything and he kept digging and eventually found out that it was too late there was quick sand and he fell through and went somewhere eh under the desert

Example 10. (1) (2) (3) (4) (5) (6) (7) (8)

and then he heard the sound of water dripping somewhere around him so he got up and went in search of the source of this sound and he began digging in the sand and after a very short time of digging he fell through some sort of hole in the sand and moments later found himself on another level of consciousness / or another plane / or another planet (9) and as he was lying on the ground

119

120 Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

(10) (11) (12) (13) (13) (14) (15) (16) (17) (18) (19) (20) (21) (22) (23)

he again heard the sound of water dripping then he found the spot where in fact water was dripping and found a small puddle on the paper and began trying to scoop it up and as he did so he created a hole in this piece of paper and again fell through the hole and went walking around this planet and ended up on top of a rather large pile of rocks so he began to climb down from the rocks and it was quite a high pile so he had di¹culty getting down and ended up falling the last bit of the way

3. The problem of providing an anchor point for a shifted TT can also be solved by adding an explicit temporal shifter such as then or a moment later, or by introducing some indeªnite anchor point through a time adverb such as suddenly. The number of adverbials used in this function increases in Quest 7, measured with respect to the total number of narrative utterances. Table 4.

English Quest 1 and Quest 7: Temporal Adverbs

Quest 1

Quest 7

89/1577 5.6%

149/1591 9.4%

Let us now summarise the ªndings for the English Quest 7 texts, where speakers are forced to adopt a perspective which they normally would not take. They try to comply with the instruction and use past tense forms. They realise that coherence cannot be achieved by using now as constant anchor for TT. This leads to attempts to establish anaphoric links. The perspective under which events are construed for verbalisation and weighted relative to one another within the hierarchical structure of the text is not fully shifted to that of a narrative anchored in the past in anaphoric terms. Information organisation follows a hybrid pattern in which features of a frame anchored in terms of the ‘experiencer’s now’ are marginally reconciled with the requirements of a past time perspective which has no place for this anchor.

Two ways of construing complex temporal structures

In order to give an indication of the contrast with past time narratives which follow an anaphoric temporal perspective, we will close this section by looking at an example of this type. There are not enough examples of selfselected past time narratives in the corpus to carry out a representative analysis (two out of a corpus of 37 Quest 1 texts show this perspective, as mentioned earlier), so its function can only be illustrative. Example 11. (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16)

eh this sort of sand man woke up ehm half buried in sand in the middle of the desert and he wanted some water so he checked his bottle there was none there and he started digging ehm / digging around where the bottle was lying to see if maybe he could gather some and instead he dug a hole and fell through our sand man friend fell through into a sort of ehm paper world where the ground seems to be made of pieces of paper he had heard some water dripping and went to ªnd it he found a little puddle

What we can see in this excerpt is that events are bounded and that the degree of phasal segmentation is lower compared to the other texts. Similarly, the ratio between situations reported under the aspect of ongoingness and situations which are viewed holistically is inversed, compared to the ‘present tense’ texts. The contexts triggering subordination (ongoing events-water dripping, states) provide a further indicator of this diŸerence (see also Carroll and Lambert in press). Although the data base is very limited, there are narratives in which speakers talk about past events using a ‘past time’ perspective. Hence, English speakers can indeed select a perspective which essentially corresponds to the only option German speakers have: the presentation of a sequence of events with an intrinsic temporal link between the single units, whereby phasal segmentation is integrated accordingly.

121

122 Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

5.2 Comparing German Quest 1 and Quest 7 Let us look ªrst at tense switches in the Quest 7 data. All 20 speakers select past tense forms as their dominant tense. Table 5.

German Quest 7: Relative frequency of tense forms

present

past

total number of narrative utterances

94 9.4%

909 90.6%

1003

Switches to the present are less frequent in the German texts as compared to the English texts. However, there is a marked degree of variation, with respect to the two tense options selected in retelling the ªlm — simple past (Präteritum 16 speakers) and Perfekt (Perfekt 4 speakers). Looking at the distribution of the switches to the present tense across subjects, we see that most of the occurrences are produced by 4 speakers, and the dominant tense for those speakers is the Perfekt. But we cannot speak of unstable patterns of TT management in this case, unlike in the English texts. Rather two options which belong to a ‘past tense’ referential frame (anaphoric) are realised in the German texts. One is a past time anchored pattern with the Präteritum as the main tense in which speakers follow what the underlying quaestio indicates: “Tell me, what happened in the ªlm?”. Example 12 and 13 illustrate this pattern. Example 12. (1) also ein kleiner mensch aus knete erwachte in der wüste ‘okay a small man made out of plasticene woke up in the desert’ (2) und sah neben sich eine wassser¶asche ‘and saw beside him a bottle of water’ (3) er hob sie auf ‘he picked it up’ (4) schüttelte sie ‘shook it’ (5) und nichts kam raus ‘and nothing came out’ (6) er setzte sich auf ‘he sat up’ (7) und äh schaute um sich ‘and looked around’

Two ways of construing complex temporal structures 123

(8) und sah nur wüste ‘and saw only desert’ (9) und ªng an zu graben ‘and began to dig’ (10) und dabei wurde das loch unter ihm so gross ‘and thereby the hole under him got so big’ (11) dass er selbst hineinªel ‘that he himself fell in’ Example 13. (1) die szene spielte in einer wüste ‘the scene takes part in the desert’ (2) das sandmännchen lag da bewegungslos ‘the sandman lay there without moving’ (3) dann wachte es auf ‘then it woke up’ (4) und hatte eine ¶asche neben sich ‘and had a bottle beside him’ (5) und ähm suchte wasser ‘and searched for water’ (6) das sandmännchen lag ähm ja im sand ‘the sandman lay in the sand’ (7) und hat nach wasser gegraben ‘and (has) dug for water’ (Perfekt) (8) und hat immer / immer tiefer gegraben ‘and (has) dug always deeper’ (Perfekt) (9) hat sich ein loch gegraben ‘and (has) dug a hole’ (Perfekt) (10) bis ähm / bis das sandmännchen selbst in das loch einsank ‘until the sandman himself into the hole sank’ (11) und verschwunden ist ‘and (has) disappeared’ (Perfekt) (12) das sandmännchen ähm befand sich nun in einer papierwüste ‘the sandman found himself in a desert’ (13) es lag wieder da ‘it lay there again’ (14) und wachte langsam auf ‘and woke up slowly’ (15) und stand auf ‘and stood up’

124 Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

(16) es herrschte ein ziemlich starker wind in der papierwüste ‘there were strong winds in the paper desert’ (17) und es ¶ogen ihm ähm sehr viele blätter an den körper ‘and there ¶ew to him many sheets on his body’ (18) und das sandmännchen ging immer weiter durch die wüste ‘and the sandman continued on through the desert’ (19) bis es irgendwo ähm eine pfütze fand ‘until it found a pool somewhere’ (20) und ähm da hat sich das sandmännchen umgeschaut ‘and there the sandman (has) looked around’ (Perfekt) (21) woher das wasser kam ‘from where the water came’ (22) am anfang hat es gedacht ‘at the beginning he (has) thought’ (Perfekt) (23) es käme von oben ‘it came from above’ (24) und dann / ja es / es / und hat nach oben geschaut ‘and (has) looked up’ (Perfekt) (25) und dann ähm hat er nach der pfütze gegraben ‘and then he (has) dug for the pool of water’ (Perfekt) (26) und er hat weiter gegraben ‘and he (has) dug on’ (Perfekt) (27) bis er wieder äh runterªel in ein loch unter der papier äh pfütze ‘until he again fell down into a hole under the paper eh pool’

Example 13 illustrates the systematicity of tense switches between Präteritum and Perfekt. Before discussing the motivation for these tense switches, we will give an example of the other basic option selected by the German speakers. Some use the Perfekt as the narrative tense with occasional switches to the present. Example 14. (1) das sandmännchen ist in eine welt unter der zweiten welt gefallen ‘the sandman has fallen into a world under the second world’ (Perfekt compatible with present perfect) (2) ähm die diesmal nicht aus papier besteht ‘which this time is not made of paper’ (3) sondern aus / also / quasi eine art steinwüste ist ‘but is almost a stone desert’ (4) und ähm muss aufpassen ‘and he has to be careful’

Two ways of construing complex temporal structures

(5) dass nicht steine von oben runterfallen ‘that stones do not fall down from above’ (6) oder musste aufpassen entschuldigung ‘or he had to be careful, sorry’ (7) und ähm ja manchmal sind auch steine von unten nach oben gedrückt worden (Perfekt) ‘and sometimes stones were pushed up from underneath’ (8) und beim zweiten mal hat dann auch so ein steinberg ihn erwischt (Perfekt) ‘and the second time he was caught up by a column of stone’ (9) der von unten aus dem boden geschossen ist (Perfekt) ‘which had shot up out of the ground’ Example 15. (1) und dann blickt sich das sandmännchen / hat sich suchend umgeblickt ‘and then the sand man looks / has looked around searchingly’ (i.e. has been looking around) (2) nichts gefunden ‘found nothing’ (3) oder / und dann hat es angefangen im sand zu graben ‘and then it (has) begun to dig in the sand’ (Perfekt) (4) wahrscheinlich auf der suche nach wasser ‘probably searching for water’ (5) und hat angefangen zu graben ‘and has begun to dig’ (Perfekt) (6) und zu graben ‘and to dig’ (7) und immer tiefer ‘and continually deeper’ (8) und irgendwann ist es mit hineingerutscht ‘and at some point he (has) slipped into it’ (Perfekt) (9) und die ¶asche mit ihm ‘and the bottle with him’

Most of the German texts are consistent in their use of tense forms in that those which exhibit tense switches remain within a constant frame of reference. This contrasts clearly with the English data. We will briefly look at the two options selected by the German speakers and then come to an interpretation of the Perfekt in its aspectual properties.

125

126 Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

(a) tense switch Präteritum – Perfekt (cf. example 13) The two forms clearly diŸer in their contribution to the temporal interpretation of a discourse.6 In the Präteritum, TT overlaps and, as a default, includes TSIT; in the Perfekt, TT follows TSIT. This is illustrated in the following excerpt: er hat sich ein loch gegraben ‘he (has) dug a hole’ (Perfekt) bis er einsank ‘until he sunk in’

TT after TSIT1 TT maintained and includes TSIT2

Intuitively the use of the Perfekt underlines the eventhood of a situation, since boundedness is one crucial, though not necessary, property of dynamic situations. In texts where the Präteritum is the dominant tense, TT management is based on an anaphoric strategy. There is no shift to an ‘experiencer-frame’. Speakers remain in the past, but they change perspective in a signiªcant way. The shift to the Perfekt occurs either together with speciªc predicate types or at particular points in the narrative macrostructure. Predicate types which trigger the shift are 2–state predicates. They move the story line and refer to the events which form the culmination of the respective scene. (b) tense switch Perfekt – Present As is well-known there are fundmental diŸerences between the German Perfekt and the English present perfect. Very often, it has to be rendered by the English simple past, as shown in the glosses for examples 13/14 above, where the temporal frame is located somewhere in the past or rather independently of any temporal anchor (the fairy tale type of indeªnite temporal reference). The relationship between predicates in the Perfect is therefore not mediated through now — as is the case in English. Rather the propositions are anaphorically related according to the strategy of a TT shift described above. Frequent use of dann (then) as a temporal connector underlines this relationship. We explain the shift to the present tense by the possibility of linking the TT for a Perfekt form to the time of utterance TU, although this is not a necessary concept for temporal linkage. The compatibility between the Perfekt and a TT = TA seems to provide the basis for a switch to the preferred perspective of a TA anchored TT, as shown in the choice of present tense in Quest 1. What we rarely get in German is a tense shift from the Präteritum (simple past) to the present tense. Texts are more consistent compared to the English data. Looking at the other features analysed for the English texts we do not ªnd

Two ways of construing complex temporal structures 127

evidence for con¶icting perspectives either. The number and type of temporal adverbs does not diŸer, the pattern of event construal remains the same and information packaging by subordination re¶ects the same event-state-hierarchy (Carroll and Lambert in press, Carroll and von Stutterheim 2003, von Stutterheim 1997). 5.3 Comparing perspective shift in English and German (Quest 7) The comparison within the two language groups revealed a contrast in the ways in which the Quest 7 texts diŸered from those in Quest 1. In the following we will give a brief overview of the ªndings for the two languages and then come to the discussion of the results. 5.3.1 Survey of the results The two groups diŸer considerably in the level of granularity for Quest 7, as shown in Table 6. Table 6. English German

Length of narrative (average 20 speakers) Quest 1

Quest 7

173.3 174.5

182.2 112.4

The English texts maintain the same level of resolution under both conditions, which includes basically the same proportion of sub-event segmentations. This is taken as an indicator that the English speakers, although using the past tense, do not actually shift the underlying deictic perspective, the ‘now of the experiencer’, as discussed earlier. Unlike in English, the German Quest 7 texts are shorter than the Quest 1 texts. Since there are no other signiªcant indicators which point to a change in perspective, and there are no indicators in the domain of temporal structure, the reason why the texts are shorter has to be sought elsewhere. The next table compares the use of adverbials (such as then, a moment later, suddenly) which mark a temporal shift, showing that although there is an overall increase in their use in Quest 7, the rate of occurrence is still lower in English. In Table 8 the results for German speakers are split with respect to the dominant tense used, and the table shows the rate of occurrence of the present tense in texts in which the Präteritum is the predominant tense form. This provides an indication of the limited role of the present in German, since

128 Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

Table 7.

Frequency of adverbials English and German Quest 1

Quest 7

English

89/1577 5.6%

149/1591 9.4%

German

247/1943 12.7%

228/1003 22.7%

Table 8. English German

Relative frequency of tense forms present tense forms

past tense forms

12.6% 2.7%

87.4% 97.2%

speakers who use both the Präteritum and Perfekt (4/20) account for the majority of present tense forms found in the German narratives, as shown in Table 9. Table 9.

“Präteritum”-speakers and “Perfekt”-speakers

German

present tense forms

past tense forms

“Präteritum”-speakers “Perfekt”-speakers

27.4% 72.3%

72.6 27.7%

The comparison, both inter- and intralinguistically, clearly illustrates the cross-linguistic diŸerences. English speakers tend to switch to the present tense much more frequently than the German speakers (with the exception of the small group of speakers using the Perfekt, as mentioned). German texts are more consistent, exhibiting an anaphoric temporal structure throughout. The occasional shifts in tense forms are not motivated by a switch to another frame and do not occur in contexts which are relevant in this respect (see Table 10). Table 10. Contexts for switches and corrections

English German

narrator phasal states viewpoint segmentation

situation reoccurs

activity protagonist environmental force

15.8% –

4.7% –

34.9% 73.9%

15.8% –

28.5% 26.0%

5.3.2 Discussion The ªgures presented above illustrate the cross-linguistic diŸerences, and we now come to the factors which determine this overall picture. In earlier studies

Two ways of construing complex temporal structures 129

on narratives we sought evidence for the diŸerent principles of construal in the linguistic means which are grammaticised in the languages in question (see Carroll and von Stutterheim 2003, von Stutterheim and Klein 2002, von Stutterheim, Nuese and Murcia Serra 2002). The relevant structural property in the case of narratives lies in the categories grammaticised in the verbal system, as mentioned earlier. The main diŸerences between the German and English verbal systems are not given by tenses but by aspectual categories. These categories are bound to viewpoint (see the deªnition in Comrie 1976, Klein 1994) and can be expected to have consequences for the construal of events in actual language production. English has a marked grammatical category for ongoingness which is not constrained with respect to temporal location, and a perfective aspect which is constrained with respect to the deictic TA. The imperfective aspect requires an external reference point, a topic time in relation to which a situation is viewed as ongoing. This can be the now of an experiencer, as outlined in Section 4. But there is no basic constraint in English which excludes the choice of another reference point, as evidenced in the spontaneous past time narratives. The decisive factor which determines the selection of one or the other option may lie in the nature of the situation. In other words, if the discourse situation suggests an external perspective (recount the content of a ªlm), and the language provides grammaticised means which can easily anchor a temporal frame based on the ‘now of the experiencer’ — speakers are more likely to avail of this option in the construal of events. From this follows the need for an external TT, independent of TSIT, anchored in terms of the now of the observer or experiencer. German speakers do not have a grammaticised imperfective aspect. The means available for aspectual marking predominantly encode the perfective viewpoint (verbal system and morphological means).7 The focus on the endpoint and post state of a situation provides the grounds for an anaphoric strategy in TT management. The interference of TA is secondary in that a post state perspective allows for a convergence of TA and TT. There are occasional switches to the deictic anchor now which, however, do not aŸect the status of the overall temporal frame. These considerations lead to the conclusion that given diŸerent grammatical systems, speakers develop speciªc principles in the construction of referential frames and with this the types of event-time relations which they incorporate. We put forward the hypothesis that speakers of languages with the imperfective as a grammaticised category have a preference for locating situa-

130 Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

tions within an externally anchored frame. Speakers of languages with no systematic aspectual system encode situations holistically, and this provides the relevant informational components for anaphoric referential progression in the temporal domain. As illustrated in the analysis, speakers of German do not set up a global referential frame (i.e., which holds for the narrative as a whole) that is anchored in respect of an external viewpoint. The grammaticised means which facilitate this on a systematic scale are absent. This cross-linguistic contrast is not trivial since speakers are, on the whole, bound by the logic of a consistent frame in order to ensure the basic requirement of text structure — its coherence. The cross-linguistic diŸerences thus lie in the overall referential frame which speakers establish, and there is a logical interrelation between the type of temporal relations selected and speciªc patterns of event construal. The language-speciªc grammatical means relate not only to what Levelt (1996) has termed microplanning, they also determine planning at the macrolevel where speakers “consider what to say”, since there are also signiªcant cross-linguistic diŸerences in the type of information which is selected for mention in the narratives (cf. Carroll and Lambert in press; von Stutterheim, Nuese and Murcia Serra 2002). These wide ranging implications of grammatical structure for information organisation in language production have to be recognised in order to explain the complex but consistent picture revealed by the empirical facts.

6.

Conclusion

In our analysis we have brought together levels which are traditionally the object of diŸerent linguistic sub-disciplines. The mediating locus is the conceptualiser, the level in language production which requires the adaptation of extra-linguistic content to linguistic structure, in particular to grammatical structure (Levelt 1989, 1996). Each grammatical system provides a speciªc format which the speaker, when trying to solve a complex verbal task, has to comply with. Just as some formal languages are better qualiªed for solving certain types of tasks, given their speciªc syntactic properties, natural languages will also re¶ect similar constraints. The general logical principles underlying all artiªcial languages are probably universal, just as the reservoir of possible categories and relations in natural language syntax can be formulated at an abstract universal level. A speciªc grammar, however, manifests a weighted selection of this elementary

Two ways of construing complex temporal structures

system and thus focuses certain categories and relations, while others are doomed to irrelevance. It is important to stress that there are probably no absolute constraints at a representational level, but it is much more eŸective to use a format which is tuned to the speciªc task. Our ªndings with English and German reveal a similar type of implication. Grammaticised categories point the way to an optimal format for processing conceptual substance. Thus the optimal format for organising information diŸers according to the grammatical system. This is what the data have shown. On the side of the speaker the optimal format for encoding time-event structures is re¶ected in a clear preference for one referential frame over another.

Notes 1. Note that these features hold for texts in everyday communication. In poetic or literary texts they can be violated (poetic license) in order to achieve eŸects which go beyond informativeness. 2. In Klein (2000a), it is argued that we actually should speak of an ‘argument-time structure’, since various time spans can be relative to certain arguments. The transitive form to close, for example, has two time spans for the direct object and one time span for the subject. Since this plays no role in the present context, we shall stick to the simpler notion of ‘temporal structure’. Note, incidentally, that the ‘two-state verbs’ in Klein (1994) represent a special — and actually a simple — case of such a temporal structure. 3. This temporal anchor can be projected onto any point on the time axis, a phenomenon which is often called secondary deixis (cf. Klein 2001). 4. A distinction must be drawn between the temporal structure of what is shown in the ªlm and a particular speaker’s representation of it, since the latter constitutes the actual speaker’s text. 5. Hopper (1995) notes that English speakers often talk about factive events in the past under the perspective of the current now. 6. Opinions vary considerably on how to analyse the German Perfekt and how it is related to the Präteritum. The problems are eminent, and we shall not try to examine them here but just look at how speakers indeed use these forms in the given task (but see Klein 2000a for a discussion and re-analysis). 7. On the role of aspectual categories in L1 acquisition (English/German) see the discussion in von Stutterheim, Nuese and Murcia Serra (2002).

131

132 Christiane von Stutterheim, Mary Carroll and Wolfgang Klein

References Bäuerle, Rainer 1979 Temporale Deixis — temporale Frage: Zum propositionalen Gebrauch deklarativer und interrogativer Sätze. Tübingen: Narr. Brown, Gillian and George Yule 1983 Discourse Analysis. Cambridge: Cambridge University Press. Carroll, Mary and Monique Lambert in press “Information structure in narratives and the role of grammaticised meanings: A study of adult French and German learners of English”. In Information structure, linguistic structure, and the dynamics of acquisition, C. Dimroth and M. Starren (eds). Amsterdam: Benjamins. Carroll, Mary and Christiane von Stutterheim 2003 “Typology and information organisation: perspective taking and languagespecific effects in the construal of events”. In Typology and Second Language Acquisition, A. Ramat (ed), 365–402. Berlin: de Gruyter. Comrie, Bernhard 1976 Aspect: An introduction to the study of verbal aspect and related problems. Cambridge: Cambridge University Press. Fabricius-Hansen, Cathrine 1986 Tempus fugit: Über die Interpretation temporaler Strukturen im Deutschen. Düsseldorf: Schwann. Garrod, Simon C. and Anthony J. Sandford 1988 “Discourse models as interfaces between language and the spatial world”. Journal of Semantics 6: 147–160. Goldberg, Adele E. 1996 Conceptual Structure, Discourse and Language. Standford: CSLI Publications. Hopper, Paul J. 1995 “The Category ‘Event’ in Natural Discourse and Logic”. In Discourse Grammar and Typology, W. Abraham, T. Givon and S. Thompson (eds), 139–153. Amsterdam: Benjamins. Kamp, Hans and Uwe Reyle 1993 From Discourse to Logic: Introduction to Modeltheoretic Semantics of Natural Language, Formal Logic and Discourse Representation Theory. Dordrecht: Kluwer Academic Publishers. Kintsch, Walter and Teun A. van Dijk 1978 “Toward a model of text comprehension and production”. Psychological Review 95: 163–182. Klein, Wolfgang 1994 Time in Language. London: Routledge. 2000a “An analysis of the German Perfect”. Language 76 (2): 358–382. 2000b “Was uns die Sprache des Rechts über die Sprache sagt”. Zeitschrift für Literaturwissenschaft und Linguistik 118: 116–148.

Two ways of construing complex temporal structures

2001

“Deiktische Orientierung”. In Sprachtypologie und sprachliche Universalien. Ein internationales Handbuch, Vol. 1/1, M. Haspelmath, E. König, W. Oesterreicher and W. Raible (eds), 575–590. Berlin: de Gruyter. Klein, Wolfgang and Christiane von Stutterheim 1987 “Quaestio und referentielle Bewegung in Erzählungen”. Linguistische Berichte 109: 163–183. Labov, William 1972 “The transformation of information in narrative syntax”. In Language in the inner city, W. Labov (ed), 354–396. Philadelphia: Philadelphia University Press. Levelt, Willem J. M. 1989 Speaking: From Intention to Articulation. Cambridge, Mass.: The MIT-Press. 1996 “Perspective Taking and Ellipsis in Spatial Descriptions”. In Language and Space, P. Bloom, M. A. Peterson, L. Nadel and M. F. Garrett (eds), 77–107. Cambridge, Mass.: The MIT-Press. Ochs, Elinor, Carolyn Taylor, Dina Rudolph and Ruth Smith 1992 “Story-telling as a theory-building activity”. Discourse Processes 15 (1): 37– 72. Partee, Barbara H. 1983 “Temporal and Nominal Anaphora”. Linguistics and Philosophy 7, 243–286. Paul, Hermann 1882 Principien der Sprachgeschichte. Jena: Niemeyer. Reichenbach, Hans 1947 Elements of Symbolic Logic. London: Free Press. Roßdeutscher, Antje 2000 “Lexikalisch gestützte formale Textinterpretation”. Arbeitsberichte des Sonderforschungsbereichs 340, 117. Tübingen (Habilitationsschrift Universität Stuttgart). Smith, Carlota T. 2000 The Domain of Tense [International Round Table “The Syntax of Tense and Aspect”]. Ms. Paris. Stutterheim, Christiane von 1997 Einige Prinzipien des Textaufbaus: Empirische Untersuchungen zur Produktion mündlicher Texte. Tübingen: Niemeyer. Stutterheim, Christiane von and Wolfgang Klein 2002 “Quaestio und l-perspectivation”. In Perspectivity and perspectivation in discourse, C. F. Graumann and W. Kallmeyer (eds), Amsterdam: Benjamins. Stutterheim, Christiane von and Monique Lambert in press “Crosslinguistic analysis of temporal perspective in text production”. In The structure of learner varieties, H. Hendricks (ed). Berlin: de Gruyter. Stutterheim, Christiane von, Ralf Nuese and Jorge Murcia Serra 2002 “Crosslinguistic differences in the conceptualisation of events”. In Information structure in a cross-linguistic perspective, B. Behrens and C. FabriciusHansen and S. Johansson (eds), 179–198. Amsterdam: Rodopi.

133

“Look here, what I am saying!” Speaker deixis and implicature as the basis of modality and future tense Thomas A. Fritz University of Passau/Bavaria

1.

Introduction

It is widely recognized among linguists that deixis plays a paramount role in the use and understanding of everyday language. Generally speaking, deixis is concerned with the relationship between language and context as re¶ected in the structure of the language itself (Levinson 1983: 54f.). A more restricted view of deixis focusses on the referential usages of closed paradigmatic sets of deictic expressions, such as morphemes or strings of morphemes (Hanks 1992: 46f.). From a semiotic point of view, these expressions can be referred to as signs. Here, I will argue that both the system of modality and the concept of future tense are based on a speaker pointing deictically to himself. I propose to show that all sentences contain at least one such deictic sign in the form of the modal morpheme. The paradigm of modal morphemes (indicative, subjunctive, imperative; modal verbs; modal adverbs) makes the pragmatic consequences of speech explicit and permits variation of the modal quality of a sentence. The indicative is known to establish two separate meanings of the sentence: the factuality of the proposition and the speaker’s belief in it. For a sentence with a ªnite indicative predicate I use the term ‘saying’, which is the basic term of this article. It has to be distinguished from ‘stating’, which only applies to the proposition of the sentence. On speech act level, ‘asserting ‘ must be considered: it implies the consequences of ‘saying’ for the speaker in terms of factuality and speaker belief. In fact, ‘saying’ bridges the gap between ‘stating’ and ‘asserting’ by regulating the

136 Thomas A. Fritz

consequences of spoken language on hearers. In the following chapters, this article is concerned with the rules that govern the process of ‘asserting’ as re¶ected by the modal morphemes on the level of ‘saying’. 2.

Speaker belief and commitment from speaker deixis

The problems which arise from an attempt to trace ‘asserting’ back to grammatical signs such as the indicative are well-known: there is no isomorphic mapping of form and function between the sign of the indicative and the communicative consequences of ‘saying’, i.e. factuality and speaker belief. Assertions only constitute a subset of the functions of the indicative, as indicatives can in fact transport all kinds of speech acts (Price 1988: 43–59). On the other hand, however, the indicative is the strongest and perhaps the only way to make an assertion. Linguistics textbooks on mood and modality qualify the indicative as unmarked in contrast to modal features that realise particular modal meanings (Palmer 1986). The markedness theory of the indicative, as it is sometimes called, is based on structuralist approaches by Roman Jakobson (1971) that link the strongest modal meaning (unlimited factuality and speaker belief) to the weakest, i.e. the unmarked modal sign. The realisation of this meaning, however, depends on all other modal signs in the sentence (and the situation) being taken into consideration by the hearer. It can only come into eŸect, if there is no marked modal sign that renders the realisation of the basic meaning impossible. The meaning of the indicative is therefore often called a default meaning: it is “accepted in default of any evidence to the contrary” (Leech 1983: 42). The most astonishing meaning of the indicative is the speaker’s belief in the factuality of the proposition. Ever since George E. Moore’s paradox (1968: 541–543), we have known that this meaning cannot be part of the proposition ‘stated’, as there is no real contradiction between the proposition and a subsequent statement of the speaker’s disbelief in it: *It is raining, but I do not believe it. Moore holds that this is just an absurd use of language. In our context, speaker belief must be regarded as a non-propositional consequence of modality, which is due to ‘saying’. Formally, it could be interpreted as an implicature meaning; H. Paul Grice (1989: 42), however, who introduced the class of ‘conventional implicatures’ for pragmatic meanings that are constantly linked to particular signs, was himself fairly critical of treating the consequences of the indicative as an implicature:

“Look here, what I am saying!” 137

“On my account, it will not be true that when I say that p, I conversationally implicate that I believe that p; for to suppose that I believe that p (or rather think of myself as believing that p) is just to suppose that I am observing the ªrst maxim of Quality on this occasion. I think that this consequence is intuitively acceptable; it is not a natural use of language to describe one who has said that p as having, for example, ‘implied’, ‘indicated’, or ‘suggested’ that he believes that p; the natural thing to say is that he has expressed (or at least purported to express) the belief that p. He has of course committed himself, in a certain way, to its being the case that he believes that p, and while this commitment is not a case of saying that he believes that p, it is bound up, in a special way, with saying that p.”

According to John Lyons (1977: 734) speaker belief is part of the commitment made by the speaker: “Commitment […] does not mean that we must believe that what we say is true; still less does it mean that the asserted proposition is in fact true. Commitment is independent of sincerity and truth; it is a matter of appropriate behaviour.”

From a deictic perspective, commitment can be regarded as the pragmatic consequence of the speaker referring deictically to himself. In fact, speaker belief is not related to the indicative alone, it is part of every instance of ‘saying’, as we would always suppose a speaker to have good reasons, when he enters into communication. Particularly, if the speaker reduces his commitment to factuality, he still has to take full responsibility for the actual amount of this reduction. He even has to account for the diŸerence between various epistemic modalisations, e.g. the stronger epistemic must be in contrast to the weaker might be. Speaker deixis therefore seems to be a basic communicative procedure that is unseparably linked to the social norms which govern the process of ‘saying’ and the commitments which result from it (Brandom 1994: ch. 3).

3.

Factuality from scalar implicature

The indicative communicates the proposition as factual, if there are no explicit signs in the sentence or the situation that exclude this meaning. Not only signs from the modal paradigm such as subjunctives or modal verbs, but also particular contexts, e.g. quotation and ªction, destroy the hearer’s right to read the indicative as factual. In the following, I will concentrate on the paradigm of modal signs that enable the speaker to vary the amount of factuality to which he wants to be committed. Such a variation of asserting is not compatible with a strictly logical view, at least not with the Fregean approach. According to Gottlob

138 Thomas A. Fritz

Frege (1967: 342–346), you can neither ‘not assert’, i.e. ‘negate’, nor ‘assert only a bit’: in his system, ‘asserting’ is an atomistic and uniform operation. On the other hand, the signs that communicate commitment to limited factuality cannot be part of the proposition, either. Ludwig Wittgenstein (1984: 517) made this point clear when he insisted that we should not misinterpret a ‘tentative claim’ as a ‘claim of tentativeness’. If modality is neither in the range of ‘asserting’, as far as this is interesting for logicians, nor belongs to ‘stating’, then it must be of a predominantly pragmatic nature. No methodology, however, has so far been developed for the pragmatics of ‘saying’: Aristotele’s and Frege’s logic systems only give rules for ‘stating’. In order to explain the ‘saying’ in the context of the paradigm of modality, pragmatic approaches have to be applied. The factual part of modality is best explained by the general theory of ‘scalar implicature’. Based on the ideas of Grice (cf. 1989: 32–37), Gazdar (1979), Horn (1989) and Levinson (1987; 2000) this theory holds that systems of semantic scales undergo regular pragmatic reasoning processes. If the scales are tight enough, a weaker sign regularly adopts the negation of the stronger sign in the paradigm as an implicature meaning. The implicature meaning becomes a sort of second meaning to the sign. In order to harmonise Jakobson’s concept of modal markedness with a pragmatic model of scalar implicatures, I interpret the asymmetrical markedness conditions of modality as semantic scales. Interestingly, this scalarity is also assumed in the classical arrangement of modalities in the logical square, which oŸers a part-wholerelationship between ‘possibility’ ‘P’ (can do) and ‘necessity’ ‘N’ (must do). The unmarked indicative (the ‘saying’ of the proposition ‘p’) (does) has to be understood as the basic sign of the system: must do ‘NECESSITY’ ‘N p’ can do ‘POSSIBILITY’ ‘P p’ does ‘INDICATIVE’ ‘p’

Figure 1. The markedness relations of modality.

“Look here, what I am saying!” 139

This does not fully comply with logical traditions, which hold that ab necesse ad esse valet consequentia (‘N p → p’) and ab esse ad posse valet consequentia (‘p → P p’). From the linguistic marker standpoint, however, the scalar order must be ‘N p > P p > p’, with the indicative as the basic element of the paradigm: must do ‘SAY NECESSARY’ ‘N p’

can do ‘SAY POSSIBLE’ negates the basic meaning of ‘NECESSARY’ and acquires the implicature meaning ‘NOT-NECESSARY’ ‘P p +> Not-N p’

does ‘SAY’ negates the implicature meaning of ‘NOT NECESSARY’ and acquires the implicature meaning ‘NECESSARY’ ‘p +> N p’

Figure 2. The scalar implicatures from ‘saying’ the signs of the modal system.

In this way, the factuality of the indicative represents the epistemic version of ‘necessity’ after two implicatures, starting from a sign marked ‘necessary’. The ªrst implicature (symbolised by the bold line and the arrow directed to the meaning that is negated) is based on the scale value of must, i.e. ‘necessity’, which results in ‘non-necessity’ as the implicature meaning of can (‘P p +> Not-N p’). In the second implicature, the negation aŸects this implicature value and allocates the strongest modal meaning, ‘necessity’ to the unmarked ‘saying’ (‘p +> N p’). These results are intuitively plausible. Aristotle was the ªrst to formulate a version of ‘possibility’, often called ‘natural possibility’, that included ‘uncertainty’. It combines the possible ‘can happen’ with the contingent ‘need not happen’. As to the epistemic ‘necessity’ of the indicative, it can be explained by the absence of any sign that carries the implicature value of ‘non-necessity’. The negations resulting from scalar implicature always aŸect the modal meaning, a propositional negation is never part of the reasoning process. This makes the attempt compatible with the Fregean concept of assertion, because all the negations concerned are either propositional or pragmatic. The most

140 Thomas A. Fritz

pragmatic meaning of all is the unlimited factuality that is linked to the unmarked ‘saying’, the indicative then acquires the value of epistemic ‘necessity’. Taken for itself, however, it is only ªniteness. Marked signs add meanings of their own, e.g. the subjunctives in Germanic languages, which are geared to expressing speaker evidentiality and hypothesis. That the value of a sign does not only depend on the sign itself, but also on the alternatives to this sign in the respective paradigm, is the basic idea of linguistic structuralism. Modelling this idea in the theory of implicature resembles a rediscovery of Roman Jakobson’s (cf. 1971: 1–4) basic thoughts on the asymmetry of sign relations. The scalar implicatures of the system of modal markedness are based on the asymmetrical relationship between signs of stronger determination and signs of weaker determination. If the scales are su¹ciently tight, the ‘saying’ of a weaker sign communicates the absence of the stronger meaning. Thus the system of modality is based on two pragmatic concepts: (1) Speaker deixis from ‘saying’ is the trigger of the speaker’s commitment to his belief in the factuality of the proposition. (2) Scalar implicature from ‘saying’ is the mechanism that determines the amount of factuality to which the speaker has to commit himself. The two concepts meet in the sign of the indicative, which is basically deictic, before it enters the pragmatic reasoning process as the sign at the centre of the scalar hierarchy of modality.

4.

Focussing modal verbs on epistemic and non-epistemic meanings

On the basis of the markedness conditions highlighted, a distinction between epistemic and non-epistemic uses of modal verbs is possible. These two ways of making sense of a modal verb can be compared best in a sentence with must: thus, He must be in the kitchen can be interpreted either as a tentative assertion by the speaker, in which the modal verb must takes on the meaning of ‘nonnecessity’, or as a sentence expressing the speaker’s commitment to unlimited factuality. We like to think of the latter use, which makes must a part of the proposition ‘stated’, as the ‘normal’ one. From the pragmatic procedures introduced here, however, it should be clear that both sentences undergo the same reasoning process.

“Look here, what I am saying!”

The two ways of using a modal verb can be described as the contradictory results of focussing in the verbal construction either on unmarked ªniteness or on the meaning of the modal verb. Whereas ‘normal’ ªnite verbs have only one modal marker, modal verbs are marked twice. The second marker, the meaning of the modal verb itself, is bound to interact with the basic modal value expressed by ªniteness alone. Epistemic uses of modal verbs integrate the modal verb into the speaker deixis. Thus, it becomes possible for the speaker to make assertions which do not commit him to full factuality.

must do ‘SAY NECESSARY’ ‘N p’ does ‘SAY’ ‘p +> N p’

must do ‘SAY NECESSARY’ ‘N p +> Not-N p’ does ‘SAY’ ‘p +> N p’

Figure 3. ‘Non-epistemic’ and ‘epistemic use’ of the modal verb must. A scalar reconstruction of the reasoning processes applied can be depicted as in Figure 3: ‘Saying necessary’ in a sentence interacts with simple ‘saying’. If the modal verb is interpreted like any other verb, it is combined with the commitment to unlimited factuality from the indicative and is used non-epistemically. The border between ‘whole’ and ‘part’ disappears, the modal verb enters into the meaning of unmarked ªnitenes (Figure 3, left). In the epistemic use, must negates the implicature value of the indicative of epistemic ‘necessity’ and aquires the meaning of ‘non-necessity’. In the ªgure on the right, this negation is again symbolized by the bold line and the arrow that stops at the border of the ‘part’. The pragmatic reasoning process then acquires a meaning that is contradictory to the basic meaning of the sign. In order to make the contradictory nature of implicature clearer, it is helpful to look more closely at the logical basis of modality. It is well-known that modality is a tautological concept. ‘Possible p’ is equivalent to ‘notnecessary not-p’, at the same time ‘necessary p’ is equated with ‘not-possible not-p’. We can hardly expect the modal reasoning of natural language to be less circular than the logical concept behind it. What we can, however, observe is that natural language views the hierarchy of ‘necessity’ and ‘possibility’ from two perspectives:

141

142 Thomas A. Fritz

‘SAY NECESSARY’ ‘N p’

‘SAY POSSIBLE’ ‘+> Not-N p’

‘SAY POSSIBLE’

‘SAY’

‘+> Not-N p’

‘+> N p’

Figure 4. The markedness perspectives of language on ‘necessity’ and ‘non-necessity’.

For natural language, the interesting opposition of the modal paradigm is the contradictory relationship between ‘necessity’ and ‘non-necessity’. This opposition is represented by the implicatures derived from two scalar hierarchies: the hierarchy on the left of Figure 4 is the hierarchy that yields ‘non-necessity’ as the implicature meaning for ‘possibility’. On the right, the reasoning starts from the basic opposition ‘possibility’ vs. ‘indicative’, which is interpreted as ‘non-necessity’ vs. ‘necessity’, as soon as the implicatures hold. We see how language manages to treat ‘non-necessity’ as part of ‘necessity’ and vice versa within the same sign system. The ‘logic’ of this reasoning process makes pragmatic use of the asymmetry of the code. The part-whole relationship between unmarked and marked signs in the modal paradigm is interpreted as contradictory. This reasoning always starts from an ‘implicatum’ that is essentially only a part of the ‘implicans’. At the end of the reasoning process, the part has adopted the position called the ‘subcontrary’ in the logical square; it is then in a contradictory position to the ‘implicans’:

‘Implicans’ contradictory

‘Implicatum’

‘Subcontrary’

Figure 5. From ‘implicatum’ to ‘subcontrary’ in the logical square.

This implicature is generally possible in logical squares, it determines the pragmatic meanings of some +> ‘not all’, sometimes +> ‘not always’, also +>

“Look here, what I am saying!” 143

‘but not only’ and their respective inversions not all +> ‘but some’, not always +> ‘but sometimes’, not only +> ‘but also’. Lawrence R. Horn (1989), who has scrutinised the pragmatics of the logical square in detail, calls this procedure an ‘R-implicature’, ‘R’ standing for relevance. Stephen C. Levinson (1987; 2000) prefers to speak of ‘Q’ instead of ‘R’, because the reasoning behind this implicature follows from the application of the Quantity maxims set up by Grice.

5.

Future and future tense

It has often been remarked that the future tense does not just play a role in the system of temporality, as many languages use the future tense not only to speak about facts in the future, but also to make modally qualiªed statements about propositions which are related to the present or the past (Comrie 1989). The latter uses resemble the epistemic variants of modal verbs, although the future tense is not, of course, formed with a modal verb in all languages: He will be in the kitchen. Er wird in der Küche sein. Il sera dans la cuisine. As the future tense can be used both for commitments to full factuality (in the future) and to limited factuality (without a clear connection to temporal reference), we can expect some pragmatic process to distinguish these contradictory uses. In Germanic languages, the future tense has no real tradition. Originally, they had no particular sign to express the future at all; formal ways of expressing the future with the verb start to be grammaticalised towards the end of the Middle Ages. Even today, all Germanic languages allow the use of present tense for the reference to future events. The relevant paradigms are, however, complex, and such a use of the present tense normally conveys additional meanings. The situation in English is complicated further by the existence of two parallel forms of the present tense, I do and I am doing. In order to avoid lengthy comparisons of the futures in various Germanic languages, I will concentrate now on the options for expressing the future in German (cf. in detail Thomas A. Fritz 2000: 169–185). An attempt to model the contradictory meanings of the werden-future in the paradigm of the marked (wird tun, i.e. ‘will do’) and the unmarked (tut, i.e. ‘does’) ‘saying’ yields a result that resembles the diŸerence between nonepistemic and epistemic modal verbs:

144 Thomas A. Fritz

wird tun ‘SAY SAYING’ ‘SAY p’

wird tun ‘SAY SAYING’ ‘SAY p +> Not-N p’

tut ‘SAY’

tut ‘SAY’

‘p +> N p’

‘p +> N p’

Figure 6. ‘Future’ and ‘epistemic’ meaning of wird tun.

In the case of commitment to full factuality, marked wird tun includes the implicature linked to the unmarked ‘saying’ and speciªes the meaning of epistemic ‘necessity’ (Figure 6, left). If the future tense is used as an epistemic modal, it negates the consequences of simple ‘saying’, which leads to ‘nonnecessity’ (right). As the German werden has no basic modal meaning such as ‘necessity’ or ‘possibility’, making it part of an epistemic process that limits the factuality of the commitment should presuppose that is has a deictic quality in itself. That there is strong evidence for the basically deictic quality of the future tense becomes obvious when we turn to the pragmatic meanings of the unmarked sign, i.e. simple ‘saying’ or the indicative. Jakobson (1971: 15) held that what is called a Q-/R-implicature today is not the only way of making sense of a markedness opposition. If the unmarkded sign takes on the meaning of the marked sign, rather than negating it, Horn (1989) and Levinson (1987; 2000) speak of an information- or I-implicature. This reasoning process makes the weaker sign usable for a more informative sign further up in the paradigm. A good example of an I-implicature is the use of unmarked ‘saying’ instead of marked ‘future tense’, as shown in Figure 7 on the left:

wird tun ‘SAY SAYING’ ‘SAY p’

wird tun ‘SAY SAYING’ ‘SAY p’

tut ‘SAY’

tut ‘SAY’

‘p +> SAY p’

‘p +> Not SAY p’

Figure 7. ‘Future’ and ‘speaker averted’ meaning of the indicative tut.

“Look here, what I am saying!” 145

Here, tut makes the explicit use of the future tense redundant; its quality of actual ‘saying’ and its pragmatic consequences make it su¹cient for this particular meaning. To show that the system of implicatures is in fact complete, there should be uses of tut that negate the explicit speaker deixis of wird tun in a Q-/R- implicature. This case makes the system of implicatures symmetrical, although the underlying markedness conditions are not. It is the use of the present tense, i.e. unmarked ‘saying’, for events in the future which may or may not materialise irrespective of any action taken by the speaker. In German, it is far more natural to say, for example, Am Freitag feiere ich meinen 36. Geburtstag (‘On Friday I celebrate my 36th birthday’) than Am Freitag werde ich meinen 36. Geburtstag feiern (‘On Friday I will celebrate my 36th birthday’), as Heinz Vater (1994: 75) has pointed out. Because the fact of the birthday is independent from the intentionality of the speaker towards it, it would be unnatural for the speaker to point at himself when referring to it. The rule generally applies to so-called ‘scheduled events’. Wording them with the werden-future would make us expect some engagement of the speaker, which could be an emotional one, for example, that he fears that it may be his last birthday. In English, the present tense for future meaning seems to be associated with the absence of explicit speaker engagement, too. To say This Friday, the train is leaving not at eight but at eight thirty is a typical way of expressing the future, if the focus lies on the facts noted in the timetable. The fourfold implicature system of the paradigm of marked and unmarked ‘saying’ does not so far explain why explicit ‘saying’ should lead to reference to the future, if interpreted in the context of full commitment. To explain this the following reasoning can be applied: The use of a marked sign such as werden has to be justiªed by the speaker. He must have a good reason for ‘saying’ more than the unmarked sign. In the case of explicit modality, this justiªcation of an extra sign can be found most easily in the facts contained in the proposition. This is how the non-epistemic meanings of modal verbs are decoded. Attributing the modality of modal verbs to the speaker and the process of ‘saying’ seems to be the second attempt, after the hearer has failed to include the modality into the proposition. Modal verbs are then interpreted as epistemic. In the case of the future tense, interpreting werden as ‘neutral future’ seems to be the ªrst justiªcation of explicit speaker deixis. If this meaning is barred, the interpretation of werden as an epistemic modal is considered.

146 Thomas A. Fritz

6.

Code and creativity, or: Don’t ‘PUSH’, I want to ‘PULL’!

The paradigm of ‘saying’ contains two elements, marked and unmarked ‘saying’. On the basis of this code, pragmatic reasoning develops four pragmatic meanings. The pragmatic interaction that we can observe is explained here from the part-whole-relationship of the signs: the unmarked sign is implied by the marked sign, but not vice versa. The two signs a and ab share a common meaning ‘A’, which is speciªed by the meaning ‘B’ in ab:

ab ‘AB’

a ‘A’

Figure 8. ‘Implicans’ ab with the meaning ‘AB’ and ‘implicatum’ a with the meaning ‘A’.

If the implication relationship is interpreted from the side of the marked sign, I will call the pragmatic reasoning a ‘PUSH procedure’. The speaker pushes a particular meaning by using the more explicit sign. Through either negation or a¹rmation of the ‘implicans’ these ‘PUSH meanings’ are contradictory to each other. If the marked sign is assumed to include the unmarked meaning, semantic ‘implication’ (A) is achieved. This is the standard interpretation of a markedness hierarchy. In our modal examples, this type of interpretation leads to non-epistemic uses of modal verbs and to the meaning of the ‘neutral’ future assigned to the future tense. If, however, the ‘whole’ pragmatically negates the ‘part’, the marked sign ab takes on the meaning ‘Not-A’. Our examples for this reasoning, which I would like to refer to as ‘marked dominance’ (B), are the epistemic uses of both modal verbs and the future tense. They negate the commitment to full factuality linked to the use of the unmarked sign.

“Look here, what I am saying!” 147

ab ‘AB’

a ‘A’

A) Implication

ab ‘AB’ +> ‘Not-A’ a ‘A’

B) Marked dominance

Figure 9. ‘PUSH-meanings’ from the use of the marked sign ab with the meaning ‘AB’ in contrast to the unmarked sign a with the meaning ‘A’.

ab ‘AB’

ab ‘AB’

a ‘A’

a ‘A’

+> ‘AB’

+> ‘Not-AB’

C) Unmarked dominance

D) Implicature

Figure 10. ‘PULL-meanings’ from the use of the unmarked sign a with the meaning ‘A’ in contrast to the marked sign ab with the meaning ‘AB’.

The full impact of the pragmatic processes becomes visible if ‘PULL meanings’ are considered. In this case, the unmarked sign derives its pragmatic meaning from the background of the paradigm, which it can again either a¹rm or negate. A¹rmation of the meaning of the marked paradigm partner produces ‘unmarked dominance’ (C). The sign then adopts a meaning for which it is actually not coded. This is the type generally referred to as I-implicature. Simple ‘saying’ with reference to the future is an example for it. The term ‘implicature’ (D) is here reserved for Q-/R-implicatures which negate the meaning of the marked sign and thus specify the meaning of the unmarked sign. This variant shows when simple ‘saying’ is used to convey that the factuality of forthcoming events is not linked to the intentionality of the speaker, the indicative then negates the explicit speaker deixis of the marked sign werden. It has been observed that ‘PULL meanings’ play an important role in our communications. Levinson (1987: 79) summarised their importance in the

148 Thomas A. Fritz

conclusion: “The less you say, the more you mean.” In fact, the most important meaning in the modal ªeld is a ‘PULL meaning’ of the ‘implicature’ (D) type. The interpretation of simple ‘saying’ as a commitment to full factuality interprets the empty sign of the indicative in contrast to the rest of the modal paradigm, which is not present in the phrase, but in our knowledge of the code. It is interesting to consider whether this ‘PULL meaning’ of commitment to full factuality can be easily replaced by a ‘PUSH stategy’. Can, in other words, the speaker achieve the consequences of simple ‘saying’ by using marked signs expressing speaker belief and commitment explicitly? An attempt to mark a sentence in this way tends to make it vulnerable to epistemic interpretations on the part of the hearer. In fact, there is no epistemic use of modal verbs that the hearer is prevented from interpreting as an instance of ‘non-necessity’, i.e. as epistemic uncertainty. The introduction of ‘objective epistemic uses’ into modality neglects the fact that the ªnal decision to interpret a modalisation lies with the hearer; she need not follow logical reasoning, if she doubts the knowledge behind the speaker’s ‘saying’. Thus, the speaker is well advised to rely on the ‘PULL mechanism’ instead of turning to ‘PUSH strategies’, when it comes to communicating commitment and factuality. There is no way of expressing this meaning as safely and eŸectively (and e¹ciently) as by simple ‘saying’ without any further signs. Being explicit in this ªeld is not a successful strategy. Observations like this should be seen in the context of the ‘movement towards explicitness’ that is about to permeate our everyday communication culture. Politicians, who are often expected to make their commitment to factuality explicit, sometimes resort to weird constructions so as to do so without appearing to be uncertain. The reason is that we normally only say that we are certain, when we are not.

7.

Markedness and implicature

To derive scalar implicatures from markedness relations gives insight into the nature of this kind of pragmatic reasoning and its link to the code. If markedness is considered to be the basis of the scalar hierarchy, it becomes evident that the method of pragmatic reasoning to be applied for a particular meaning depends on the marking structure of the code. As marking is typically asymmetric, there are two ways of coding the semantic structure. We resort here to Jakobson’s example of ‘male’ and ‘female’ animals and use a lion and a lioness. Here the ‘empty sign’ in respect to gender is lion, lioness is marked ‘female’. It is

“Look here, what I am saying!” 149

always possible to use the unmarked sign lion in the sense of ‘unmarked dominance’ (C) (I-implicature) in order to refer to a ‘female’ animal of this species. On the other hand, we obtain the enriched ‘implicature’ (D) (Q-/Rimplicature) meaning of ‘male’ for lion, if it negates the marked sign lioness pragmatically:

lioness ‘FEMALE LION’

lioness ‘FEMALE LION’

lion ‘LION’

lion ‘LION’

‘ +> FEMALE LION’

‘ +> MALE LION’

Figure 11. Pragmatic meanings of the unmarked lion in contrast to marked lioness.

If, however, the meanings were marked in the opposite way and we had a sign such as *lioner marked ‘male’ in the paradigm with the unmarked lion, the types of implicatures would change. The pragmatic meaning of ‘male lion’ for the unmarked lion would be obtained by ‘unmarked dominance’ (C) (Iimplicature), the value of ‘female lion’ would be linked to ‘implicature’ (D) (Q-/R-implicature):

*lioner ‘MALE LION’

*lioner ‘MALE LION’

lion ‘LION’

lion ‘LION’

‘ +> MALE LION’

‘ +> FEMALE LION’

Figure 12. Pragmatic meanings of the unmarked lion in contrast to marked *lioner.

From this, we can conclude that the two contradictory ways of making sense of the unmarked sign re¶ect, in fact, the markedness conditions of the paradigm. An inversion of markedness links the same meaning to pragmatic negation instead of a¹rmation, and vice versa. Introducing markedness into the system of scalar implicatures therefore makes one kind of reasoning technically obsolete. It can always be expressed by markedness inversion.

150 Thomas A. Fritz

To sum up, every sign in connection with the paradigm can produce contradictory meanings. Markedness inversion can additionally change these meanings. It has to be stressed, however, that all the meanings described here (apart from the ‘implication’ (A)) are only possible, because the signs posit a ‘nil’. It is the very absence of any scrap of concrete meaning in the unmarked sign that enables speakers and hearers to apply pragmatic focussing processes. These are not subject to logical rules, because the sign actually possesses neither of the meanings it adopts; it merely makes use of them. Asymmetry of the code thus becomes a condition of creative sign use. A wide range of pragmatic meanings can be explained directly from the structure of the code. Alas, it is not enough to know the code. You must know the creative ways in which it can be used.

References Brandom, Robert B. 1994 Making It Explicit. Reasoning, Representing, and Discursive Commitment. Cambridge/Mass.: University Press. Comrie, Bernard 1989 “On identifying future tenses”. In Tempus – Aspekt – Modus. Die lexikalischen und grammatischen Formen in den germanischen Sprachen W. Abraham and Th. Janssen (eds), 51–63. Tübingen: Niemeyer Verlag. Frege, Gottlob 1967 Kleine Schriften. Ed. by Ignacio Angelelli. Hildesheim: Olms Verlag. Fritz, Thomas A. 2000 Wahr-Sagen. Futur, Modalität und Sprecherbezug im Deutschen. Hamburg: Meiner Verlag. Gazdar, Gerald 1979 Pragmatics. Implicature, Presupposition and Logical Form. New York: Academic Press. Grice, H. Paul 1989 Studies in the Way of Words. Cambridge/Mass.: Harvard University Press. Hanks, William F. 1992 “The indexical ground of deictic reference”. In Rethinking context. Language as an interactive phenomenon. A. Duranti and Ch. Goddwin (eds), 43–76. Cambridge: University Press. Horn, Lawrence R. 1989 A Natural History of Negation. Chicago and London: The University of Chicago Press. Jakobson, Roman 1971 Selected Writings. Vol. 2: Word and Language. The Hague: Mouton.

“Look here, what I am saying!”

Leech, Geoffrey N. 1983 Principles of Pragmatics. London: Longman. Levinson, Stephan C. 1983 Pragmatics. Cambridge: University Press. Levinson, Stephen C. 1987 “Minimisation and conversational inference” In The Pragmatic Perspective. J. Verschueren (ed.), 61–129. Amsterdam and Philadelphia: John Benjamins. Levinson, Stephen C. 2000 Presumptive Meanings. The Theory of Generalized Conversational Implicature. Cambridge/Mass.: MIT Press. Lyons, John 1977 Semantics. 2 Vol. Cambridge: University Press. Moore, George E. 1968 “A Reply to my Critics” In The Philosophy of G.E. Moore. P.A. Schilpp (ed.), 533–677. 3rd ed. La Salle/Illinois: Open Court. Palmer, Frank R. 1986 Mood and Modality. Cambridge: University Press. Price, Huw 1988 Facts and the Function of Truth. Oxford: Basil Blackwell. Vater, Heinz 1994 Einführung in die Zeit-Linguistik. 3rd ed. Hürth-Efferen: Gabel Verlag. Wittgenstein, Ludwig 1984 Werkausgabe. Vol. 1. Frankfurt/Main: Suhrkamp Verlag.

151

The ‘subjective’ eŸects of negation and past subjunctive on deontic modals The case of German dürfen and sollen Tanja Mortelmans University of Antwerp

1.

German mood and modal verbs as deictic modal expressions

In German, both mood markers (Indikativ, Konjunktiv I and II1) and modal verbs can be considered as deictic expressions, insofar as they locate a state of aŸairs in a particular epistemic region relative to the deictic center. According to Bühler (1999 [1934]: 102), the deictic center (die Origo des Zeigfeldes) can be captured by the so-called Zeigwörter (‘pointing words’) ich (‘I’), jetzt (‘now’) and hier (‘here), referring to a personal, temporal and local dimension of the deictic center respectively. It is striking that a modal qualiªcation (something like wirklich or faktisch) is lacking in Bühler’s characterization (and for that matter, also in a number of other in¶uential studies on deixis — see Diewald 1999: 173). Diewald relates the absence of a modal speciªcation to the highly abstract nature of the semantic domain of modality, which — in contrast to the (also abstract) domain of time — does not have a physical correlate (Diewald 1999: 174). An alternative explanation for the fact that the deictic center is not characterized in terms of its epistemological status (or its relation to reality) could focus on the highly intertwined nature of the temporal and the modal domain; perhaps a temporal qualiªcation by necessity involves a modal one: following Brisard, tenses do not just signal “temporal location, but a general epistemic stance” (Brisard 1999: 288).2 For the purpose of this paper, however, I will abstract from the possibly close interrelation of time and modality and accept the existence of a relatively independent modal dimension, which I will characterize in terms of Langacker’s

154 Tanja Mortelmans

so-called basic epistemic and dynamic evolutionary model (Langacker 1991: 242– 243 and 275–278). 1.1 Langacker’s (1991) basic epistemic model and dynamic evolutionary model The basic epistemic model (see Figure 1) makes use of three theoretical notions — known reality, immediate reality, and irreality — in order to capture a speaker’s basic conception of reality. Known reality (also referred to as basic reality or simply reality) can be deªned as the set of situations that a given conceptualizer (C) accepts as being real; it “represents the history of what actually happened” (Achard 1998: 42). The notion of immediate reality acknowledges the fact that reality is generally conceived as dynamic, as constantly evolving and growing: immediate reality, then, refers to the latest stage of reality’s evolution and functions as a vantage point, from which a given conceptualizer (the speaker) views things. Finally, irreality comprises everything outside (known) reality. Irreality (Known) Reality Immediate Reality C

Figure 1. Langacker’s basic epistemic model

This basic epistemic model does not represent the only possible conception of reality that speakers entertain. A more intricate conception is captured by the so-called dynamic evolutionary model (see Figure 2), which includes the speaker’s knowledge of the possible (future) course of reality and adds the notions of potential and projected reality (see Langacker 1991; Achard 1998: 43). Both potential and projected reality are by deªnition part of irreality. As their names suggest, potential reality refers to all future paths that reality is not precluded from following, i.e. that are potential, whereas projected reality refers

Creating subjectivity: negation and Konj. II with root modals

to those future developments which can be projected or anticipated by the conceptualizer with considerable conªdence. Potential Reality Reality

Present Reality

Projected Reality

C

Figure 2. Langacker’s dynamic evolutionary model

Whereas the meaning of the German moods can best be characterized relative to the basic epistemic model, the semantics of the modals seems to rely more on the dynamic evolutionary model. 1.2 The deictic meaning of the German moods Indikativ, Konjunktiv I and Konjunktiv II Location in reality represents the unmarked or default value, since reality is the epistemic sphere in which the speaker ªnds herself: “[…] the speaking ego typically regards his/her own current state of being as real, as factual, so that this value in the modal-deictic dimension appears as the unmarked one” (Diewald 1999: 175, my translation). In German, those states of aŸairs that are viewed as real, are marked by the indicative mood, the “Normalmodus” (Heidolph, Flämig, and Motsch 1981: 522). The moods Konjunktiv I and II, on the other hand, both mark a deviation from the default value; by deªnition, they are taken to locate a state of aŸairs outside the speaker’s conception of known reality, albeit in diŸerent ways. 1.2.1 The German Konjunktiv I as a modal deictic marker Following Diewald (1999), I will consider the present subjunctive (or Konjunktiv I), whose function mainly consists in marking indirect speech, as a quotative marker indicating an origo shift: it is not the actual speaker, but a reported speaker who is responsible for the content of the clause. The Konjunktiv I is generally assumed to re¶ect the use of an indicative in direct speech (see for instance Duden 1998: 783), which implies that the reported

155

156 Tanja Mortelmans

speaker has originally located the state of aŸairs within (her conception of) known reality.3 It should be emphasized that both conceptions of reality, i.e. the one entertained by the reported speaker and the one entertained by the reporting speaker, may very well be (and in practice often are) compatible with each other. This, however, need not be the case. With a present subjunctive, the reporting speaker is not taken to commit herself to the validity of the reported speaker’s conception of reality, as the following examples show. (1) a.

Unrichtig ist die Behauptung, daß ich Incorrect is the assertion that I Anwalt sei. [Frankfurter Rundschau, 17.03.1999] lawyer be-KonjI ‘Incorrect is the assertion that I am a lawyer.’ b. Es stimmt nicht, daß die Friedensbewegung It is.correct not, that the peace movement unnötig geworden sei. [SZ, 07.09.1995] unnecessary become-pp be-KonjI ‘It is not true that the peace movement has become unnecessary.’

1.2.2 The German Konjunktiv II as a modal deictic marker The past subjunctive (or Konjunktiv II) is generally attributed two main functions: it either expresses irreality — in a broad sense (see for instance Zifonun, HoŸmann, and Strecker 1997; Fabricius-Hansen 2000; Duden 1998) — or it marks indirect speech. In its latter function, the Konjunktiv II is typically taken to replace ambiguous forms of the default present subjunctive (Duden 1998: 785), i.e. the Konjunktiv II comes in to replace forms which could be interpreted as either indicative or subjunctive. I will not go into this reportive use of the past subjunctive (which is generally considered to be less typical), and focus on the past subjunctive signalling irreality instead. According to Diewald (1999: 185–187), the Konjunktiv II doesn’t express irreality in a direct or immediate way, but rather signals nonfactuality by referring to conditions that are not fulªlled (see also Kasper 1987); it is the unfulªlled condition that brings about the nonfactuality of the state of aŸairs. Apart from its deictic function, the Konjunktiv II is therefore said to exhibit a phoric or textual component, as it points to a condition which is either textually available or can be reconstructed from the linguistic or situational context. This phoric reference to a condition explains why in traditional German reference grammars like Duden the meaning or use of the past subjunctive is often illustrated in unreal conditionals. The

Creating subjectivity: negation and Konj. II with root modals 157

unreal conditional can be regarded as the past subjunctive’s prototypical grammatical environment, in which it unfolds the full potential of its meaning (see Mortelmans (2000: 196–199) for a more profound discussion of the meaning of the past subjunctive in conditionals).4 (2) a.

Wäre ich an seiner Stelle gewesen, be-KonjII I in his place been-pp, hätte ich gehandelt. have-KonjII I act-pp ‘If I had been in his position, I would have taken action.’ b. Das wäre schon, wenn ewiger Friede That be-KonjII nice if everlasting peace herrschte. prevail-KonjII ‘That would be nice, if everlasting peace existed.’ (examples taken from Duden 1998: 162)

1.2.3 The possible inclusion of clausal negation as a modal deictic marker Interestingly, Diewald (1999) includes a short note on clausal negation, “the semantic value of which partakes in the make-up of the marked modal values” (Diewald 1999: 180; my translation) in her discussion of the German moods. Clausal negation, most often expressed by the negation particle nicht in combination with the indicative of the ªnite verb, indicates that a speaker regards a state of aŸairs as nonfactual, i.e. it attributes the indicative’s opposite semantic value to a state of aŸairs. Negation could therefore be claimed to enter the paradigm of modal deictic expressions, although, of course, from a formal point of view it does not count as a mood.5 Langacker (1991: 134) also stresses the “close association” of clausal negation in English with the tense and modality predications of a ªnite clause, which is not only witnessed by their phonological proximity (e.g. isn’t, didn’t, shouldn’t), but also by their semantic similarity. In the course of this paper, I will try to shed some more light on the relationship between clausal negation and markers of modality (both mood and modal verbs), more speciªcally, on the way in which they interact with each other. 1.3 The German modals as deictic expressions Traditionally, six verbs are taken to belong to the German core modals: müssen, können, sollen, wollen, dürfen, and mögen. These verbs share a number of morphological, syntactic and semantic properties, which — taken together —

158 Tanja Mortelmans

make them stand out from other auxiliaries. First, the modal verbs are preteritpresents,6 i.e. original past tense forms with a present meaning; second, they combine with an inªnitive without zu, third, they are semantically ambiguous to the extent that they possess both a root7 and an epistemic meaning. The semantic domain covered by the root modals is rather large and heterogeneous: it extends from volition (wollen, mögen) to ability (können), permission (können, dürfen), obligation (sollen) and real-world necessity (müssen). The meanings expressed by the German modals in their epistemic usage seem to converge more, as they can all be associated with an assessment of likelihood by the speaker. Diewald (1999) only takes the epistemically used German modals — which are shown to exhibit a higher degree of grammaticalization than their non-epistemic ancestors — as being truly deictic. The epistemic modals are said to be no part of the state of aŸairs, but to represent the relationship between the speaker and the proposition, which is characterized in terms of a “sprecherbasierte Faktizitätsbewertung [a speaker-based assessment of factuality]” (Diewald 1999: 14). By contrast, the root modals are claimed to present a condition of the sentence subject, in which case the modal verb is part of the coded scene and thus functions as a Nennwort (Bühler 1999: 103). Diewald divides the non-deictic root meanings into three groups: root modals express either deontic, dispositional or volitional modality. Deontic modality, typically expressed by the modals sollen and dürfen, is reserved for usages having to do with intersubjective relationships in the domains permission, obligation, order and prohibition (Diewald 1999: 74). The modals können and müssen are taken to express dispositional modality, i.e. they refer to internal or external abilities and dispositions of the subject (Diewald 1999: 76). The modal status of volitional modality (covered in German by the verbs wollen and mögen) is somewhat of a borderline case (it is excluded by van der Auwera and Plungian (1998), for instance). For reasons of simplicity, it will not be considered any further in the course of this paper. 1.3.1 The relationship between the modal’s scope and root/epistemic modality The semantic distinction between root and epistemic modality has a number of formal correlates, one of them pertaining to the scope of the modal. Modals expressing root modality are prototypically associated with narrow scope, which means that the modal predicates a condition on the subject (see 3a), whereas the epistemic reading, exempliªed in (3b), evokes a wide-scope inter-

Creating subjectivity: negation and Konj. II with root modals 159

pretation, since the modal modiªes the entire proposition (Nordlinger and Traugott 1997: 301–304; Diewald 1999: 77). (3) a. You must play this ten times over. b. He said: “I must have a temperature”.

Wide-scope interpretations, however, also arise with root modals, in which case they “emphasize the particular event that should be brought about” (Nordlinger and Traugott 1997: 303). A number of cases can be discerned here. First, passivization is a syntactic means to invoke a wide-scope reading, as it is no longer the subject of the sentence to which the obligation (or any other modal condition, for that matter) can be attributed. (4) In dem freiheitlichen Europa der Zukunft soll ja niemand mehr vertrieben werden.8 ‘In the free Europe of the future no one shall be expelled any more.’

On a semantic level, inanimate subjects in combination with stative or process verbs also give rise to a wide-scope reading of the modal. Since the subject has no agentive properties, the modal by necessity has scope over the whole proposition. (5) Es muß und soll deshalb eine vordringliche Aufgabe der neuen Kooperation sein, zu verhindern, […]. ‘It must and should therefore be a priority task of the new co-operation to prevent [...].’

Wide-scope interpretations of non-epistemic modals fulªl an important bridging function towards the development of an epistemic meaning. Nordlinger and Traugott (1997), for instance, assume that the wide-scope deontic interpretation of ought to paved the way for the epistemic reading of the modal through the conventionalization of a conversational implicature.9 The notion of scope also ªgures prominently in Diewald’s (1999) account of the rise of the epistemic meanings of the German modals, since the possibility of a wide-scope interpretation is taken to function as a crucial Vorbedingung (‘precondition’) for the development of an epistemic meaning (Diewald 1999: 367). 1.3.2 Speaker involvement in root modals Diewald’s distinction between non-deictic and deictic modality seems to be clear for cases like the following — examples and paraphrases are taken from Diewald (1999).

160 Tanja Mortelmans

(6) a.

Sie musste sich hinsetzen, denn ihre Beine waren schwach und wollten ihr nicht gehorchen. ‘She had to sit down, because her legs were shaky and didn’t want to obey her.’ Paraphrase: Sie war gezwungen sich hinzusetzen, denn... ‘She was forced to sit down, because...’ b. Er muss eine Vorahnung davon gehabt haben. ‘He must have had a premonition.’ Paraphrase: Ich halte es für wahrscheinlich, dass er eine Vorahnung gehabt hat. ‘I think it is probable that he has had a premonition.’

Rejecting Bybee, Perkins, and Pagliuca’s (1994: 179) category of speaker-oriented modality, Diewald does not accord any special status to speaker-oriented non-epistemic usages of the German modals. Directive usages of the modals are treated as indirect speech acts (Diewald 1999: 36), which in her view are likely to arise given that the root modals code a condition of the subject, which is the result of a directive that has been performed: the root modals “represent a condition, which arises from a performed directive for the receiver of that directive” (Diewald 1999: 125; my translation). In an appropriate context, therefore, the assertion that a directive speech act has been performed may give rise to an (indirect) directive interpretation. (7) Du mußt deine Hausaufgaben machen → Mach deine Hausaufgaben ‘You must do your homework’ → ‘Do your homework!’

However, since directive usages point to some element of the deictic center (typically the speaker) as a locus of potency and are thus linked to the immediate speech situation, I would prefer to consider them as intermediate cases between genuinely descriptive, non-deictic usages, and epistemic (deictic) ones, in which the modal functions as a Zeigwort. This is in line with Goossens (1999, 2000), who draws attention to the fact that the rise of epistemic must in Early Modern English took place at a time when subjective deontic usages of must, i.e. speaker-oriented usages in which the speaker associates herself with the modal force, began to outnumber the non-subjective root usages of the verb, the latter being clearly dominant in Middle-English. Goossens (2000: 167) concludes that “subjectiªcation in the participant-external, more speciªcally the deontic, area paved the way for the development of the (subjectiªed) epistemic sense”.

Creating subjectivity: negation and Konj. II with root modals

Note that directive usages like (7) typically situate the state of aŸairs expressed in the inªnitive (deine Hausaufgaben machen) within irreality, or, to be more precise, within potential reality (see section 1.1): it is not real at the moment of speaking, but has the potential of becoming real at a later stage. Localization within potential reality is a feature that speaker-oriented root modals share with their epistemic counterparts. An epistemic modal can be taken to express that a state of aŸairs cannot be assessed as either real or unreal at the time of speaking, but may turn out to be real in a later stage. In this connection it should also be noted that a speaker-oriented root modal cannot appear in the past tense, which typically evokes an implicature of actualization. In (8), the state of aŸairs (all sein Geld hergeben, einer Ehekrise meiner Eltern beiwohnen) can therefore not be said to be situated within potential reality. (8) a.

Er musste all sein Geld hergeben. ‘He had to hand over all his money.’ b. Zuhause durfte ich dann einer Ehekrise meiner Eltern beiwohnen. [example taken from Diewald 1999, p. 131] ‘At home I could witness a matrimonial crisis between my parents.’

According to Diewald (1999), a speaker-oriented interpretation of a root modal is merely a matter of context: given the appropriate circumstances, a root modal can take up a directive meaning. It has been shown, however, that the proportion of speaker-oriented usages associated with a particular root modal might increase over time, so that it by default comes to express a speaker-oriented meaning. This process of subjectiªcation, whereby “meanings become increasingly based in the speaker’s subjective belief state/attitude towards the proposition” (Traugott 1989: 311) has been witnessed for deontic must (Goossens 1999, 2000) and ought to (Nordlinger and Traugott 1997). 1.3.3 The eŸect of negation and past subjunctive on a speaker-oriented reading In the next sections, I will provide an analysis of the German modals dürfen and sollen in order to establish if and to what extent these verbs have developed a subjective-deontic reading. I will focus on two markers which seem to enhance a speaker-oriented interpretation, namely the past subjunctive and clausal negation. These two elements are mentioned by Diewald (1999: 219) in her discussion of subjective-epistemic können, which has a clear preference for either negated contexts (kann nur in 9a) or for the past subjunctive (könnte in 9b).

161

162 Tanja Mortelmans

(9) a.

Das kann nur dort verloren gegangen sein hier doch nicht. ‘That can only have got lost there but not here.’ b. Er könnte Foch und Falkenhayn vor Augen gehabt haben. ‘He could have had Foch and Falkenhayn in mind.’

Diewald links this preference of epistemic können to appear in combination with either negation or a past subjunctive to the phoric character of these two markers: “they express the dependence upon a previously mentioned element” (Diewald 1999: 221; my translation). As phoric markers, negation and past subjunctive are said to cancel or neutralize the inherently reactive character of können, which is by itself more or less incompatible with an epistemic interpretation. The possibility expressed by können always implies that its bearer (the subject of können) has somehow initiated its existence; asserting a possibility is — in Diewald’s view — a reactive move, since the speaker asserting it is taken to react to the subject’s initiative (Diewald 1999: 159).10 The reactive component inherent in können prevents a straightforward epistemic use of the verb, which is typically unconditional (unbedingt), referring only to the speaker as an authority of assessment. On a very abstract level, then, Diewald seems to assume that the reactive nature of können is neutralized — in an almost homeopathic process — by a similar feature of phoric reference to a previous condition expressed by either negation or Konjunktiv II. Instead of focusing on the phoric character of negation and past subjunctive, however, I would prefer to highlight their inherently speaker-oriented nature — which is, by the way, also explicitly stated by Diewald: “Indem der Sprecher auf die Relevanz einer zusätzlichen Bedingung hinweist, wird der Sprecher als Instanz der Bewertung hervorgehoben [as the speaker points to the relevance of an additional condition [by using either a past subjunctive or a negation marker], the speaker as authority of assessment is emphasized]” (Diewald 1999: 221). Negation and past subjunctive can therefore be considered as means to enable or at least facilitate a speaker-oriented interpretation, not only in the epistemic, but also — as I hope to show — in the root domain.

2. Speaker-oriented usages of the German modals: the case of deontic dürfen and sollen In the upcoming sections, I will concentrate on the German modal sollen in its indicative and past subjunctive usages. Before embarking on this, however, I will brie¶y comment on some speaker-oriented deontic usages of the modal

Creating subjectivity: negation and Konj. II with root modals 163

dürfen (English ‘may’) in order to show that negation can play an important role in inducing a speaker-oriented interpretation. 2.1 The modal verb dürfen as a “reactive” modal As with können, Diewald (1999) attributes a reactive component to the semantics of dürfen: when a speaker uses dürfen to express a permission, she is taken to respond to a previous wish of the addressee of the permission (typically the subject): “dürfen […] reacts to a previous communicative move, which is aimed at receiving a directive (the permission) to perform the action” (Diewald 1999: 132; my translation). In a force dynamic account (see Talmy 1988, 2000) one could say that the original initiative with dürfen lies in the hands of the addressee of the permission, typically the sentence subject, who has somehow asked to receive it. The original locus of potency is therefore not to be identiªed with the speaker. Negating permission, however, can be regarded as a clear intervention by the speaker and is therefore likely to eŸect a rearrangement of the force distribution. 2.1.1 The use of dürfen in negated and a¹rmative contexts The indicative of dürfen generally exhibits a deontic meaning, paraphrasable as ‘to have the permission, the right to do x’. Interestingly, indicative dürfen typically occurs in negated contexts: an analysis of 150 indicative instances of dürfen (see Mortelmans 1999: 714) revealed that only 24 of them appeared in an a¹rmative sentence.11 In negated contexts, dürfen invites both descriptive and directive interpretations. Example (10), taken from a manual for lorry drivers, hosts a considerable number of instances of descriptive dürfen: the speaker simply describes existing regulations. Note that each instance of dürfen in this section is negated. (10) Die Ladung des Fahrzeuges ist gemäß nachstehenden Bestimmungen unterzubringen: Die Breite der Ladung darf nicht mehr als 2,50 m betragen. […] Beide Lichter dürfen nicht blenden. Die Länge von Fahrzeug und Ladung zusammen darf 20 m, die Höhe 4 m nicht überschreiten. Lange Gegenstände dürfen niemals nach vorne über das ziehende Fahrzeug hinausragen. […] Fahnen, Schilder und Leuchten dürfen nicht höher als 1,55 m über der Fahrbahn angebracht werden. […] Das Fahrzeug darf nie überladen sein.12

In directive environments,13 however, negated instances of dürfen typically invite a speaker-oriented, directive interpretation, in which the speaker urges

164 Tanja Mortelmans

the addressee not to perform a particular action. Consider the following instances, all taken from a survey of 66 du darfst occurrences, 41 of which were negated.14 (11) a.

Geh weg, du darfst meiner Mami nicht weh tun! ‘Go away, you mustn’t hurt my mummy!’ b. Nein, Emanuel, da darfst du dich nicht einmischen. ‘No, Emanuel, you mustn’t interfere in that.’ c. Nicht sprechen, Michael, du darfst jetzt nicht sprechen. ‘Don’t speak, Michael, you mustn’t speak now.’ d. “Das darfst Du nicht riskieren”, riet der Stürmer eindringlich ab. ‘ “You mustn’t risk that”, the striker insistently advised.’

Affirmative instances containing darfst, on the other hand, seem to be more compatible with a non-directive, descriptive interpretation, in which the role of the speaker is somewhat less prominent. An interesting observation in this light is the fact that 5 out of 25 affirmative instances appear in a conditional context — whereas only 1 of the 41 negated instances features in the apodosis of a conditional. In these cases, the permission expressed by darfst depends on a condition, the realization of which lies in the hands of the sentence subject. In terms of force distribution, one could say that in (12) both subject and speaker are somehow involved as loci of potency in the modal relation, whereas in (11) the main force is clearly with the speaker. (12) a.

Die […] Abgeordnete […] erklärte, die Koalition gehe bei der Diskussion um den Ladenschluß nach dem Motto vor: “Wenn Du mir am Samstag eine Stunde gibst, darfst Du am Donnerstag abend bis 20.30 Uhr ausweiten” […]. [Mannheimer Morgen, 11.05.1989] ‘The member of parliament stated that the coalition in its discussion of the shop closing time proceeded according to the motto: “If you give me an hour on Saturday, you can stay open till 8.30 p.m. on Thursday [...].”’ b. Natürlich darfst Du heute das Fernsehen ausfallen lassen, wenn Dir Deine Hausaufgaben wichtiger sind! [Mannheimer Morgen, 14.11.1995] ‘Of course you can drop television today, if your homework is more important to you.’ c. Wenn du mir das Rätsel lösen kannst, darfst du meine Tochter heiraten. [Mannheimer Morgen, 24.12.1995] ‘If you can solve the riddle, you are allowed to marry my daughter.’

Creating subjectivity: negation and Konj. II with root modals 165

2.2 The modal verb sollen as a “non-reactive” modal Together with dürfen, the German modal sollen is taken to express deontic modality. In contrast to dürfen, however, Diewald (1999: 128) assumes that sollen does not indicate a reactive move of the speaker. Sollen expresses that a directive has been uttered, which is initiated by the speaker and is independent of any previous communicative moves of the directive’s addressee. Whereas dürfen is anaphoric — it points to a previous speech act — sollen is more of a cataphoric marker, as it refers to what is to follow (Diewald 1999: 129). With respect to its non-reactive nature, sollen is said to resemble müssen; the main diŸerence between the latter verbs pertains to the fact that sollen requires an entity capable of communication (typically a person) as a modal force, whereas the locus of potency evoked by müssen is generally undetermined and vague. In contrast to dürfen, then, nothing seems to prevent a speaker-oriented interpretation of sollen. Such an interpretation seems even more likely than with müssen, given that sollen is taken to imply that the locus of potency is human and can therefore very well be equated with the speaker.15 In order to see whether this prediction is borne out, I conducted an analysis of soll and sollte occurrences compiled from a part of the so-called Wendekorpus, provided by the Institut für deutsche Sprache. The search was restricted to occurrences in the records of the German house of parliament (labelled Bundestagsprotokolle 2.Hj. 89). This part of the Wendekorpus mainly contains spoken and writtento-be-spoken material taken from speeches by the members of the Bundestag. The indicative form soll occurred 129 times, past subjunctive sollte was found 117 times.16 2.2.1 The indicative form soll: a corpus analysis Given that a directive, speaker-oriented interpretation is most likely to occur in a declarative sentence, it is slightly surprising to ªnd that indicative soll, unlike past subjunctive sollte, does not occur that often in a declarative (see Table 1). The quantitative dominance of indicative soll in the corpus is due to its relatively high frequency in interrogatives (31 soll vs. 2 sollte) and — to a minor extent — in subclauses (48 soll vs. 28 sollte). Table 1. The frequency of soll and sollte in various clause types declarative interrogative subordinate clause Total number

soll

sollte

50 31 48 129

77 2 28 107

166 Tanja Mortelmans

As is shown in Table 2, the indicative of sollen appears with a considerable frequency in conditional subclauses (Example 13a), in complement clauses introduced by the conjunction daß (Example 13b), in relative clauses (Example 13c) and in indirect interrogatives (Example 13d). The modal verb generally does not invoke a speaker-oriented interpretation; instead, soll as a Nennwort typically describes the existence of a modality (obligation, intention), whereby the locus of potency is not the speaker, but a previously mentioned entity. Table 2. Indicative soll in subclauses conditional complement (daß) relative clause indirect interrogative Total number

7 15 14 12 48

(13) a.

Beide Länder haben erkannt, daß ihr System reformiert werden muß, wenn es den Herausforderungen der Zukunft gerecht werden soll, […]. ‘Both countries have acknowledged that their system has to be reformed, if it is to meet the challenges of the future.’ b. Sie schlagen [...] vor, daß im Bereich von Wissenschaft und Technologie ein interministerieller Ausschuß gebildet werden soll, der die wissenschaftlich-technologische Zusammenarbeit organisieren soll. ‘They propose that in the ªeld of science and technology an interministerial commission should be formed, which is to organize the scientiªc-technological co-operation.’ c. Vaclav Havel, der im Oktober den Friedenspreis des deutschen Buchhandels erhalten soll […]. ‘Vaclav Havel, who is to receive the German Book Trade’s Peace Prize in October [...].’ d. Wenn das kein Zynismus ist, dann frage ich mich, wie man Zynismus definieren soll. ‘If that isn’t cynicism, then I ask myself how on earth one should deªne it.’

Occasionally, i.e. in only 4 (out of 48) instances, the speaker can be taken to be personally involved. This personal involvement is overtly marked, though, either in the main clause or in the subclause itself.

Creating subjectivity: negation and Konj. II with root modals 167

(14) a.

Ich bin der Ansicht, Deutschland soll wiedervereinigt werden. ‘I am of the opinon that Germany should be reunited.’ b. Uns allen ist der Wunsch sehr nahe, daß dieses Haus Wirklichkeit werden soll. ‘We all share the wish that this house should become reality.’ c. Man kann gar nicht genug Klarheit darüber schaŸen, daß die Westgrenze Polens nach unserem politischen Willen dauerhaften Bestand haben soll. ‘One can’t be too clear on the issue that Poland’s western border in accordance with our political will should have long-lasting existence.’

In 3 other instances, a speaker-oriented interpretation does not seem unlikely either, but since there are no contextual clues, it is difficult to decide whether the speaker simply reports on a given intention or plan, or really associates herself with it. This can be taken to suggest that soll by itself is not very well capable of transporting a speaker-oriented meaning, but needs additional markers to unambiguously convey it. (15) Es [das Zusammengehörigkeitsgefühl] wird für die Zukunft unserer Nation, dieser geteilten Nation, die zusammenwachsen soll, ganz wichtig werden. ‘It [the feeling of belonging together] will become extremely important for the future of our nation, this divided nation, which is to/should merge into one.’

In 31 occurrences, indicative soll features in an interrogative sentence, which is often to be interpreted as either a rhetorical or a deliberative question, for which the speaker does not really seek an answer. Again, the speaker cannot be taken to associate herself with the modal force: soll typically refers to existing obligations or intentions, the origin of which is either undifferentiated (16a–b) or can be traced to the addressee of the question (16c). (16) a.

Woher soll eine Gesellschaft die Kraft für Veränderungen nehmen, wenn ein großer Teil [...] das Land verläßt? ‘Where should a society take the power for changes, if a large part leaves the country?’ b. Aber die Frage bleibt doch: Wer soll das eigentlich bezahlen? ‘But the question remains: Who is to pay for that?’ c. Soll denn unsere Bevormundung gleich die andere ablösen, unter der diese Menschen leiden?

168 Tanja Mortelmans

‘Is it really our intention to replace one form of imposition, under which these people suŸer, by another (namely ours)?’

If we finally look at the declarative main clauses containing indicative soll, we must conclude that here a speaker-oriented interpretation seems to be dominant. In 33 out of 50 declarative occurrences, the speaker as a locus of potency seems to partake in the modal relationship, an observation, however, which should be handled with some care. First, 3 speaker-oriented occurrences contain an instance of quotative sollen, which, according to Diewald (1999), represents an inherently deictic (epistemic) usage of the verb. Like the present subjunctive, the speaker signals that another speaker has assessed the proposition as factual, the reporting speaker, however, does not commit herself to the factuality of the proposition (see Diewald 1999: 225–226). (17) Geld [...] soll in bezug auf Ungarn im Spiel gewesen sein [...]. ‘Money is said to have played a role with respect to Hungary.’

This leaves us with 30 speaker-oriented deontic occurrences of indicative sollen, which — under the assumption that deontic modality prototypically invokes a narrow-scope reading of the modal — can be expected to have an animate subject and an agentive verb as complement in the majority of the cases. This expectation, however, is not borne out: half of the verbs in the infinitive denote either a state (11 instances) or an uncontrollable process (4 instances), both invoking a wide-scope deontic interpretation. Moreover, in a considerable number of cases the deontic reading seems to be intertwined with a future one, so that sollen comes to express an intention of the speaker with respect to the realization of a state of affairs (for this future sollen, see also Diewald 1999: 281). (18) a.

Wir wollen mit unseren Nachbarn das gemeinsame europäische Haus bauen und bewohnen. Jeder soll darin eine Wohnung haben, die seinen Bedürfnissen entspricht. ‘We want to build and inhabit a common European house with our neighbors. In this house, everyone shall have a place to live in that meets their needs.’ b. [...] denn dieser Zusammenschluß freier Staaten soll sich auf der Grundlage der Freiheit und des Selbstbestimmungsrechts vollziehen. ‘ [...] because this union of free states is to take place on the basis of freedom and the right to self-determination.’

Creating subjectivity: negation and Konj. II with root modals 169

Passive occurrences, which are found 6 times, invoke a wide-scope deontic interpretation as well. Again, the deontic reading is sometimes mingled with a future one. (19) Ich will hier eindeutig zusichern: Es ist nicht die Absicht der Bundesregierung, das Thema [...] zu vertagen. Wir brauchen natürlich Zeit für vernünftige Gespräche. Aber das Thema soll auf keinen Fall zu einem späteren Zeitpunkt behandelt werden. ‘I want to assure you unambiguously: it is not the intention of the federal government to adjourn this issue. Of course we need time for sensible discussions. But this issue is in no way to be treated at a later moment.’

This leaves only 9 occurrences, in which speaker-oriented indicative soll appears in a prototypically deontic context, i.e. the subject is human and the verb in the infinitival complement refers to an action. Note that the subject is preferably indefinite (6 out of 9 occurrences) and that 4 occurrences are negated. I will come back to these observations in the discussion of sollte. (20) a.

Das ist eine Aufgabe der Bundesregierung. Das soll man doch nicht dem [...] Senat zuschieben. ‘That is a task of the federal government. One shouldn’t pass it on to the senate.’ b. Jetzt soll niemand sagen, das eine habe nichts mit dem anderen zu tun. ‘At this point no one should say that these things have got nothing to do with one another.’

By way of a preliminary conclusion, I would like to stress the fact that the alleged prototypically deontic modal sollen occurs only rarely in a typical deontic context. Instead, indicative soll can be shown to be compatible with a wide array of contexts and usages, which on the whole do not seem to favor a speaker-oriented interpretation of the verb. Having established these facts, we will now proceed to soll’s past subjunctive counterpart sollte. 2.2.2 The past subjunctive form sollte: a corpus analysis Diewalds’s view on past subjunctive sollte is somewhat ambivalent. On the one hand, she treats sollte — together with möchte — as a verb form, “whose splitting-oŸ [from the original modal lexeme] has already gone a long way” (Diewald 1999: 198; my translation). On the other hand, Diewald claims that the semantics of sollte, which in the literature is mostly associated with a judgmental value and a weaker imperative force than both indicative sollen and

170 Tanja Mortelmans

müssen, can be explained in a fully compositional way. Whereas the indicative of sollen is said to express the subject’s condition of having received a directive, the past subjunctive form is taken to refer to a condition that aŸects the validity, relevance or factuality of the subject’s state: “Der Zustand des Sollens gilt, wenn und soweit die phorisch indizierte Bedingung gilt [the state of sollen is valid if and insofar the phorically indicated condition is valid]” (Diewald 1999: 199). In the following sentence, therefore, the existence or validity of the directive, expressed by the soll-element, is claimed to depend upon a condition, as is indicated by the Konjunktiv II; only if this condition — which according to Diewald could be something like “disharmony in a marriage is no good thing, if one is too insistent, this might lead to disharmony” —is accepted, the directive is said to be valid. (21) Die Dame sollte ihren Mann nicht zu sehr drängen. ‘The lady should not be too insistent on her husband.’

Diewald claims that conditions of this kind can be found for all occurrences of sollte in her corpus (Diewald 1999: 201). Two remarks, however, should be made in this connection. First, past subjunctive sollte in practice does not occur that often in an overt conditional construction: I found only 6 examples in my material.17 And, second, if it does, it is typically combined with a factual (instead of a hypothetical) protasis, the validity of which is unquestioned.18 This can be taken to imply that even in conditionals, the validity of the modal relation (expressed in the main clause) is not subjected to an open condition, but seems to be taken for granted by the speaker. (22) a.

Wenn wir, meine Damen und Herren, sagen, wir wollen unser Beispiel und Modell nicht aufdrängen, so sollte doch [...] niemand glauben, [...]. ‘If we, ladies and gentlemen, say that we do not want to impose our example and model, no one should believe [...].’ b. Niemand sollte Genugtuung darüber empªnden, wenn Menschen in so großer Zahl ihr Zuhause aufgeben wollen. ‘No one should feel satisªed about the fact that so many people want to give up their homes.’ c. Denn auf die Frage [...] sollte man sich wohl einlassen, wenn man, wie die Bundesregierung, dazu auŸordert, die Ausreise nicht über die diplomatischen Vertretungen zu suchen. ‘For one should get involved in that issue, if one, like the federal government, asks people not to leave the country via the embassies.’

Creating subjectivity: negation and Konj. II with root modals

Rather than positing an almost always implicit condition for usages of sollte, I would claim that sollte is a subjective-deontic modal marker, which typically refers to the speaker as a locus of potency. This implies that the past subjunctive element within sollte has lost its original conditional meaning and has fused with the modal into a new unit, which cannot be analyzed in a fully compositional way, i.e. sollte cannot be adequately described as a genuine past subjunctive of sollen (anymore). A number of observations relating to the corpus material sustain this view. On a superªcial level, the relatively high frequency of past subjunctive sollte (107 sollte vs. 129 soll) can be regarded as a ªrst indication of the fact that sollte does not function as a genuine past subjunctive: being a marked mood, the past subjunctive is generally expected to occur far less often than the unmarked indicative (Jäger 1971 and Bausch 1979 have shown that past subjunctive verb forms account for only 4 to 6 per cent of all ªnite verb-forms in present-day German). In the corpus at hand, however, the diŸerence between the number of soll and sollte occurrences is not that big.19 The syntactic behavior of sollte also points into the same direction (see table 1 above). First, sollte shows a clear preference for declarative sentences: 71 % of all sollte-instances appear in a declarative, which is the province of only 38 % of the indicative occurrences. Conversely, sollte is extremely rare in questions, both direct and indirect ones, whereas soll-questions accounted for about a quarter of all indicative instances. Assuming, however, that sollte is a speaker-oriented marker, we should not be surprised at the fact that it does not frequently appear in interrogatives, which are typically addressee-oriented. Finally, sollte occurs considerably less often in subclauses than its indicative counterpart (26 % sollte vs. 37 % soll). Table 3. Past subjunctive sollte in subclauses conditional complement (daß) relative clause indirect interrogative Total number

8 13 6 1 28

In contrast to indicative sollen, a clear majority of the subordinate sollteoccurrences (22 out of 28) invites a speaker-oriented interpretation, whereby the verb generally carries a deontic meaning. In the complement clauses (23a–d), we generally ªnd explicit expressions of the speaker’s involvement, which is not the case, however, for the relative clauses (23e–g). In the latter, the

171

172 Tanja Mortelmans

speaker undoubtedly associates herself with the obligation (recommendation), without however overtly marking her personal involvement. (23) a.

Deshalb meinen wir, daß man [..] im mittelständischen Bereich ansetzen sollte [...]. ‘That’s why we think that one should start in the middle class.’ b. Gleichwohl sind wir der Meinung, daß […] geholfen werden sollte. ‘Nevertheless we think that one should help.’ c. Ich denke, daß man deshalb von dem Beginn einer zweiten Etappe […] sprechen sollte. ‘I think that therefore one should speak of the beginning of a second phase.’ d. Ich glaube, daß sich die Europäische Gemeinschaft nicht mit jenen Kräften verbinden sollte, die [...] nur den Status quo stabilisieren können. ‘I believe that the European Community should not join forces with those who are only able to maintain the status quo.’ e. Und das ist eine Leistung der Bevölkerung [...] in der ganzen DDR, […] die nicht etwa selbstgefällige Triumphgefühle auslösen sollte. ‘And that is an achievement of the people in the whole GDR, which should not provoke any complacent feelings of triumph.’ f. Denn er hat den Schuß Realismus in die außenpolitische Debatte gebracht, ohne den man über die Ergebnisse [...] nicht sprechen sollte. ‘Because he has brought the necessary dash of realism in the foreign policy debate, without which one shouldn’t speak about the results.’ g. Sie haben das mit einer Würde und mit einer sprachlichen Kraft getan, die auch für uns Maßstäbe setzen sollte [...]. ‘They have done that with such dignity and verbal power, which should set standards for us as well […].’

Past subjunctive sollte also appears in a conditional protasis (see Example 24), which I consider to be a speaker-oriented use as well, to the extent that the speaker can be said to localize the hypothetical event expressed in the protasis within potential reality. Note that sollte — like English conditional should — does not carry a deontic meaning here. (24) Sollte sich eine Mehrheit für zwei deutsche Staaten entscheiden, so werden wir dies selbstverständlich respektieren. ‘Should a majority decide in favor of two German states, we will of course respect this.’

Creating subjectivity: negation and Konj. II with root modals 173

If we focus on the declarative sollte-occurrences, a number of similar observations can be made. First, each instance seems to be speaker-oriented to some extent (whereas a considerable number of the indicative declarative instances of sollen — 17 out of 50 — were not). Compare in this respect (25a) with (25b) and note that replacing non-subjective soll by sollte in (25a) does not seem to be a very likely option. (25) a.

Einig sind Sie sich also alle in dem Ziel, die DDR einzuverleiben. Die kapitalistische Marktwirtschaft soll ihr Comeback in der DDR ªnden. Der Schlußpunkt ist dann die Wiedervereinigung. ‘So you all agree on the goal to annex the GDR. The capitalist economy is to ªnd its comeback in the GDR. The end point will be the re-uniªcation.’ b. Nur, eines sollte man auch in dieser Aktuellen Stunde nicht verschweigen. ‘But there is one thing that even at this current moment should not be concealed.’

With past subjunctive sollte, the speaker involvement is only sometimes explicitly marked by expressions like ich denke, meiner Meinung nach or meine ich. (26) Das ist ein Kon¶ikt […], der dringend einer öŸentlichen Debatte bedarf. Sie sollte meiner Meinung nach im Abgeordnetenhaus von Berlin geführt werden, nicht hier. ‘This is a con¶ict that is in urgent need of a public debate. This should, in my view, be done in the house of parliament in Berlin, not here.’

In contrast to indicative soll, past subjunctive sollte is combined with an animate subject and an agentive verb in almost half of the cases (38 out of 77 declarative occurrences). As with soll, the animate subject is preferably indeªnite,20 as Table 4 shows. Note, however, that for every man soll occurrence in the corpus, we ªnd about 5 instances of man sollte. Table 4. The nature of the subjects with sollte in declaratives subject

sollte

man (‘one’) niemand (‘no one’) wer (‘who’) jeder (‘everybody’) der eine oder der andere (‘someone, somebody’) manch einer (‘many a person’) Total number of indeªnite subjects

15 6 3 2 1 1 28

Total number of deªnite subjects

10

174 Tanja Mortelmans

Another observation pertains to the relatively high frequency of negated occurences: almost one third (i.e. 25 out of 77) of the declaratives containing sollte is negated, which more or less resembles the proportion of negated occurrences with speaker-oriented soll in declaratives (4 out of 9). Again, it should not be ignored that for each negated occurrence of soll, we ªnd about 5 negated occurrences of sollte. Passive complements are found in 13 sollteoccurrences; this again re¶ects the amount of passives with the indicative (6 out of 30). In view of these ªndings, then, it seems very di¹cult to posit a clear distinction between deontic, speaker-oriented usages of soll and deontic, speakeroriented usages of sollte. They share a preference for indeªnite subjects, and they are to more or less the same extent negated and passivized. In the following occurrences, therefore, indicative soll and past subjunctive sollte seem to be simply interchangeable. (27) a.

Die Deutschen haben — auch dies soll hier noch einmal gesagt werden — […] ein Recht auf Selbstbestimmung. ‘The Germans have — this also should be said here once again — the right to self-determination.’ b. Meine Damen und Herren, wir haben — das sollte man sagen — […] doch […] den nötigen Schuß Realismus zu beachten. ‘Ladies and gentlemen, we still — one should say that — have to observe the necessary dash of realism.’

Speaker-oriented, deontic sollte instances can mainly be said to clearly outnumber the indicative ones. This does not mean, however, that every instance of soll can be replaced by sollte. In the wide-scope future-deontic usages, exempliªed in (18) and (19), replacing soll by sollte seems to induce a shift in meaning, whereby the speaker does not so much look into the future, but concentrates on what is going on right now.

3. Negation and past subjunctive in combination with sollen: A new interpretation How can it be accounted for that past subjunctive sollte is generally preferred over indicative soll to express a speaker-oriented obligation? What role should we attribute to negation in this respect, bearing in mind that one third of the inherently subjective sollte-occurrences is negated? Note also that indicative

Creating subjectivity: negation and Konj. II with root modals 175

soll is negated in 12 instances (out of a total number of 129), 8 of which are clearly speaker-oriented. Roughly, the answer to these questions could be sketched in the following way. In contrast to Diewald (1999), I assume that the modal verb sollen is anaphoric as well, i.e. it typically refers to a previous communicative move which has given rise to the modal relation of obligation. It is well-known that sollen as a main verb meant something like to owe, i.e. the subject is obliged to do something (to pay something back), because someone else did something for him or her. It is sollen’s anaphoric character that prevents the speaker from functioning as a straightforward locus of potency. Unlike können and dürfen, however, the original initiative does not lie with the subject, but with a third entity — diŸerent from both subject and speaker.21 Speaker-oriented usages of sollen, therefore, have a tendency towards negation and past subjunctive, as both represent speaker-oriented interventions in an otherwise more objective scene. (28) a.

Du sollst es tun. (weil ein Dritter es will) ‘You are to do it.’ (because a third person wants you to) b. Du sollst/solltest es nicht tun. (weil ich es will) ‘You shouldn’t do it.’ (because I want you to)

The tendency towards negation is less outspoken with sollen than with dürfen, for at least two reasons. First, the conceptual move from a sentence-external locus of potency (a third person) towards the (also external) speaker as locus of potency in the case of sollen is less drastic than the move from a sentenceinternal modal force (the subject) towards the speaker in the case of dürfen. In other words, on the basis of its semantics, sollen is less in need of additional speaker-oriented markers than dürfen. Second, sollen allows both internal and external negation; in fact, it often seems di¹cult to decide whether negation in the case of sollen aŸects the inªnitival complement or the modal. In the German literature on this subject, sollen is usually said to invite internal negation: “Es wird ein Gebot ausgesprochen, etwas nicht zu tun [It expresses an order not to do something]” (Diewald 1999: 132, see also there for other references on internal negation with sollen). Fritz (1997: 55) also claims that internal negation is more common with sollen, but concedes that external negation is sometimes possible as well. His example sentence for external negation Du sollst nicht [so dumm lachen] (‘You shouldn’t laugh so foolishly’), however, seems to be compatible with internal negation as well: Was du tun sollst ist: nicht so dumm lachen. I therefore side with Coates, when she writes that with root should in English “it is immaterial whether the modal predication or the main predica-

176 Tanja Mortelmans

tion is seen as being negated: semantically it makes no diŸerence” (Coates 1983: 64). Since a speaker-oriented interpretation of negation only arises with widescope external negation, it does not come as a surprise that the link between speaker-oriented soll(te) and negation is less strong. As far as the subjective character of the Konjunktiv II in combination with sollen is concerned, we can point to other past subjunctive forms of the modals, which have acquired speaker-oriented meanings as well. Both könnte and dürfte, for instance, exhibit an epistemic meaning, which — especially in the case of epistemic dürfte — seems to resist a compositional analysis. (29) Es dürfte unmöglich sein zu beweisen, daß im folgenden Satz das Dativobjekt das Akkusativobjekt “determiniere”. [Engel, U. (1970), Forschungsberichte des IDS 5, p. 14] ‘It is probably impossible to prove that in the following sentence the dative object determines the accusative object.’

An epistemic reading of sollte, however, is marginal (Diewald 1999: 202–203): in present-day German, sollte seems to have specialized as a subjective deontic marker. Note, however, that the necessary precondition for an epistemic meaning to arise — the possibility of a wide-scope reading of the modal — is fulªlled (through passivization, or the combination of inanimate subjects with stative verbs, both evoking a wide-scope reading). Whether this means that sollte is on its way towards the development of a full-¶edged epistemic meaning, is hard to say, though. Finally, it should be noted that one German modal has remained conspicuously absent in the previous discussion: müssen, which is almost exclusively used in a¹rmative contexts (Diewald 1999: 160; Mortelmans 1999: 241) and whose past subjunctive form (müsste) generally does not trigger a speaker-oriented interpretation (Mortelmans 1999: 400–401). In view of the argumentation put forward in this paper, this can only be taken to mean one thing: müssen is not anaphoric or reactive and therefore by itself very well capable of transporting speaker-oriented meanings (Diewald 1999: 224). And indeed, it can be observed that indicative muß is often used as a subjective marker, both in the root (Mortelmans 1999: 759) and the epistemic realm (Diewald 1999: 215–225). (30) a.

Petra Kelly ruft noch in derselben Nacht ihre innigste Freundin an, Erika Heinz in Calw: “Du mußt mir helfen. Du mußt sofort kommen. Noch heute Nacht!” [Spiegel (1993) 26, p. 126] ‘On the same night, Petra Kelly calls her best friend, Erika Heinz in Calw: “You must help me. You must come immediately. Tonight!”’

Creating subjectivity: negation and Konj. II with root modals 177

b. Doch es muß unendlich viel Mühe gekostet haben. [Die ZEIT, 19.04.85, p.70] ‘But it must have cost a tremendous amount of time and eŸort.’

4. Conclusion: a systematic account of modal verbs and mood in German Diewald (1999) argues in favor of an integration of the epistemically used German modals into the more strongly grammaticalized mood system, since epistemic modals and the moods Indikativ and Konjunktiv can be considered as modal deictic expressions with a similar function, i.e. they express the speaker’s epistemic stance with respect to a state of aŸairs. The on the whole less grammaticalized root modal uses are discarded by Diewald: as Nennwörter, nonepistemic modals are said to be part of the propositional content of the sentence, whereby speaker-oriented interpretations only arise in an indirect way. In the course of this paper, however, I have argued that a number of similarities between root modals — more speciªcally their speaker-oriented usages — and epistemic modals can be detected, which calls Diewald’s exclusion of the root modals as deictic expressions into question. First, both speakeroriented root modals and epistemic modals refer to the deictic center (prototypically the speaker) as a locus of potency. Second, the state of aŸairs expressed in the inªnitive is typically non-actual, as it is localized in either potential or projected reality. In other words, both speaker-oriented root and epistemic meanings can be described in terms of the dynamic evolutionary model. Third — and this can be considered as the main issue of this contribution — root and epistemic modals interact with negation and the past subjunctive in a similar way, i.e. negation and past subjunctive enhance the tendency towards a speakeroriented interpretation. These observations therefore call for an integrated account of mood and modal verbs, which should not concentrate too much on the conceptual diŸerence between root and epistemic meanings (see also Langacker 1991: 272), but focus on the relationship to the deictic center and on the way in which this relationship is linguistically construed. Speaker-oriented usages of indicative sollen, for instance, tend to occur in contexts in which the speaker objectiªes her own involvement by bringing it onto the linguistic scene (see Examples 14a–c), whereas the speaker using past subjunctive sollte generally does not explicitly mark her personal involvement. The latter can therefore to claimed to be inherently subjective. As the German root modals

178 Tanja Mortelmans

provide interesting borderline cases between descriptive uses, on the one hand, and inherent speaker-oriented ones, on the other, an analysis of the German modals in terms of deictic expressions should not a priori restrict itself to their epistemic usage.

Notes 1. The status of the imperative as a genuine mood marker or Modus is rather controversial. It is excluded from the category of mood in Diewald (1999: 172); van der Auwera and Plungian (1998: 83) “consider distinctions such as imperative […] as pertaining to illocutionary type [rather] than to modality”, a position also taken by Langacker (1991: 503), who treats imperative as one of the basic sentence types, on a par with declarative and interrogative. 2. It is well-known that tense and mood categories tend to overlap. One could for instance refer to the temporal and modal functions of German werden and English will, the con¶ation of tense and mood in the German Konjunktiv I and II (mood markers built on a present and past verb stem respectively) or the mainly modal use of the original past tense forms would, could, should, might (and must!) in English. For an alternative explanation, however, see Leiss 1992. 3. In fact, Diewald (1999) attributes a double function to the Konjunktiv I: apart from signalling a change of perspective, the Konjunktiv I is also said to characterize the relationship between the reported speaker and the proposition as [+factual], i.e. the reporting speaker asserts that the reported speaker has indeed said what is quoted (Diewald 1999: 182). The latter function is taken to be the truly deictic one; note that in this analysis the Konjunktiv I is associated with a [+factual] deictic value, which is rather awkward. 4. The conditional interpretation, whereby the nonfulªllment of a particular condition triggers the non-factuality of the state of aŸairs, is to be viewed as a default interpretation, i.e. the most likely interpretation of a past subjunctive verb form, given that no other contextual clues are at hand. The past subjunctive does not necessarily trigger a counterfactual interpretation, though. A well-known exception is the so-called polite or deferential use of the past subjunctive (Ich hätte eine Frage, lit. I would have a question). 5. Arguments in favor of integrating negation in the modal system are provided by Lehmann (1991: 1), who claims that negation is sometimes expressed on a par with other modal markers, i.e. by means of auxiliaries resembling modal verbs or in¶ectional markers resembling other mood markers. Compare in this respect also Honda (1996), who discusses a number of languages in which the negative marker is also used to mark irrealis events or events unrealized for the speaker, such as imperatives, conditional protases, and sentences referring to future states of aŸairs (Honda 1996: 190). 6. Strictly speaking, the modal wollen does not count as a preteritpresent, as it is derived from an original optative form (Diewald 1999: 181). The only other surviving preteritpresent that does not belong to the group of modal verbs is wissen. It contrasts with the modal

Creating subjectivity: negation and Konj. II with root modals 179

verbs on both a syntactic and a semantic level. When used as an auxiliary, wissen does not take a bare inªnitive, but an inªnitive with zu. It can then be paraphrased as know how to, be able to (e.g. sich zu benehmen wissen: ‘know how to behave oneself’) so that it resembles können. Unlike können (and the other German modals), however, wissen does not exhibit an epistemic or epistemic-like meaning. 7. Following Coates (1983), Sweetser (1990) and Langacker (1991), amongst others, I will use the notion root modality as a cover term for non-epistemic modality. 8. Note that word-by-word translations will not be provided for the German examples to come. The English translations are mainly intended to capture the meaning of the modal verb rather than to exactly render the German original. 9. “[…] if something is socially or morally required and expected then, given the assumption that people are generally concerned to do what is socially and morally correct, it will probably be done” (Nordlinger and Traugott 1997: 313). 10. This resembles Achard’s (1998) description of the French modal pouvoir in its ability and possibility sense: “the subject is [...] most strongly associated with the locus of potency, precisely because of her intentionality” (Achard 1998: 149). 11. Diewald is slightly surprised at the high percentage of negated dürfen in her corpus material: out of 20 non-deictic indicative occurrences, only 7 are a¹rmative (Diewald 1999: 132). 12. Example taken from Anweisung für Fahrer von Dienstwagen, 01.03.1962, p. 11. For reasons of space, this example is not translated, its main function being an illustration of dürfen’s outspoken preference for negation. 13. A directive environment can be characterized as a speech situation that easily invites a directive interpretation. Expressions in which the subject (addressee) will have to function as agent of the action speciªed in the inªnitive qualify as such (e.g. you must do that). 14. The data were collected from a number of electronic corpora provided by the Institut für deutsche Sprache (http://corpora.ids-mannheim.de/~cosmas/). A set of 72 occurrences was compiled, 66 of which appeared in a declarative sentence. 15. Diewald notes that müssen can take up a deontic meaning of obligation as well, which is, however, only a contextually induced variant of its basic dispositional meaning (Diewald 1999: 154). 16. In 10 of these 117 occurrences, sollte is to be analyzed as a past indicative rather than a past subjunctive; these instances will be left undiscussed. 17. Occurrences of sollte in the protasis of a conditional are excluded for the time being (but see Section 2.2.2, Example (24)). 18. Since the non-factuality of the protasis can be taken to represent the default case (Comrie 1986, Akatsuka 1985), conditionals whose protasis is factual, are marked. 19. Note that muss in the same corpus occurred no less than 439 times, its past subjunctive counterpart müsste, however, only 22 times. 20. Nordlinger and Traugott (1997: 308) draw attention to the fact that prototypical narrow-scope deontics have a deªnite subject. To the extent that sollte favors indeªnite,

180 Tanja Mortelmans

generic subjects like man or niemand, it cannot be considered as a prototypical narrowscope deontic modal. 21. Compare also DWB (16, 1468): “grundbedeutung von sollen ist die einer verp¶ichtung oder eines zwanges, der auf einem fremden willen beruht [basic meaning of sollen is one of obligation or compulsion that is based on other people’s will].”

References Achard, Michel 1998 Representation of Cognitive Structures. Syntax and Semantics of French Sentential Complements [Cognitive Linguistics Research 11]. Berlin and New York: de Gruyter. Akatsuka, Noriko 1985 “Conditionals and the epistemic scale”. Language 61(3): 625–639. Bausch, Karl-Heinz 1979 Modalität und Konjunktivgebrauch in der gesprochenen deutschen Standardsprache. München: Hueber. Brisard, Frank 1999 A critique of localism in and about tense theory. Ph.D. dissertation, Department of Germanic Linguistics and Literature, University of Antwerp. Bühler, Karl [1934] 1999 Sprachtheorie: Die Darstellungsfunktion der Sprache. Mit einem Geleitwort von Friedrich Kainz. Stuttgart: Lucius & Lucius. Bybee, Joan L., Perkins, Revere and Pagliuca, William 1994 The evolution of grammar: tense, aspect, and modality in the languages of the world. Chicago: University of Chicago Press. Coates, Jennifer 1983 The Semantics of the Modal Auxiliaries. London and Canberra: Croom Helm. Comrie, Bernard 1986 “Conditionals: a typology”. In On Conditionals, E. C. Traugott, A. ter Meulen, J. Snitzer Reilly and C. A. Ferguson (eds), 77–99. Cambridge, London and New York: Cambridge University Press. Diewald, Gabriele 1999 Die Modalverben im Deutschen: Grammatikalisierung und Polyfunktionalität [Reihe Germanistische Linguistik 208]. Tübingen: Niemeyer. Duden, Grammatik der deutschen Gegenwartssprache 1998 Herausgegeben von der Dudenredaktion. Bearbeitet von Peter Eisenberg […]. 6. neu bearb. Auflage [Der Duden in 12 Bänden Bd. 4]. Mannheim, Leipzig, Wien, Zürich: Dudenverlag. DWB = Deutsches Wörterbuch 1854ff. Jakob und Wilhelm Grimm. 33 Bde. Leipzig: Hierzel [Reprint München 1984].

Creating subjectivity: negation and Konj. II with root modals

Fabricius-Hansen, Cathrine 2000 “Die Geheimnisse der deutschen würde-Konstruktion”. In Deutsche Grammatik in Theorie und Praxis, R. Thieroff, M. Tamrat, N. Fuhrhop and O. Teuber (eds), 83–96. Tübingen: Niemeyer. Fritz, Gerd 1997 “Historische Semantik der Modalverben. Problemskizze — exemplarische Analysen — Forschungsüberblick”. In Untersuchungen zur semantischen Entwicklungsgeschichte der Modalverben im Deutschen, G. Fritz and T. Gloning (eds) [Reihe Germanistische Linguistik 187], 1–157. Tübingen: Niemeyer. Goossens, Louis 1999 “Metonymic bridges in modal shifts”. In Metonymy in Language and Thought, K.-U. Panther and G. Radden (eds) [Human Cognitive Processing 4], 193– 210. Amsterdam: Benjamins. 2000 “Patterns of meaning extension, “parallel chaining”, subjectification, and modal shifts”. In Metaphor and Metonymy at the Crossroads: A Cognitive Perspective, A. Barcelona (ed.) [Topics in English Linguistics 30], 149–170. Berlin and New York: de Gruyter. Heidolph, Karl Erich, Flämig, Walter and Motsch, Wolfgang 1981 Grundzüge einer deutschen Grammatik. Berlin: Akademie. Honda, Isao 1996 Negation: A cross-linguistic study. Ph.D. dissertation, Faculty of the Graduate School of State, University of New York at Buffalo. Jäger, Siegfried 1971 Der Konjunktiv in der deutschen Sprache der Gegenwart: Untersuchungen an ausgewählten Texten. München: Hueber. Kasper, Walter 1987 “Konjunktiv II und Sprechereinstellung”. In Satzmodus zwischen Grammatik und Pragmatik. Referate anläßlich der 8. Jahrestagung der Deutschen Gesellschaft für Sprachwissenschaft, Heidelberg 1986, J. Meibauer (ed.), 96–113. Tübingen: Niemeyer. Langacker, Ronald W. 1991 Foundations of Cognitive Grammar. Volume II: Descriptive Application. Stanford: Stanford University Press. Lehmann, Christian 1991 “Strategien der Situationsperspektion”. Sprachwissenschaft 16: 1–26. Leiss, Elisabeth 1992 Die Verbalkategorien des Deutschen: ein Beitrag zur Theorie der sprachlichen Kategorisierung [Studia linguistica Germanica 31]. Berlin: de Gruyter. Mortelmans, Tanja 1999 Die Modalverben ‘sollen’ und ‘müssen’ im heutigen Deutsch unter besonderer Berücksichtigung ihres Status als subjektivierter ‘grounding predications’. Ph.D. dissertation, Department of Germanic linguistics and literature, University of Antwerp.

181

182 Tanja Mortelmans

2000

“Konjunktiv II and epistemic modals in German. A division of labour”. In Constructions in Cognitive Linguistics, A. Foolen and F. van der Leek (eds) [Current Issues in Linguistic Theory 178], 191–215. Amsterdam and Philadelphia: Benjamins. Nordlinger, Rachel and Traugott, Elizabeth Closs 1997 “Scope and development of epistemic modality: evidence from ought to”. English Language and Linguistics 1(2): 295–317. Sweetser, Eve 1990 From etymology to pragmatics: metaphorical and cultural aspects of semantic structure [Cambridge studies in linguistics 54]. Cambridge: Cambridge University Press. Talmy, Leonard 1988 “Force dynamics in language and cognition”. Cognitive Science 2: 49–100. 2000 Toward a Cognitive Semantics. Volume I: Concept Structuring Systems. Cambridge, Massachusetts and London, England: The MIT Press Traugott, Elizabeth C. 1989 “On the rise of epistemic meanings in English: an example of subjectification in semantic change”. Language 65(1): 31–55. van der Auwera, Johan and Plungian, Vladimir A. 1998 “Modality’s semantic map”. Linguistic Typology 2: 79–124. Zifonun, Gisela, Hoffmann, Ludger and Strecker, Bruno 1997 Grammatik der deutschen Sprache. Volume 3. Berlin/New York: de Gruyter.

Part 3: Person and Text

Politeness distinctions in second person pronouns Johannes Helmbrecht University of Erfurt

1.

Introduction

Politeness distinctions in personal pronouns of the second person are a common phenomenon in European languages. Nevertheless, these distinctions have received attention in linguistic typology only sporadically. The present paper aims to ªll this gap. After a discussion of the deictic nature of personal pronouns and their formal properties (Section 2), an overview will be given on the range of data which can be found in the languages of the world with respect to politeness distinctions in second person pronouns (Section 3). Section 4 will show that politeness is a functional domain of language use; two major politeness strategies will be discussed which may well point to the emergence of politeness distinctions in pronouns. The frequency of politeness distinctions and the areal distribution of the diŸerent historical sources of polite pronouns in the languages of the world will be presented in Section 5. There are certain large areas in the world where politeness distinctions occur, which allows the conclusion that politeness distinctions in pronouns are not only functionally conditioned but also areally by means of linguistic and cultural contact.

2.

Theoretical background

Personal pronouns belong to the deictic expressions of a language. Speakers use personal pronouns in order to direct the attention of the addressee to one or more human individuals who may be present or absent in the actual speech situation. Usually, this act of reference is an integral part of a speech act, or more generally of a communicative exchange.

186 Johannes Helmbrecht

There is a wide variety of linguistic means a speaker may use in order to refer to one or more people that he wants to talk about or he wants to talk to. Expressions such as the tall guy with the red hair or this woman over there are noun phrases used to pick a speciªc human being out of a range of referential possibilities in a certain speech situation. In addition, these expressions indicate that the individuals referred to are presumably not known to the addressee or that they are not activated at the time of speaking in the mind of the addressee. Therefore, the speaker adds speciªcations to facilitate the addressee’s interpretation in order to understand who is meant by the speaker. These speciªcations may be linguistic or gestural or both. An expression like the woman over there, for instance, is usually accompanied by a certain body position of the speaker — he or she faces or turns slightly towards the intended referent, and in most cases a slight gesture with the eye or arm helps the addressee identify the referent. Other linguistic means for referring to people are “honoriªc terms of address” such as Mister, Madam, which are often combined with a proper name, “relational nouns” such as father, uncle which imply a certain kinship relation to someone else who has to be obligatorily speciªed by possessive pronouns in some languages, “proper names” such as Tom Smith, and last but not least “personal pronouns”. The distinctive features of personal pronouns may be summarized as follows. 1. Personal pronouns are inherently referential expressions. They do not occur with modifying and determining elements that are otherwise necessary to form referential noun phrases. Adjectives, attributive demonstratives, and articles are dependent elements (in terms of dependency grammar) in a noun phrase with a lexical noun as head and fulªll certain functions there. 2. The referent of a personal pronoun can be identiªed only with respect to the actual speech event, i.e. according to the speech act roles he or she performs during the speech event. Personal pronouns were classiªed as “indexical symbols” in semiotics exactly because of this trait. They have a speciªc semantic content, the speech act roles, and at the same time they are pointers, i.e. they establish a reference which can be decoded only with respect to the actual speaker who uses the pronoun in his utterance (cf. Jakobson 1971[1957], and Helmbrecht 2002 for a more detailed discussion of this categorization). What makes personal pronouns (and deictics in general) a special category of linguistic expression is that they encode a certain type of relation, the relation between the “origo” (cf. Bühler 1982[1934]) and an intended “referent”. Every act of pointing presupposes a commonly shared point of departure,

Politeness distinctions in second person pronouns 187

the indexical ground (cf. Hanks 1990, 1992), from which the pointing starts. This is the origo of gestural as well as linguistic pointing. The addressee of the pointing needs to be aware of the indexical ground of pointing; otherwise s/he is neither able to interpret the pointing itself nor to correctly identify the intended target. The relation between origo and target encoded in personal pronouns (and other deictic expressions) re¶ects a basic cognitive principle, the conceptualization of scenes along the notions ªgure and ground (cf. Langacker 1990; Dirven & Verspoor 1998). The origo of the pointing is the presupposed ground of the reference to an entity in the focus of attention. The referent is the ªgure in this relation. It is the default assumption that the origo of a pointing act coincides with the speaker of the same act. However, the origo may be shifted to some other person, a process which then has to be marked by certain linguistic means (cf. Bühler 1982: 102–120). The target of the act of pointing is the intended referent. The intended referent is usually characterized semantically by the pointing expression. Demonstrative pronouns, for instance, may characterize the referent as male or female, as human or non-human, or as animate or inanimate. Personal pronouns characterize the referent as well with respect to the speech act role and the size of the respective speaker and hearer groups. Deictic expressions are, in addition, characterized by the encoding of a “deictic relation” between the origo and the intended referent. This deictic relation is semantically characterized in demonstrative pronouns. Categories such as proximal versus distal, or visible versus non-visible indicate the type of relation between the origo and the intended referent (cf. Hanks 1992: 51). The speciªc deictic relation between the origo and the referent in personal pronouns of the ªrst and second person coincides with the characterization of the referent itself. The addressee is the hearer of the utterance in relation to the speaker, who is the indexical ground for the second person singular reference. The relational structure of deictic expression is summarized in Figure 1. The indexical ground of the deictic reference is the origo in Figure 1. As already mentioned above, the default origo for pronominal reference is the speaker. Personal pronouns of the ªrst and second person singular refer to the respective speech act participating individuals by characterizing the intended referent according to his or her performance of a certain speech act role. Who is meant can be determined only with regard to the origo, the actual speaker. Pronouns such as I and you encode not only the speech act role of the intended referent, but also the relation of the intended referent to the origo, the speaker of the actual speech act. Both the characterizing semantics and the deictic

188 Johannes Helmbrecht

NON-SPEECH ACT PARTICIPANT

Deictic Relation REFERENT (Figure)

ORIGO (Indexical Ground)

SPEAKER (Default indexical ground)

ADDRESSEE

Figure 1. The relational structure of personal pronouns

relation to the origo are conventionally (i.e. semantically) encoded in personal pronouns. It is especially the encoded relation to the origo (indexical ground) that makes personal pronouns (and other deictics) a particular class of expressions in language. 3. There are two basic speech act roles, speaker and hearer, which are opposed to a negatively deªned non-person category (cf. Benveniste 1956). Personal pronouns encode the singular roles as well as various combinations of these roles in the languages of the world. The identiªcation of the referent of a ªrst or second person personal pronoun in general does not pose a problem for the addressee. A communicative event is a highly structured situation with clear acoustic, perceptual, and gestural signs about who is performing which role. The structure of a communicative event together with the rules of the use of personal pronouns is learned empractically (cf. Bühler 1982[1934]) from the beginning of language acquisition. It is therefore hardly surprising to ªnd a parallelism between cognitive and linguistic structure. From the point of view of semantic structure, there are some similarities between kinship nouns such as father, polite nominal forms of address such as excellence, and personal pronouns. These expressions construe diŸerent social roles in terms of a kinship relation or in terms of a social relation between the addressee and the speaker. Their primary function, however, is not the identiªcation of the referent — as is the case with speech act roles designating

Politeness distinctions in second person pronouns 189

pronouns — but rather the attribution of respect to the addressee due to the social rank of the role expressed. Nouns designating social ranks are possible historical sources for personal pronouns (cf. Lehmann 1995: 39–42). 4. Personal pronouns are shifters (cf. Jakobson 1971[1957]), i.e. they have a systematically changing reference depending on the change of speech act roles during a conversation. This feature of pronoun usage is re¶ected in the systematic correlation between the diŸerent pronominal forms within a single paradigm. In this respect, pronouns contrast sharply with proper names, which have a unique reference throughout all instances of discourse. Since speech act roles are the deªning semantic features of personal pronouns, the number of forms in a paradigm is necessarily limited. Proper names and nouns employed for reference to people do not have such limitations; they usually belong to an open class of words in a language. Reference to absent or present human participants in a speech event is certainly the primary function of personal pronouns. Another aspect of the reference of personal pronouns, however, is what has been termed social deixis in the literature (cf. Levinson 1983: 89–94; Fillmore 1997). There are many languages in which diŸerent pronominal forms may be used to indicate diŸerent degrees of respect. In principle, there are two types of honoriªcation: referent honoriªcs (cf. Levinson 1983: 90, Shibatani 1998: 342–44), i.e. polite reference to a third person, and addressee honoriªcs, i.e. polite reference to an addressee of the speech act. The latter are the subject of this paper. Only those forms displaying at least some of the distinguishing characteristics of personal pronouns are considered.

3.

Pronominal forms of polite address — the data

There is a remarkable variety of linguistic means to express some degree of politeness with respect to the addressee. The linguistic strategies speakers use to account for the various social relationships between themselves and the hearers in a communicative event may lead to the conventionalization of nonprototypical usages of already existing personal pronouns in a language. For example, speakers of French and other languages in Europe and many other languages in the world use second person plural pronouns to address the hearer politely. The exact historical date, political region, and circumstances under which French people started to use vous to address higher ranking or

190 Johannes Helmbrecht

foreign persons is not known, although it is plausible to assume that this convention was developed during the Middle Ages around the courts of the French crown and the aristocracy. The polite use of vous became a conventional rule. This usage then introduced a new distinction into the paradigm of personal pronouns without increasing or changing the actual number of forms. The second person singular reference received a distinction between familiar/equal versus formal/polite, a distinction not made in second person plural reference in French. The fact that so many languages conventionalize the polite use of the second person plural pronoun for the reference to a single addressee does not mean that the pragmatic rules underlying the usage of second person plural pronouns are the same in all languages. Politeness rules may diŸer considerably across languages. The sets of people who are considered to deserve polite address by means of a second person plural pronoun (or other honoriªc second person pronoun) may vary signiªcantly from language to language. In addition, the polite form and the corresponding familiar form may be used either asymmetrically or symmetrically. In the former case, one of the interlocutors has to give the polite form whereas the other uses the familiar form in return. In the latter case, both use the same level of politeness. And thirdly, these underlying pragmatic rules of the use of polite forms may change over time within a single language. For example, the usage of the second person plural polite pronoun you in the early period of Modern English was generalized to such a degree that the original second person singular pronoun thou became obsolete. Finally the familiar/polite distinction disappeared in the English pronoun system leaving the former second plural polite pronoun as the sole pronoun for second person reference in the system. The linguistic strategies to account for the social relations between interlocutors may lead to the intrusion of new pronominal forms. In Spanish, e.g. the forms usted/ustedes, which are derived historically from a complex term of polite address — vuestra merced meaning ‘your grace’, became an integral part of the pronominal paradigm. They are used for second person singular and plural polite address. Here, the introduction of politeness distinctions in the paradigm of personal pronouns was achieved by the introduction of new forms. The grammaticalization of certain nouns — mostly designating high social status — to pronouns can be observed frequently in Far East Asian languages where the linguistic expression of politeness is much more pervasive than in European languages. In Table 1, I have compiled a list of pronominal forms and their diachronic sources used in languages as addressee honoriªcs. For each

Politeness distinctions in second person pronouns

usage type that may lead to a change in the pronoun system with respect to politeness distinctions, one illustrative example is given. Table 1. Pronominal forms for polite addressee reference1 Target

Source

Example

2nd sg hon

2nd pl 3rd sg 3rd pl 1st pl Dem (2nd) Re¶. Status term

French: tu/vous Italian: tu/lei German: du/Sie Ainu: aoka (< a- (1st transitive actor) + oka (exist.pl) Sinhalese: oyaa Hungarian: maga (2nd sg hon) / maguk (2nd pl.hon) Spanish: Usted (2nd sg hon < vuestra merced ‘Your grace’)

Third person singular pronouns as well as third person plural pronouns are used to express respect. First person plural pronouns (particularly inclusive forms), demonstrative pronouns, re¶exive pronouns, and nouns designating social status or social rank are used for polite reference to the addressee. A distinct word class of personal pronouns as we know them from European languages is sometimes di¹cult to discern in Far Eastern Asian languages. In Japanese, Vietnamese, and other languages of that area, there are numerous words which are used in a way similar to European pronouns, namely to refer to the speaker, hearer, a non-speech act participant or combinations of these referential categories. In addition, these words distinguish the sex of the speaker or addressee, his or her social rank, the level of style of the conversation and the social distance between the interlocutors. These distinctions are di¹cult to systematize and represent in an ordered paradigm with pairs of oppositions as is usually done with personal pronouns in other languages. Rather, these words display many grammatical features typical of nouns in these languages. Therefore, attempts were made to classify and to describe these words by means of a pronoun — noun continuum rather than as a single class of words (cf. Sugamoto 1989). Another characteristic of these languages is that their speakers generally try to avoid the use of pronouns in discourse. The intended referent, no matter whether this is the speaker, hearer or some third person, often has to be pragmatically inferred from the situational or linguistic context of the utterance. Because English and other European languages have personal pronouns in syntactic positions where these languages have none, this phenomenon has been dealt with under the rubric of zero anaphora (cf. e.g. Li & Thompson 1979). This phenomenon should not be confused with what has been called pro

191

192 Johannes Helmbrecht

drop in the generative tradition. Pro drop languages indicate the referent of the subject by means of a bound pronominal a¹x on the verb, i.e. the syntactic/ morphological place of the pronominal reference is diŸerent, but there is a pronominal expression. Nevertheless, a closer look at words used for person reference in these Far Eastern Asian languages reveals that some of them come close to the traditional notion of personal pronouns and these are covered in the present investigation. There are words in Japanese and in Vietnamese which are predominantly used e.g. for speaker reference or hearer reference and their morphosyntactic behavior diŸers in signiªcant ways from the behavior of regular nouns. In Japanese, for example, nouns are generally not pluralized. They are free to receive a plural or singular interpretation depending on the context, but may optionally take a plural marker. Some of the putative personal pronouns such as ªrst person watashi are intrinsically singular and have to be pluralized obligatorily if used as a ªrst person plural pronoun; e.g. watashi is pluralized by a su¹x -tachi (watashi-tachi ‘we’; cf. Sugamoto 1989: 276). Kinship terms such as Vietnamese anh ‘elder brother’ also have certain diŸerences in their grammatical behavior if they are used for pronominal reference. In this pronominal usage they are not allowed to take nominal classiªers which are possible in their nominal usage (cf. Cooke 1968: 125).

4.

Functional interpretation

Politeness is a functional domain of language and language use. Linguistic research in pragmatics, anthropology and sociolinguistics of the last decades has shown that the dimension of participation (cf. Hymes 1974; Duranti 1985, 1997; GoŸman 1981; Levinson 1983; a.o.), i.e. the question of who is talking to whom in which situation, is crucial for the understanding of the linguistic and cognitive structure of expressions and grammatical categories. This is especially true for the social relations among the participants in speech situations. Speakers, hearers, bystanders, audience and others not present but who are talked about are not simply senders, receivers, or subjects of a message, but rather they are normally connected to each other within a system of social roles and relations and cultural practices. There are two social parameters that seem to be of particular signiªcance with respect to politeness. One involves the relative social rank, power and prestige of the individual members within a society, the other involves the

Politeness distinctions in second person pronouns 193

social distance, i.e. the degree of intimacy and familiarity relative to one another.2 The relative social rank of an individual depends on variables such as age, sex, kinship relation, heritage, wealth, political power etc. and the social roles that the individual is licensed to play. The parameter of relative social distance describes the fact that people who are close relatives or friends, or who belong to the same peer groups etc., are usually closer to each other than unrelated people who do not know each other on an every-day basis. It is obvious that these parameters and values play a role as conceptual categories and determine the choice of linguistic means for polite address. But the question remains what exactly politeness is and what linguistic strategies are used to express respect. A very powerful theory of the nature of politeness is the concept of face proposed by Brown & Levinson (1987). Face is the public self-image every adult member of a society wants to claim for him- or herself. Face is an important part of the personal identity, it includes a knowledge of the social position of oneself as well as the social position of others. Brown & Levinson assume “that the mutual knowledge of members’ public self-image or face, and the social necessity to orient oneself to it in interaction, are universal” (Brown & Levinson 1987: 62). They hypothesize that there are basically two aspects of face, negative face, which is “the basic claim to territories, personal preserves, rights to nondistraction — i.e. to freedom of action and freedom from imposition” (Brown & Levinson 1987: 61). On the other hand, positive face crucially includes the desire that this self-image should be appreciated and approved by the other interactants (cf. Brown & Levinson 1987: 62). In short, negative face includes all wants which have to do with the maintenance of the self-assigned autonomy of a person, while positive face, on the other hand, includes all wants which have to do with the positive evaluation and acceptance of oneself by others. Face is interpreted by the authors not as a psychological entity, but as face wants of the speaker and hearer. There are numerous acts which may threaten these face wants. For example, orders and requests may threaten the addressee’s negative face wants, because these speech acts might urge him to respond in a positive way i.e. to do what was ordered or requested although he does not like it. The positive face wants of the addressee are threatened if the speaker expresses disapproval, criticism or complaints about the addressee. Now, since these face-threatening acts (hence FTAs) occur all the time and are unavoidable to some degree, speakers may follow certain linguistic strategies to combine these FTAs with redressive actions. These redressive actions constitute politeness. Their purpose is to minimize the imposition of the addressee if an FTA is

194 Johannes Helmbrecht

unavoidable. Corresponding to the two types of face wants, there are two types of redressive actions, i.e. two types of politeness; negative politeness and positive politeness. Positive politeness comprises all the redressive strategies which appeal to the positive face wants of the addressee. An example may illustrate this point. If you want to ask someone for a favor, it might be helpful to highlight in advance that you belong to the same set of people who share speciªc wants, goals and beliefs. Linguistically, the one who intends to direct the request to someone else may use in-group markers such as the same slang or dialect variety as the addressee, nicknames which are normally used only among close friend or colleagues, or he may include the addressee in the designated activity by using a ªrst person plural inclusive pronoun. An example for this strategy is given in (1). The situation is as follows. A customer asks for ten regular stamps in a post o¹ce. In response, the employee oŸers a prefabricated booklet instead of stamps taken from a larger sheet of stamps, asking the following question. (1) Nehmen wir auch das kleine Briefchen hier ? ‘Do we also take this little booklet (with stamps) here ?’

The employee used the ªrst person plural pronoun including himself in the activity expressed, although his role in this situation is obviously the opposite. His role is to sell the stamps and not to buy them. But this type of inclusion is a strategy to express positive politeness. The employee expresses that he shares the interest and perspective of the customer. It is plausible to imagine that this polite use of ªrst person plural pronouns in German, which is not conventionalized, may become a convention in German or was the starting point for the development of a second person polite pronoun out of a ªrst inclusive plural form in other languages. This is perhaps what happened in Ainu, an isolate language of Japan, in which one can observe an extension of the ªrst plural pronoun to the second person polite meaning (cf. Table 1 above). Negative politeness includes all linguistic strategies that diminish the degree of potential imposition of an FTA on an addressee. Speakers try to avoid a direct and clear expression of the FTA. This avoidance strategy includes, among other things, a lowering of the illocutionary force of the speech act and an avoidance of a direct reference to the addressee. The latter aspect is particularly important for the rise of politeness distinctions in pronouns. The use of second person plural pronouns in order to refer to a single individual could have been motivated by the reluctance to address the hearer directly. An FTA usually forces an addressee to act in a way intended by the

Politeness distinctions in second person pronouns 195

speaker. A speaker X who is not licensed to commit an FTA toward hearer Y because of his social position with respect to Y or because of the personal relation between them may use the second person plural pronoun instead of the singular one in order to dissolve the obligation imposed on Y among a plurality of addressees. Speaker X diminishes Y’s responsibility to act in the requested way by addressing a plurality of hearers. The reference to the single addressee can easily be inferred from the situational context and does not pose a problem for the interlocutors. If this usage of the second person plural pronoun is repeated frequently, it may become a convention of language use. However, as soon as there is a rule in a language to use the second person plural pronoun for polite address, this choice is obligatory if the pragmatic and sociolinguistic conditions are fulªlled. Then, people do no longer use this pronoun to make an indirect reference. The French second plural pronoun vous, for instance, is used like a normal second person singular pronoun. But how do the second person plural pronouns acquire polite meanings historically? The assignment of polite meaning to personal pronouns is certainly an epiphenomenon of the linguistic practices outlined above, i.e. it is an unintended result of this pronoun usage. At some point in history, hearers of a social rank higher than the speakers’ are repeatedly addressed in the second person plural in order to minimize FTAs. The second person plural pronoun now receives a polite meaning and this polite usage will become conventionalized. The original motivation for this usage disappears and eventually, this process — which is an invisible-hand process in terms of the theory of linguistic change by Keller (1994) — leads to the introduction of a new politeness distinction in the pronominal paradigm.3 Further, the asymmetrical use of the second person plural polite pronoun — i.e. the social superior receives the polite second plural pronoun but uses the singular pronoun to address the inferior — may develop into a symmetrical usage later, e.g. among strangers, since speakers who do not know each other cannot be certain about their relative social position and may therefore address each other slightly more politely than the rules may require. The avoidance strategy of speakers is the main functional source for the emergence of politeness distinctions in pronouns. The processes by which diŸerent forms acquire a politeness meaning vary, however. Owing to lack of space, I cannot give a detailed analysis of the other sources of politeness distinctions given in Table 1 (cf. Helmbrecht 2001, 2002).

196 Johannes Helmbrecht

5.

Frequency and areal distribution

A statistical survey of the frequency and geographical distribution of politeness distinctions in pronouns shows (a) that about 25% of all languages have politeness distinctions in pronouns and (b) that the occurrence of politeness distinction is not randomly distributed but shows certain geographical foci. In Table 2, an overview of the relative frequencies of the diŸerent techniques to express politeness is given. The frequencies are calculated from a representative sample of 100 languages that has been established independently of the research topic. It turns out that the use of the second person plural pronoun is the most common technique to address politely among the languages of the sample. Less frequent is the use of third person plural pronouns, re¶exive pronouns, and kinship and status nouns. Some of the techniques, such as the use of third person singular pronouns, are not attested in the languages of the sample. The majority of languages showing politeness distinction in second person pronouns use pronominal sources to express these distinctions. Only a third of these languages, that is 8% of all the languages examined, use status or kinship nouns. The languages which use kinship nouns for polite address are mostly located in the East and Southeast Asian area. Table 2. Linguistic techniques and frequency Target

Source

Examples

2nd sg hon

French, Persian, Yoruba, Nama, Kannada, Fijian, etc., (not found e.g. in central America) 3rd sg Italian, Hungarian, Indonesian 3rd pl German, Hindi, Kannada, Luvale, Tagalog 1st pl Nahuatl, Ainu Dem Sinhalese Re¶. Imbabura Quechua, Hungarian, Hindi Kinship term Indonesian, Vietnamese, Burmese, Thai Status term Dutch, Spanish, Rumanian, Burmese, Vietnamese, Thai

Frequency

2nd pl

16 % n.a.4 5% n.a. n.a. 3% 2% 6%

In Table 3, the languages of the sample are grouped according to their membership in diŸerent honoriªc types. Four types are distinguished: 1. languages which have no politeness distinctions in their pronouns, 2. languages which show the European type binary politeness distinction, 3. languages which show two and more degrees of politeness distinctions in their pronoun system, and 4. languages whose speakers in general avoid the use of pronouns for second person polite address.

Politeness distinctions in second person pronouns 197

The ªrst parameter is self-explanatory. Languages grouped under this feature in my database have no personal pronouns in their paradigms used to express diŸerent degrees of respect or intimacy toward the addressee. Around 75% of the sample have no politeness distinctions in personal pronouns. As can be seen from Table 3, there are certain large geographical areas where politeness is not a category of personal pronouns at all, e.g. North and South America, Papua New Guinea, Australia, and perhaps Africa. There are languages in Africa showing binary politeness contrasts. These binary politeness distinctions may be the result of language contact with the colonial languages such as French, German, and Dutch. This question certainly needs further investigation. The second parameter “Binary politeness distinction” covers all languages with a European T/V type of politeness distinctions. These languages have a paradigmatic opposition between at least one intimate or familiar pronoun of address and another one expressing respectful address. Such a binary distinction may also be expressed by several distinct pronouns. It is criterial that pronouns do not indicate more than one politeness distinction. Table 3 below shows that there are two culturally and historically connected areas where this honoriªc type is prevailing, Europe and to a lesser degree the Near and Middle East. The third parameter — “Several politeness distinctions” — covers all languages exhibiting two or more degrees of politeness within a pronominal paradigm. These systems are rare cross-linguistically. Languages with this feature do occur in Meso-America (e.g. Nahuatl) and in Southeast Asia (e.g. Java) they are, however, frequent in India. It seems to be a deªning linguistic feature of this area. Many languages there have two degrees of politeness expressed by a second person plural pronoun and an even more polite third person pronoun or re¶exive pronoun. The fourth parameter — “Pronoun avoidance” — is perhaps the most di¹cult of the four and deviates terminologically from the previous ones. The names of the ªrst three parameters describe a categorical feature of the pronominal paradigms to be investigated. The term “pronoun avoidance”, however, describes a strategy of pronoun usage which has some eŸect on the shape of the respective paradigm. Languages such as Japanese, Burmese, Korean, Thai and others in that area have a strong sensitivity to politeness within their grammars. Speakers generally avoid using personal pronouns when addressing other persons and rather use status and kinship terms, titles and other complex nominal expressions. The eŸect of this strategy is that there are rarely polite pronouns of address. If there are second person pronouns, they are used to address equals or inferiors. Polite forms of address (for addressing superiors)

198 Johannes Helmbrecht

Table 3. Distribution over global regions and continents Global regions/ continents 100 Languages in total

No politeness distinction

74

Europe

English

Near and Middle East

Abkhaz, Hebrew, Lezgian

India

Manipuri

Central Asia

Mongolian, Burushaski, Chukchee

Southeast Asia

Jakaltek, Mezquital Otomi, Rama, Copainala Zoque

Papua

Alamblak, Asmat, Bukiyip, Daga, Dani, Imonda, Kewa, Maibrat, Pasi

Australia

Goonyiandi, Kayardild, Mangarayi, Martuythunira, Maung, Tiwi, WangaaybuwanNgiyamba

Paciªc

Luvukaleve, Rapanui

South America

Africa

18

3

6

Persian, Georgian, Turkish Hindi, Kannada Chinese Tagalog

Mixtec

Fijian

Acoma, Cree, Inuit, Karok, Keres, Kiowa, Koasati, Kootenai, Lakhota, Makah, Maricopa, Oneida, Slave, Wichita, Yaqui Apurina, Barasano, Quechua Guaraní, Hixkaryana, Kraho, Mapuche, Mataco, Mura-Piraha, Sanuma, Warao, Wari, Yagua Amele, Egyptian, Bagirmi, Grebo, Hausa, Krongo, Lango, Malagasy, Oromo, Senoufo, Songhay, Swahili, Tamazight, Zulu

Several Pronoun politeness avoidance distinctions

Basque, Finnish, French, German, Greek, Russian, Spanish

Chamorro, Hmong Njua, Paiwan

Central America

North America

Binary politeness distinction

Luvale, Nama, Sango, Yoruba

Burmese, Indonesian, Japanese, Korean, Thai, Vietnamese

Politeness distinctions in second person pronouns 199

mostly do not belong to the class of personal pronouns in these languages. Hence there are lexical gaps in the group of pronominal forms with respect to second person politeness categories.

6.

Conclusions

The relational structure of personal pronouns includes two poles, the origo and the referential target. The origo is the cognitive ground for an act of pointing, the referent of the pointing is the ªgure. The peculiarity of personal pronouns is that they have lexicalized this relation. The default cognitive ground for pronominal reference is the speaker and the diŸerent types of targets semantically coded in pronominal paradigms are the speaker, the hearer, third persons, and combinations of these categories. The principal deictic nature of pronouns re¶ects the cognitive representation of the speech event, i.e. the relation between two individuals in a dyadic conversation. However, personal pronouns also re¶ect social relations. Speakers have clear concepts of the social relations in their society and their position in this network. Politeness distinctions in personal pronouns take up aspects of these social relations and the cultural practices which are associated with them. The investigation of politeness distinctions in personal pronouns has shown that the emergence of linguistic distinctions may be explained functionally on the basis of pronoun usage with respect to certain underlying strategies of positive and negative politeness. Pronouns of diŸerent person and number categories and other referring expressions receive politeness meanings secondarily via the conventionalization of their usage in FTAs with some redressive eŸorts. The typological study of the geographical distribution of honoriªc types has shown that politeness distinctions in pronouns are areal phenomena. This leads to the conclusion that the existence of politeness distinctions in personal pronouns of a speciªc language can be explained in two ways. One explanation may be functional, taking the politeness strategies of pronoun usage into account. The other way refers to contact-induced borrowing. The fact that there are certain large geographic areas where politeness distinctions in pronouns occur very frequently suggests that linguistic practices of polite address as well as the forms for polite address can be subject to borrowing.

200 Johannes Helmbrecht

Notes 1. The following abbreviations are used: 1st, 2nd, 3rd = first, second, third person, sg = singular, pl = plural, hon = honorific, Dem = demonstrative pronoun, Refl = reflexive pronoun, FTA = face threatening act. 2. These parameters were introduced by Brown, R. & Gilman, A. (1972) in their pioneering study on the polite use of personal pronouns. 3. This theory describes language change as the result of local processes of language use, which comprises two aspects. The ªrst is intentional. Speakers start to use a new expression or way of speaking in certain contexts for certain communicative reasons. If other speakers adopt this habit for whatever reason, this new way of speaking may become conventionalized. The eŸect on the language system, however, may be very diŸerent from what the speakers intended to perform with the new expression. If speakers avoid a direct reference in certain types of social situations — e.g. inferior speaking to superior — using a second plural pronoun (with a singular addressee reference), features of the context “superiority of the addressee” may ªnally becomes a semantic feature of the second plural pronoun. This has never been the intention of the speakers, but is a causal result of this new type of pronoun usage. 4. Since these languages do not belong to the 100 language[s] sample from which the percentages of occurrence are calculated, the occurrence of the third person singular pronoun as a polite form of addressee reference has to be marked as “not attested” in this table.

References Benveniste, Emile 1956 “La nature des pronoms”. In For Roman Jakobson. The Hague: Mouton. Brown, Penelope & Levinson, Stephen C. 1987 Politeness. Cambridge: CUP. Brown, R. and A. Gilman 1972 “The pronouns of solidarity and power”. In Language and social context, Giglioli, P (ed.), 252–82. Harmondsworth: Penguin. Bühler, Karl 1982[1934] Sprachtheorie. Stuttgart, New York: Gustav Fischer (UTB). Cooke, Joseph R. 1968 Pronominal Reference in Thai, Burmese, and Vietnamese. Berkeley, CA and Los Angeles, CA: University of California Press. Dirven, René and Marjolijn Verspoor 1998 Cognitive exploration of language and linguistics. Amsterdam: Benjamins. Duranti, Alessandro 1985 “Sociocultural Dimensions of Discourse”. In Handbook of Discourse Analysis, vol. 1: Disciplines of Discourse, T. A. V. Dijk (ed.), 193–230. New York: Academic Press.

Politeness distinctions in second person pronouns 201

1997 Linguistic Anthropology. Cambridge: CUP. Ferguson, Charles. A. 1996 “Individual and Social in Language Change: Diachronic Changes in Politeness Agreement in Forms of Address”. In Sociolinguistic Perspectives — Papers on Language in Society 1959–1994, Huebner, Thom (ed.), 227–240, Oxford: Oxford University Press. Fillmore, Charles J. 1997 Lectures on Deixis. Stanford, Calif. : CSLI Publ. Goffman, Ervin 1981 Forms of Talk. Philadelphia: University of Pennsylvania Press. Haase, Martin 1994 Respekt: Die Grammatikalisierung von Höflichkeit. München: LINCOM EUROPA. Hanks, William 1990 Referential Practice. Chicago: University of Chicago Press 1992 “The indexical ground of deictic reference”. In Rethinking context: language as an interactive phenomenon, Duranti, Alessandro & Goodwin, Charles (eds.), 43–77. Cambridge: Cambridge University Press. Head, Brian F. 1979 “Respect Degrees in Pronominal Reference”. In Universals of Human Language. Vol. 3., Greenberg, J. (ed.), 151–211. Stanford: Stanford University Press. Helmbrecht, Johannes 2001 “Politeness Distinctions in Personal Pronouns”. To appear in: World Atlas of Language Structures, Comrie, Bernard & Matthew Dryer & David Gil & Martin Haspelmath (eds.). Leipzig: Max Planck Institut für evolutionäre Anthropologie. 2002 Personal Pronouns — Form, Function, and Usage. Habilitationsschrift, University of Erfurt. Hymes, Dell 1974 Foundations in Sociolinguistics: An Ethnographic Approach. Philadelphia: University of Pennsylvania Press. Jakobson, Roman 1971[1957] “Shifters, Verbal Categories, and the Russian Verb”. In Selected Writings II, Jakobson, Roman 130–147. The Hague: Mouton. Keller, Rudi 1994 Sprachwandel. Tübingen/ Basel: A. Francke Verlag (UTB). 1995 Zeichentheorie. Tübingen/ Basel: A. Francke Verlag (UTB). Langacker, Ronald W. 1990 Concept, Image, and Symbol: the Cognitive Basis of Grammar. Berlin: Mouton de Gruyter. Levinson, Stephan, C. 1983 Pragmatics. Cambridge: CUP.

202 Johannes Helmbrecht

Lehmann, Christian 1995 Thoughts on Grammaticalization. München: LINCOM. Li, Charles N. and Sandra A. Thompson 1979 “Third Person Pronouns and Zero-Anapher in Chinese Discourse”. In Syntax and Semantics 12: Discourse and Syntax. Givón, T. (ed.), 311–335, New York: Academic Press. Malsch, Derry L. 1987 “The grammaticalization of social relationship: the origin of number to encode deference”. In Papers from the 7th International Conference on Historical Linguistics, Ramat, Anna Giacalone & Onofrio Carrube & Guiliano Bernini (eds.), 405–418, Amsterdam: Benjamins. Shibatani, Masayoshi 1998 “Honorifics”. In Concise encyclopedia of pragmatics, Mey, Jacob L. (ed.), 341–350, Amsterdam [a. o.]: Elsevier. Simon, Horst J. 1997 “Die Diachronie der deutschen Anredepronomina aus der Sicht der Universalienforschung”, in: Sprachtypologie und Universalienforschung (STUF) 50,3: 267–281. Stolz, Thomas 1992 “Höflichkeit im klassischen Aztekisch: Scheinaktanten”. In Z. Phon. Sprachwiss. Kommun.forschung (ZPSK) 45,3: 252–278. Sugamoto, N. 1989 “Pronominality: a noun-pronoun continuum.” In: Linguistic categorization. Proceedings of an International Symposium in Milwaukee, Wisconsin, April 10–11, 1987, Corrigan, Roberta L. & Eckman, Fred & Noonan, Michael (eds.), 267–291. Amsterdam & Philadelphia: J. Benjamins (CILT, 61).

Deictic use of demonstrative pronouns in the Rigveda* Katharina Kupfer München

1.

Introduction

This contribution describes the pronominal deictic system of Old Indic, focusing on its proximal deictic demonstrative pronoun idám and the neutral, or der-deictic, pronouns. An analysis of their use in the Rigveda raises the question of whether the proªling function of all demonstratives is of the same onomasiological quality in this language. If it can be shown that der-deictic demonstrative pronouns constitute a functional class of their own, the statement that demonstratives in general are opaque deictic expressions might be questioned. In the course of the argument the demonstrative pronouns will be particularly examined for the way in which the vantage point, i.e., possible or actual origo, is conceptualized.

2.

Corpus

The corpus is the Rigveda, a collection of the oldest Indian texts. Dating from the second millennium B.C., it consists of originally oral poetry and is handed down in Vedic Sanskrit. The Rigveda is made up of ten books, which are subdivided into hymns. These contain a total of about twenty-seven thousand sentences. The meter of the hymns is based on the number of syllables in a line.

204 Katharina Kupfer

3.

The pronominal system

The pronominal system of Vedic Sanskrit includes personal, re¶exive and demonstrative pronouns. The pronouns which have been claimed in the literature to be demonstratives are: a- (816 instances in the corpus), adás (59), avá(3), ena- (102), etád (341), idám (1397), im (219), tád (3346), tyád (212). The system of demonstrative pronouns is divided into enclitic (a-, ena-, im) and accented (adás, avá-, etád, idám, tád, tyád) pronouns. It would seem that only the accented pronouns can be used deictically. The other demonstrative pronouns play only minor roles in the local deictic system: The defective paradigm of ena- as well as im (and the pronoun sim) belong to the suppletive paradigms of the third person pronoun. All of these pronouns are used in the accusative only. The pronoun tyád is used only for participants which are considered by the speaker to be speciªc and deªnite, and it cannot be used for local deixis. This article will focus on the two pronouns which occur most often in the corpus, and which play a central role in the demonstrative paradigm of Vedic Sanskrit: tád and idám. They have been chosen, since tád illustrate the problem of der-deixis, while idám will help to show the diŸerence between derdeixis and proximal deixis.1 3.1 The pronouns idám and tád Besides the above-mentioned tád, there is also the demonstrative pronoun etád, which shows the same stem formation and in¶ection as tád except for the e- in its paradigm. Both will be discussed below in detail. Table 1 shows the paradigm of the accentuated pronoun idám, as well as that of the paradigm of the stems a-/i-/ena- (underlined).2

Deictic use of demonstrative pronouns in the Rigveda 205

sg. f.

sg. n.

du. m. du. f.

acc. nom.

ayám

iyám

idám

imám

´ imam

idám

ena´ anéna ena asmaí ásmai asmai ´ asm at asmat asyá/ásya asya asmín asmin

aya´ anáya aya´ asyaí asyai

ena´ anéna

´ ima/ imé imaú ima´ imé imaú ´ abhyam

asmaí

-

-

-

´ asy ah asyah ´ asyah asyah ´ asyam asyam

´ asmat

-

-

-

-

-

loc.

gen.

abl.

dat.

sg. m.

instr.

Table 1.

asyá ayóh imásya asmín ayóh

-

du. n. imé -

-

pl. m.

pl. f.

pl. n.

imé

´ imah

´ im a´ im ani

´ im an

´ imah

ima´ ´ im ani ebhíh

´ hih ab abhíh ebhih abhih ebhyáh abhyáh ebhyah abhyah ebhíh

-

esam esú esu

´ asam asam asú asu

-

´ esam esam esú

It can be seen that many forms within the paradigm are identical except for the accent. Of the 816 instances of the enclitic forms found in the Rigveda, ten are in the instrumental case, 149 in the dative, six in the ablative, 597 in the genitive and 50 in the locative. The two paradigms are functionally split in such a way that text cohesion is associated with enclitic forms and deixis with accentuated forms. The clitics express such varied semantic roles as instrument, recipient, source, place and — in the case of the genitive — possession, partition, and experiencer.3 3.2 Deictic use of idám The ªrst example illustrates the local deixis of the pronoun idám; the divisions shown correspond to the lines of the stanza quoted. (1)4 RV 06.047.31 amuh aja prati a´ a. a´ on dist:n.a.pl.f drive: pr.imp.2.sg against on ´ vartaya imah/ turn:caus:pr.imp.2.sg prx:n.a.pl.f ‘Fetch those over there, make these here turn back!’ vavaditi / b. ketumát dundubhíh bright:abl.sg.m.n drum:i.pl.m speak:ints:pr.3.sg ‘(The) drums are repeating (it) clearly.’

206 Katharina Kupfer

c.

sám ásvaparnah cáranti with horse.wing:n.a.pl.m.f move:pr.3.pl nah nárah / 1.pl.a.d.g. man:n.a.pl.m ‘(They who have) horses for wings are coming together:’ ´ indra rathínah jayantu // d. asmakam 1.pl.g Indra:v.sg charioteer:n.a.pl.m.f conquer:pr.imp.3.pl ‘our men. (Sc. God) Indra, (our) charioteers shall conquer!’

The context in song RV 06.47 (resp. Example (1)) is war. At the beginning of stanza 31, cattle are referred to by the feminine forms of the demonstrative pronouns. The cows of the enemy shall be conquered, and one’s own cows shall stay in the territory of the reciting priest and his lord. The demonstrative pronoun idám, expressing proximal deixis, refers to the cows of the priest’s side. The question of anaphoric use arises in particular in the last stanza of a song, where the probability that a participant has been already mentioned is greater than in the ªrst stanzas of a hymn. This is illustrated by the following example, stanza sixteen: (2) RV 03.029.08, 15–16 ´ hotar své a. sida sit:pr.imp.2.sg priest:v.sg.m refl:l.sg ´ / uloké cikitvan world:l.sg.m knowing:n.sg.m ‘Sit down, hotar(-priest), as one knowing his own place!’ b. sadáya yajñám suk®rtásya sit:caus:imp.2.sg sacriªce:a.sg.m well.done:g.sg.m.n yónau / womb:l.sg.m ‘Put (the) sacriªce in (the) womb of a well-done (deed)!’ ´ ´ devan havísa c. devavih enjoyment.for.gods:n.sg.m.f.n god:a.pl.m oblation:i.sg.n yajasi / sacriªce:pr.subj.2.sg ‘(As one who has) enjoyments for (the) gods you shall worship (the) gods with an oblation.’ d. ágne b®rhát yájamane ªre:v.sg.m high:n.a.sg.n sacriªce:part.pr.l.sg.m.n váyah dhah// energy:n.a.sg.n put:aor.subj.2.sg “(Sc. God) Agni, you shall put high energy in one (who) is sacriªcing.’

Deictic use of demonstrative pronouns in the Rigveda 207

amitrayúdhah marútam iva ªghting.with.enemies:n.a.pl.m.f Marut:g.pl.m like ´ prayah/ onset:n.a.pl.f ‘(They [sc. the family of the Kusikas] are) ªghting (the) enemies, like onsets of (the) Marut(-gods);’ ´ bráhmanah vísvam b. prathamajah ªrstborn:n.a.pl.m.f.n prayer:g.abl.sg.n ít viduh / all:n.a.sg.nemph know:pf.3.pl ‘(sc. the god Agni is the) ªrstborn of (the holy) word: (as such the people) know him (sc. Agni, i.e. the ªre) well indeed.’ ´ c. dyumnávat bráhma kusikasah a´ shining:n.a.sg.n prayer:n.a.sg.n Kusika:n.pl.m on irire / set.in.motion:caus:pf.3.pl ‘(The) Kusikas have made a brilliant prayer,’ d. éka-ekah dáme agním 1:n.sg.m-1:n.sg.m house:l.sg.m.n ªre:a.sg.m sám idhire // with kindle:pf.3.pl ‘(since) they have kindled (the) Fire, each one at home (and yet) together.’

a.

a.

yád adyá tva pra yatí when today 2.sg.a forward go:part.pr.l.sg.m.n yajñé asmín / sacriªce:l.sg.m prx:sg.l.m.n ‘After we have chosen (a) you (sc. god Agni) today during this ongoing sacriªce’ b. hótar cikitvah priest:v.sg.m knowing:v.sg.m ávr® nimahi ihá / prt:choose:pr.1.pl.mid here ‘here, ye knowing hotar(-priest),’ c. dhruvám ayah ªrm:n.a.sg.n prt-sacriªce:aor.2.sg utá asamisthah/ dhruvám ªrm:n.a.sg.n and prt-labor:aor-2.sg ‘you made oŸerings constantly and you have recited (the texts, sc. as our priest) constantly.’

208 Katharina Kupfer

d. pra janán forward know:part.pr.n.sg.m ´ úpa yahi vidvan know:part.pf.act.n.sg.m towards go:imp.pr.2.sg sómam // soma:a.sg.m ‘Come with insight, (as one who is) knowing, towards (our) soma (-sacriªce)!’

The sacriªce is mentioned ªrst in stanza ªve and again in stanza eight. Both passages contain a general statement about the sacriªce: the god Agni is the ensign of the sacriªcial ªre (stanza ªve) and the divine priest (stanza eight). Until the last stanza there is no further reference to the sacriªce. But then the perceived object becomes a diŸerent one, while the viewer, i.e. the performing priest, remains the same. He now refers to the actual, ongoing sacriªce; it appears in the same sentence with the word for “today”. Furthermore, this stanza gives a reason for the god Agni to attend the sacriªce of the reciting priest and his lord and not to go to the sacriªce of a rival or an enemy. Although the participant ‘sacriªce’ is already mentioned and therefore given, the use of the deictic pronoun idám makes it clear that the spatial interpretation is intended. 3.2 Anaphoric use of aThe enclitic stems a-/i- (and suppletively the defective stem ena-), on the other hand, are used only for text cohesion, although the stems a-/i- historically belong — as is well known — to the same paradigm as the deictic pronoun idám.5 (3)

RV 07.086.01 ´ a. dhira tú asya mahina´ wise:n.a.pl.n but anph:g.sg.m.n great:i.sg.m.n ´ / janumsi birth:n.a.pl.n ‘But intelligent (are the) generations (only) by his might,’ b. ví yáh tastámbha ródasi asunder rel:n.sg.m prop:pf1.3.sg.2.pl world:n.a.du.f cit urvi´ / even wide:n.a.du.f ‘who has propped asunder even (the) two wide worlds.’

Deictic use of demonstrative pronouns in the Rigveda 209

´ prá nakam r® svám nunude forward sky:a.sg.m high:a.sg.m push:pf.1.3.sg.mid brhántam / high:a.sg.m ‘(He) has pushed away (the) high, lofty ªrmament’ náksatram papráthat ca d. dvita´ doubly.so star:n.a.sg.n spread:pf.1.3.sg and ´ // bhuma earth:n.a.sg.n ‘(and the) star as well (sc. as the morning and evening star) and (he has spread) out (the) earth.’

c.

Hymn 86 (i.e. example (9)) is dedicated to the god Varuna. His divine domin´ and he is ion is often alluded to by the expression “occult power” (maya), omniscient. In the ªrst stanza of this hymn he is referred to by the form asya, the genitive singular masculine (and neuter) of the stem a-. The god Varuna is a ruler, not a divine priest, as the god Agni (in example (2)). Therefore Varuna must not be thought of as being present at the actual sacriªce, and the poet can refer to him by a pronoun which is not used deictically. As one can see from example (3), the pronoun asya introduces a participant with a relative construction. The relative clause is a restricting, establishing one.6 The relative clause is used for grounding, since it has the function of giving a clue to the listening audience as to how the dependence of the antecedent is conceptualized within the main clause. From the context of someone who has propped asunder even the two wide worlds, the ancient hearer was able to conclude that the god Varuna is meant. The stems a-/i- are typically used to refer to the most salient participant of the discourse, as in the given example. The second example of the enclitic stem a- concerns the ritual substance soma, which is, on the one hand, a kind of juice, and on the other hand, a god, i.e. the personiªcation of that juice and the plant it is made from. 7 (4) RV 08.094.04 a. ásti sómah ayám sutáh / cop:pr.3.sg soma:n.sg.m prx:n.sg.m press:part.pf.n.sg.m ‘This soma(-juice) here is pressed.’ b. píbanti asya marútah / drink:pr.3.sg anph:g.sg.m.n Marut:n.a.pl.m ‘(The) Marut(-gods) drink of it’

210 Katharina Kupfer

c.

´ utá svarajah asvína // and self-ruling:n.a.pl.m.f Asvin:n.a.d.m ‘and (the two) self-ruling Asvin(-gods).’

It is possible, of course, to refer deictically to the soma drink, which is used in the ongoing sacriªce. This is the case in the ªrst part of the stanza in the phrase sómah ayám.8 However, the important event in this stanza is the invocation of two speciªc groups of gods. Only the Maruts and the Asvins shall drink of the soma and none of the other gods, who are invited to drink in the following stanzas. The crucial point in the second line of this stanza (píbanti asya marútah) is that no longer the soma juice is proªled (as it is in the ªrst line), but the invited gods. Therefore one must translate asya with the third-person pronoun to avoid a local setting in the translation. In summary, deixis and text cohesion are clearly kept distinct by the presence or absence of accent within the forms of idám and the paradigm of the two enclitic stems a-/i-. The function of the two forms of the pronoun is proximal deixis for the accented forms of idám and marking of the most salient participant for the stems a-/i-. By the same criteria, the function of the enclitic pronouns enad, im and sim can be described as text cohesion, and that of the demonstrative pronouns etád, adás (and the relic form avá-) as diŸerent kinds of deixis. Taking the demonstrative pronoun tád into account as well (cf. below), it is shown that this principle works in an iconic way. 3.4 Deictic use of tád The demonstrative pronoun tád raises a problem concerning the kind of deixis it expresses. Tád is etymologically connected to the German pronouns der, die, das, the deixis of which is normally described as der-deixis.9 This kind of deixis can be described as a neutral deixis, as it can only be deªned in a negative way: it is neither proximal nor distal. The problem connected with this kind of deixis will be discused in detail below (in section four). To complete the description of text cohesion involving the demonstrative pronouns treated in this article, the next example will illustrate that, in opposition to the demonstrative pronoun idám, the demonstrative pronoun tád can be used to achieve text cohesion. The following example shows anaphoric use of the demonstrative pronoun tád.

Deictic use of demonstrative pronouns in the Rigveda

(5) RV 02.001.13–15 ´ ´ agne adityasa a. tvam 2.sg.a ªre:v.sg.m Aditya:n.pl.m asyàm / mouth:n.a.sg.m ‘O Agni, (the) sons of (the goddess) Aditi have made you (the) mouth,’ ´ ´ b. tvam jihvam súcayah 2.sg.a tongue:a.sg.f shining:n.pl.m.f cakrire kave / make:pf.3.pl.mid wise:v.sg.m ‘(the) brilliant ones (have made) you (the) tongue, o wise poet.’ ´ ´ ratis acah adhvarésu sascire / c. tvam 2.sg.a generous:n.a.pl.m.f sacriªce:l.pl.m accompany:pf.3.pl.mid ‘(They who are) dispensing gifts (sc. the &wp1,92;ditya-gods) follow you at (the) sacriªce.’ ´ havíh adanti a´ d. tvé devah 2.sg.l gods:n.pl.m oblation:n.a.sg.n eat:pr.3.pl on hutam // sacriªce:part.pf.n.a.sg.n ‘In you (the) gods eat (an) oŸering.’ tvé agne vísve am®r®´tasah 2.sg.l ªre:v.sg.m all:n.pl.m immortal:n.pl.m adrúhah / non.deceitful:n.pl.m ‘All (the) immortal gods, (the) nondeceitful (eat it) in you.’ ´ devah havíh b. asa´ mouth:i.sg.n gods:n.pl.m oblation:n.a.sg.n a´ hutam / adanti eat:pr.3.pl on sacriªce:part.pf.n.a.sg.n ‘(The) gods eat (the) oŸered oblation through (your) mouth.’ c. tváya mártasah svadante asutím / 2.sg.i mortal:n.pl.m sweeten:pr.3.pl brew:a.sg.f ‘(The) mortals sweeten (the) brew with you.’ d. tvám gárbhah virúdham 2.sg.m womb:n.sg.m plant:g.pl.f jajñise súcih // generate:pf.3.pl.mid shining:n.sg.m.f ‘You, (the) glowing, are born as (the) fruit of (the) womb of (the) plants.’ a.

211

212 Katharina Kupfer

´ tvám tan sám ca práti 2.sg.n dem:a.pl.m with and towards ca asi majmána / and be:pr.2.sg greatness:i.sg.m ‘You are equal to them and on a par with (them) in majesty.’ b. ágne sujata prá ca deva ªre:v.sg.m well.born:v.sg.m before and god:v.sg.m ricyase / leave:pr.2.sg.mid ‘O noble god Agni, you excel (them),’ c. p®rksáh yád átra mahina´ bringing.food:n.sg.m when here great:i.sg.m.n ví te bhúvat / asunder 2.sg.g.d be:aor.subj.3.sg ‘when (you who) bring food expand your majesty here:’ ´ p®rhivi´ ródasi d. ánu dyava after heaven:n.a.du.m earth:n.a.du.f world:n.a.du.f ubhé // both:n.a.du.m.f.n ‘over heaven, (d) earth — (over) both worlds.’ a.

´ (A.PL.M.) In stanza 15 the demonstrative pronoun tád occurs in the form t an and refers to the gods who were mentioned in the foregoing text. In stanza 13 the gods were introduced as a new participant. At this point they appear only in a mythological context. The gods had elected the sacriªcial ªre as their priest at some time in the past, and because of this act the ªre-god Agni is now, among other things, the mouth of the gods when a sacriªce is celebrated. In stanza 15 the demonstrative pronoun tád is used in a generic statement. Agni, is more ´ The view powerful than the other gods who are referred to by the form t an. that the use of the demonstrative pronoun tád in stanza 15 is anaphoric instead of deictic is supported by the predicate, in which a generic statement is expressed.10 Other examples of anaphoric use of the demonstrative pronoun tád can easily be added (cf. Kupfer 2002:212–217). The next step will be to show that the demonstrative pronoun tád expresses local deixis as well, since it has been argued above that the proximal demonstrative pronoun idám can only be used deictically, and the same is true of the distal demonstrative pronoun adás.11 The remaining discussion will therefore concentrate on the question of local deixis to prove a deictic use of the demonstrative pronoun tád. The hymn 14 in

Deictic use of demonstrative pronouns in the Rigveda 213

the next example is dedicated to the god Agni and is about the oŸerings that are presented to him. (6)

RV 03.014.07 ´ kavikrato yani a. túbhyam daksa 2.sg.d ªt:v.sg.m wise:v.sg.m rel:n.a.pl.n ima´ / prx.n.a.pl.n ‘For you (are), o expert, o wise (priest), these (things here), which’ b. déva mártasah adhvaré god:v.sg.m mortal:n.pl.m sacriªce:l.sg ákarma / prt:do:aor.1.pl ‘(we) mortals did, o god, at (the) sacriªce.’ c. tvám vísvasya suráthasya 2.sg.n all:g.sg.m.n well.chariot:g.sg.m.n bodhi / wake:aor.imp.2.sg ‘You, be aware of every (sc. man who has a) good chariot!’ d. sárvam tád agne all:n.a.sg.n dem:n.a.sg:n ªre:v.sg.m am®rta svada ihá // immortal:v.sg.m sweeten:pr.imp.2.sg here ‘All this here, o Agni, o immortal, make tasty!’

In the example quoted, which is the last stanza of the song, the sacriªcial procedure is summarized. God Agni is asked to render the sacriªce edible for himself and then for the other gods through his nature (ªre) and in his function as the divine priest. For locating the statement sárvam tát agne am®rta svadehá in space and time, the adverb ihá “here” is used. Another possibility would have been to use the demonstrative pronoun idám. To rule out the possibility that the demonstrative pronoun tád is generally used as a personal pronoun,12 and must also be considered such in the last example quoted, the syntactic conditions shall be looked at with respect to a possible zero subject. As one can see from asnavat “he shall attain” in the next example, Vedic Sanskrit is a pro-drop language: the patron, who is the institutor of the sacriªce has not been introduced into the discourse in the ªrst two stanzas of this hymn.

214 Katharina Kupfer

(7) RV 01.001.03 a. agnína rayím asnavat ªre:i.sg.m wealth:a.sg.m attain:pr.subj.3.sg ‘With (god) Agni (the worshipping patron) shall attain wealth,’ b. pósam eva divé-dive / prosperity:a.sg.m just day sky:l.sg.m-day sky:l.sg.m ‘just prosperity every day,’ c. yasásam virávattamam // honoured:a.sg.m.f. abounding.in.heroes:superl.a.sg.m.n ‘honoured (one), one abounding in heroic (sons).’

The comparison of example (1a), without a second personal pronoun alongside the imperative, and example (6c), with such a pronoun, also shows that pro-drop is given. The following example illustrates on the one hand deictical use of the demonstrative pronoun tád and on the other hand the lexical contrast between tád and idám. (8) RV 06.009.04 a. ayám hóta prathamáh pásyata prx:n.sg.m priest:n.sg.m ªrst:n.sg.m see:pr.imp.2.pl imám / prx:a.sg.m ‘This (one here is the) ªrst hotar(-priest). Look at this (one here)!’ b. idám jyótih amr®´tam prx:n.a.sg.n light:n.a.sg.n immortal:n.a.sg.n mártyesu / mortal:l.pl.m.n ‘This (here) (sc. he, god Agni) (is the) immortal light among (the) mortals.’ c. ayám sá jajñe prx:n.sg.m dem:n.sg.m generate:pf.3.sg a´ ní sattah / dhruváh ªrm:n.sg.m on down sit:part.pf.n.sg.m ‘This (one here) has been born (as) such, (as the) permanent (one), (who) has been set down,’ várdhamanah // d. ámartyah tanva¡ immortal:n.sg.m body:i.sg.f grow:part.pr.mid.n.sg.m ‘(the) immortal, growing on his own.’

The construction is such that the adjective dhruváh is in apposition to the identifying clause ayám sá jajñe.13 Since the subject is expressed by the demon-

Deictic use of demonstrative pronouns in the Rigveda 215

strative pronoun idám and the apposition of the adjective there is no open slot in the sentence to be ªlled with a further nominative.14 The pronoun tád in this sentence is therefore not a personal pronoun (nor a sentence connector), but a demonstrative pronoun. The demonstrative pronoun idám refers to the burning ªre of the sacriªce. In the ªrst two lines the god Agni is identiªed with a certain social status (as a hotar-priest) and with an object (the divine light). In line (c) the divine ªre is identiªed by a property, expressed by an adjective in apposition. By the reference to the burning ªre the referent of the proximal deictic pronoun idám and the referent of the demonstrative pronoun tád are contrasted. 3.5 The demonstrative pronoun etád Etád is etymologically connected to the demonstrative pronoun tád. Like tád, its deixis is neither proximal nor distal; however, etád is used for marking a change in the information status of a participant.

loc. gen. abl. dat. instr.

acc. nom.

Table 2. sg. m.

sg. f.

sg. n.

du. m. du. f.

esá

esa´

etád

etáu

-

du. n. -

pl. m.

pl. f.

pl. n.

eté

eta´h

eta´ eta´ni

esáh etám

eta´m

etád

-

-

-

eta´n

eta´h

eténa

etáya

eténa

-

-

-

-

-

eta´ eta´ni etébhih

-

-

-

-

-

-

etébhyah -

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

etásyam -

-

-

-

-

-

-

-

The demonstrative pronoun etád occurs 341 times in the corpus. Of these instances, four are in the instrumental case, one in the dative and one in the locative, that is, etád is almost completely restricted to expressing subject and direct object. The demonstrative pronoun etád has two functions: ªrst, it can be used to contrast two participants, as in the following example. Second, it can also draw the attention of the hearer to a not yet mentioned or new participant. The latter is the case in instances where the demonstrative pronoun etád is the ªrst word

216 Katharina Kupfer

of a hymn (e.g. RV 04.045.01), where its function as a focus marker is out of the question. (9) RV 02.014.02 yáh apáh a. ádhvaryavah priest :n.v.pl.m rel:n.s.m water:a.pl.f ´ vavrivamsam / cover:part.pf.act.a.sg.m ‘Adhvaryu(-priests)! (He) who has struck (the) water-encompassing’ ´ b. v®rtrám jag hana asánya iva V®rtra:a.sg.m strike:3.sg.pf.act thunderbolt:i.sg.f like v®rksám / tree:a.sg.m ‘(demon) V®rtra with (his) thunderbolt, like (a) tree’ c. tásmai etám bharata dem:d.sg.m.n foc:a.sg.m bear:2.pl.imp ´ / tadvas aya longing.for.that:d.sg.m.n ‘— to him, (who is) longing for that, bear this!’ d. esáh índrah arhati d.foc:n.sg.m Indra:n.sg.m deserve:3.sg.act pitím asya // drink:a.sg.f anph:g.sg.m.n ‘This (god) Indra deserves (the) drink of it.’

In the stanza quoted the speaker refers to two participants, the god Indra and the soma drink, which have several things in common. Both can be worshipped as gods, both have similar or identical epithets, and both are imagined as being present during the recital of the stanza. The speaker therefore faces the di¹culty of making references to the two clear for the hearer. If the second participant, soma, had already been introduced into the discourse, the speaker could use enam or sim, which are forms of the third-person accusative. For the soma, which is introduced in this sentence, the speaker has to use either the proximal demonstrative pronoun idám, or, marking the shift in information relevance, the demonstrative pronoun etád. The two demonstrative pronouns tád and etád are used to refer to the participant Indra in the example, thus demonstrating the parallelism of the two deicitc demonstratives as well.

Deictic use of demonstrative pronouns in the Rigveda 217

4.

Some theoretical remarks

The system of demonstrative pronouns in Vedic can be described fully satisfactorily — and easily — for the proximal and for the distal demonstrative pronoun. But one has to present several arguments to show that the demonstrative pronouns tád and etád must also be described in terms of deixis and are not personal pronouns. As argued above in 3.4, one can arrive at a complete description if the syntax is also taken into account. But one is still left with the problem of why der-deictic demonstratives are hard to describe. Local deixis is a necessary, but not su¹cient condition for the description of such demonstratives. In the following section the claim will be pursued that the crucial point is really the kind of deixis: that type of der-deixis which restricts itself to localization in space and time is problematic to describe. This becomes especially clear if one considers its metaphorical use. The theoretical problems involving demonstratives are discussed discerningly by De Mulder, who points out (1996: 36) that it is the combination of a demonstrative (as ground) and the descriptive content of a noun phrase (as ªgure), which “allows the hearer to localize the referent.” The assumption, however, that demonstratives are opaque deictic expressions is problematic and can be questioned (cf. De Mulder 1996: 32; 44 fn. 5). Taking the Vedic demonstrative pronouns as an example, one can see that the proximal deictic pronoun idám or the distal deictic adás have a lexical content, i.e. “nearness” or “distance” to the deictic center. This type of demonstrative pronoun has a proªled meaning15 and should therefore not be described as displaying opaque deixis. Demonstrative pronouns with der-deixis raise a speciªc conceptual problem, because der-deixis has no positive deictic center. The deictic center of a proximal demonstrative is the nearness to the origo, but the Vedic demonstrative pronoun tád (like the corresponding ones in other languages) lacks such a center. What is left is the contrasting function which is deªned negatively: it is neither proximal nor distal. I would therefore conclude that der-deictic demonstrative pronouns only prepare the setting within which localization can take place. The formal means for localization can be an adverb, a localizing particle, a noun phrase or a gesture. Der-deictic demonstratives strengthen De Mulder’s claim that demonstratives are opaque deictic expressions. Only regarding der-deictic demonstratives makes a second classiªcation by further criteria necessary. Leaving the possibility of a syntactic criterion like

218 Katharina Kupfer

distribution — which is one possible way to solve the problem — aside, I shall attempt a solution by taking an expanded view of deixis. This point has already been noted and criticized by Himmelmann (1997:45–90). He points out (1997:88), that the context as well as the indexical ground for deictic reference will be created only in the particular interaction between speaker and addressee. Also according to Benveniste (1971: 217), one has to assume that all deictic expressions — I, here, now, as well as this and that — refer to instances de discours. All these deictic expressions, of course, show similarities in that they are grounding predications. This criterion has already been accepted above, but other than in the case of der-deictic pronouns, speciªcally I, here and now, the discourse referent is not described by classiªcation. Benveniste (1971: 226) explains this fact as follows: “It is in the instance of discourse in which I designates the speaker that the speaker proclaims himself as the ‘subject’. ... Language is so organized that it permits each speaker to appropriate to himself an entire language by designating himself as I.”16 In contrast to this, one important feature of demonstrative pronouns lies in their use in classifying predications (cf. De Mulder (1996: 33) with further references). Summing up this discussion, it is crucial that the indexical ground is created in the discourse by speaker and addressee and that demonstrative pronouns are used for classiªcation. The contrasting function of demonstrative pronouns has already been pointed out. This function as well as classiªcation imply a selection between discourse referents. The above-mentioned problem of discerning between anaphoric and deictic use of a der-deictic demonstrative pronoun is involved in the sense that in the case of anaphora, as well as in the case of the contrasting function of der-deictic demonstrative pronouns, a choice is made between discourse referents which have been introduced in the discourse. Anaphora is deªned by this criterion. Concerning der-deixis, however, the selection involves not only discourse referents that have been introduced, but also inferable referents (cf. Kupfer (2002:27)). The condition for inferability is the creation of an indexical ground by speaker and addressee.

5.

Conclusion

Vedic Sanskrit uses the presence or absence of accent for deictic or anaphoric purposes in an iconic way. The pronoun tád raises conceptual problems, as it represents der-deixis, which is neutral concerning localization in space and

Deictic use of demonstrative pronouns in the Rigveda 219

time. The problem of whether demonstrative pronouns are localizing elements per se is a problem limited to der-deictic pronouns and results from the fact that they are neutral deictic expressions. A complete description of derdeictic demonstrative pronouns has to take into account the fact that the grounding of these expressions is done indexically in the discourse by speaker and addressee.

Notes * I would like to thank Jürgen Bohnemeyer and Friedrich Lenz for their helpful comments on an earlier version of this paper. 1. For a more complete description of the distribution and the onomasiological use of the Vedic pronouns just listed cf. Kupfer (2002). 2. The latter forms are often referred to in the literature as ‘defective pronoun a-’ and considered separate from the lemma idám. The stem ena- belongs suppletively to the stems a-/i- (cf. Kupfer (2002:64–74) and the three stems are part of the paradigm of the third person pronoun (Kupfer (2002:75–79). 3. For more details (esp. for the direct object) cf. below 3.3. The subject belonging to this suppletive paradigm is either the nominative of the demonstrative pronoun tád or a zero subject. 4. The words of the Indic text are given without sandhi. 5. The stem ena- is etymologically connected to the word for the cardinal number ‘one’, cf. Kupfer (2002:332–336, 374–376). 6. The function of establishing relative clauses in general is described by Himmelmann (1997: 38, 78) and for old IE languages by Lühr (2000: 77–88). 7. More examples with a similar construction are found in Oldenberg (1907). He also discusses the question of whether the pronoun a- can be used adjectivally, which, in contrast to the demonstrative pronoun idám, is not possible, cf. Kupfer (2002:52–54), too. 8. Nominal sentences normally show the copula as zero. In example (4) a marked construction is used, cf. Jamison (1990). 9. The Vedic demonstrative pronoun tád shows der-deixis as well as the German pronoun, cf. below or, in more detail, Kupfer (2002:165–212). 10. The copula must be used in the second person, while it would be zero in the third person if something about the known reality were stated. 11. On the problem of possible anaphoric uses of the demonstrative pronoun adás cf. Kupfer (2002:89–91), but no example of this use can be found in the entire corpus. 12. If the demonstrative pronoun tád were in fact a third-person pronoun, it would be used anaphorically in (6) to refer to “these things” in line (6a).

220 Katharina Kupfer

13. This is one of the examples that show the tendency of the demonstrative pronoun tád to become a sentence connector. This does not happen until the period of Vedic Prose, cf. the summarized arguments given in Hock (1997). 14. The translation into German can be done literally: “Dieser wurde als der geboren, als (der) Feststehende.” The meaning of the sentence is that god Agni is born as the personiªcation of everlasting worship, since he is the divine priest. For a detailed explanation cf. Thieme (1964:20). 15. I do not use “meaning” as a deªnition of truth conditions as discussed by De Mulder (1996: 31). If it should be deªned by a (necessary, but not su¹cient) truth condition, then: “the demonstrative process is the very perception of the demonstrative” (De Mulder (1996: 37), with the conclusion that demonstratives are self-referential. 16. Langacker (1991: 318–319) distinguishes between two other classes: proªled (I, here, now) and unproªled (yesterday, last year). The problem he discusses concerns only the class of expressions, “in which some fact of the ground is put onstage and proªled.”

References Sources R.gveda 1994

2000

Rig Veda: A Metrically Restored Text with Introduction and Notes. Edited by Barend A. van Nooten, Gary B. Holland. (Harvard Oriental Series 50). Cambridge, Mass./London: Harvard University Press. R.gveda-Samhita. On the basis of the edition by Th. Aufrecht, Bonn 1877 (2nd ed.), entered by H.S. Ananthanarayana, Austin / Texas; TITUS version with corrections by Fco. J. Martínez García, synoptically arranged with the metrically restored version by B. van Nooten and G. Holland and the “Padapatha” version by A. Lubotsky, by Jost Gippert, Frankfurt a/M, 31.1.1997 / 28.2.1998 / 24.6.1998 / 22.10.1999 / 1.6.2000: http://titus.fkidg1. uni-frankfurt.de/texte/etcs/ind/aind/ved/rv/mt/rv.htm

Bibliography Benveniste, Émile 1971 Problems in General Linguistics. Coral Gables, Fla.: Univ. of Miami Press. DeMulder, Walter 1996 Demonstratives as locating expressions. In: Martin Pütz, René Dirven (eds.), The Construal of Space in Language and Thought, 29-48. (Cognitive Linguistics Research 8.) Berlin/New York : Mouton de Gruyter.

Deictic use of demonstrative pronouns in the Rigveda 221

Himmelmann, Nikolaus 1997 Deiktikon, Artikel, Nominalphrase. Zur Emergenz syntaktischer Struktur. (Linguistische Arbeiten 362). Tübingen: Max Niemeyer Verlag. Hock, Hans H. 1997 Nexus and “extraclausality” in Vedic, or “sa-ªgé” all over again: A historical (re)examination. In: Hans Henrich Hock (ed.), Historical, Indo-European, and Lexicographical Studies: a Festschrift for Ladislav Zgusta on the Occasion of his 70th Birthday, 49-87. (Trends in linguistics. Studies and monographs 90). Berlin/New York: Mouton de Gruyter. Jamison, Stephanie W. 1990 The tense of the predicated past participle in Vedic and beyond. IndoIranian Journal 33: 1–19. Kupfer, Katharina 2002 Die Demonstrativpronomina im Rigveda. (Europäische Hochschulschriften XXI 244). Frankfurt a.M.: Peter Lang. Langacker, Ronald W. 1991 Concept, image, and symbol. The cognitive basis of grammar. (Cognitive Linguistics Research 1). Berlin & New York: Mouton de Gruyter. Lühr, Rosemarie 2000 Der Nebensatz und seine Konkurrenten in der Indogermania: Der altindische Relativsatz. Historische Sprachforschung 113: 71–87. Lubotsky, Alexander 1997 A R.gvedic Word Concordance. I: A-N. II: P-H. (American Oriental Series 82, 83). New Haven, Connecticut: American Oriental Society. MacDonell, Arthur A. 1916 A Vedic Grammar for Students. Oxford: Oxford University Press. Oldenberg, Hermann 1907 Vedische Untersuchungen. 20. Die enklitischen Formen des Pronominalstamms a-. Zeitschrift der Deutschen Morgenländischen Gesellschaft 61: 825– 830. Thieme, Paul 1964 Gedichte aus dem Rig-Veda. (Unesco-Sammlung repräsentativer Werke, asiatische Reihe). Stuttgart: Philipp Reclam jun. Wackernagel, Jacob & Albert Debrunner 1930 Altindische Grammatik. III: Nominal¶exion – Zahlwort – Pronomen. Göttingen: Vandenhoeck & Ruprecht.

Towards a uniªed model of domain-bound reference* Manfred Consten University of Jena, Germany

1.

Introduction

In many ‘discourse spaces’ there are possible links to both a deictic and an anaphoric interpretation of domain-bound reference, i.e. a form of reference depending on knowledge taken from the situation (cf. Langacker 1998; 1999: 262).1 Most surveys of how deªnite reference is received are restricted to one of these two kinds of reference (e.g. Schwarz 2000). This means they take a decision for granted which the hearer has yet to make. Bellmann (1990: 244, my translation) criticises: According to theory pronouns have to be either anaphoric or deictic. This dichotomy is also suggested by the traditional Generative Grammar: while anaphoric pronouns arise by pronominalization during transformation, deictic ones are generated already in the deep structure. […] But it may turn out that such a dichotomy does not exist in a strict sense in spoken language […]. Quite the opposite, spoken language is characterized particularly by the potential merge of deictic and anaphoric elements.

Moreover, theoretical considerations suggest that diŸerent kinds of anaphora and deixis should not be seen as a dichotomy but rather in gradual terms, and that they can be accounted for by one uniªed explanatory model of domainbound reference. In order to give an appropriate redeªnition of anaphora and deixis, the following issues must be addressed: indirect anaphora (cf. section 3), ‘indirect deixis’ (my term; cf. Section 4) and pronominal reference in sign languages. Overlaps or inconsistencies between anaphoric and deictic processes occur, for instance, with the reception of text-picture-relations (cf. section 5). Instead of the traditional anaphora-deixis dichotomy, I propose a

224 Manfred Consten

gradual model of domain-bound reference and will investigate pronominal reference as the most typical case of domain-bound reference (cf. section 6).

2.

Deªnitions of deixis and anaphora

For more than 2000 years there have been attempts to distinguish between deixis and anaphora (cf. also Windisch 1869, Bosch 1983, Lenz 1997). According to Windisch (1869a: 54) the concept of anaphoric pronouns has its origin in the work of the Sanskrit grammarian Panini. The term anaphora is derived from Greek and corresponds to the Sanskrit term anvâdêça, meaning “renewed reference” (ibid). The classical Indian approach does not aim at establishing an anaphoric word class of its own: the same pronominal forms were to be used in an anaphoric as well as a demonstrative way, making the oldest theory of anaphora one of the most progressive. Modern linguists, however, prefer to take into account the works of ancient Greek grammarians, especially Appolonius Dyskolus (2nd century AD). Lenz (1997: 17) criticizes that some investigators, for example Ehlich (1982), pass Apollonius’ ideas oŸ as their own creation. Apollonius already found all criteria used by modern linguists to distinguish between anaphora and deixis. He deªnes the function of pronouns as “pointing back” and “pointing at”. First and second personal pronouns are deictic, third personal pronouns are anaphoric or deictic. Anaphora presupposes that the referent is known. Deixis, on the other hand, presupposes that the referent is present (cf. Lenz 1997: 14f). Windisch (1869), who seems to be the ªrst modern Indo-European linguist to ‘dig up’ Appolonius’ work again, interprets him as follows. Deixis is “the immediate (ªrst) pointing at an object which is present in the outside world or which can be thought to be present”, anaphora, however, means “to refer back to an object that had already been mentioned in the text“. (Windisch 1869b: 394, my translations) Real deixis can be distinguished from anaphora by the fact that it directly refers to the real object in the outside world which has been unknown until now or which has at least not until now been introduced in speech, whereas anaphora takes place when an object which has already been mentioned before is taken up again by a pronoun. (Windisch 1869a: 52)

As can be seen, the deªnitions of anaphora and deixis mingle two sets of criteria (cf. Bosch 1983: 6), ªrst the status of a referent in discourse (cf. 2.1) and second

Towards a unified model of domain-bound reference 225

the distinction between the physical and the textual domain of common accessibility (cf. 2.2). 2.1 Discourse-based deªnitions A discourse based deªnition of anaphora and deixis deªnes the two terms with respect to the state of discourse (Ehlich 1979, 1982): Referring to something already known — maintenance — corresponds to anaphora; focussing a new referent — establishment — corresponds to deixis (cf. Webber 1979). In using an anaphor, the speaker’s instruction to the hearer is somehow the exact opposite of the instruction he gives by using a deictic element. He instructs the hearer to treat a previously verbalized element as remaining in focus. (Ehlich 1982: 329)

Appolonius’ concept of “speech” involves (implicitly) the modern idea of a level of conceptualization where referents can be introduced and retrieved. The discourse world is not independent of a ‘physical world’, but it is neither identical with the latter nor a subset of it: while talking to somebody, I can see people, trees, cars etc. without them playing any role in the discourse. By means of a deictic reference, however, they can be included in the discourse world at any time. Thus, on the one hand, a ‘discourse world’ is a selection of the things from the ‘physical world’ that are relevant to discourse. On the other hand, things which are not (any longer) physically perceptible, but which the speaker and/or hearer remember, can be (re-)introduced into the discourse world. Appolonius regards the latter process as anaphoric; a reference to things of the physical world which until now were not part of the discourse world is regarded as deictic. We will call such approaches discourse based. They are supported by Ehlich (1982), by text linguists (Noordman/Vonk 1998: 237, 244) and by supporters of the Discourse Representation Theory (DRT, e.g. Asher 1993).2 Thus, discourse-based accounts are, roughly speaking, based on the pair of criteria ‘maintenance’ (of a referent known before) versus ‘establishment’ (of a new referent).

2.2 ‘Phoric’ deªnitions The term ‘phoric’ is derived from Greek ferein which means “carrying, referring”. Here deixis and anaphora are distinguished by the domain of reference in which the referent is accessible: Reference to a textually present object (one that

226 Manfred Consten

was mentioned before) is anaphoric reference; reference to a physically present object, i.e. an object being perceived by the senses, is deictic reference. One major proponent of this approach is Bühler (1934). He distinguishes — irrespective of syntactic word classes — pointing expressions and symbolic expressions. These words stand for diŸerent ways of identifying objects in two diŸerent “domains” (Zweifeldertheorie “two domain theory”, cf. Lenz 1997: 24): (a) Symbolic expressions constitute meaning in the Symbolfeld “domain of symbols, symbolic ªeld” — they are symbols and do not depend on situational information. (b) Pointing expressions are deictic signals which acquire meaning in the Zeigfeld “domain of pointing, deictic ªeld” from case to case (Bühler 1934: 104, 149). Deixis and anaphora diŸer only in the Zeigfeld. They are ‘modi of pointing’ (Bühler 1934: 80), meaning that they are diŸerent kinds of deixis. Bühler diŸerentiates: (a) deixis ad oculos et aures (referent in the domain of common accessibility of speaker and hearer); (b) deixis at the phantasma (“situations of recollection and imagination similar to perception”, Bühler 1934: 133, my translations); (c) anaphoric pointing “to something which is to be looked for and found in places in speech” (Bühler 1934: 121). The anaphoric Zeigfeld is the discourse. The fact that anaphora is considered to be a special case of deixis indicates a common ground of anaphora and deixis which is described in all phoric approaches (namely that both identify a referent by ‘pointing’). Thus I include Bühler in the ‘phoric’ approaches. The criterion of being known does not explicitly occur in Bühler’s theory. Furthermore, deixis in the sense of introducing something in a discourse world does not ªt into his terminology since for him ‘discourse world’ is a relevant notion in the anaphoric domain only. In other words, within the phoric approaches we have to distinguish between a broad deªnition of deixis (deixis including anaphora) and a narrow one (deixis versus anaphora) (cf. Ehrich 1992: 9). I will use the latter deªnition. Although Bühler stresses the close relationship between anaphora and deixis, he makes an essential distinction — or this distinction occurred to him because no modern concept of reference had been available to him: Anaphoric reference points at a certain passage, not at the object which is named by the symbolic expression in this passage (Bühler 1934: 388–390, cf. Ehrich 1992: 9 and Lenz 1997: 29). Anaphoric expressions are accordingly not referential,3 in contrast to deictic ones, which are used to ‘point at referents’. Bühler’s term “syntactic pointing” (1934: 388) also suggests the idea of a purely text-internal anaphoric relation. Therefore, to Bühler deixis ad oculos and deixis in the phantasma (both referential) are closer to each other than deixis in the phantasma and non-referential anaphora (both missing a physically present

Towards a unified model of domain-bound reference 227

referent). Thus, Bühler postulates rather diŸerent processes for anaphoric and deictic reference, contrary to his obvious intention to describe the diŸerent ‘modi of pointing’ as homogeneously as possible. Consequently, to Karl Bühler there seems to be no gradual transition from anaphora to deixis. August Bühler (1988: 291, my translation), for example, fails to notice this circumstance as he summarizes: This classiªcation of kinds of pointing is based on degrees [!] of sensory presence of the thing pointed at: in the ªrst case [deixis ad oculos et aures] the thing pointed at is in the domain of acoustic or visual perception. In the second case [anaphora] it is in the domain of perception of text. Finally, in the third case [deixis at the phantasma] the thing pointed at is present only in memory or imagination.

Thus, with a consequent classiªcation according to “grades of sensory presence” anaphoric reference appears to be ‘in between’ physical deixis and deixis at the phantasma. 2.3 Problems It follows from the deªnitions of deixis presented above that there is a categorical distinction between reference to present things and reference to absent things: the former belongs to the domain of deixis ad oculos, and Bellmann (1990: 158, my translation) critically remarks: “In practice, in deictic reference the borderline between personal and local deixis seems to be blurred and unclear.” On the other hand, referring to absent things is ‘deixis in the phantasma’ or an anaphoric maintenance — it does not show any local deictic relations. What happens, however, if a referent present in the beginning moves away from the Zeigfeld in the course of the discourse? This question plays an important role in investigating sign languages and it is linked with the distinction between syntactic and topographic space. It is hard to decide whether a speaker points at a verbally located index (i.e. anaphorically) or whether he or she points at the referent itself (i.e. deictically) since index and referent can coincide spatially. (For critical discussions see Lillo-Martin/Klima 1990, Liddell 1990, Keller 1998, Wrobel 2002). Moreover, there are tricky cases which have to be categorized as a mixture between deixis ad oculos and anaphora in Bühler’s system — or which are both at the same time, namely “in cases when a place is introduced by the means of language which is used as a new origo afterwards, e.g. After 100 meters you will come to an intersection. You can see the theater from here.” (Buhl 1997: 55, my translation).

228 Manfred Consten

In short, the ªrst problem of Bühler’s (1934) approach is the psychologically questionable view that anaphors are not referential. The second problem is that it does not seem to be possible to distinguish between the textual and the non-textual parts of the current discourse space. Such cases show the necessity to presume a level of conceptualization in which a referent is located independently of its presence or absence in the Zeigfeld. Nevertheless, it shouldn’t make a categorical diŸerence when referring to a part of such a conceptualization whether the referent is physically present, present in the text or non-existent. This is the basic idea of modern theories of discourse representation, such as Cognitive Grammar, DRT or text linguistic accounts. Approaches to this can already be found in Bühler’s theory, which is contrary to his distinction in which deixis is referential and anaphora stands for something else. Bühler (1934: 121f) infers a cognitive discourse representation from the anaphoric and deictic usability of the same expressions. This discourse representation is structured in analogy to the visible environment of speaker and hearer. This, however, suggests a gradual transition between anaphora and deixis at the phantasma. In both cases the referent is to be found in a cognitive representation where linguistic and extra-linguistic knowledge can interact to diŸerent extents. Does this mean then that the discourse-based deªnition (2.1) necessarily complements the phoric one? The fact that anaphora and deixis are each deªned twice using two sets of criteria seems to suggest that these sets of criteria can be deduced from each other or that they at least correlate with one other. The double deªnition by Appolonius results automatically from the consideration that something that is not yet known from the discourse would have to be physically present in order to be introduced into the discourse. On the other hand, in the anaphoric case, Appolonius could only imagine that something present — “known” — in the imagination or memory of speakers and hearers came in through the preceding verbal discourse. This does not include phenomena like being known by ‘common ground’ or uniqueness so that deªnite reference such as the sun, the Pope would have to be anaphoric if the referent is not present.4 If “anaphoric” is deªned as “known since textually mentioned before” and deictic as “introduced into the discourse from the physical world”, both deªnitions coincide. Nevertheless, these features are independent of each other. In some cases they lead to contradictory results and I therefore oppose the point of view that we are dealing with “ways of viewing one and the same phenomenon“ (Lenz 1997: 16, my translation). The respective criteria do not have to be fulªlled simultaneously; they might even contradict each other. This is illus-

Towards a unified model of domain-bound reference 229

trated by the following problematic, yet very simple, example: Reference to physically present referents does not always need to introduce a new element — you can also refer to an object which is physically present several times. According to the discourse-based deªnition (2.1), such maintenance would be anaphoric because it takes up objects that were mentioned before and are known. According to the phoric deªnition (2.2), however, this would be a case of deixis since the referent is physically present. Thus, a consistent distinction between anaphora and deixis is problematic in terms of ‘phoric’ deªnitions — according to diŸerent discourse spaces — as well as if one includes a discourse representation. In the next sections I will turn to anaphoric and deictic phenomena which are ‘in between’ the simple cases of anaphora and deixis ad oculos, namely anaphora without co-reference (3) and deixis without referents in the Zeigfeld (4). In section 5 I will come back to the problems of distinguishing anaphora from deixis.

3.

Indirect anaphora

Anaphora without an explicitly co-referring antecedent (indirect, associative or contiguous anaphora) has recently become a central topic in Cognitive Grammar (van Hoek 1995, 1997; Langacker 1998, 1999) as well as in text linguistics (Fraurud 1992, Cornish 1999, Schwarz 2000, Consten 2001a,b). 3.1 Indirect anaphora as a common means of establishing coherence Some approaches view anaphoric reference without an antecedent as a mere form of inference-based reference. Müsseler (1995: 55f) gives the following example: Susan went to the surgery. He gave her tablets for her headache. Indirect anaphora is seen as conceptual based inference (Hirst/Brill 1980), “bridging” (Clark 1978), “inferables” (Prince 1981), “extensions“ (of the surface of the text) or (with Sichelschmidt/Günther 1990) “elaboratives“ (cf. Hellman 1996). These approaches are rejected. Instead, widened approaches to anaphora propose diŸerent types of indirect anaphora to do justice to its variants involving diŸerent types of both semantic and conceptual knowledge. A general deªnition of anaphora is supposed to include direct, indirect and borderline cases (cf. Schwarz 2000): anaphors are linguistic expressions that are used to relate a referent to a so-called anchor already mentioned in text or discourse.

230 Manfred Consten

In Cognitive Grammar approaches, the notion of ‘anchor’ is not limited to anaphoric cases: it is deªned as “the content already part of the CDS [current discourse space]” (Langacker 1999: 264f). Felicitous pronominal reference is explained in terms of a ‘reference point’ that constitutes a referential dominion (van Hoek 1995, 1997; Langacker 1999: 237): A pronoun portrays its referent as being immediately accessible in the current, shared discourse context […]. To be accessible is to be in the dominion of a currently active reference point. (Langacker 1999: 235)

In case of direct anaphora, this reference point is coded by an antecedent coreferring with the anaphor (ibid.), but such an explicit pronoun-antecedent relationship is not the typical case of anaphora (Cornish 1999). On the contrary, indirect anaphorics are a common means of establishing textual coherence. Cf. the examples from Consten (2001a): (1)′′ She wanted to unlock the door of her friend’s apartment when she noticed that she had lost the key. (1′)′ She wanted to see her friend when she noticed that she had lost the key. (1′′) Her friend lived in an apartment. This apartment had a door which you had to open before entering. She wanted to unlock this door. To unlock the door, she needed a key. She noticed that she had lost this key.

In (1) key is an indirect anaphor anchored by door or by the VP unlock the door. As this link is conventional, (1) will be seen as a coherent text which is easy to understand. Version (1′) is more di¹cult to resolve, as the hearer has to infer a less conventional link, i.e. the link from friend to key. This requires individual knowledge or some imagination; nevertheless such strongly underspeciªed texts are common, especially in oral communication. Thus, resolving indirect anaphorics is a common hearer-based process of creating coherence. Version (1′′), on the other hand, is strongly overspeciªed, thus most cohesive and uneconomic, which by no means leads to an optimal, easy resolution but rather to unacceptability. 3.2 Gradual categorisation of direct and indirect anaphora The following examples will show that there are gradual transitions between ‘direct’ and ‘indirect’ anaphoric reference. They are arranged in a way that the portion of involved extra-linguistic knowledge becomes increasingly greater. The cases (a.–d.) distance themselves more and more from traditional anaphora and become increasingly more ‘indirect’. Nevertheless, the examples are

Towards a unified model of domain-bound reference 231

all within the domain of anaphora since we ªnd an associated ‘anaphoric anchor’ in the pretext. All examples are concerned with pronominal indirect anaphora. Although there are restrictions imposed on this kind of anaphora (cf. van Hoek 1995) and although it is rarely dealt with in the literature, it can be found in all varieties of speech. a. Antecedent and anaphor of different lexical categories (2) Bremsen ging nicht, weil die nur heiss wurden, Braking was impossible because they became hot ohne dass der Wagen langsamer wurde. without the car getting slower. (Newspaper Express 12.5.1998)

The anchoring expression bremsen (“brake” = V inf. and N pl.) is homophonous with the noun die Bremsen which would be expected as the co-referring antecedent; thus it is very close to direct anaphora. Nevertheless, we cannot presume co-reference because the anchor refers to a verbal concept and the anaphoric expression to a nominal one. The following example is of a similar kind. In the English translation the anchoring verb schneien “to snow” and the expected noun Schnee “the snow” are homophonous; in German they are recognisably morphologically related. (3) Es hat so schön geschneit, und es ist auch kalt genug, dass er It has snowed so beautifully, and it’s cold enough for it liegen bleibt. to settle. (oral comm.)

In (2) and (3) the process of resolving anaphors is conveyed by linguistic (lexical) knowledge and not by inferences, although there is no co-reference. Such cases could be called “morphologically based indirect anaphora”. b. Direct anaphora without gender-agreement — or indirect anaphora? (4) Madonna hat ein neues Spielzeug [n.] — Er [m.] ist gerade 18! Madonna has got a new toy — He’s just 18! (Newspaper headline Bild 15.3.1996: 12)

It is not clear whether this example is a case of direct or indirect anaphora. As an argument for direct anaphora we can state that toy and he denote the same referent. Moreover, in German, where the gender of an expression and the sex of the referent can diŸer, it is not unusual that antecedent and anaphor do not

232 Manfred Consten

agree in gender. Sometimes — for example (4) — only the pronoun speciªes the sex of the referent. Thus, missing gender agreement does not necessarily speak against the classiªcation as direct anaphoric reference. On the other hand, situational knowledge — a feature of indirect anaphora — is necessary in order to interpret the anaphor correctly. A link has to be established between toy — a prototypically inanimate object — and an 18-year-old male referent. This happens through inference, i.e. by activating assumptions about the referent Madonna. In this way the usage of the personal pronoun in (4) is a mixture of direct and inference-based anaphoric reference and appears to be witty and playful. c. Semantically based indirect anaphora — the anaphoric expression belongs to the argument structure of the anchoring verb. (5) “Ich wäre wunschlos glücklich, wenn Sie nicht immer auf dem “I’d be completely happy if you didn’t always park on the Lehrerparkplatz parken würden. Das nächste mal lasse ich ihn teacher’s parking lot. Next time I will have it abschleppen.” towed away.” (Beverly Hills 90210, TV-soap, German RTL)

Here we are clearly dealing with an indirect anaphoric expression since the referent denoted by the pronoun (it — “car“) was not mentioned before. Nevertheless, there is a strong anchorage in the speech structure: the referent belongs to the argument structure of the anchoring verb park and can be expected. Schwarz (2000: 100, my translation) writes on such cases: “By activating the mental frame of the verb in the lexicon the thematic role, which is not explicitly mentioned, […] is the implicit referent for the indirect anaphor in the mental text base. The anchoring takes place by assigning semantic roles.” We can therefore speak of semantically based indirect anaphora. d. Inference-based indirect anaphora Inference-based pronominal anaphorics are rather rare but by no means impossible. In the following example, tabooing the referent seems to be the main reason for the fact that it is not denoted by a lexical NP:5 (6) Blasenschwäche? Dafür Inconturina SR. Incontinence? Use Inconturina SR. Oft kommt es ungewollt: im Schlaf, Beruf, beim Sport. Often it comes when you don’t want it: while sleeping, at work, doing sports.

Towards a unified model of domain-bound reference 233

Mit Inconturina SR hält es einfach besser! Inconturina SR makes it hold better! Man geht wieder zur Toilette, wann man will; nicht dann, wenn es You go to the bathroom when you want to, not when it kommt. comes. (Newspaper advertisement)

The verbal degree of anchoring — the only referent points are coded by bladder and bathroom — is considerably weaker than with semantically based indirect anaphora; the involved part of extra-linguistic knowledge is correspondingly larger.

4.

‘Indirect Deixis’

In this section we will apply the dichotomy of direct and indirect reference to deixis. My hypothesis is that in analogy to anaphora this dichotomy can be made for deixis and that it is of a gradual character in both cases. In the case of direct reference, the referent is immediately accessible. In the anaphoric case, it is explicitly mentioned in the pretext; in the case of direct deictic reference, it is part of the immediate extra-linguistic environment. Cf. Langacker’s example (1999: 239): An appropriate reference point is sometimes provided by the non-linguistic context. If we are walking down the street, for example, and see a man running after a departing bus, I can felicitously utter He’ll never catch it, even though neither the man nor the bus has been mentioned in the previous discourse.

For reasons of simplicity I conªne myself in the following examples to reference in the visual domain, i.e. Bühler’s deixis ad oculos, as well as to text-picturerelations as a special case of visual deixis. I have described indirect anaphora above as referring to a reference point or ‘anaphoric anchor’ in the pretext allowing the hearer to ªnd the referent. Analogously, by indirect deixis I understand deictic reference where the referent cannot be seen directly but can be found indirectly by using visual features of the current discourse space. I call such features ‘deictic anchors’. An every-day example of this is: (7) [pointing gesture into an empty o¹ce:] Ist der nicht da? Is he not here? (attested oral example)

234 Manfred Consten

The deictic anchor is the o¹ce the speaker is pointing at. If speaker and hearer share the knowledge that this is the o¹ce of Professor V., Professor V. can easily be inferred as the referent for the pronoun he. Consequently, here Professor V.’s empty room has the same status as an anaphoric anchor: it constitutes a referential domain where the intended referent can be found. A gradual occurrence of the feature direct/indirect results from the fact that the relation between anchor and referent can be of varying strength and that it can be inferred with the help of diŸerent cognitive processes. Example (7) might be a rather simple case; it can be solved by drawing an obvious conclusion. Referring to a garment lying around, which anchors a pronominal reference to its owner, is a similar case: (8) [Motioning with one’s head to a scarf hanging on the hat-stand:] DIE wollte ich auch noch anrufen. HER wanted I as well to phone “I wanted to phone HER as well.” (attested oral example; the referent had left her scarf there)

The following are some more attested examples of indirect deixis: (9) [Speaker and hearer are standing in front of an abstract picture in the speaker’s ¶at:] DEN kenne ich persönlich. HIM know I personally “I know HIM personally.” (The referent is the painter. If the picture had shown a person, there would have been an interesting ambiguity). (10) [Speaker takes a book from the shelves in the hearer’s ¶at:] DEN kann ich überhaupt nicht ausstehen HIM can I not at all stand “I can’t stand HIM at all.” (The referent is the author of the book.)

Whereas the previous examples can be understood with the help of general knowledge (namely the knowledge that rooms and garments have owners or that works of art have a creator), the following example is based on individual knowledge which speaker and hearer only share within a small community: (11) [Hearer pulls a certain face on a whim. Speaker:] Ich weiß gar nicht, ob der noch lebt. I don’t know at all whether he is still alive. (The referent is a former teacher of the hearer and the speaker and the grimace is an imitation of this teacher.)

Towards a unified model of domain-bound reference 235

All the cases mentioned above have in common that visual features of the current discourse space function as reference points, i.e. they anchor a verbal expression. In the following examples — text-picture-relations from newspaper headlines — the relation between the textual and the visual level is a bit more complex. (12) “He was watching a plane taking oŸ” (Newspaper Bild 27.7.1998)

(13) “He just wanted to get pizza” (Newspaper Express 9.9.2000)

236 Manfred Consten

In these examples, the referent for he — the respective driver of the car — is inferred from visual components. In addition, the picture and text form a whole because the text supplies an explanation or a background for the facts that are only visually shown.

5.

Deictic or anaphoric?

5.1 Text and picture So far I have tried to show that anaphoric and deictic reference are similar to each other in that they are both linked to a reference point (or ‘anchor’), helping the hearer to identify the referent. If an expression is anchored in the textual domain, it is anaphoric; if it is anchored in the non-textual domain, it is deictic. In both cases the relationship between anchor and referent can be more or less immediate. Additionally, it can involve diŸerent kinds of knowledge. Now I will return to the question whether anaphora and deixis can be clearly distinguished at all (2.3). My hypothesis is that the categorization of anaphora versus deixis is a gradual one. My examples are mainly text-picturerelations in newspaper articles. Here visual anchors (represented as pictures) and textual anchors are arranged in a similar way so that we can draw conclusions about their interaction. There is a rich body of semiotic research on text-picture-relations (Cassirer 1923, Peirce 1932). Pictures, as well as texts, are seen as signs of reference, meaning they represent things (basically Goodman 1968: 5). Thus we can speak of co-reference in an arrangement of text and picture as in (14a) as well as in (14b):

(14) a. Boris Becker (14) b. Boris Becker… The German tennis star

The psychological and psycholinguistic investigation into resolving text-picture-relations, on the other hand, seems to be at a loss: “The relationship between verbal and nonverbal information falls between complete integration and complete separability” (Bartlett/ Till/ Levy 1980: 447). It is clear that there is an interaction and “that language and picture have mutual access to each

Towards a unified model of domain-bound reference 237

other” (Zimmer 1983: 21f, my translations) but it is not clear how this interaction works: “Thus we can stress that up to now models for the interaction of both systems have neither been developed in perceptual psychology nor in psycholinguistics” (Zimmer 1983: 8). A linguistic model that integrates the in¶uence of visual information as a determinant of text understanding is required. 5.2 Anaphoric and deictic, anaphoric versus deictic Processing a domain-bound expression means that non-textual (visual) and textual information interact; that is — according to the traditional deªnition — there is deictic and anaphoric processing. Both kinds of information can contribute to the establishment of reference. As the following text-picturestructures will show, they can be in coalition as in (15) or in competition as in (16), (19) and (21). (15)

Quetting gets life Judge: You shot her in the middle of the heart. small text: Quetting will be imprisoned for at least 15 years […] for the murder of Verena Ritzinger (small picture). (Newspaper Express 8.8.2000)

238 Manfred Consten

Here, the picture and text each refer to the same referent. The reader will link both sources of information to make a uniªed representation. But is the picture information processed in a deictic way? The graphic arrangement (the little picture in the ongoing text immediately after the pronoun) gives rise to questions regarding its deictic character since deixis is supposed to be a reference to the outside of the text level: although the picture is processed visually and not linguistically, it seems to be part of the text and thus it could be supposed to lie within the anaphoric domain. Consequently, we are dealing with a cross between deixis and anaphora. In contrast to (15), deictic and anaphoric information are in competition in the following examples: The hearer has to decide in favor of either the anaphoric or the deictic interpretation: (16)

Towards a unified model of domain-bound reference 239

Dead boy in the freezer. He did it (lit. He was it) (Newspaper Bild 5.9.2000) Well, obviously he is the man on the picture and not the dead boy, so the deictic interpretation is the right one.

The ambiguity in (16) is resolved by extra-linguistic conceptual knowledge, a process which is also typical of ambigous pure anaphora; cf. the examples (17), (18) and Hirst (1981: 49): “It seems that an anaphor resolver will need just about everything it can lay its hands on.” (17) Die Grapefruit-Hälften werden dick mit braunem Zucker bestreut und kommen unter den Grill, bis er geschmolzen ist. The grapefruit halves are sprinkled with a lot of brown sugar and then they are put under the grill until it has melted. (Magazine Bild-Woche 10/1996) (18) Allein in Bukarest lassen sich täglich 100 bis 150 Opfer von Hundebissen behandeln. Tierschützer haben verhindert, dass die Straßenhunde von den Behörden getötet werden. Seitdem vermehren sie sich rasant. In Bucharest alone 100 to 150 victims of dog bites are treated daily. Animal rights campaigners have prevented the authorities from killing the street dogs. Since then they have been breeding rapidly. (Newspaper Bild 21.6.96)

Linguists occasionally take stress as a criterion in order to distinguish between anaphorically and deictically used pronouns (cf. Cornish 1996: 24,32). In example (19) the pronoun printed in italics is stressed. Stress suggests the deictic reading — the referent for er “he” is the chef in the picture and not the famous chef mentioned in the headline. (19) Cologne without famous chef …this is why he has to cook for Clinton & Co (Newspaper Express 28.5.1999)

240 Manfred Consten

Nevertheless, in many cases stress is not a su¹cient criterion for a distinction between anaphora and deixis since anaphoric pronouns can also be stressed for emphasis or to contrast them to something else. In the next example the ambiguous pronoun is unstressed. At the same time I would like to show that ambiguity between the deictic and anaphoric versions cannot only be found in newspaper articles but also in all areas of discourse provided that suitable links can be found in the physical environment as well as in the pretext. (20) [Speaker 1 points at the TV screen where a woman can be seen:] “Is this Constanze Lazil?“ — [Speaker 2:] “Do you know her?“ (Dialogue in the movie Jesus from Montreal, Kan/F 1989).

Possibly speaker 2 wants to answer the question in the a¹rmative — using an indirect speech act. In this case the deictic and anaphoric versions are identical. If the woman shown on TV, to whom speaker 1 deictically refers, is not Constanze Lazil, there is an anaphoric and a deictic reading with diŸerent referents: the anaphoric one would be “Do you know Constanze Lazil?“; the deictic one would be “Do you know the woman on TV?“. A similar example is quoted by Ariel (1996: 17): (21) A and B are sitting at the bar of a pub, discussing the bartender. A: …When I was sitting here yesterday with Chaim, he told me that he thinks he is gay. B: Looks bewildered. A: Well, you know how Chaim is, he says that about everyone, that they’re gay.

Here we are dealing with a misunderstanding with regard to the pronoun in he is gay because B understands it in an anaphoric way (referring to Chaim) but it was intended to be deictic (referring to the bartender) — as becomes clear in the last sentence.

6.

Towards a uniªed model of domain-bound reference

As a result of the sections 3–5, textual information (corresponding to anaphoric reference) and non-textual information (corresponding to deictic reference) should be treated analogously — as van Hoek’s and Langacker’s notion of ‘reference points’ (that can be provided by the linguistic or by the nonlinguistic context) suggests. The following reasons support such a claim:

Towards a unified model of domain-bound reference 241

a. Pictures can be part of a text, cf. (15); b. Non-textual and textual information can contribute to the resolution of reference by being in coalition, cf. (14a), (15); they can be on opposite sides in the competition between readings, cf. (16), (19), (.20) c. Processes of resolving ambiguous reference work in a similar way, cf. (16), (17),(18); d. In both cases there is a gradual transition from immediate resolution of reference to an indirect one (e.g. inference-based) (cf. section 3 for indirect anaphora and 4 for indirect deixis). These results lead me to the hypothesis that instead of an anaphora-deixis dichotomy, textual and non-textual information should be handled in one uniªed model since they can interact in various ways. I sum up the criteria for a classiªcation of domain-bound reference discussed here in a two-axes model (Figure 22 below). The ªeld built up by the two axes is supposed to facilitate a classiªcation of anaphoric and deictic usage while expressing a gradual occurrence and combination of both features ‘accessibility of the referent’ and ‘textual versus non-textual anchoring’. In other words, the model shows no distinctive boxes.6 First, we will look into the axis “anchoring (textual — non-textual)”. This axis is concerned with the problem of how the hearer knows whether a referential expression is co-referring with an antecedent or whether it refers to something else (disjointedness).7 The extreme poles of the feature “anchoring“ are, on the one hand, ‘direct anaphora’ (i.e. a relation between an antecedent and an anaphoric expression referring to the same entity) and, on the other hand, physical deixis like a demonstrative connected with a pointing gesture to a present referent. To what degree is this axis gradual? Even if we could categorically distinguish between textual and extra-textual information (example (15) is evidence against this view!), their respective parts in each current discourse space are diŸerent. We ªnd a gradual transition between total textual dominance and total dominance of non-textual information. We could think of the ‘exact middle’ of this axis as a current discourse space where textual and visual information can contribute to an interpretation in equal parts. The vertical axis deals with the “accessibility of the referent”. This term was introduced by the ‘accessibility theory’ (cf. Ariel 1990, 1994, 1996) and is also used in Cognitive Grammar approaches (cf. van Hoek 1997; Langacker 1999: 235). Ariel’s theory abandons the distinction between co-referent and disjoint investigated by Levinson (1987) and starts out from searching a refer-

242 Manfred Consten

ent in a discourse representation: “Addressee is to search among his mental representations for an entity whose accessibility to him is indicated by the speciªc expression.” (Ariel 1996: 20) Ariel gives a scale based on word classes which is supposed to correspond with diŸerent degrees of accessibility. The usage of the devices of reference corresponding to this hierarchy is taken into account by the speaker: “Speakers […] choose their referring expressions by taking into consideration the degree of accessibility of the mental entity for the addressee“ (Ariel 1996: 20). Although the scale given (Ariel 1996: 21) is more diŸerentiated than Levinson’s (1987), it ªnally follows a grading from a low to a high descriptive content as well.8 Ariel integrates items like “stressed pronoun + gesture” in her linear scale. Such an expression is supposed to express a lower accessibility of the referent than a pronoun without a gesture and a higher one than full lexical NPs. This point is questionable: The diŸerence between the deictic use and the anaphoric use of an expression cannot be deªned in terms of immediacy — in both cases the referent can be immediately accessible. The diŸerence rather is a question of diŸerent domains of accessibility. I consequently plead for separating these two axes. Both anaphoric and deictic reference can be ‘direct’ or ‘indirect’ (cf. Section 3 and 4). In the ªrst case the referent is immediately accessible, either because it has already been introduced by an expression in the pretext — or because it exists in the physical environment. In all the other cases it must be found by means of anaphoric anchors in the discourse or by means of deictic anchors in the physical environment — by more or less exhaustive cognitive processes.9 (22)

accessibility: immediate coreferential anaphora physical deixis 14b 15 14a

2 3 anchor: textual

non-textual 4 5 12,13

10

6 indirect anaphora

7 8,9 11 indirect deixis

non-immediate

Towards a unified model of domain-bound reference 243

Although I discussed some of the previous examples in terms of this scheme, I did not intend to assign the examples to speciªc points on the two axes proposed above. I rather wanted to show that thinking in terms of categories of deixis and anaphora is not appropriate. Theoretical considerations and the analysis of natural language examples made it clear that the existing categories can be abandoned in favour of a gradual cross classiªcation. I hope my uniªed model of domain-bound reference helps to eliminate the terms deixis and anaphora from the vocabulary of cognitive linguistics. At least linguists might feel increasingly comfortable not using them…

Notes * I would like to thank Jördis Beyer (Jena), Sonja Eisenbeiss (Nijmegen), Friedrich Lenz (Giessen/ Passau) and Adrian Simpson (Jena) for their considerable help. 1. In Langacker’s theory, ‘current discourse space’ is deªned as follows: “Any kind of conceptualization is capable of being evoked as a basis for the characterization of linguistic elements. This conceptual base includes any relevant aspects of the social, communicative and discourse context. It thus includes both the ground and the current discourse space. The ground (G) consists of the speech event, its participants, and its immediate circumstances. The current discourse space (CDS) is deªned as the mental space comprising those elements and relations construed as being shared by the speaker and hearer as a basis for communication at a given moment in the ¶ow of discourse” (Langacker 1998: 3). 2. DRT (cf. e.g. Kamp/Reyle 1993) seems to be an attractive framework for the integration of anaphora and deixis in a uniªed model: “The DRT account highlights the close connection between deictic and anaphoric uses of pronouns. Deictic reference, as opposed to anaphoric reference (in so far as this is a real distinction), uses as referents those nonlinguistically described objects that are part of the discourse context, while anaphoric reference is dependent upon a linguistically expressed antecedent. As a cognitive theory, however, DRT postulates a common form of representation for information derived from diŸerent sources. DRT is thus able to exploit the similarities between deictic and anaphoric reference.” (Asher 1993: 231). Without giving an exhaustive analysis of DRT, I would like to mention the following points against DRT: anaphora and deixis are strictly distinguished from each other by presuming that anaphors are dependent on an antecedent Furthermore, there are serious di¹culties in the ªeld of inference-based anaphora and borderline cases between anaphora and deixis. The examples dealt with in DRT are rather simple and they represent only a small part of natural language phenomena. Although DRT, in principle, oŸers the possibility to deal with anaphora and deixis in one single approach, it does not make psychologically plausible statements about anaphoric or deictic processes by which an element gets into a discourse structure, particularly when verbal and non-verbal information is competing with each other (cf. 5.). For more critical remarks concerning DRT, see Löbner (1996).

244 Manfred Consten

3. This view becomes clear using the terminology “endophora” (which corresponds to anaphora) versus “exophora” (which corresponds to deixis) by Halliday/Hasan (1976: 33). Although this terminology reveals the common ground of ‘phoric’ reference, it diŸerentiates between references ‘within’ and ‘to the outside’ of the text level. 4. The discussion about such cases was especially led by Hawkins (1978). His term “larger situation use” makes it clear that he regards the discourse relevant knowledge of speaker and hearer as part of the situation. It is contrasted with “associative anaphora”. Associative anaphora is also based on the knowledge of speaker and hearer but uses a link in the pretext (the village... the church) and is ‘indirectly anaphoric’ in this respect; cf. 3. 5. The same is true for an Australian nursery rhyme quoted in Hirst (1981: 6): The boy stood on the burning deck/ picking his nose like mad He rolled it into little balls/ and threw it at his dad. 6. By “gradual” I mean that one feature has (inªnitely) many expressions which can be arranged on a scale. In the two-axes model each of the axes allow a gradual classiªcation. 7. Regarding this question Levinson (1987: 384) has introduced a hierarchical scale: The “more minimal” a form is, the more the co-reference reading is preferred, that is, NP < pronoun < Ø (similar Langacker 1999: 240f) . Ariel (1996) criticises that some cases of ambiguity (cf. (20),(21)) cannot be explained. Full lexical NPs, which according to Levinson (1987) should not be anaphoric, are, nevertheless, often used in this way (cf. Ariel (1996: 16f) and Fraurud (1996) with (very diŸerent) corpus dates according to which between 32 and 81% of all deªnite full lexical NPs are interpreted as co-referent). I therefore do not intend to make predictions based on word classes. 8. In addition, van Hoek (1997) and Langacker (1999) pay attention to an antecedent’s status as subject or object: “The natural path subject > object > oblique re¶ects the salience of clausal participants with respect to ªgure/ground alignment (primary ªgure > secondary ªgure > ground)” (Langacker 1999: 241f). 9. A concept of ‘immediacy’ including deixis as well as anaphora can be found in C. Lyons (1999: 160f), who extends the term ostension from a purely “situational ostension” (i.e. deixis) to a “textual-situational ostension”. Both “textual ostension” and “situational ostension” have in common that the referent is “immediately accessible”. In contrast to C. Lyons’s ostension we ªnd the inference-based reference which he, however, does not assign to the poles anaphoric/ deictic any longer. Instead of the division ‘ostension’ versus ‘inference’ my sketch expresses graduality with regard to this feature as well. In other words, it is not possible to ªnd distinct categories between the poles ‘immediate’ (direct) and ‘non-immediate’ (‘indirect’), as the examples in 3.2 were supposed to show.

Towards a unified model of domain-bound reference 245

References Apollonius Dsykolos [1912] “Apollonii Dyscoli de constructione libri quattuor”. In Grammatici Graeci II, G. Uhlig (ed.). Leipzig: Teubner (Reprint 1965 Hildesheim: Olms). Ariel, Mira, 1990 Accessing NP Antecedents. [Croom Helm Linguistics Series] London: Routledge. 1994 “Interpreting anaphoric expressions: A cognitive versus a pragmatic approach”. Journal of Linguistics 30/1. 3–42. 1996 “Referring Expressions and the +/– Coreference Distinction”. In Reference and Referent Accessibility. T. Fretheim and J. Gundel (eds), 13–36. Amsterdam/ Philadelphia: Benjamins. Asher, Nicolas 1993 Reference to Abstract Objects in Discourse. Dordrecht: Kluwer. Bartlett, J., Till, R. and J. Levy 1980 “Retrieval characteristics of complex pictures: effects of verbal coding”. Journal of Verbal Learning and Verbal Behaviour 19. 430–449. Bellmann, Günter 1990 Pronomen und Korrektur. Zur Pragmalinguistik der persönlichen Referenzformen. Berlin: de Gruyter. Buhl, Heike 1997 Wissenserwerb und Raumreferenz. Tübingen: Niemeyer. Bühler, August 1988 “Karl Bühlers Theorie der Deixis”. In Karl Bühler‘s Theory of Language, A. Eschbach (ed.) [Viennese Heritage/ Wiener Erbe, Bd. 2.] 287–299. Amsterdam/ Philadelphia: John Benjamins. Bühler, Karl 1934 (31982) Sprachtheorie. Die Darstellungsfunktion der Sprache. Stuttgart: Gustav Fischer. Bosch, Peter 1983 Agreement and anaphora. London: Academic Press. Cassirer, Ernst 1923 Philosophie der symbolischen Formen, Bd. 1: Sprache. Berlin. Clark, Herbert 1978 “Inferring what is meant”. In Studies in the Perception of Language, W. Levelt and G. Flored d’Arcais (eds), 295–322. Chichester: Wiley. Consten, Manfred 2001a “Indirect anaphora and deixis in texts: Domain-bound reference and coherence”. Poster presentation at International Workshop on Reference & Coherence. Utrecht / NL, 11.01.2001. 2001b “Anaphorische und deiktische Referenz“. Doctoral thesis, University of Jena. To appear 2004. Tübingen: Niemeyer.

246 Manfred Consten

Cornish, Francis 1996 “‘Antecedentless’ anaphors: deixis, anaphora, or what?”. Journal of Linguistics 32. 19–41. 1999 Anaphora, Discourse, and Understanding: Evidence from English and French. Oxford: Clarendon Press. Ehlich, Konrad 1979 Verwendungen der Deixis beim sprachlichen Handeln. Frankfurt/M.: Peter Lang. 1982 “Anaphora and Deixis: Same, Similar, or Different?”. In Speech, Place and Action, R. Jarvella and W. Klein (eds), 315–339. Chichester: Wiley. 1983 “Deixis und Anapher“. In Essays on Deixis, G. Rauh (ed), 79–97. Tübingen: Narr. Ehrich, Veronika 1982 “Da and the system of spatial deixis in German”. In Here and there, J. Weissenborn and W. Klein (eds), 43–63. Amsterd./Phil.:Benjamins. 1992 Hier und Jetzt. Studien zur lokalen und temporalen Deixis im Deutschen. Tübingen: Niemeyer. Fraurud, Kari 1992 “Processing noun phrases in natural discourse”. Ph.D. thesis, University of Stockholm. 1996 “Cognitive Ontology and NP Form”. In Reference and Referent Accessibility. T. Fretheim and J. Gundel (eds), 65–88. Amsterdam/ Philadelphia: Benjamins. Greene, Steven, Richard Gerrig, Gail McKoon and Roger Ratcliff 1994 “Unheralded Pronouns and Managment by Common Ground. Journal of Memory and Language 33. 511–526. Goodman, Nelson 1968 Languages of Art: An Approach to a Theory of Symbols. Indianapolis. Halliday, Michael and Hasan, Ruqaiya 1976 Cohesion in English. London: Longman. Hawkins, John 1978 Definiteness and Indefiniteness. London: Croom Helm. Hellman, Christina 1996 The ‘price tag’ on Knowledge Activation in Discourse Processing. In Reference and Referent Accessibility. T. Fretheim and J. Gundel (eds), 193–212. Amsterdam/ Philadelphia: Benjamins. Hirst, Graeme 1981 Anaphora in natural language understanding. Berlin: Springer. Hirst, W. and Brill, G. 1980 “Contextual aspects of pronoun assignment”. Journal of Verbal Learning and Verbal Behavior 19. 168–175. Kamp, Hans, and Reyle, Uwe 1993 From discourse to logic. Dordrecht: Kluver.

Towards a unified model of domain-bound reference 247

Keller, Jörg 1998 Aspekte der Raumnutzung in der Deutschen Gebärdensprache. Hamburg. Langacker, Ronald 1998 Grounding, coding and discourse. Essen: LAUD A 465. 1999 Grammar and conceptualization. Berlin: de Gruyter. Lenz, Friedrich 1997 Diskursdeixis im Englischen. Tübingen: Niemeyer. Levinson, Stephen 1987 “Minimization and conversational inference”. In The pragmatic perspective: Proceedings of the International Pragmatics Conference. J. Verschueren and M. Papi (eds), 61–129. Amsterdam/ Philadelphia: Benjamins. Liddell, Scott 1990 “Four functions of a focus. Reexamining the structure of space in ASL”. In Sign Language Research: Theoretical Issues, C. Lucas (ed), 176–198. Washington: Gaullaudet Univ. Press. Lillo-Martin, D. and E. Klima 1990 “Pointing out differences: ASL pronouns in syntactic theory”. In Theoretical issues in sign language researchk, P. Siple and S. Fischer (eds). Chicago University Press. Löbner, Sebastian 1996 “Associative Anaphora”. In The Proceedings of the Workshop on Indirect Anaphora (IndiAna Workshop), Lancaster University. Lyons, Christopher 1999 Definiteness. Cambridge: CUP. Müsseler, Jochen 1995 “Focussing and the process of pronominal resolution”. In Focus and coherence in discourse processing, G. Rickheit and C. Habel (eds), 53–74. Berlin: de Gruyter. Noordman, Leo and Wietske Vonk 1998 “Discourse Comprehension”. In Language Comprehension: A Biological Perspective, A. Friederici (ed), 229–262. Heidelberg: Springer. Peirce, Charles 1932 Collected Papers, Vol. 2. Cambridge/Mass.: Harvard University Press. Prince, Ellen 1981 “Towards a taxonomy of given-new information”. In Radical Pragmatics, P. Cole (ed), 223–256. New York: Academic Press. Schwarz, Monika 2000 Indirekte Anaphern in Texten. Studien zur domänen-gebundenen Referenz und Kohärenz im Deutschen. Tübingen: Niemeyer. Sichelschmidt, Lorenz and Udo Günther 1990 “Interpreting anaphoric relations during reading: Inspection time evidence”. Journal of Semantics 7. 321–345.

248 Manfred Consten

van Hoek, Karen 1995 “Conceptual reference points: A cognitive grammar account of pronominal anaphora constraints”. Language 71. 310–340. 1997 Anaphora and Conceptual Structure. Chicago/ London: University of Chicago Press. Webber, B. 1979 A Formal Approach to Discourse Anaphora. New York/ London: Garland. Windisch, Ernst 1869a “Untersuchungen über den Ursprung des Relativpronomens in den idg. Sprachen”. Leipzig: Melzer. 1869b “Untersuchungen über den Ursprung des Relativpronomens in den idg. Sprachen”. In Studien zur griech. u. lat. Grammatik II. 1–2, Curtius (ed) 201–419. Leipzig: Hirzel. Wrobel, Ulrike 2002 Referenz in Gebärdensprachen: Raum und Person. Research report, Institute for Phonetics and Verbal Communication; Munich (FIPKM), Vol. 37. Zimmer, Hubert 1983 Sprache und Bildwahrnehmung. Franfurt/M: Haag und Herchen.

Deixis and speech situation revisited The mechanism of perceived perception Heiko Hausendorf Universität Bayreuth

1.

Introduction

It is obvious that the meaning of deictic expressions is bound to the actual speech situation in which they are orally produced. Accordingly, there is a close relationship between deixis and speech situation which provides the starting point at least for most linguistic approaches towards the understanding of deixis. Given this starting point, it is surprising that while the scope of deictic phenomena has continously been extended and diŸerentiated, the notion of ‘speech situation’ still follows an everyday understanding of face-to-face interaction conceptualized in terms of descriptive characteristics instead of theoretically explicated assumptions about face-to-face interaction as a subject in its own right:1 It is amazing how the theory of deixis, presumably a pragmatic domain par excellence, is operating under a concretistic conception of the speech situation that has remained virtually waterproof to the interactionally-oriented research of the past decades. (…) A situation is a complex social construct and is not deªnable on the basis of inspection with the naked eye. (Fuchs 1992: 34)

Following this argument, deictic expressions and the utterances in which they appear cannot easily be separated from their ‘situational context’, but have to be considered a part of face-to-face interaction and accordingly a part of the speech situation itself. The task then is to account theoretically and empirically for the way in which deixis contributes “to the creation and modiªcation of (…) interactionally relevant contexts” (Pekarek 1998: 127–128). The present paper is meant to take a further step in this direction. But the scope of

250 Heiko Hausendorf

‘interactionally relevant contexts’ will be restricted to the situational context in the narrow sense: all of what is accessible to the participants by means of sensory perception. It is this kind of ‘situational context’ which the phenomenon of deixis has to be related to — as I will argue in the following. In doing so, I will concentrate on relatively clear cases of what has been called primary deixis. Primary deixis is meant to refer to those expressions by means of which some facets of the “ground” are put “onstage”, i.e. aspects of the “immediate circumstances” of the speech event.2 The following piece of data, taken from an oral explanation of a board and dice game (called ‘Maleªz’), illustrates this kind of deictic phenomena: (1) und eh (.) nachher wenn jemand also meinetwegen < hier steht> < und ne zwei macht darf er da drauf> < und darf sich diesen turm wechnehmen und darf jemand eh>< also darf> den woanders hinsetzen (238, 34f.) and eh (.) afterwards when someone well let’s say <<pointing to a circle> stands here> <<pointing to another circle> and gets a two he’ll be allowed to go on there> < and will be allowed to take away this tower and will be allowed someone eh> < well will be allowed> to put this somewhere else

This fragment illustrates the demonstratio ad oculos et ad aures that Bühler has introduced as the primary mode of deixis (Bühler [1934] 1965: 80) and from which all of the other deictic modes could be derived. Many other authors have followed this approach taking this demonstratio as the core phenomenon of deixis (see 4 below). Note that it does not easily ªt into the traditional subcategories of local, personal and temporal deixis. Although there might be a bias towards local and object deixis (especially in contrast to temporal deixis), the demonstratio, in principle, lies beyond these common distinctions: it is not the nature of the referent (space, time, person), but the nature of the referential act itself that is crucial for the demonstratio ad oculos et ad aures.3 I will take this seemingly clear case of deixis as the empirical starting point to make clear the way deixis contributes to the social construction of the speech situation. We will then be led to a more precise deªnition of the demonstratio and its relation to other ‘secondary’ modes of deixis. The questions that will be raised in detail are the following: –

What constitutes the ‘situation’ that obviously comes into being in discourse when participants demonstrate with the help of eyes and ears? What

Deixis and speech situation 251

–

–

proper notion of ‘situation’ was implied when, for instance, Bühler introduced the expression ‘situated discourse’ (“situationsnahe Rede”)? Many diŸerent approaches to deixis agree that the speech situation is not objectively given as a physical entity, but has to be viewed as a social as well as a cognitive construal depending on the participants’ activities of attention direction.4 The task would then be to explain that and how the situation or the “ground” (s. above) can be derived from the participants’ attention and perception activities.5 And how can we account for the meaning of demonstratives in such cases? Our everyday understanding would lead us to say that such expressions ‘point’ to something ‘given’ in the situation. But what is the semantic mechanism that lies behind this practical reasoning? This practical reasoning of demonstratives as ‘pointers’ obviously fails to explain that and why ostensive gestures (used together with demonstratives) cannot be left out in most cases without any problems.6 And last but not least: oral instructions (like the one illustrated above) and face-to-face interaction in general emerge and proceed as multi-channel communication systems: participants are required not only to speak or to listen, but to point and to look, to shape and to take. The notion of demonstratio ad oculos et ad aures already emphasises this connection between diŸerent communication channels. The question, therefore, is: How can the role of deixis be related to face-to-face interaction as a multichannel communication system? “Human communication proceeds simultaneously in more than one channel”, as Bar-Hillel (1954) has already pointed out to stress the characteristics of “indexical expressions” in natural languages.7 Demonstration therefore has to do with the intertwining of diŸerent communication channels and the task would accordingly be to account precisely for the involved channels of communication and for the role deictics play with regard to the intertwining of these diŸerent channels (see 3 below).

My ideas on these questions have emerged from empirical analyses of audio and video data from oral instructions (cf. Klann-Delius et al. 1985; Hausendorf 1995). In the present paper, these data and ªndings will not be covered in any detail. But I will start oŸ with some examples taken from these data just to stress the point that oral instructions have to be considered as acts of multi-channel communications depending upon the participants’ sensory perception.

252 Heiko Hausendorf

I will then introduce my central argument: The speech situation has to be grasped in terms of the participants’ sensory perception and the role of deixis is to bring about the transition from individual acts of perception to mutually shared perception. The deictic mechanism that allows for this transition is the mechanism of perceived perception. It is, as will be pointed out, the medium of mutally shared sensory perception in terms of perceived perception in which the situation acquires its communicative relevance. The speech situation, accordingly, appears to be what conversation analysis would like to call an ‘interactive achievement’: socially constructed in the course of the interaction instead of physically given (cf., for instance, the contributions in Duranti and Goodwin (eds.) 1992). Although this thesis has emerged from a social interactionist point of view, it seems to be very well compatible with cognitive semantics in general (s. note 4 above) and with new cognitive linguistic approaches on deixis in particular as far as, for instance, the role of perceptual interpretative strategies and the emphasis on constructive activities are concerned.8 Concepts about mental strategies to construe and organize the perceptual scene (“Zeigfeld” in Bühler’s terms) and the assumption (to be presented in this article) that the situation is a social-interactive achievement in terms of mutually shared perception indeed seem to complement each other. To conclude this article, I will, additionally, turn to some consequences for the linguistic conception of primary deixis and — contrary to most of the recent research on deixis — I will argue for a narrow concept of deixis bound to sensory perception while other aspects of situational and contextual dependence should be related to a broader notion of ‘situatedness’ (see 4 and 5 below).9

2.

Deixis in oral instructions

Oral instructions that are given for practical purposes, in general, depend upon the participants’ sensory perception. They are, in this sense, multichannel communications. This becomes obvious when you turn to such cases of deixis which regularly occur in this type of discourse. The following examples serve to illustrate the intertwining of diŸerent communicative channels and the diŸerent deictic expressions that are made use of to organize this intertwining in discourse. The data stem from game explanations (of ‘Maleªz’, see above Section 1), which can be viewed as oral instructions in so far as they are concerned with the

Deixis and speech situation 253

how of practical performances: using (parts of) the body and objects like dice and tokens to make moves on a special board marked with diŸerent routes from certain starting points to a certain goal or ªnishing point. The examples are typical of explanations of this type of game. They are, however, not restricted to this domain, but can be expected in nearly every context in which the how of practical performances has to be taught and learned. In these data it is not necessary to search carefully for deictic expressions: speech of this type is replete with deictics and pointing gestures in nearly every instruction sequence. The following accordingly represents only a small sample: (2) das spiel kenn ich ja das kenn ich /tippt auf Spielkasten; Blickrichtungswechsel zum Spielkasten/ that game I do know yes I know that /tapping on the Maleªz case; change of gaze direction towards the case/ (3) ja paß mal auf kannst dir also son püppchen da aussuchen /zeigt auf Püppchen im Spielkasten/ o.k. look you can choose eh such a ªgure there /pointing to the ªgures in the case/ (4) und da is das ziel /zeigt auf Zielfeld auf Spielbrett/ and that’s the ªnish of the game /pointing to the ªnishing-point on the board/ (5) ich kann aber hier eins zwei drei vier fünf sechs so ne /setzt Püppchen sechs Felder vor/ but I can here one two three four vive six like this can’t I /moving the ªgure six circles ahead/ (6) hier sind die barrikaden /zeigt kurz auf Barrikadepositionen eines imaginären Spielbretts/ eh stellen wir uns mal [vor] hier wär das Brett /formt ein imaginäres Spielbrett auf dem Tisch/ here are the obstacles /pointing to the position of obstacles on an imaginary board/ eh imagine the board would be here /tracing an imaginary board on the table/

All of these examples illustrate the demonstratio ad oculos et ad aures (perhaps excluding (6), see below) accompanied by pointing gestures (“illustrators” in the sense of Ekman and Friesen 1969). Turning to common formal characteristics, (2) and (3) can be taken as examples of object-deixis, (4) as an example

254 Heiko Hausendorf

of local deixis, (5) as an example of modal deixis and maybe (6) as an example of what Bühler called “deixis at phantasma”. Regarding more closely the contextual embedding in discourse, another set of classiªcation criteria arises that may shed some light on the functional aspects of these forms. In (2) the discourse context can be identiªed by the task of focusing: Imagine the participants sitting around the table. On the table there are some sweets, juice, glasses, paper and pencils for drawing, and the Maleªz case containing the play materials (board, dice, tokens and little white blocks serving as ‘obstacles’ (“Barrikaden”) in the course of the game). Given this context, (2) serves to focus the participants’ attention towards the case on the table, which can easily be recognized as that containing the Maleªz materials. This is the ªrst time within the interaction that perceiving activities (looking at the case) become relevant which can be directly connected with the following explanation. There are many diŸerent forms in the data performing this function of focusing, but in general a demonstrative and a pointing gesture are involved. To be sure, the other deictic devices illustrated in examples (3) to (6) are also connected with perceiving activities and in this sense also ‘focus’ the participants’ attention. But in examples (3) to (5) the perceiving activities seem to be already prepared and ‘pre-adjusted’ in favour of the perception of the board, and (6) appears to be a diŸerent case since it illustrates some kind of ‘asif perception’ (see below). Therefore, the tasks that are performed with these forms go beyond the (simple and basic) task of focusing. In example (3) the functional context can be described by the task of identifying the materials (here: the ªgures): Given focusing activities as illustrated by (2), more ambitious forms of attention direction can take place that depend upon a connection between special lexical items or their ‘periphrases’ (Püppchen (pieces), weißer Klotz (white block), weiße Dinger (white pieces)) and the visual, sometimes also tactile, perception of objects and their size, shape, consistency, colour and other qualities. This overlap between code and message — to pick up Jakobson’s terms (see below Section 4) — becomes especially emphasized by the preceding “so ein” (‘such a’) which regularly occurs in identifying procedures.10 Turning to example (4), the discourse function can be identiªed by the task of localizing certain places and points on the board: an already introduced space for actions (the game board) is divided into certain inner spaces (the starting positions, the circles, the winning point) which are named and pointed to. In example (5) we ªnd the function of illustrating possible moves with the ªgures on the board: The participants’ attention gets directed towards the kind

Deixis and speech situation 255

of actions that can be performed with the tokens on the board. Here again, sensory perception (in general: visual, possibly tactile perception) is connected with verbal information about the game, in this case information about the rules of the games in terms of what is allowed and what is forbidden. The knowledge of the game is demonstrated (instead of explicated), and that exactly is the context in which the task of illustrating becomes relevant. Concerning the verbal forms, the modal adverb (“so”) regularly occurs. Furthermore, a kind of ªne-tuned synchronization is evident between counting and the kinetic process of hand movement: moving the piece from circle to circle the participants simultaneously count each of these circles. There can be no doubt that this rhythmic counting also serves as a deictic device in so far as it directs our attention towards the perceiving of the practical performance.11 There is, ªnally, example (6) where the function is to imagine the board on the table — the situation being that the ‘real’ materials were not available and the explainers were accordingly unable to use the materials within the explanation. Activities like (6) create a mutually shared as-if perception of non-visible objects and their local characteristics. This as-if perception obviously allows the explainer to carry on with his usual explanation strategy although the presupposition of this strategy (the presence of the materials) is missing! This might be the reason why the deixis at phantasma so often occurs in everyday explanation of space descriptions or route directions. Neither the documented forms nor their functions are restricted to the explanation of games but can be — and already have been — found in other sorts of oral instructions, too.12 Oral instructions obviously do not only depend and perhaps do not even primarily depend on what can be heard but on what can be seen, touched and moved. This is the context in which deixis regularly appears. Oral instructions are, therefore, not only ‘oral’ but also bodily or somatic events that make use of the participants’ diŸerent sensory perception abilities.13 In this way, oral instructions beneªt from the properties of face-toface interaction — as can easily be shown when comparing oral with written instructions.

3.

Deixis and the interactive achievement of the speech situation

Oral instructions are — as I have tried to show — common cases of what Bühler has called ‘situated discourse’: Participants obviously make use of what can be taken from the situation by senses, i. e. what is accessible by means of

256 Heiko Hausendorf

human sensory perception, when they have to explain and to demonstrate the how of practical performances. The striking presence of deictic expressions and pointing gestures in these cases gives us strong empirical evidence that every explanatory account on deixis has to explicate the notion of ‘(speech) situation’. Given this assumption, it is surprising that most of the approaches to deixis, of course, do use the notion of situation, but fail to explicate its meaning beyond our everyday understanding. The situation is treated as a device for varying items of the speech acts derived from space, time and (involved) persons. The same holds true for the understanding of face-to-face interaction, which is usually taken as some sort of ‘situational background’ of the utterances in question (see Section 1 above). It was no coincidence that GoŸman — who re-discovered the everyday encounters of interaction as an important subject of sociological and linguistic research — spoke about the “neglected situation” (more than 35 years ago). He argued — in slightly diŸerent terms — that the situation is not an objectively given entity derivable from data of space, time and persons, but something that emerges as a social system sui generis whenever people ªnd themselves in one another’s presence. Going back to the ‘situation’, therefore, means to deal with this small social system and its communicative appearances (GoŸman 1964). And there is no reason, GoŸman concluded, that we can abstract speech from this social system and its communicative appearances. Rather, we have to take speech as being part of these communicative appearances. The question then runs as follows: How is the situation achieved by and within interactive moves and what is the contribution speech provides for this interactive achievement of the ‘speech situation’? Following this line of thinking, we are directly guided to the role of deixis, since deixis — as I will point out in the following — is the semiotic ‘tool’ par excellence to contribute (verbally as well as nonverbally) to the achievement of the so-called speech situation. Contrary to most linguistic descriptions of deictic meaning, I do not take the situation as the resource by means of which the ‘open slot’ of a deictic expression is ‘ªlled’, but rather the other way round consider the situation as what is created by means of these deictic expressions. Their meaning must accordingly be derived from the mechanism the interactive achievement of the situation depends upon. In what follows, I will develop this point of view by introducing three assumptions concerned with the connection between situation, perception and deixis. Let us start oŸ with the very notion of ‘situation’: What does it mean bearing in mind the given examples of situated discourse (in Section 2)?

Deixis and speech situation 257

What does the situation used by the instruction consist of? The answer is rather simple and leads to my ªrst thesis: The situation consists of what is visually perceivable by the participants.14 ‘Situated discourse’, therefore, means that the participants make use of what they can see when they are explaining and instructing. The participants’ visual perception is accordingly used as one communication channel among other channels. Note that visual perception is considered the most important domain of sensory perception in this thesis. That does not mean, however, that the other forms of aural, tactile, olfactory and gustatory perception should fundamentally be excluded. Of course, they have to be (and they can in fact be) integrated in this understanding of situation. Nearly everything accessible by human senses can become a part of the social situation, no matter if it is heard, seen, touched, felt, smelt … or whatever kind of sensory perception might be involved.15 Rooting back the situation in this way to what is sensually accessible sheds some light on the possible communication channels participants are allowed to make use of in face-to-face interaction — and do indeed use simultaneously, at least with regard to aural and visual perception. Re¶ecting the notion of situation from this point of view, some questions arise at once. One of these is: What are the conditions under which certain phenomena of visual perception are included to be a part of the social situation whereas others are obviously excluded? From the point of view of ‘context analysis’ (Bateson/Birdwhistell): How do perceiving activities acquire “maintrack status” in the ongoing interaction? Sure enough, it is only a small slice of what the participants visually perceive during the encounter that indeed becomes relevant for the on-going interaction. So, what are the cues that organize this selection process?16 To put the same question in another more technical way: How can we account for the transition from single perceiving activities to mutually shared perception? This is considered to be a problem not only for observers and (conversation) analysts but also for the interaction system itself. For interaction, in fact, starts to emerge whenever single perceiving activities can be transformed into mutually shared perception. The constitution of mutually shared perception is interaction. But what are the mechanism and ‘helpers’ that lie behind this process? The answer is not as complicated as the question might sound. The data presented make it easy to see that and how single perceiving activities become mutually shared ones and as such gain interactive ‘main-track’ status: It is deixis — verbal as well as nonverbal deixis — that contributes essentially to

258 Heiko Hausendorf

this transition. Pointing to something accessible by reception senses instantaneously establishes the channel of perception for communicative purposes. Two qualifying remarks are necessary in order to avoid objections that might be made to this point of view. Especially the second point is worth discussing at some length since it will have some consequences for the linguistic conception of deixis. To begin with, one could, of course, object that there are also other ways to transform perceiving activities into communicative moves, for instance those belonging to what Bühler called the ‘symbolic ªeld’ of language: Accordingly, there are the verba sentiendi by means of which perceiving activities can be verbalized in themselves. The verbalization of a certain perception (*I see you got a new sofa, didn’t you?) indeed establishes its communicative relevance and, in this way, can be considered as functionally equivalent with deictic means (like: *That’s new, isn’t it?). Pointing is, nevertheless, in most situations considerably more economical and eŸective than verbalizing perceptions. The latter appears to be restricted to certain contextual conditions under which the fact of sensory perceptions is to be highlighted or emphasized in itself. The second objection one could probably make contrarily sticks to contextual conditions under which the relevance of sensory perception seems to be ‘naturally’ given as a self-evident matter of fact. For one could argue that there is a minimum of ‘mutually shared perception’, a minimum of joint attention that can be taken for granted in everyday encounters due to anthropological and culture-speciªc expectabilities and achievements — and established without the ‘help’ of verbal or gestural ‘shifters’. Indeed, the participants perceiving each other as being perceived by themselves is the conditio sine qua non of face-to-face interaction achieved and kept alive by means of bodily communication (mutual gaze, body posture and face direction) as well as of social organisations and arrangements of gatherings, encounters and meetings. This minimum of mutually shared perception and joint attention that is necessary for the emergence and existence of face-to-face interaction could be and is in fact referred to when those present begin to speak. This minimum of mutually shared perception, accordingly, serves as the ‘situational background’ which natural languages systematically make use of17 — without the need for special devices that attract attention towards sensory perception as a communication channel. It is that which is automatically taken for granted as long as the participants are obviously present in the full sense of human sensory perception capabilities, i.e. awake and not sleeping or ‘handicapped’ by whatever might reduce sensory perception.

Deixis and speech situation 259

There can be no doubt, however, that even this minimum of mutally shared perception is not simply ‘given’ but is indeed an important social as well as cognitive achievement, an ontogenetically as well as phylogenetically ambitious result of social and cognitive development. And we have good reason to assume that these developments (both the cognitve and the social) depend upon deictic devices in the narrow sense we have introduced: semiotic and especially nonverbal devices to establish the relevance of perceiving activities for communicative purposes. Given these developments as ontogenetic and phylogenetic achievements, participants are, of course, no longer aware of them but tend to take them for granted. Finally, there is a further argument that can be added concerning this assumption of a minimum of mutually shared perception, namely that there are many social encounters and activity types in which this minimum does not su¹ce to meet the local conversational requirements. Whenever sensory perception is to be extended or diŸerentiated in order to make use of what can be seen, heard, smelt or touched in the physical environment, deictic devices can be expected to make sure that these perceiving activities become mutually shared: take, for instance, oral instructions (examples given in 2 above), or imagine any other context in which the participants’ sensory perceptions become interactively relevant. Apart from these qualiªcations, I would propose to consider deixis as a device whose main function is to ‘help’ perceiving activities to become mutually shared communicative moves. In a sense, this appears to be a transition from semiotically occupied channels of communication to originally nonsemiotical ones. But how exactly can we describe the mechanism that lies behind this transition from situation to interaction and the typically deictic device that organises this transition? To answer this question, we have to look for general conditions of interactive moves — and are then led to a somewhat strange formulation: Deixis allows visual perception to be perceived in itself! The circularity of perception that is expressed in this thesis is indeed crucial for interaction. For interaction as a social reality sui genesis starts when the participants’ perceiving each other can be perceived in itself (“wenn wahrgenommen werden kann, daß wahrgenommen wird”, as Luhman (1984: 560) has put it). Co-presence (‘Anwesenheit’) is accordingly nothing but perceived perception, as GoŸman (1964: 135) has already argued. Mutual gaze, for instance, can su¹ce to achieve this circularity of perception. The well-known adjacency pairs that appear at the beginning of encounters in general secure

260 Heiko Hausendorf

and establish this circularity of perception, which is all the more ¶uid as long as it depends solely upon mutual gaze. Visual perception activities are, in general, neither expectable nor accountable. This is due to the fact that visual perception — leaving aside the special case of mutual gaze — cannot be reliably observed in itself. Human ocular behaviour can only to some extent be observed in itself. Therefore, it fails to fulªl the requirements of social interaction: It is not reliable for participants. And this is precisely the point where deixis comes into use: By means of deictic expressions and pointing gestures that can reliably be heard and seen by the participants, a certain visual perception in itself becomes perceivable in an expectable way — and for precisely this reason, it becomes part of that suction and swirl of perceiving perceptions which constitute interaction as a social reality in its own right: Whenever the participants’ perceiving each other can be perceived in itself, communication cannot be avoided but has irrevocably started (although participants are, of course, free to be silent, to look away from the vis-à-vis or to act ‘self-orientedly’ as obviously as possible).18 Accordingly, there is reason to assume that deixis brings about a situation in which visual perception can be perceived in itself — some sort of ‘perception’s little helpers’: that is the function deictics appear to fulªl in face-to-face interaction. Perception (not only, but in many cases primarily: visual perception) becomes perceivable in itself: this is a key phenomenon, furthermore, to explain that and how communication is allowed to make use of the participants’ situational surroundings. That leads to my last thesis: It is the medium of perceived perception by means of which the ‘situation’ is interactively achieved. The term ‘situation’ is quoted here to indicate that this term should be understood as an abbreviational way to refer to certain aspects of a social construction process instead of referring to a physically given entity. As a complex chiŸre for a social construction the ‘situation’ is – –

dynamic rather than static, locally produced rather than globally given.

For this reason the situation cannot simply be documented by means of audio and video recordings but has to be empirically reconstructed by means of ªnegrained analysis of communicative moves by means of which the situation is socially constructed.19 Accordingly, we had better drop the idea that deictic expressions ‘point’ to something. No doubt deictic devices draw and direct the participants’ attention to something. But deictic expressions such as demon-

Deixis and speech situation 261

stratives are neither pointers nor localizers by themselves: Instead it makes sense to say that they start a process of demonstration and in this sense ‘ring a bell’ (Kleiber 1983). And the question then is: how does this demonstration work? How exactly can we account for the success of such activities of drawing and directing someone’s attention? The idea of pointing expresses our lay understanding, which is useful and necessary for practical purposes that we follow as participants, but is misleading when we want to discover the mechanism behind this practise: the social mechanism of perceived perception (see above). Going back to this social mechanism we are able to explain that demonstration consists of visual perception that has become perceivable in itself by means of verbal or/and nonverbal deixis. To sum up, it is due to deixis that the communicative construction of this ‘situation’ is so tremendously ‘successful’ in everyday face-to-face interaction: – – –

again and again, as economically as possible and highly inconspicuously.

That is the role of deixis in oral communcation as it comes into focus from an interactionist point of view. What then are the consequences that can be drawn from this point of view with regard to the linguistic understanding and grammatical conception of deixis?

4.

Some suggestions for the lingustic conception of deixis

The consequences for the linguistic conception of deixis can be explicated in terms of two suggestions, which will be discussed in the following: 1. re-deªning the demonstratio ad oculos et ad aures, 2. arguing for a narrow concept of deixis and a broader notion of ‘situatedness’. These suggestions lead us to locate deixis at the interface of language as a semiotic code and sensory perception as a nonsemiotic communication channel (s. 5 below). 4.1 Re-deªning the demonstratio ad oculos et ad aures To begin with, the foregoing assumptions guide us to a re-deªnition of what Bühler has not accidentally called the demonstratio ad oculos et ad aures and

262 Heiko Hausendorf

what is usually taken as the paradigm for ‘echte’ deixis (Klein 1978). ‘Echte’ deixis — or to use Harweg’s (1990b) distinction: ‘strong’ deixis — has to be restricted to such cases in which it is empirically provable that elements of a semiotic code (be it verbal or nonverbal) are used to make sensory perception become perceivable in itself. To be sure, this does not automatically hold true for all expressions typically treated as deictic ones: for instance, not every use of a local adverb like ‘here’ implies that sensory perception becomes perceivable in itself — although there is reason to assume that a reference to the situation (in the sense of that minimum of mutually shared perception, see above Section 3) is indeed implied whenever participants use this local adverb). Numerous endeavours to deªne this primary mode of deixis have gone back to ‘accompanying’ pointing gestures as the crucial criterion, or more precisely have emphasized the question as to whether the use of a deictic expression requires an accompanying pointing gesture or not20 and many authors have considered this diŸerence to be a crucial and basic one.21 On the one hand, the focus on gestures does indeed express the idea that there are lots of empirical cases in which the minimum of mutually shared perception (the minimum of joint attention taken for granted in face-to-face interaction) does not su¹ce to meet the communicative demands and is, therefore, ‘extended’ and ‘developed’. On the other hand, the ªxation on gestures nevertheless appears to be too rough and too shallow. It is easy to see that treating the presence of gestures as the necessary and su¹cient criterion obviously fails to grasp the phenomenon in question: Neither does the demonstratio ad oculos et ad aures depend in each and every case on pointing gestures, nor do such pointing gestures exclusively occur in the context of the demonstratio ad oculos in particular and deictic expressions in general. Contrary to this idea, it is argued here that verbal and gestural deixis are both in principle devices in their own right relating to diŸerent semiotic media of social interaction. Both ªnger pointing and a simultaneously used verbal deictic like ‘this’ turn the participants’ act of looking at an object into an interactive move. For this purpose, they are functionally equivalent semiotic devices.22 They can occur together, but they need not. Instead of postulating obligatory pointing gestures (which might hold true for certain deictics but not for all cases to be included), one should focus on what is constitutive for both pointing gestures and deictic expressions: the construction of joint attention in the sense of perceived sensory perception. Instead of gestures, it is the role of (visual) perception that has accordingly to be discussed if we wish to proceed with the deªnition of ‘echte’ or ‘strong’ deixis.

Deixis and speech situation 263

The concept of perceived perception (developed in Section 3 above) is meant to be a further step in this direction. 4.2 Deixis vs. situatedness Following the suggested re-deªnition of the demonstratio ad oculos et ad aures, I propose a distinction between deixis in this narrow sense of perceived sensory perception, on the one hand, and a broader notion of ‘situatedness’ on the other.23 Even Bühler considered much more linguistic items to be deictic ones than those which are in fact strictly bound to sensory perception — and so did most of the subsequent research. Deixis has become a synonym not only for ‘situatedness’ but also for context-dependence of natural language in general. As such it has become a catchword for very diŸerent relationships between linguistic items and contextual structures reaching far beyond the traditional deªnition of deixis as a referential relationship including participants, time and place (Auer 1988: 264–266; Blühdorn 1995: 113–114). And even the latter, canonical deªnition of deixis already goes beyond the demonstratio ad oculos et ad aures, in so far as the necessity of sensory perception is generally neglected as a delimitating force within this deªnition. It is for this reason that I would like to make a distinction between a variety of diŸerent ‘situatedness’ phenomena bound to a minimum of mutually shared perception that can be taken for granted in face-to-face interaction, on the one hand, and deictic phenomena in the sense of perceived perception on the other. This notion of situatedness diŸers from the one Auer (1988) has introduced. While Auer uses the term to refer to a wider pragmatic level of ‘situated’ vs. ‘displaced’ discourse, I take this notion in order to indicate that and how the grammar of natural languages (in numerous ways) refers to a minimum of mutually shared perception.24 Whenever this minimum is to be extended and diŸerentiated, deixis in the very sense of demonstration comes into use.

5.

Conclusion: Deixis at the interface between language and perception

Combining these suggestions with classical linguistic approaches, we are led to locate the phenomenon of deixis (in the narrow sense of perceived perception) at the interface between language as a semiotic code and sensory perception as a nonsemiotical channel of communication. This might be a deeper sense of Jakobson’s (1971) tricky ‘double structures’ by means of which he deªned

264 Heiko Hausendorf

deictics (which he called ‘shifters’) in the overlap between code and message. Only in the case of perceived sensory perception by means of deictics do we have this strong sense of ‘shifting’ between code and message, or between “Text“ and “Situation“ as Weinrich (1993) has put it in his “text grammar”. In other cases of ‘situatedness’ we have perhaps formerly deictic shifters which are now already grammaticalized: these could be treated — as Weinrich accordingly does — in terms of ‘semantic characteristics’ belonging to the morphematic and lexical structure of the code itself. ‘Situatedness’ is in these cases a grammatical category diŸerently marked in diŸerent languages. It is a notion by means of which the close relationship between grammar and face-to-face interaction in terms of a taken-for-granted minimum of mutually shared attention can be expressed. Deixis in the narrow sense of perceived perception, however, operates at the interface between code and message where sensory perception is required as a channel of communication in its own right.

Notes 1. Cf., for instance, Fillmore 1975: 92 and Lyons 1977: 637 for deªnitions of the prototypical standard situation for deixis. Wunderlich 1971: 174–179, for instance, suggested a “pragmatic notation” for the “idealized speech situation” which was explicitly equated with the “idealized verbal contact between persons”. In these approaches the ‘speech situation’ does not attract attention in itself, but rather appears to be a vehicle in order to argue for a wider notion of linguistic competence including ‘pragmatic’ components as well as semantic and syntactic ones (cf. also the critical remarks made by Ehlich 1979: 117–119). 2. The term of “ground” stems from Langacker 1990: 318–319, who distinguishes two broad classes of deictics according to the question whether the ground remains “oŸstage as an implicit, unproªled reference point” or is put “onstage” and “proªled”. I will argue that an adequate understanding of deixis has to be derived from the latter class of deictics while the ªrst one can be related to a broader concept of ‘situatedness’ (see 4 and 5 below). 3. Although time is obviously not perceivable by senses, sensory perception can, for instance, be used to achieve the simultaneousness of diŸerent actions. It seems to be this context of co-ordination and ªne-tuning that is, indeed, relevant for the use of temporal deictics within the demonstratio ad oculos et ad aures (cf. Bühler’s discussion of what he called Augenblicksmarken (1965: 102). Apart from such special contexts of ‘momentary orientation’ the deictic conceptualization of time appears to be derived (metaphorically?) from that of space. 4. Cf., for instance, Langacker 1990: 315, who discusses this ‘conceptualist’ theory of meaning (…“an expression’s meaning cannot be reduced to an objective characterization of the situation described”) from the point of view of cognitive semantics.

Deixis and speech situation 265

5. Cf., for instance, Ehlich 1979: 770–771, who accordingly takes the “deictic procedure” as an act of orientation (“Orientierungshandlung”). 6. Cf. de Mulder 1996 who discusses in detail the thesis (known from Kleiber) that demonstratives are neither pointers nor localizers; see also section 3 below. 7. Cf. Cheang 1990: 57–63 for further comments. 8. Cf. the contributions published in Pütz and Dirven 1996, Part A. 9. Auer 1988 has suggested a similar delimitation: “Recent research in the ªeld has tended to continuously widen the scope of deixis. Contrary to this tendency, I will argue for a narrow, traditional concept of deixis as related (…) to participants, time and place. At the same time, I will argue for the necessity to distinguish deixis from the much wider notion of situated (vs. displaced) language.” As will be shown, the proposed understanding of deixis as belonging to the participants’ sensual ªeld is even more rigorous (see section 4 below). 10. Identifying is perhaps the most prominent deictic procedure illustrating the interplay of ‘naming’ and ‘pointing’ which has always been associated with deixis (cf., for instance, the notion of ‘ostensive deªnition’ from Quine and the discussion in Weinrich 1993: 444-445; the notion of ‘deictic procedure’ (“deiktische Prozedur”) stems from Ehlich 1979 (see note iv), cf. Zifonun, HoŸmann and Strecker 1997: 311-316 for further comments). 11. Note that such devices are not restricted to instructive purposes: they typically occur as well in critical phases of the game when it is necessary to avoid the impression of having cheated. 12. Cf., for instance, Brünner 1987 who found similar forms and functions in the context of instructions at work. 13. The notion of ‘somatic communication’ stems from Scollon and Scollon 1995, who introduce it as a more adequate term than ‘oral communication’ for covering the characteristics of face-to-face interaction. 14. This is the reason why the role of visual perception (and what is visually perceivable) was stressed right from the beginning of occidental reasoning about ‘deixis’ (cf. Apollonios Dyskolos, cited from Harweg 1990b: 307). Whenever the speech situation is explicitly connected with perception, the role of deixis is related to the co-ordination of the participants’ acts of perception (cf., for instance, Harweg 1990b; Ehlich 1979: 770; Zifonun, HoŸmann and Strecker 1997: 312-313). But neither is this co-ordination process explained in detail, nor is it made clear how these co-ordination activities contribute to the communicative construction of the speech situation (see below in this section for both questions). 15. It is, therefore, just to ease the argumentation (and to keep in touch with our data, in which visual perception is indeed the most prominent nonverbal communication channel (see examples in Section 2)), that we focus on visual perception in the following. 16. The same question was raised by Schnelle 1990 with regard to computational aspects of language comprehension in face-to-face situations; cf. also Cheang 1990 who analyses visual perceiving activities from an ‘organismic’ point of view. 17. This idea was explicitly accounted for by Bühler when speaking about the ‘deictic ªeld’ (“Zeigfeld”) of language and has continued to be the starting point for the linguistic

266 Heiko Hausendorf

approaches to deixis (see Section 1 above). The question, however, is how to deal with this kind of ‘situatedness’, particularily: whether we really need a special theory of ‘deixis’ (which can, strictly speaking, not easily be restricted but tends to comprise more and more linguistic phenomena) or should rather argue for a narrow concept that draws upon more than that minimum of mutually shared perception we are currently talking about (see Section 4 below for these linguistic consequences). 18. This is the sense of Watzlawick, Beavin and Jackson’s (1967) well-known axiom of “the impossibility of not communicating” which has, accordingly, to be derived from the regularities of face-to-face interaction (Hausendorf 1992: 52-55). 19. Note that notions such as ‘visibility’ or ‘accessibility’ have to be viewed as socially constructed, too. What becomes obvious in special situational arrangements (referential object out of sight or behind the participants) in fact holds true in general: the visibility/ accessibility of referential objects is interactively achieved by communicative moves (using verbal and nonverbal deictic devices) — although these moves and their outward deictic forms can be extremely inconspicuous (gaze, a pointing ªnger, a demonstrative action). 20. Linguists tend to treat nonverbal communication primarily as being complementary to language. Pointing gestures, accordingly, appear to be some sort of ‘concomitant’ of certain deictic expressions. Cf. Harweg 1990b: 295 as an example of such a ‘verbalist’ point of view: “Zeigegesten sind bekanntlich eine referenzdeterminierende Begleiterscheinung oder wie ich vorziehe zu sagen: ein referenzdeterminierender Bestandteil einer bestimmten Teilklasse von Deiktika.” There is no doubt that the expression ‘deictics’ in this statement refers exclusively to verbal deictics, which seem to be supported by gestures, while gestures seem to be triggered by verbal deictics. 21. Cf. Harweg 1990b: 295 for a systematic account; Harweg himself has emphasized the notion of “strong” deixis to cover this distinction but has nevertheless suggested focusing on the ‘possibility of common visual perception’ (rather than on accompanying gestures) in order to deªne “strong deictics” (1990a; 1990b). 22. One could, of course, claim that ªnger pointing additionally provides information about a particular region within the Zeigfeld (de Mulder 1996: 37). But this eŸect also depends upon the mobilization of visual perception. It therefore makes sense to say: Even pointing gestures do not point but rather ‘ring a bell’ (see above 3): Looking is required! 23. According to this distinction, one would assume that deictic systems in natural languages encode as a basic feature items like ‘joint attention’, ‘visibility’ or ‘accessibility’ and that such features would be needed for semantic descriptions of deictic systems (cf. Özyürek/Kita and Meira, this volume). 24. This is the sense in which the term ‘situation’ is used in Weinrich’s “Text Grammar” (1993: 21), which seeks to describe the numerous grammatical appearances of ‘situatedness’ in terms of semantic characteristics (1993: 1081–1088).

Deixis and speech situation 267

References Auer, Peter 1988 “On Deixis and displacement”. Folia Linguistica XXII: 263–292. Bar-Hillel, Yehoshua 1954 “Indexical expressions”. Mind 63: 359–379. Blühdorn, Hardarik 1995 “Was ist Deixis?” [What is deixis?] Linguistische Berichte 156: 109–142. Brünner, Gisela 1987 Kommunikation in institutionellen Lehr-Lern-Prozessen. Diskursanalytische Untersuchungen zu Instruktionen in der betrieblichen Ausbildung. [Communication in Institutional Processes of Teaching and Learning. Discourse Analytic Studies on Instructions in Factory Training.] Tübingen: Narr. Bühler, Karl 1965 Sprachtheorie. Die Darstellungsfunktion der Sprache [Theory of Language: the Representational Function of Language.] Stuttgart: Fischer. First published in 1934. Cheang, Kiseang 1990 Semantik der Deixis. Eine organismische Analyse sprachlicher Deixis. [Semantics of Deixis. An Organismic Analysis of Verbal Deixis.] Opladen: Westdeutscher Verlag. De Mulder, Walter 1996 “Demonstratives as locating expressions”. In: The Construal of Space in Language and Thought, Martin Pütz and René Dirven (eds.), 29–48. Berlin/ New York: Mouton de Gruyter. Duranti, Alessandro and Goodwin, Charles (eds.) 1992 Rethinking Context: Language an an Interactive Phenomenon. Cambridge: Cambridge University Press. Ehlich, Konrad 1979 Verwendungen der Deixis beim sprachlichen Handeln. Linguistisch-philologische Untersuchungen zum hebräischen deiktischen System. [Uses of Deixis in Verbal Action. Linguistic and Philological Studies on the Hebrew Deictic System.] (Forum Linguisticum 24.) Frankfurt/Main/Bern/Las Vegas: Lang. Ekman, Paul and Friesen, Wallace P. 1969 “The repertoire of nonverbal behavior: categories, origins, usage, and coding”. Semiotica 1, 49–98. Fillmore, Charles J. 1975 Santa Cruz Lectures on Deixis 1971. Bloomington, Indiana: Indiana University Linguistics Club. Fuchs, Anna 1992 Essays on Deixis. Köln: Arbeiten des Kölner Universalien-Projektes (=akup) Nr. 88.

268 Heiko Hausendorf

Goffman, Erving 1964 “The neglected situation”. In: Directions in Sociolinguistics: The Ethnography of Communication, John J. Gumperz and Dell Hymes (eds.), 133–136. (Special Issue of American Anthropologist.) New York. Harweg, Roland 1990a “Starke, schwache und superschwache Deixis”. [Strong, weak and very weak deixis.] In: Studien zur Deixis [Studies on Deixis], Roland Harweg, 239–257. Bochum: Brockmeyer. 1990b “Zeiggesten, Wahrnehmung und starke Deiktika”. [Pointing gestures, perception and strong deictics.] In: Studien zur Deixis [Studies on Deixis], Roland Harweg, 295–307. Bochum: Brockmeyer. Hausendorf, Heiko 1992 Gespräch als System. Linguistische Aspekte einer Soziologie der Interaktion. [Talk as System. Linguistic Aspects of a Sociology of Interaction.] Opladen: Westdeutscher Verlag. 1995 “Deixis and orality: Explaining games in face-to-face interaction”. In: Aspects of Oral Communication, Uta M. Quasthoff (ed.), 181–197. (Research in Text Theory 21.) Berlin/New York: de Gruyter. Jakobson, Roman 1971 Shifters, verbal categories, and the Russian verb. In: Selected Writings. Volume II, Roman Jakobson, 130–147. The Hague: Mouton. First published in 1950/57. Klann-Delius, Gisela et al. 1985 Untersuchungen zur Entwicklung von Diskursfähigkeiten am Beispiel von Spielerklärungen.[Studies on discourse acquisition in the case of explaining games.] Berlin: Linguistische Arbeitsberichte. Kleiber, Georges 1983 “Les démonstratifs démontrent-ils? Sur le sens des adjectifs et pronoms démonstratifs”. [The demonstratives: do they point? On the sense of demonstrative adjectives and pronouns.] Le francais moderne 51/2: 99–117. Klein, Wolfgang 1978 “Wo ist hier? Präliminarien zu einer Untersuchung der lokalen Deixis”. [Where is here? Preliminaries to an analysis of local deixis.] Linguistische Berichte 58: 18–40. Langacker, Ronald W. 1990 Concept, Image, and Symbol. The Cognitive Basis of Grammar. (Cognitive Linguistics Research 1) Berlin/New York: Mouton de Gruyter. Luhmann, Niklas 1984 Soziale Systeme. Grundriß einer allgemeinen Theorie. [Social Systems. Outline of a General Theory.] Frankfurt/Main: Suhrkamp. Lyons, John 1977 “Deixis, space and time”. In: Semantics. Volume 2, John Lyons, 636–724. Cambridge: Cambridge University Press.

Deixis and speech situation 269

Pekarek, Simona 1998 “Deixis and the interactional construction of context”. In: University of Pennsylvania Working Papers in Linguistics. Volume 5.1, Alexis Dimitriadis, Hikyoung Lee, Christine Moisset and Alexander Williams (eds.), 127–138. (Proceedings of the 22nd Annual Penn Linguistics Colloqium.) Pütz, Martin and René Dirven (eds.) 1996 The Construal of Space in Language and Thought. (Cognitive Linguistics Research 8) Berlin/New York: Mouton de Gruyter. Schnelle, Helmut 1990 Vorwort. [Foreword.] In: Semantik der Deixis. Eine organismische Analyse sprachlicher Deixis, Kiseang Cheang, 9–10. [Semantics of Deixis. An Organismic Analysis of Verbal Deixis.] Opladen: Westdeutscher Verlag. Scollon, Ron and Suzanne Scollon 1995 “Somatic communication: How useful is ‘Orality’ for the characterization of speech events and cultures?” In: Aspects of Oral Communication, Uta M. Quasthoff (ed.), 19–29. (Research in Text Theory 21) Berlin, New York: de Gruyter. Watzlawick, Paul, Janet H. Beavin and Don D. Jackson 1967 Pragmatics of Human Communication. A Study of Interactional Patterns, Pathologies, and Paradoxes. Norton: New York. Weinrich, Harald 1993 Textgrammatik der deutschen Sprache. [Text Grammar of the German Language] Mannheim: Duden. Wunderlich, Dieter 1971 “Pragmatik, Sprechsituation und Deixis”. [Pragmatics, speech situation and deixis.] Zeitschrift für Literaturwissenschaft und Linguistik 1: 153–190. Zifonun, Gisela, Hoffmann, Ludger and Strecker, Bruno 1997 “Deixis und situative Orientierung”. [Deixis and situational orientation.] In: Grammatik der deutschen Sprache. Band 1. [Grammar of the German Language. Volume 1.], Gisela Zifonun, Ludger Hoffmann and Bruno Strecker (eds.), 311 365. Berlin, New York: de Gruyter.

Index

6H-model 75, 77 A accessibility 32, 33, 225, 226, 241, 242, 245, 246 addresse effects 3 honorifics 189, 190 -based 8 analogical deixis 88, 89, 91 anaphora 191, 218, 223–233, 236, 238– 241, 243–248 see also indirect anaphora anaphoric strategy 112, 114, 126, 129 use 206, 208, 210, 212, 242 anchor 101, 103, 104, 106–109, 111–114, 116, 120, 126, 129, 229–231, 233–236 see also temporal anchor antecedent 209, 229–231, 241, 243, 244 assertion 106, 136, 139, 140, 156, 160 B binary politeness distinction 196, 197 Brazilian 3, 7–11, 29, 31, 32, 34, 37, 39 Bühler 18, 69, 71–75, 77, 80, 81, 83, 88, 89, 153, 158, 186–188, 226–228, 233, 250– 252, 254, 255, 258, 261, 263–265 C classification 89, 217, 218, 227, 232, 241, 243, 244 co-reference 229, 231, 236, 244 cognition 70, 92 cognitive grammar 228–230, 241 conceptual relation 79, 80, 84, 89

conceptualisation vii-x, 133 conceptualiser 130 contrasting function 217, 218 conversational dyad 13, 16, 19–21, 32, 33, 35 D definite reference 223, 228 deictic marker 155–157 relation 187 strategy 109, 115–118 see also der-deictic demonstratio ad oculos et ad aures 250, 251, 253, 261–264 demonstrative pronoun 203, 204, 206, 210, 212–219 questionnaire 5, 8, 11 systems 3 usage 9 deontic meaning 171 modality 158, 165, 168 der-deictic 203, 217–219 directive 160, 161, 163–165, 170 discourse 13, 15, 36, 38, 39, 99, 105, 118, 126, 129, 189, 191, 209, 213, 216, 218, 219, 223–230, 233, 235, 240–242, 250– 252, 254–257, 263 representation 99, 225, 228, 229, 242 space 228, 233, 235, 241, displacement 42, 74, 75, distal 3–5, 8–10, 17, 26, 29, 32, 34, 35, 187, 210, 212, 215, 217 domain 56, 57, 127, 130, 153, 158, 162, 185, 192, 223–227, 231, 233, 236–238,

272 Index

240, 241, 249, 253, 257 domain-bound reference 223, 224, 240, 241 dürfen 153, 157, 158, 161–163, 165, 175 E epistemic meaning 158, 159, 176 modality 158 event construal 112, 114, 127, 130 F face-to-back 13, 18, 19, 24, 26, 27, 33, 35 face-to-face 13, 16, 18, 19, 24, 26–30, 33– 37, 249, 251, 255–258, 260–264 interaction 69, 249, 251, 255–258, 260–264 Fillmore 22, 189, 201, 264 focus 24, 29, 52, 53, 55, 56, 61, 97, 129, 145, 156, 161, 173, 177, 216, 225, 254, 261, 262 force dynamics 51, 66, 182 future tense 101, 135, 143–146 G German 31, 41, 46-63, 77, 78, 85, 98–100, 105–107, 109, 112–115, 121, 122, 124– 130, 143–145, 153, 153–177, 194, 197, 210, 219, 220, 231, 232, 236, 246 gestures 38, 69, 70, 72, 78, 81, 83, 88, 89, 251, 253, 256, 260, 262 see also pointing gestures grammaticised categories 131 Grice 136, 138, 143 H hearer 13, 15–22, 24–27, 29–36, 42, 55, 56, 69, 100, 136, 137, 145, 148, 187–189, 191–195, 209, 215–217, 223, 225, 226, 228, 230, 233, 234, 236, 238, 241, 243 honorific terms of address 186 I illustrators 253

imagination-oriented deixis 78, 80, 83, 88, 89, 91 immediacy 242, 244 imperfective aspect 104, 117, 129 implication 131, 146, 150 implicature 55, 56, 135–145, 147–151, 159, 161 indexical expressions 251, 267 indicative 88, 135–142, 144, 147, 148, 155– 157, 162, 163, 165–171, 173, 174, 176, 177 indirect anaphora 223, 229–233, 241 deixis 223, 233, 234, 241 inference 151, 229, 232, 241, 243 inferior 195, 200 information organisation 97, 115, 120, 130, 132 inside space 27, 31 intended object 76–82, 84, 85 interactive achievement 252, 255, 256 intrinsic 61, 63, 71–75, 79, 81, 84, 90, 118, 121 J joint attention 258, 262, 266 L Lakoff 42, 56 Langacker 20, 21, 29, 32, 33, 37, 41, 60, 81, 153–155, 157, 177, 187, 220, 221, 223, 229, 230, 233, 240, 241 Levinson 38, 135, 138, 143, 144, 147, 189, 192, 193, 241, 242 local deixis 73, 75, 77, 87, 90, 204, 205, 212, 217, 227, 254, 268 dimension 75, 77, 83–85, 153 locative adverbs 7, 9 M macrostructure 118, 126 markedness 136, 138–140, 144–146, 148– 150

Index 273

medial 3–7, 10, 17, 26, 29, 32, 34, 35, 37 metaphor 181 metonymy 44, 51, 181 modal verb 140, 141, 143, 158, 163, 165, 166, 175 modality 135, 136, 138–141, 145, 148, 151, 153, 157–160, 165, 166, 168 mood 53, 136, 151, 153, 155, 157, 171, 177, 178 motion verbs 40, 41, 43, 62, 64, 67 multi-channel communication system 251 mutually shared perception 252, 257–259, 262, 263, 266 N necessity 73, 138–142, 144, 148, 153, 158, 159, 193, 228, 263, 265 negation 138, 139, 141, 146, 149, 150, 153, 157, 161–163, 174–179, 181 neutral deixis 210 O observer’s perspective 82 Old Indic xii, 203 opaque deixis 217 origo 18, 27, 42, 69–73, 75–82, 84–90, 153, 155, 186–188, 199, 203, 217, 227 instantiation 90 -allocating act 87 see also secondary origo outside space 21, 27, 30, 35 P paradigm 14, 15, 18, 20, 27, 135, 137–140, 142–150, 157, 189–191, 195, 197, 204, 205, 208, 210 past subjunctive 153, 156, 157, 161, 162, 165, 169–179 past tense 115, 117, 118, 120, 122, 127, 128, 158, 161, 178 perceived perception 249, 252, 259–261, 263, 264 perception see also mutually shared perception

perceived perception sensory perception visual perception perfective aspect 129 Perfekt 100, 106, 122–126, 128, 131 person-oriented 3, 13, 15, 16, 21, 28–30, 35 personal dimension 84, 86, 87, 91 pronoun 186, 188, 213–215, 232 perspective 17, 18, 20, 23, 27, 31, 33, 34, 73, 74, 76, 82, 109, 114–118, 120, 121, 126, 127, 129 see also observer’s perspective phasal decomposition 112 pointers 186, 251, 261, 265 pointing 16, 69–72, 78, 81–84, 89, 135, 186, 187, 199, 224, 226, 227, 234, 241, 250, 253, 254, 256, 258, 260–262, 265 pointing gestures 89, 253, 256, 260, 262, 266 politeness 185, 189–196, 199–201 distinctions 197 distinctions (frequency) 185 see also positive and negative politeness binary politeness distinctions Portuguese 3, 7–11, 19, 29, 31, 32, 34, 36– 40 positive and negative politeness 199 possibility 138, 139, 141, 142, 144, 159, 162, 176, 213, 217 pragmatic universal 35 present tense 101, 115, 116, 118, 121, 122, 126–128, 143, 145 primary deixis 250, 252 origo 86–88, 90 progressive 63, 98, 99, 104, 106, 108, 110, 111, 224 pronominal reference 187, 192, 199–201, 223, 224, 230, 234 Pronoun 186, 188–192, 194–197, 199, 202–206, 208–210, 212–220, 224, 230, 232, 234, 238–240, 242

274 Index

proximal 3, 4, 8, 10, 17, 19, 26, 29, 32, 34, 35, 187, 203, 204, 206, 210, 212, 215–217 deixis 204, 206, 210 Q quaestio 97, 98, 107, 122, 133 R reference point 4, 9, 73, 116, 129, 230, 233, 236 referent honorifics 189 referential frame 109, 110, 112, 114, 118, 122, 130, 131 relative clause 166, 171, 209 relatum 76–80, 84, 85 retelling 107, 108, 114, 122 Rigveda 203, 205, 221 root modality 158, 179 route description 78 S scalar implicature 137–140 scalarity 138 schema 42, 43 secondary origo 86, 87, 89, 90 sensory perception 250–252, 255–259, 261–264 shifters 189, 201, 258, 264 shifting 22, 23, 72, 73, 87, 264 side-by-side 13, 18, 19, 26–28, 30, 32–35 situated discourse 251, 255–257 situatedness 252, 261, 263, 264, 266 situational context 71, 72, 156, 195, 249, 250 social deixis 189 roles 188, 192, 193 sollen 153, 157, 158, 161, 162, 165, 166, 168–171, 173–177, 180, 181 Spanish 13–19, 26–36, 38–40, 65, 190

speaker 3–10, 15–24, 26, 27, 29–36, 42–44, 47, 53, 55, 56, 69–90, 105–109, 117, 130, 135–137, 140, 145–148, 154–178, 186– 188, 191–193, 195, 204, 216, 218, 219, 225–228, 234, 240, 242–244 deixis 135–137, 140, 141, 145, 147 speaker-based 158 speaker-side region 31, 35 speech act roles 186, 188, 189 subjectification 160, 161 subjective 60, 71, 153, 160, 161, 171, 173, 174, 176, 177 subjunctive 135, 153, 155–157, 161, 162, 165, 168–179 superior 195, 200 T temporal anchor 101, 103, 104, 106–109, 112, 126, 131 dimension 85 tense switch 114, 126 text cohesion 205, 208, 210 the self 193 three-term system 26, 29, 31, 32 time of situation 101, 103, 104, 106, 110, 112 Tiriyó 3–5, 9–11 topic time 102, 103, 105–108, 110, 112, 113, 115, 129 management 110 two-term system 3, 31, 33 U utterance 70, 72, 81, 83–85, 88, 90, 101, 105–108, 110, 116, 126, 186, 187, 191 V Vedic Sanskrit 203, 204, 213, 218 visual perception 81, 227, 257, 259–261

In the PRAGMATICS AND BEYOND NEW SERIES the following titles have been published thus far or are scheduled for publication: 1. WALTER, Bettyruth: The Jury Summation as Speech Genre: An Ethnographic Study of What it Means to Those who Use it. Amsterdam/Philadelphia, 1988. 2. BARTON, Ellen: Nonsentential Constituents: A Theory of Grammatical Structure and Pragmatic Interpretation. Amsterdam/Philadelphia, 1990. 3. OLEKSY, Wieslaw (ed.): Contrastive Pragmatics. Amsterdam/Philadelphia, 1989. 4. RAFFLER-ENGEL, Walburga von (ed.): Doctor-Patient Interaction. Amsterdam/Philadelphia, 1989. 5. THELIN, Nils B. (ed.): Verbal Aspect in Discourse. Amsterdam/Philadelphia, 1990. 6. VERSCHUEREN, Jef (ed.): Selected Papers from the 1987 International Pragmatics Conference. Vol. I: Pragmatics at Issue. Vol. II: Levels of Linguistic Adaptation. Vol. III: The Pragmatics of Intercultural and International Communication (ed. with Jan Blommaert). Amsterdam/Philadelphia, 1991. 7. LINDENFELD, Jacqueline: Speech and Sociability at French Urban Market Places. Amsterdam/Philadelphia, 1990. 8. YOUNG, Lynne: Language as Behaviour, Language as Code: A Study of Academic English. Amsterdam/Philadelphia, 1990. 9. LUKE, Kang-Kwong: Utterance Particles in Cantonese Conversation. Amsterdam/Philadelphia, 1990. 10. MURRAY, Denise E.: Conversation for Action. The computer terminal as medium of communication. Amsterdam/Philadelphia, 1991. 11. LUONG, Hy V.: Discursive Practices and Linguistic Meanings. The Vietnamese system of person reference. Amsterdam/Philadelphia, 1990. 12. ABRAHAM, Werner (ed.): Discourse Particles. Descriptive and theoretical investigations on the logical, syntactic and pragmatic properties of discourse particles in German. Amsterdam/Philadelphia, 1991. 13. NUYTS, Jan, A. Machtelt BOLKESTEIN and Co VET (eds): Layers and Levels of Representation in Language Theory: A functional view. Amsterdam/Philadelphia, 1990. 14. SCHWARTZ, Ursula: Young Children’s Dyadic Pretend Play. Amsterdam/Philadelphia, 1991. 15. KOMTER, Martha: Conflict and Cooperation in Job Interviews. Amsterdam/Philadelphia, 1991. 16. MANN, William C. and Sandra A. THOMPSON (eds): Discourse Description: Diverse Linguistic Analyses of a Fund-Raising Text. Amsterdam/Philadelphia, 1992. 17. PIÉRAUT-LE BONNIEC, Gilberte and Marlene DOLITSKY (eds): Language Bases ... Discourse Bases. Amsterdam/Philadelphia, 1991. 18. JOHNSTONE, Barbara: Repetition in Arabic Discourse. Paradigms, syntagms and the ecology of language. Amsterdam/Philadelphia, 1991. 19. BAKER, Carolyn D. and Allan LUKE (eds): Towards a Critical Sociology of Reading Pedagogy. Papers of the XII World Congress on Reading. Amsterdam/Philadelphia, 1991. 20. NUYTS, Jan: Aspects of a Cognitive-Pragmatic Theory of Language. On cognition, functionalism, and grammar. Amsterdam/Philadelphia, 1992. 21. SEARLE, John R. et al.: (On) Searle on Conversation. Compiled and introduced by Herman Parret and Jef Verschueren. Amsterdam/Philadelphia, 1992.

22. AUER, Peter and Aldo Di LUZIO (eds): The Contextualization of Language. Amsterdam/Philadelphia, 1992. 23. FORTESCUE, Michael, Peter HARDER and Lars KRISTOFFERSEN (eds): Layered Structure and Reference in a Functional Perspective. Papers from the Functional Grammar Conference, Copenhagen, 1990. Amsterdam/Philadelphia, 1992. 24. MAYNARD, Senko K.: Discourse Modality: Subjectivity, Emotion and Voice in the Japanese Language. Amsterdam/Philadelphia, 1993. 25. COUPER-KUHLEN, Elizabeth: English Speech Rhythm. Form and function in everyday verbal interaction. Amsterdam/Philadelphia, 1993. 26. STYGALL, Gail: Trial Language. A study in differential discourse processing. Amsterdam/ Philadelphia, 1994. 27. SUTER, Hans Jürg: The Wedding Report: A Prototypical Approach to the Study of Traditional Text Types. Amsterdam/Philadelphia, 1993. 28. VAN DE WALLE, Lieve: Pragmatics and Classical Sanskrit. Amsterdam/Philadelphia, 1993. 29. BARSKY, Robert F.: Constructing a Productive Other: Discourse theory and the convention refugee hearing. Amsterdam/Philadelphia, 1994. 30. WORTHAM, Stanton E.F.: Acting Out Participant Examples in the Classroom. Amsterdam/Philadelphia, 1994. 31. WILDGEN, Wolfgang: Process, Image and Meaning. A realistic model of the meanings of sentences and narrative texts. Amsterdam/Philadelphia, 1994. 32. SHIBATANI, Masayoshi and Sandra A. THOMPSON (eds): Essays in Semantics and Pragmatics. Amsterdam/Philadelphia, 1995. 33. GOOSSENS, Louis, Paul PAUWELS, Brygida RUDZKA-OSTYN, Anne-Marie SIMONVANDENBERGEN and Johan VANPARYS: By Word of Mouth. Metaphor, metonymy and linguistic action in a cognitive perspective. Amsterdam/Philadelphia, 1995. 34. BARBE, Katharina: Irony in Context. Amsterdam/Philadelphia, 1995. 35. JUCKER, Andreas H. (ed.): Historical Pragmatics. Pragmatic developments in the history of English. Amsterdam/Philadelphia, 1995. 36. CHILTON, Paul, Mikhail V. ILYIN and Jacob MEY: Political Discourse in Transition in Eastern and Western Europe (1989-1991). Amsterdam/Philadelphia, 1998. 37. CARSTON, Robyn and Seiji UCHIDA (eds): Relevance Theory. Applications and implications. Amsterdam/Philadelphia, 1998. 38. FRETHEIM, Thorstein and Jeanette K. GUNDEL (eds): Reference and Referent Accessibility. Amsterdam/Philadelphia, 1996. 39. HERRING, Susan (ed.): Computer-Mediated Communication. Linguistic, social, and cross-cultural perspectives. Amsterdam/Philadelphia, 1996. 40. DIAMOND, Julie: Status and Power in Verbal Interaction. A study of discourse in a closeknit social network. Amsterdam/Philadelphia, 1996. 41. VENTOLA, Eija and Anna MAURANEN, (eds): Academic Writing. Intercultural and textual issues. Amsterdam/Philadelphia, 1996. 42. WODAK, Ruth and Helga KOTTHOFF (eds): Communicating Gender in Context. Amsterdam/Philadelphia, 1997. 43. JANSSEN, Theo A.J.M. and Wim van der WURFF (eds): Reported Speech. Forms and functions of the verb. Amsterdam/Philadelphia, 1996.

44. BARGIELA-CHIAPPINI, Francesca and Sandra J. HARRIS: Managing Language. The discourse of corporate meetings. Amsterdam/Philadelphia, 1997. 45. PALTRIDGE, Brian: Genre, Frames and Writing in Research Settings. Amsterdam/Philadelphia, 1997. 46. GEORGAKOPOULOU, Alexandra: Narrative Performances. A study of Modern Greek storytelling. Amsterdam/Philadelphia, 1997. 47. CHESTERMAN, Andrew: Contrastive Functional Analysis. Amsterdam/Philadelphia, 1998. 48. KAMIO, Akio: Territory of Information. Amsterdam/Philadelphia, 1997. 49. KURZON, Dennis: Discourse of Silence. Amsterdam/Philadelphia, 1998. 50. GRENOBLE, Lenore: Deixis and Information Packaging in Russian Discourse. Amsterdam/Philadelphia, 1998. 51. BOULIMA, Jamila: Negotiated Interaction in Target Language Classroom Discourse. Amsterdam/Philadelphia, 1999. 52. GILLIS, Steven and Annick DE HOUWER (eds): The Acquisition of Dutch. Amsterdam/ Philadelphia, 1998. 53. MOSEGAARD HANSEN, Maj-Britt: The Function of Discourse Particles. A study with special reference to spoken standard French. Amsterdam/Philadelphia, 1998. 54. HYLAND, Ken: Hedging in Scientific Research Articles. Amsterdam/Philadelphia, 1998. 55. ALLWOOD, Jens and Peter Gärdenfors (eds): Cognitive Semantics. Meaning and cognition. Amsterdam/Philadelphia, 1999. 56. TANAKA, Hiroko: Language, Culture and Social Interaction. Turn-taking in Japanese and Anglo-American English. Amsterdam/Philadelphia, 1999. 57 JUCKER, Andreas H. and Yael ZIV (eds): Discourse Markers. Descriptions and theory. Amsterdam/Philadelphia, 1998. 58. ROUCHOTA, Villy and Andreas H. JUCKER (eds): Current Issues in Relevance Theory. Amsterdam/Philadelphia, 1998. 59. KAMIO, Akio and Ken-ichi TAKAMI (eds): Function and Structure. In honor of Susumu Kuno. 1999. 60. JACOBS, Geert: Preformulating the News. An analysis of the metapragmatics of press releases. 1999. 61. MILLS, Margaret H. (ed.): Slavic Gender Linguistics. 1999. 62. TZANNE, Angeliki: Talking at Cross-Purposes. The dynamics of miscommunication. 2000. 63. BUBLITZ, Wolfram, Uta LENK and Eija VENTOLA (eds.): Coherence in Spoken and Written Discourse. How to create it and how to describe it.Selected papers from the International Workshop on Coherence, Augsburg, 24-27 April 1997. 1999. 64. SVENNEVIG, Jan: Getting Acquainted in Conversation. A study of initial interactions. 1999. 65. COOREN, François: The Organizing Dimension of Communication. 2000. 66. JUCKER, Andreas H., Gerd FRITZ and Franz LEBSANFT (eds.): Historical Dialogue Analysis. 1999. 67. TAAVITSAINEN, Irma, Gunnel MELCHERS and Päivi PAHTA (eds.): Dimensions of Writing in Nonstandard English. 1999. 68. ARNOVICK, Leslie: Diachronic Pragmatics. Seven case studies in English illocutionary development. 1999.

69. NOH, Eun-Ju: The Semantics and Pragmatics of Metarepresentation in English. A relevance-theoretic account. 2000. 70. SORJONEN, Marja-Leena: Responding in Conversation. A study of response particles in Finnish. 2001. 71. GÓMEZ-GONZÁLEZ, María Ángeles: The Theme-Topic Interface. Evidence from English. 2001. 72. MARMARIDOU, Sophia S.A.: Pragmatic Meaning and Cognition. 2000. 73. HESTER, Stephen and David FRANCIS (eds.): Local Educational Order. Ethnomethodological studies of knowledge in action. 2000. 74. TROSBORG, Anna (ed.): Analysing Professional Genres. 2000. 75. PILKINGTON, Adrian: Poetic Effects. A relevance theory perspective. 2000. 76. MATSUI, Tomoko: Bridging and Relevance. 2000. 77. VANDERVEKEN, Daniel and Susumu KUBO (eds.): Essays in Speech Act Theory. 2002. 78. SELL, Roger D. : Literature as Communication. The foundations of mediating criticism. 2000. 79. ANDERSEN, Gisle and Thorstein FRETHEIM (eds.): Pragmatic Markers and Propositional Attitude. 2000. 80. UNGERER, Friedrich (ed.): English Media Texts – Past and Present. Language and textual structure. 2000. 81. DI LUZIO, Aldo, Susanne GÜNTHNER and Franca ORLETTI (eds.): Culture in Communication. Analyses of intercultural situations. 2001. 82. KHALIL, Esam N.: Grounding in English and Arabic News Discourse. 2000. 83. MÁRQUEZ REITER, Rosina: Linguistic Politeness in Britain and Uruguay. A contrastive study of requests and apologies. 2000. 84. ANDERSEN, Gisle: Pragmatic Markers and Sociolinguistic Variation. A relevance-theoretic approach to the language of adolescents. 2001. 85. COLLINS, Daniel E.: Reanimated Voices. Speech reporting in a historical-pragmatic perspective. 2001. 86. IFANTIDOU, Elly: Evidentials and Relevance. 2001. 87. MUSHIN, Ilana: Evidentiality and Epistemological Stance. Narrative retelling. 2001. 88. BAYRAKTAROG
LU, ArFn and Maria SIFIANOU (eds.): Linguistic Politeness Across Boundaries. The case of Greek and Turkish. 2001. 89. ITAKURA, Hiroko: Conversational Dominance and Gender. A study of Japanese speakers in first and second language contexts. 2001. 90. KENESEI, István and Robert M. HARNISH (eds.): Perspectives on Semantics, Pragmatics, and Discourse. A Festschrift for Ferenc Kiefer. 2001. 91. GROSS, Joan: Speaking in Other Voices. An ethnography of Walloon puppet theaters. 2001. 92. GARDNER, Rod: When Listeners Talk. Response tokens and listener stance. 2001. 93. BARON, Bettina and Helga KOTTHOFF (eds.): Gender in Interaction. Perspectives on femininity and masculinity in ethnography and discourse. 2002 94. McILVENNY, Paul (ed.): Talking Gender and Sexuality. 2002. 95. FITZMAURICE, Susan M.: The Familiar Letter in Early Modern English. A pragmatic approach. 2002. 96. HAVERKATE, Henk: The Syntax, Semantics and Pragmatics of Spanish Mood. 2002.

97. MAYNARD, Senko K.: Linguistic Emotivity. Centrality of place, the topic-comment dynamic, and an ideology of pathos in Japanese discourse. 2002. 98. DUSZAK, Anna (ed.): Us and Others. Social identities across languages, discourses and cultures. 2002. 99. JASZCZOLT, K.M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume 1. 2003. 100. JASZCZOLT, K.M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume 2. 2003. 101. LUKE, Kang Kwong and Theodossia-Soula PAVLIDOU (eds.): Telephone Calls. Unity and diversity in conversational structure across languages and cultures. 2002. 102. LEAFGREN, John: Degrees of Explicitness. Information structure and the packaging of Bulgarian subjects and objects. 2002. 103. FETZER, Anita and Christiane MEIERKORD (eds.): Rethinking Sequentiality. Linguistics meets conversational interaction. 2002. 104. BEECHING, Kate: Gender, Politeness and Pragmatic Particles in French. 2002. 105. BLACKWELL, Sarah E.: Implicatures in Discourse. The case of Spanish NP anaphora. 2003. 106. BUSSE, Ulrich: Linguistic Variation in the Shakespeare Corpus. Morpho-syntactic variability of second person pronouns. 2002. 107. TAAVITSAINEN, Irma and Andreas H. JUCKER (eds.): Diachronic Perspectives on Address Term Systems. n.y.p. 108. BARRON, Anne: Acquisition in Interlanguage Pragmatics. How to do things with words in a study abroad context. n.y.p. 109. MAYES, Patricia: Language, Social Structure, and Culture. A genre analysis of cooking classes in Japan and America. 2003. 110. ANDROUTSOPOULOS, Jannis K. and Alexandra GEORGAKOPOULOU (eds.): Discourse Constructions of Youth Identities. n.y.p. 111. ENSINK, Titus and Christoph SAUER (eds.): Framing and Perspectivising in Discourse. n.y.p. 112. LENZ, Friedrich (ed.): Deictic Conceptualisation of Space, Time and Person. 2003. 113. PANTHER, Klaus-Uwe and Linda L. THORNBURG (eds.): Metonymy and Pragmatic Inferencing. n.y.p. 114. KÜHNLEIN, Peter, Hannes RIESER and Henk ZEEVAT (eds.): Perspectives on Dialogue in the New Millennium. n.y.p.